Affiliations
Department of Family Medicine, Boston University School of Medicine/ Boston Medical Center, Boston, Massachusetts
Given name(s)
Michael K.
Family name
Paasche‐Orlow
Degrees
MD, MA, MPH

Predictors of Medication Adherence

Article Type
Changed
Mon, 05/22/2017 - 18:40
Display Headline
Predictors of medication adherence postdischarge: The impact of patient age, insurance status, and prior adherence

In the outpatient setting, medication adherence (defined as percentage of prescribed medication doses taken by a patient during a specific time period) ranges between 40% and 80% for chronic conditions.1 During acute care hospitalization, changes are often made to patients' medication regimens, which can be confusing and contribute to nonadherence, medication errors, and harmful adverse events.2 Indeed, it is estimated that almost half of patients encounter a medication error after discharge, and approximately 12%17% experience an adverse drug event after returning home.36 It is likely that some of these adverse events may be the result of medication nonadherence.7 Improved patientprovider communication, systems to reconcile prehospitalization and posthospitalization medications, as well as development of mechanisms to enhance adherence, may prevent many of these errors and have become new targets for quality improvement.4, 8 Although postdischarge medication adherence is a crucial target for avoiding adverse events and rehospitalization, few studies have focused on understanding its incidence and predictors, in particular, patient demographic factors such as age and insurance status.911

In addition, few studies have looked at general and posthospital adherence in a population where health literacy is measured, an important area because medication changes during hospitalization may be particularly confusing for patients with low health literacy.11, 12 Health literacy is defined as the degree to which an individual has the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions.13 Prior outpatient research shows that low health literacy is associated with poor patient understanding of the medication regimen and instructions for medication use, which may contribute to postdischarge medication nonadherence.14, 15 Understanding the factors associated with postdischarge medication adherence could help refine interventions that are oriented toward improving transitions in care, patient safety, and reducing unnecessary rehospitalization.

We report here on factors associated with postdischarge medication adherence using data from the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL‐CVD) study.16

METHODS

Study and Participants

PILL‐CVD was a federally funded, 2‐site randomized controlled trial using pharmacist‐assisted medication reconciliation, inpatient pharmacist counseling, low‐literacy adherence aids, and telephone follow‐up that aimed to decrease rates of serious medication errors after hospital discharge.16 The study targeted patients with cardiovascular disease (hospitalized on cardiology or general medical or geriatric services for acute coronary syndromes [ACS] or acute decompensated heart failure [ADHF]) at 2 large academic hospitals, Brigham and Women's Hospital (BWH) and Vanderbilt University Hospital (VUH).

Subjects were eligible for enrollment if they met criteria for ACS or ADHF, were likely to be discharged to home as determined by the primary medical team at the time of study enrollment, and took primary responsibility for administering their medications prior to admission (caregivers could be involved in medication management after discharge). Exclusion criteria included severe visual or hearing impairment, inability to communicate in English or Spanish, active psychiatric illness, dementia, delirium, illness too severe to participate, lack of a home phone number, being in police custody, or participation in another intensive medication adherence program (eg, due to renal transplant).

Out of 6416 patients originally screened for possible enrollment, 862 were randomly assigned to receive usual care or usual care plus the intervention, and 851 remained in the study.16 Both the main study and this secondary data analysis were approved by the Institutional Review Boards of each site.

Baseline Measures

Following informed consent and study enrollment, a variety of baseline data were collected on study participants from medical records and patient interview, including primary language, demographic information (age, race, insurance status, income, and education level), cognition (through administration of the 05‐point MiniCog scale),17 and level of health literacy (through use of the 036‐point short form of the Test of Functional Health Literacy in Adults [s‐TOFHLA] scale).18 Baseline information was also collected on medication use, including number of preadmission medications, measurement of self‐reported adherence prior to admission (using the Morisky scale, a validated 04‐point questionnaire shown to correlate with disease control and indicative of general patterns of adherence),19 and a medication understanding score, adapted from other instruments, which quantifies understanding of the indication, dose, and frequency of up to 5 randomly selected preadmission medications on a 03‐point scale.16, 20, 21

Outcome Measures

Outcomes were collected 30 days postdischarge through a structured questionnaire, administered by telephone. Only patients who completed this call are included in the present analysis. Postdischarge medication adherence was assessed by asking patients to report the number of days out of the previous week they had taken each medication from their postdischarge regimen exactly as prescribed.22 A score was calculated for each medication as the proportion of adherent days (eg, if a patient reported missing 2 days of a medication in the previous week, then adherence would be 5/7 or 71%). A global postdischarge adherence score was then derived for each patient by averaging the adherence score for all regularly scheduled medications. This quantitative measure focused on adherence to medications patients knew they should be taking and did not measure medication discrepancies (sometimes termed unintentional nonadherence).

Analysis

Patient characteristics were summarized and reported using simple descriptive statistics. Candidate predictors of postdischarge medication adherence were chosen a priori from patient characteristics assessed during hospital admission. These included patient age, gender, race, ethnicity, marital status, insurance, years of education, presence of primary care physician (PCP), study site, number of preadmission medications, medication understanding, baseline adherence, cognition, and health literacy. Unadjusted results were calculated using univariable linear regression, with each patient's adherence score as the dependent variable and each predictor as the independent variable. Adjusted results were then derived using multivariable linear regression with all the candidate predictors in the model.

Lastly, because of missing data for some predictors, in particular baseline adherence and medication understanding, multiple imputation techniques were used to impute missing data and increase statistical power.23 We used the Markov Chain Monte Carlo (MCMC) method for multiple imputation, which generally assumes that the data came from a normal distribution and that the missing data are missing at random. Because of the essentially normal distribution of the data, and because the amount of missing data was so small (<1% for almost all variables, 5% for baseline adherence, and 8% for medication understanding), we expected little bias and present the complete case analysis, which maximized statistical power.

Two‐sided P values <0.05 were considered significant, and SAS version 9.2 (Cary, NC) was used for all analyses.

RESULTS

Table 1 shows descriptive baseline patient characteristics of study sample (responders) as well as nonresponders at 30 days. For the responders, the mean age of the 646 patients was 61.2 years, 94.7% were insured, and 19.3% had inadequate or marginal health literacy. Patients were prescribed an average of 8 preadmission medications. Most patients (92.3%) had a regular PCP prior to admission. Nonresponders had nonsignificant trends towards having lower health literacy, medication understanding, and baseline medication adherence.

Baseline Characteristics
CharacteristicTotal N, 30‐Day RespondentsValueTotal N, NonrespondentsValue
  • Abbreviations: PCP, primary care physician; SD, standard deviation; s‐TOFHLA, short form of the Test of Functional Health Literacy in Adults. *03, with 3 indicating better understanding. 036, with higher scores indicating higher health literacy. 04, with 4 indicating higher baseline adherence. 05, with higher scores indicating better cognition; a score <3 indicates dementia.

Age, mean in yr (SD)64661.2 (13.5)4555.4 (14.3)
Gender, N (percentage)646 45 
Female272 (42.1)18 (40.0)
Male374 (57.9)27 (60.0)
Race, N (percentage)643 45 
White511 (79.5)32 (71.1)
Black104 (16.2)11 (24.4)
Other28 (4.4)2 (4.4)
Ethnicity, N (percentage)639 45 
Hispanic24 (3.8)1 (2.2)
Not Hispanic615 (96.2)44 (97.8)
Marital status, N (percentage)646 45 
Married/cohabitate382 (59.1)20 (44.4)
Separated/divorced118 (18.3)11 (24.4)
Widowed81 (12.5)5 (11.1)
Never married65 (10.1)9 (2.0)
Insurance type, N (percentage)646 45 
Medicaid53 (8.2)5 (11.1)
Medicare270 (41.8)13 (28.9)
Private289 (44.7)19 (42.2)
Self‐pay34 (5.3)8 (17.8)
Years of education, mean in yr (SD)64314.0 (3.1)4513.3 (2.7)
Presence of PCP prior to admission, N (percentage)646 45 
Yes596 (92.3)38 (84.4)
No50 (7.74)7 (15.6)
Site, N (percentage)646 45 
Site 1358 (55.4)8 (17.8)
Site 2288 (44.6)37 (82.2)
No. of preadmission medications, mean no. (SD)6417.8 (4.8)457.7 (5.4)
Medication understanding score, mean (SD)*5972.4 (0.5)402.2 (0.62)
Health literacy (s‐TOFHLA) score, mean (SD)64229.1 (8.9)4526.0 (12.0)
Baseline adherence (SD)6132.7 (1.1)452.4 (1.2)
MiniCog score, N (percentage)646 45 
Demented63 (9.8)5 (11.1)
Not demented583 (90.2)40 (88.9)

The average postdischarge adherence score was 95% (standard deviation [SD] = 10.2%), and less than 10% of patients had an adherence score of less than 85%; overall the distribution was left‐skewed. Table 2 illustrates crude and adjusted parameter estimates for variables in the model. Table 3 shows significant findings in the fully adjusted model, which used multiple imputation techniques to account for missing data.

Crude and Adjusted Measurements
PredictorCrude Parameter Estimate (Beta) With 95% Confidence IntervalsP ValueAdjusted Parameter Estimate (Beta) With 95% Confidence IntervalsP Value
  • NOTE: For crude estimates, value is category vs absence of parameter in univariable testing. For adjusted estimates of categorical variables, value is each category compared to referent category. Beta‐coefficient represents absolute change in adherence (eg, 0.010 for age means a 1% absolute increase in adherence for every 10 yr increase in patient age). Abbreviations: PCP, primary care physician; Ref, referent; s‐TOFHLA, short form of the Test of Functional Health Literacy in Adults.

Age per 10 yr0.010 (0.007, 0.020)<0.00010.010 (0.002, 0.020)0.018
Male gender0.012 (0.004, 0.028)0.1370.003 (0.014, 0.020)0.727
Race/ethnicity    
White0.011 (0.009, 0.031)0.266RefRef
Black0.017 (0.038, 0.005)0.130.006 (0.017, 0.030)0.598
Other0.010 (0.029, 0.049)0.5990.017 (0.027, 0.062)0.446
Hispanic/Latino0.005 (0.037, 0.047)0.8030.036 (0.013, 0.085)0.149
Marital status    
Married/cohabitate0.006 (0.011, 0.022)0.500RefRef
Separated/divorced0.005 (0.025, 0.016)0.6640.009 (0.014, 0.031)0.446
Widowed0.001 (0.023, 0.025)0.9220.013 (0.039, 0.013)0.338
Never married0.009 (0.035, 0.018)0.5150.004 (0.033, 0.025)0.784
Insurance type    
Private0.008 (0.008, 0.024)0.347RefRef
Medicaid0.046 (0.075, 0.018)0.0020.026 (0.058, 0.007)0.121
Medicare0.012 (0.004, 0.028)0.1380.002 (0.023, 0.018)0.844
Self‐pay0.027 (0.062, 0.008)0.1350.029 (0.073, 0.015)0.202
Years of education0.003 (0.0003, 0.005)0.0280.0001 (0.003, 0.003)0.949
Presence of PCP prior to admission0.007 (0.022, 0.037)0.6300.002 (0.032, 0.036)0.888
Site0.050 (0.065, 0.034)<0.00010.038 (0.056, 0.021)<0.0001
No. of preadmission medications0.0003 (0.002, 0.001)0.6840.0001 (0.002, 0.002)0.918
Medication understanding score per point0.007 (0.009, 0.023)0.3900.006 (0.011, 0.023)0.513
Health literacy (s‐TOFHLA) score per 10 points0.0006 (0.008, 0.01)0.8970.003 (0.008, 0.01)0.644
Baseline adherence per point0.023 (0.016, 0.031)<0.00010.017 (0.009, 0.024)<0.0001
Cognitive function0.004 (0.022, 0.031)0.7570.008 (0.019, 0.036)0.549
Significant Results in Adjusted Analyses With Multiple Imputation
PredictorParameter Estimate (Beta) With 95% Confidence IntervalsP Value
  • NOTE: Total observations, 646; 67 with missing values. All variables adjusted for gender, race, cognitive function, number of preadmission medications, marital status, health literacy score, medication understanding score, presence of primary care physician (PCP), years of school, Hispanic/Latino ethnicity. Abbreviations: Ref, referent.

Age per 10 yr0.010 (0.004, 0.020)0.004
Insurance type  
PrivateRefRef
Medicaid0.045 (0.076, 0.014)0.005
Medicare0.010 (0.030, 0.010)0.333
Self‐pay0.013 (0.050, 0.025)0.512
Site0.036 (0.053, 0.019)<0.0001
Baseline adherence per point0.016 (0.008, 0.024)<0.0001

Intervention arm was of borderline statistical significance in predicting postdischarge adherence (P = 0.052), and so was removed from the final model. Study site, age, insurance, and baseline adherence were the only significant independent predictors of postdischarge adherence in the fully adjusted model (Table 3). For example, for every 10‐year increase in age, patients had, on average, an adjusted 1% absolute increase in their adherence score (95% confidence interval [CI] 0.4% to 2.0%). For every 1‐point increase in baseline medication adherence (based on the Morisky scale), there was a 1.6% absolute increase in medication adherence (95% CI 0.8% to 2.4%). In unadjusted analyses, patients with Medicaid were less adherent with medications after discharge than were patients with private insurance. This difference became nonsignificant in adjusted analyses, but when analyses were repeated using multiple imputation techniques, the results again became statistically significantMedicaid insurance was associated with a 4.5% absolute decrease in postdischarge adherence compared with private insurance (95% CI 7.6% to 1.4%). Study site (specifically, Brigham and Women's Hospital) was also a significant predictor of greater postdischarge medication adherence. Years of education was a significant predictor of adherence in unadjusted analyses, but was not an independent predictor when adjusted for other factors. When baseline adherence was removed from the multiple imputation model, there were no changes in which factors were significant predictors of adherence.

DISCUSSION

In this study, we found that low baseline adherence, younger age, Medicaid insurance, and study site were significant predictors of lower 30‐day medication adherence. Of particular interest is our finding regarding baseline adherence, a simple measure to obtain on hospitalized patients. It is notable that in our study, education was not an independent significant predictor of postdischarge adherence, even when baseline adherence was removed from the model. The same is true for medication understanding, cognitive function, and health literacy.

Older patients appeared more adherent with medications in the month after hospital discharge, perhaps reflecting increased interaction with the healthcare system (appointments, number of physician interactions), a greater belief in the importance of chronic medication management, or a higher level of experience with managing medications. A similar relationship between age and adherence has been shown in outpatient studies of patients with hypertension, diabetes, and other chronic diseases.2427

Medicaid patients may be less likely to remain adherent because of the plan's limited coverage of medications relative to patients' ability to pay. For example, Medicaid in Tennessee covers the first 5 generic medications at no cost to the patient but has co‐payments for additional medications and for brand name drugs. Medicaid in Massachusetts has co‐payments of $1 to $3 for each medication. Alternatively, Medicaid insurance may be a marker for other patient characteristics associated with low adherence for which we were not fully able to adjust.

Site differences were also notable in this study; these differences could have been due to differences in insurance coverage in Tennessee versus Massachusetts (which has near‐universal coverage), differences in types of insurance (eg, fewer patients at Brigham and Women's Hospital had Medicaid than at Vanderbilt), cultural and geographic differences between the 2 locations, or other differences in transitional care between the 2 sites.

This study corroborates previous literature on medication adherence (specifically unintentional nonadherence) in the outpatient setting,4, 811 for example, on the association of younger age with low adherence in certain populations. On the other hand, it may contrast with previous literature which has sometimes shown a relationship between patient education or health literacy and medication adherence.14, 15, 2835 However, previous studies have not focused on the transition from inpatient to outpatient settings. Perhaps intensive medication education in the hospital, even under usual care, mitigates the effects of these factors on postdischarge adherence. Finally, baseline adherence seems to correlate with postdischarge adherence, a finding which makes intuitive sense and has been previously reported for specific medications.36

There are several limitations to this study. Although large, the study was performed at only 2 clinical sites where most patients were white and fairly well‐educated, perhaps because patients admitted to a tertiary care center with ACS or ADHF are more affluent than general medical inpatients as a whole; this may limit generalizability. Postdischarge medication adherence might have been higher than in other patient populations given the nature of the population, possible loss‐to‐follow‐up bias, and the fact that half of the subjects received an intervention designed to improve medication management after discharge; such low rates of nonadherence in our study may have reduced our ability to detect important predictors in our models. In addition, the period of follow‐up was 30 days, thus limiting our findings to short‐term postdischarge medication adherence. Postdischarge medication adherence was based on patient self‐report, which not only assumed that the patient was still managing his/her own medications after discharge, but may also be susceptible to both recall and social acceptability bias, which might overestimate our adherence scores, again limiting our ability to detect important predictors of nonadherence. However, other studies have shown a good correlation between self‐reported medication adherence and other more objective measures,37, 38 and recall was only for 7 days, a measure used previously in the literature39, 40 and one designed to reduce recall bias. Systematic underreporting in certain patient populations is less likely but possible.

In the future, research should focus on targeting patients who have low baseline adherence to evaluate the effects of various interventions on postdischarge medication outcomes. Repeating the study in a population with a high prevalence of low health literacy might be illuminating, given that previous studies have shown that patients with low health literacy have less ability to identify their medications and have less refill adherence.29, 30

In conclusion, in patients hospitalized with cardiovascular disease, predictors of lower postdischarge adherence include younger age, Medicaid insurance, and low baseline adherence. It may be prudent to assess baseline adherence and insurance type in hospitalized patients in order to identify those who may benefit from additional assistance to improve medication adherence and medication safety during transitions in care.

Acknowledgements

Meeting Presentations: SGIM New England Regional Meeting, oral presentation, Boston, MA, March 4, 2011; and SGIM National Meeting, poster presentation, Phoenix, AZ, May 6, 2011. Dr Schnipper had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Disclosures: Financial support was provided by R01 HL089755 (NHLBI, Kripalani), K23 HL077597 (NHLBI, Kripalani), K08 HL072806 (NHLBI, Schnipper), T32HP10251‐02 (Cohen), and by the Division of General Medicine, Massachusetts General Hospital and the Harvard Medical School Fellowship in General Medicine and Primary Care (Cohen). Dr Kripalani is a consultant to and holds equity in PictureRx, LLC, which makes patient education tools to improve medication management. PictureRx did not provide materials or funding for this study. All other authors disclose no relevant or financial conflicts of interest.

Files
References
  1. Osterberg L,Blaschke T.Adherence to medication.N Engl J Med.2005;353(5):487497.
  2. Coleman EA,Smith JD,Raha D,Min SJ.Posthospital medication discrepancies: prevalence and contributing factors.Arch Intern Med.2005;165(16):18421847.
  3. Cua YM,Kripalani S.Medication use in the transition from hospital to home.Ann Acad Med Singapore.2008;37(2):136141.
  4. Moore C,Wisnivesky J,Williams S,McGinn T.Medical errors related to discontinuity of care from an inpatient to an outpatient setting.J Gen Intern Med.2003;18(8):646651.
  5. Forster AJ,Murff HJ,Peterson JF,Gandhi TK,Bates DW.The incidence and severity of adverse events affecting patients after discharge from the hospital.Ann Intern Med.2003;138(3):161167.
  6. Forster AJ,Murff HJ,Peterson JF,Gandhi TK,Bates DW.Adverse drug events occurring following hospital discharge.J Gen Intern Med.2005;20(4):317323.
  7. Schnipper JL,Kirwin JL,Cotugno MC, et al.Role of pharmacist counseling in preventing adverse drug events after hospitalization.Arch Intern Med.2006;166(5):565571.
  8. Vira T,Colquhoun M,Etchells E.Reconcilable differences: correcting medication errors at hospital admission and discharge.Qual Saf Health Care.2006;15(2):122126.
  9. Hassan M,Lage MJ.Risk of rehospitalization among bipolar disorder patients who are nonadherent to antipsychotic therapy after hospital discharge.Am J Health Syst Pharm.2009;66(4):358365.
  10. Mansur N,Weiss A,Hoffman A,Gruenewald T,Beloosesky Y.Continuity and adherence to long‐term drug treatment by geriatric patients after hospital discharge: a prospective cohort study.Drugs Aging.2008;25(10):861870.
  11. Kripalani S,Henderson LE,Jacobson TA,Vaccarino V.Medication use among inner‐city patients after hospital discharge: patient‐reported barriers and solutions.Mayo Clin Proc.2008;83(5):529535.
  12. Lindquist LA,Go L,Fleisher J,Jain N,Friesema E,Baker DW.Relationship of health literacy to intentional and unintentional non‐adherence of hospital discharge medications.J Gen Intern Med.2012;27(2):173178.
  13. Office of Disease Prevention and Health Promotion, US Department of Health and Human Services.Healthy People 2010. Available at: http://www.healthypeople.gov/Document/pdf/uih/2010uih.pdf. Accessed February 15,2012.
  14. Davis TC,Wolf MS,Bass PF, et al.Literacy and misunderstanding prescription drug labels.Ann Intern Med.2006;145(12):887894.
  15. Kripalani S,Henderson LE,Chiu EY,Robertson R,Kolm P,Jacobson TA.Predictors of medication self‐management skill in a low‐literacy population.J Gen Intern Med.2006;21(8):852856.
  16. Schnipper JL,Roumie CL,Cawthorn C, et al;for the PILL‐CVD Study Group.Rationale and design of the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL‐CVD) study.Circ Cardiovasc Qual Outcomes.2010;3(2):212219.
  17. Borson S,Scanlan JM,Watanabe J,Tu SP,Lessig M.Simplifying detection of cognitive impairment: comparison of the Mini‐Cog and Mini‐Mental State Examination in a multiethnic sample.J Am Geriatr Soc.2005;53(5):871874.
  18. Nurss JR.Short Test of Functional Health Literacy in Adults.Snow Camp, NC:Peppercorn Books and Press;1998.
  19. Morisky DE,Ang A,Krousel‐Wood M,Ward HJ.Predictive validity of a medication adherence measure in an outpatient setting.J Clin Hypertens (Greenwich).2008;10(5):348354.
  20. Marvanova M,Roumie CL,Eden SK,Cawthon C,Schnipper JL,Kripalani S.Health literacy and medication understanding among hospitalized adults.J Hosp Med. In press.
  21. Marvanova M,Roumie CL,Eden SK,Cawthon C,Schnipper JL,Kripalani S.Health literacy and medication understanding among hospitalized adults.J Hosp Med.2011;6(9):488493.
  22. Toobert DJ,Hampson SE,Glasgow RE.The summary of diabetes self‐care activities measure: results from 7 studies and a revised scale.Diabetes Care.2000;23(7):943950.
  23. Rubin DB.Multiple Imputation for Nonresponse in Surveys.New York, NY:John Wiley 1987.
  24. Hinkin CH,Hardy DJ,Mason KI, et al.Medication adherence in HIV‐infected adults: effect of patient age, cognitive status, and substance abuse.AIDS.2004;18(suppl 1):S19S25.
  25. Wong MC,Jiang JY,Griffiths SM.Factors associated with antihypertensive drug compliance in 83,884 Chinese patients: a cohort study.J Epidemiol Community Health.2010;64(10):895901.
  26. Wong MC,Kong AP,So WY,Jiang JY,Chan JC,Griffiths SM.Adherence to oral hypoglycemic agents in 26,782 Chinese patients: a cohort study.J Clin Pharmacol.2011;51(10):14741482.
  27. Gazmararian J,Jacobson KL,Pan Y,Schmotzer B,Kripalani S.Effect of a pharmacy‐based health literacy intervention and patient characteristics on medication refill adherence in an urban health system.Ann Pharmacother.2010;44(1):8087.
  28. Kalichman SC,Ramachandran B,Catz S.Adherence to combination antiretroviral therapies in HIV patients of low health literacy.J Gen Intern Med.1999;14(5):267273.
  29. Gazmararian JA,Kripalani S,Miller MJ,Echt KV,Ren J,Rask K.Factors associated with medication refill adherence in cardiovascular‐related diseases: a focus on health literacy.J Gen Intern Med.2006;21(12):12151221.
  30. Persell SD,Osborn CY,Richard R,Skripkauskas S,Wolf MS.Limited health literacy is a barrier to medication reconciliation in ambulatory care.J Gen Intern Med.2007;22(11):15231526.
  31. Chew LD,Bradley KA,Flum DR,Cornia PB,Koepsell TD.The impact of low health literacy on surgical practice.Am J Surg.2004;188(3):250253.
  32. Gatti ME,Jacobson KL,Gazmararian JA,Schmotzer B,Kripalani S.Relationships between beliefs about medications and adherence.Am J Health Syst Pharm.2009;66(7):657664.
  33. Fang MC,Machtinger EL,Wang F,Schillinger D.Health literacy and anticoagulation‐related outcomes among patients taking warfarin.J Gen Intern Med.2006;21(8):841846.
  34. Paasche‐Orlow MK,Cheng DM,Palepu A,Meli S,Faber V,Samet JH.Health literacy, antiretroviral adherence, and HIV‐RNA suppression: a longitudinal perspective.J Gen Intern Med.2006;21(8):835840.
  35. Platt AB,Localio AR,Brensinger CM, et al.Risk factors for nonadherence to warfarin: results from the IN‐RANGE study.Pharmacoepidemiol Drug Saf.2008;17(9):853860.
  36. Muntner P,Mann DM,Woodward M, et al.Predictors of low clopidogrel adherence following percutaneous coronary intervention.Am J Cardiol.2011;108(6):822827.
  37. Shi L,Liu J,Fonseca V,Walker P,Kalsekar A,Pawaskar M.Correlation between adherence rates measured by MEMS and self‐reported questionnaires: a meta‐analysis.Health Qual Life Outcomes.2010;8:99.
  38. Shi L,Liu J,Koleva Y,Fonseca V,Kalsekar A,Pawaskar M.Concordance of adherence measurement using self‐reported adherence questionnaires and medication monitoring devices.Pharmacoeconomics.2010;28(12):10971107.
  39. Grant RW,Devita NG,Singer DE,Meigs JB.Polypharmacy and medication adherence in patients with type 2 diabetes.Diabetes Care.2003;26(5):14081412.
  40. Grant RW,Devita NG,Singer DE,Meigs JB.Improving adherence and reducing medication discrepancies in patients with diabetes.Ann Pharmacother.2003;37(7–8):962969.
Article PDF
Issue
Journal of Hospital Medicine - 7(6)
Publications
Page Number
470-475
Sections
Files
Files
Article PDF
Article PDF

In the outpatient setting, medication adherence (defined as percentage of prescribed medication doses taken by a patient during a specific time period) ranges between 40% and 80% for chronic conditions.1 During acute care hospitalization, changes are often made to patients' medication regimens, which can be confusing and contribute to nonadherence, medication errors, and harmful adverse events.2 Indeed, it is estimated that almost half of patients encounter a medication error after discharge, and approximately 12%17% experience an adverse drug event after returning home.36 It is likely that some of these adverse events may be the result of medication nonadherence.7 Improved patientprovider communication, systems to reconcile prehospitalization and posthospitalization medications, as well as development of mechanisms to enhance adherence, may prevent many of these errors and have become new targets for quality improvement.4, 8 Although postdischarge medication adherence is a crucial target for avoiding adverse events and rehospitalization, few studies have focused on understanding its incidence and predictors, in particular, patient demographic factors such as age and insurance status.911

In addition, few studies have looked at general and posthospital adherence in a population where health literacy is measured, an important area because medication changes during hospitalization may be particularly confusing for patients with low health literacy.11, 12 Health literacy is defined as the degree to which an individual has the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions.13 Prior outpatient research shows that low health literacy is associated with poor patient understanding of the medication regimen and instructions for medication use, which may contribute to postdischarge medication nonadherence.14, 15 Understanding the factors associated with postdischarge medication adherence could help refine interventions that are oriented toward improving transitions in care, patient safety, and reducing unnecessary rehospitalization.

We report here on factors associated with postdischarge medication adherence using data from the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL‐CVD) study.16

METHODS

Study and Participants

PILL‐CVD was a federally funded, 2‐site randomized controlled trial using pharmacist‐assisted medication reconciliation, inpatient pharmacist counseling, low‐literacy adherence aids, and telephone follow‐up that aimed to decrease rates of serious medication errors after hospital discharge.16 The study targeted patients with cardiovascular disease (hospitalized on cardiology or general medical or geriatric services for acute coronary syndromes [ACS] or acute decompensated heart failure [ADHF]) at 2 large academic hospitals, Brigham and Women's Hospital (BWH) and Vanderbilt University Hospital (VUH).

Subjects were eligible for enrollment if they met criteria for ACS or ADHF, were likely to be discharged to home as determined by the primary medical team at the time of study enrollment, and took primary responsibility for administering their medications prior to admission (caregivers could be involved in medication management after discharge). Exclusion criteria included severe visual or hearing impairment, inability to communicate in English or Spanish, active psychiatric illness, dementia, delirium, illness too severe to participate, lack of a home phone number, being in police custody, or participation in another intensive medication adherence program (eg, due to renal transplant).

Out of 6416 patients originally screened for possible enrollment, 862 were randomly assigned to receive usual care or usual care plus the intervention, and 851 remained in the study.16 Both the main study and this secondary data analysis were approved by the Institutional Review Boards of each site.

Baseline Measures

Following informed consent and study enrollment, a variety of baseline data were collected on study participants from medical records and patient interview, including primary language, demographic information (age, race, insurance status, income, and education level), cognition (through administration of the 05‐point MiniCog scale),17 and level of health literacy (through use of the 036‐point short form of the Test of Functional Health Literacy in Adults [s‐TOFHLA] scale).18 Baseline information was also collected on medication use, including number of preadmission medications, measurement of self‐reported adherence prior to admission (using the Morisky scale, a validated 04‐point questionnaire shown to correlate with disease control and indicative of general patterns of adherence),19 and a medication understanding score, adapted from other instruments, which quantifies understanding of the indication, dose, and frequency of up to 5 randomly selected preadmission medications on a 03‐point scale.16, 20, 21

Outcome Measures

Outcomes were collected 30 days postdischarge through a structured questionnaire, administered by telephone. Only patients who completed this call are included in the present analysis. Postdischarge medication adherence was assessed by asking patients to report the number of days out of the previous week they had taken each medication from their postdischarge regimen exactly as prescribed.22 A score was calculated for each medication as the proportion of adherent days (eg, if a patient reported missing 2 days of a medication in the previous week, then adherence would be 5/7 or 71%). A global postdischarge adherence score was then derived for each patient by averaging the adherence score for all regularly scheduled medications. This quantitative measure focused on adherence to medications patients knew they should be taking and did not measure medication discrepancies (sometimes termed unintentional nonadherence).

Analysis

Patient characteristics were summarized and reported using simple descriptive statistics. Candidate predictors of postdischarge medication adherence were chosen a priori from patient characteristics assessed during hospital admission. These included patient age, gender, race, ethnicity, marital status, insurance, years of education, presence of primary care physician (PCP), study site, number of preadmission medications, medication understanding, baseline adherence, cognition, and health literacy. Unadjusted results were calculated using univariable linear regression, with each patient's adherence score as the dependent variable and each predictor as the independent variable. Adjusted results were then derived using multivariable linear regression with all the candidate predictors in the model.

Lastly, because of missing data for some predictors, in particular baseline adherence and medication understanding, multiple imputation techniques were used to impute missing data and increase statistical power.23 We used the Markov Chain Monte Carlo (MCMC) method for multiple imputation, which generally assumes that the data came from a normal distribution and that the missing data are missing at random. Because of the essentially normal distribution of the data, and because the amount of missing data was so small (<1% for almost all variables, 5% for baseline adherence, and 8% for medication understanding), we expected little bias and present the complete case analysis, which maximized statistical power.

Two‐sided P values <0.05 were considered significant, and SAS version 9.2 (Cary, NC) was used for all analyses.

RESULTS

Table 1 shows descriptive baseline patient characteristics of study sample (responders) as well as nonresponders at 30 days. For the responders, the mean age of the 646 patients was 61.2 years, 94.7% were insured, and 19.3% had inadequate or marginal health literacy. Patients were prescribed an average of 8 preadmission medications. Most patients (92.3%) had a regular PCP prior to admission. Nonresponders had nonsignificant trends towards having lower health literacy, medication understanding, and baseline medication adherence.

Baseline Characteristics
CharacteristicTotal N, 30‐Day RespondentsValueTotal N, NonrespondentsValue
  • Abbreviations: PCP, primary care physician; SD, standard deviation; s‐TOFHLA, short form of the Test of Functional Health Literacy in Adults. *03, with 3 indicating better understanding. 036, with higher scores indicating higher health literacy. 04, with 4 indicating higher baseline adherence. 05, with higher scores indicating better cognition; a score <3 indicates dementia.

Age, mean in yr (SD)64661.2 (13.5)4555.4 (14.3)
Gender, N (percentage)646 45 
Female272 (42.1)18 (40.0)
Male374 (57.9)27 (60.0)
Race, N (percentage)643 45 
White511 (79.5)32 (71.1)
Black104 (16.2)11 (24.4)
Other28 (4.4)2 (4.4)
Ethnicity, N (percentage)639 45 
Hispanic24 (3.8)1 (2.2)
Not Hispanic615 (96.2)44 (97.8)
Marital status, N (percentage)646 45 
Married/cohabitate382 (59.1)20 (44.4)
Separated/divorced118 (18.3)11 (24.4)
Widowed81 (12.5)5 (11.1)
Never married65 (10.1)9 (2.0)
Insurance type, N (percentage)646 45 
Medicaid53 (8.2)5 (11.1)
Medicare270 (41.8)13 (28.9)
Private289 (44.7)19 (42.2)
Self‐pay34 (5.3)8 (17.8)
Years of education, mean in yr (SD)64314.0 (3.1)4513.3 (2.7)
Presence of PCP prior to admission, N (percentage)646 45 
Yes596 (92.3)38 (84.4)
No50 (7.74)7 (15.6)
Site, N (percentage)646 45 
Site 1358 (55.4)8 (17.8)
Site 2288 (44.6)37 (82.2)
No. of preadmission medications, mean no. (SD)6417.8 (4.8)457.7 (5.4)
Medication understanding score, mean (SD)*5972.4 (0.5)402.2 (0.62)
Health literacy (s‐TOFHLA) score, mean (SD)64229.1 (8.9)4526.0 (12.0)
Baseline adherence (SD)6132.7 (1.1)452.4 (1.2)
MiniCog score, N (percentage)646 45 
Demented63 (9.8)5 (11.1)
Not demented583 (90.2)40 (88.9)

The average postdischarge adherence score was 95% (standard deviation [SD] = 10.2%), and less than 10% of patients had an adherence score of less than 85%; overall the distribution was left‐skewed. Table 2 illustrates crude and adjusted parameter estimates for variables in the model. Table 3 shows significant findings in the fully adjusted model, which used multiple imputation techniques to account for missing data.

Crude and Adjusted Measurements
PredictorCrude Parameter Estimate (Beta) With 95% Confidence IntervalsP ValueAdjusted Parameter Estimate (Beta) With 95% Confidence IntervalsP Value
  • NOTE: For crude estimates, value is category vs absence of parameter in univariable testing. For adjusted estimates of categorical variables, value is each category compared to referent category. Beta‐coefficient represents absolute change in adherence (eg, 0.010 for age means a 1% absolute increase in adherence for every 10 yr increase in patient age). Abbreviations: PCP, primary care physician; Ref, referent; s‐TOFHLA, short form of the Test of Functional Health Literacy in Adults.

Age per 10 yr0.010 (0.007, 0.020)<0.00010.010 (0.002, 0.020)0.018
Male gender0.012 (0.004, 0.028)0.1370.003 (0.014, 0.020)0.727
Race/ethnicity    
White0.011 (0.009, 0.031)0.266RefRef
Black0.017 (0.038, 0.005)0.130.006 (0.017, 0.030)0.598
Other0.010 (0.029, 0.049)0.5990.017 (0.027, 0.062)0.446
Hispanic/Latino0.005 (0.037, 0.047)0.8030.036 (0.013, 0.085)0.149
Marital status    
Married/cohabitate0.006 (0.011, 0.022)0.500RefRef
Separated/divorced0.005 (0.025, 0.016)0.6640.009 (0.014, 0.031)0.446
Widowed0.001 (0.023, 0.025)0.9220.013 (0.039, 0.013)0.338
Never married0.009 (0.035, 0.018)0.5150.004 (0.033, 0.025)0.784
Insurance type    
Private0.008 (0.008, 0.024)0.347RefRef
Medicaid0.046 (0.075, 0.018)0.0020.026 (0.058, 0.007)0.121
Medicare0.012 (0.004, 0.028)0.1380.002 (0.023, 0.018)0.844
Self‐pay0.027 (0.062, 0.008)0.1350.029 (0.073, 0.015)0.202
Years of education0.003 (0.0003, 0.005)0.0280.0001 (0.003, 0.003)0.949
Presence of PCP prior to admission0.007 (0.022, 0.037)0.6300.002 (0.032, 0.036)0.888
Site0.050 (0.065, 0.034)<0.00010.038 (0.056, 0.021)<0.0001
No. of preadmission medications0.0003 (0.002, 0.001)0.6840.0001 (0.002, 0.002)0.918
Medication understanding score per point0.007 (0.009, 0.023)0.3900.006 (0.011, 0.023)0.513
Health literacy (s‐TOFHLA) score per 10 points0.0006 (0.008, 0.01)0.8970.003 (0.008, 0.01)0.644
Baseline adherence per point0.023 (0.016, 0.031)<0.00010.017 (0.009, 0.024)<0.0001
Cognitive function0.004 (0.022, 0.031)0.7570.008 (0.019, 0.036)0.549
Significant Results in Adjusted Analyses With Multiple Imputation
PredictorParameter Estimate (Beta) With 95% Confidence IntervalsP Value
  • NOTE: Total observations, 646; 67 with missing values. All variables adjusted for gender, race, cognitive function, number of preadmission medications, marital status, health literacy score, medication understanding score, presence of primary care physician (PCP), years of school, Hispanic/Latino ethnicity. Abbreviations: Ref, referent.

Age per 10 yr0.010 (0.004, 0.020)0.004
Insurance type  
PrivateRefRef
Medicaid0.045 (0.076, 0.014)0.005
Medicare0.010 (0.030, 0.010)0.333
Self‐pay0.013 (0.050, 0.025)0.512
Site0.036 (0.053, 0.019)<0.0001
Baseline adherence per point0.016 (0.008, 0.024)<0.0001

Intervention arm was of borderline statistical significance in predicting postdischarge adherence (P = 0.052), and so was removed from the final model. Study site, age, insurance, and baseline adherence were the only significant independent predictors of postdischarge adherence in the fully adjusted model (Table 3). For example, for every 10‐year increase in age, patients had, on average, an adjusted 1% absolute increase in their adherence score (95% confidence interval [CI] 0.4% to 2.0%). For every 1‐point increase in baseline medication adherence (based on the Morisky scale), there was a 1.6% absolute increase in medication adherence (95% CI 0.8% to 2.4%). In unadjusted analyses, patients with Medicaid were less adherent with medications after discharge than were patients with private insurance. This difference became nonsignificant in adjusted analyses, but when analyses were repeated using multiple imputation techniques, the results again became statistically significantMedicaid insurance was associated with a 4.5% absolute decrease in postdischarge adherence compared with private insurance (95% CI 7.6% to 1.4%). Study site (specifically, Brigham and Women's Hospital) was also a significant predictor of greater postdischarge medication adherence. Years of education was a significant predictor of adherence in unadjusted analyses, but was not an independent predictor when adjusted for other factors. When baseline adherence was removed from the multiple imputation model, there were no changes in which factors were significant predictors of adherence.

DISCUSSION

In this study, we found that low baseline adherence, younger age, Medicaid insurance, and study site were significant predictors of lower 30‐day medication adherence. Of particular interest is our finding regarding baseline adherence, a simple measure to obtain on hospitalized patients. It is notable that in our study, education was not an independent significant predictor of postdischarge adherence, even when baseline adherence was removed from the model. The same is true for medication understanding, cognitive function, and health literacy.

Older patients appeared more adherent with medications in the month after hospital discharge, perhaps reflecting increased interaction with the healthcare system (appointments, number of physician interactions), a greater belief in the importance of chronic medication management, or a higher level of experience with managing medications. A similar relationship between age and adherence has been shown in outpatient studies of patients with hypertension, diabetes, and other chronic diseases.2427

Medicaid patients may be less likely to remain adherent because of the plan's limited coverage of medications relative to patients' ability to pay. For example, Medicaid in Tennessee covers the first 5 generic medications at no cost to the patient but has co‐payments for additional medications and for brand name drugs. Medicaid in Massachusetts has co‐payments of $1 to $3 for each medication. Alternatively, Medicaid insurance may be a marker for other patient characteristics associated with low adherence for which we were not fully able to adjust.

Site differences were also notable in this study; these differences could have been due to differences in insurance coverage in Tennessee versus Massachusetts (which has near‐universal coverage), differences in types of insurance (eg, fewer patients at Brigham and Women's Hospital had Medicaid than at Vanderbilt), cultural and geographic differences between the 2 locations, or other differences in transitional care between the 2 sites.

This study corroborates previous literature on medication adherence (specifically unintentional nonadherence) in the outpatient setting,4, 811 for example, on the association of younger age with low adherence in certain populations. On the other hand, it may contrast with previous literature which has sometimes shown a relationship between patient education or health literacy and medication adherence.14, 15, 2835 However, previous studies have not focused on the transition from inpatient to outpatient settings. Perhaps intensive medication education in the hospital, even under usual care, mitigates the effects of these factors on postdischarge adherence. Finally, baseline adherence seems to correlate with postdischarge adherence, a finding which makes intuitive sense and has been previously reported for specific medications.36

There are several limitations to this study. Although large, the study was performed at only 2 clinical sites where most patients were white and fairly well‐educated, perhaps because patients admitted to a tertiary care center with ACS or ADHF are more affluent than general medical inpatients as a whole; this may limit generalizability. Postdischarge medication adherence might have been higher than in other patient populations given the nature of the population, possible loss‐to‐follow‐up bias, and the fact that half of the subjects received an intervention designed to improve medication management after discharge; such low rates of nonadherence in our study may have reduced our ability to detect important predictors in our models. In addition, the period of follow‐up was 30 days, thus limiting our findings to short‐term postdischarge medication adherence. Postdischarge medication adherence was based on patient self‐report, which not only assumed that the patient was still managing his/her own medications after discharge, but may also be susceptible to both recall and social acceptability bias, which might overestimate our adherence scores, again limiting our ability to detect important predictors of nonadherence. However, other studies have shown a good correlation between self‐reported medication adherence and other more objective measures,37, 38 and recall was only for 7 days, a measure used previously in the literature39, 40 and one designed to reduce recall bias. Systematic underreporting in certain patient populations is less likely but possible.

In the future, research should focus on targeting patients who have low baseline adherence to evaluate the effects of various interventions on postdischarge medication outcomes. Repeating the study in a population with a high prevalence of low health literacy might be illuminating, given that previous studies have shown that patients with low health literacy have less ability to identify their medications and have less refill adherence.29, 30

In conclusion, in patients hospitalized with cardiovascular disease, predictors of lower postdischarge adherence include younger age, Medicaid insurance, and low baseline adherence. It may be prudent to assess baseline adherence and insurance type in hospitalized patients in order to identify those who may benefit from additional assistance to improve medication adherence and medication safety during transitions in care.

Acknowledgements

Meeting Presentations: SGIM New England Regional Meeting, oral presentation, Boston, MA, March 4, 2011; and SGIM National Meeting, poster presentation, Phoenix, AZ, May 6, 2011. Dr Schnipper had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Disclosures: Financial support was provided by R01 HL089755 (NHLBI, Kripalani), K23 HL077597 (NHLBI, Kripalani), K08 HL072806 (NHLBI, Schnipper), T32HP10251‐02 (Cohen), and by the Division of General Medicine, Massachusetts General Hospital and the Harvard Medical School Fellowship in General Medicine and Primary Care (Cohen). Dr Kripalani is a consultant to and holds equity in PictureRx, LLC, which makes patient education tools to improve medication management. PictureRx did not provide materials or funding for this study. All other authors disclose no relevant or financial conflicts of interest.

In the outpatient setting, medication adherence (defined as percentage of prescribed medication doses taken by a patient during a specific time period) ranges between 40% and 80% for chronic conditions.1 During acute care hospitalization, changes are often made to patients' medication regimens, which can be confusing and contribute to nonadherence, medication errors, and harmful adverse events.2 Indeed, it is estimated that almost half of patients encounter a medication error after discharge, and approximately 12%17% experience an adverse drug event after returning home.36 It is likely that some of these adverse events may be the result of medication nonadherence.7 Improved patientprovider communication, systems to reconcile prehospitalization and posthospitalization medications, as well as development of mechanisms to enhance adherence, may prevent many of these errors and have become new targets for quality improvement.4, 8 Although postdischarge medication adherence is a crucial target for avoiding adverse events and rehospitalization, few studies have focused on understanding its incidence and predictors, in particular, patient demographic factors such as age and insurance status.911

In addition, few studies have looked at general and posthospital adherence in a population where health literacy is measured, an important area because medication changes during hospitalization may be particularly confusing for patients with low health literacy.11, 12 Health literacy is defined as the degree to which an individual has the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions.13 Prior outpatient research shows that low health literacy is associated with poor patient understanding of the medication regimen and instructions for medication use, which may contribute to postdischarge medication nonadherence.14, 15 Understanding the factors associated with postdischarge medication adherence could help refine interventions that are oriented toward improving transitions in care, patient safety, and reducing unnecessary rehospitalization.

We report here on factors associated with postdischarge medication adherence using data from the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL‐CVD) study.16

METHODS

Study and Participants

PILL‐CVD was a federally funded, 2‐site randomized controlled trial using pharmacist‐assisted medication reconciliation, inpatient pharmacist counseling, low‐literacy adherence aids, and telephone follow‐up that aimed to decrease rates of serious medication errors after hospital discharge.16 The study targeted patients with cardiovascular disease (hospitalized on cardiology or general medical or geriatric services for acute coronary syndromes [ACS] or acute decompensated heart failure [ADHF]) at 2 large academic hospitals, Brigham and Women's Hospital (BWH) and Vanderbilt University Hospital (VUH).

Subjects were eligible for enrollment if they met criteria for ACS or ADHF, were likely to be discharged to home as determined by the primary medical team at the time of study enrollment, and took primary responsibility for administering their medications prior to admission (caregivers could be involved in medication management after discharge). Exclusion criteria included severe visual or hearing impairment, inability to communicate in English or Spanish, active psychiatric illness, dementia, delirium, illness too severe to participate, lack of a home phone number, being in police custody, or participation in another intensive medication adherence program (eg, due to renal transplant).

Out of 6416 patients originally screened for possible enrollment, 862 were randomly assigned to receive usual care or usual care plus the intervention, and 851 remained in the study.16 Both the main study and this secondary data analysis were approved by the Institutional Review Boards of each site.

Baseline Measures

Following informed consent and study enrollment, a variety of baseline data were collected on study participants from medical records and patient interview, including primary language, demographic information (age, race, insurance status, income, and education level), cognition (through administration of the 05‐point MiniCog scale),17 and level of health literacy (through use of the 036‐point short form of the Test of Functional Health Literacy in Adults [s‐TOFHLA] scale).18 Baseline information was also collected on medication use, including number of preadmission medications, measurement of self‐reported adherence prior to admission (using the Morisky scale, a validated 04‐point questionnaire shown to correlate with disease control and indicative of general patterns of adherence),19 and a medication understanding score, adapted from other instruments, which quantifies understanding of the indication, dose, and frequency of up to 5 randomly selected preadmission medications on a 03‐point scale.16, 20, 21

Outcome Measures

Outcomes were collected 30 days postdischarge through a structured questionnaire, administered by telephone. Only patients who completed this call are included in the present analysis. Postdischarge medication adherence was assessed by asking patients to report the number of days out of the previous week they had taken each medication from their postdischarge regimen exactly as prescribed.22 A score was calculated for each medication as the proportion of adherent days (eg, if a patient reported missing 2 days of a medication in the previous week, then adherence would be 5/7 or 71%). A global postdischarge adherence score was then derived for each patient by averaging the adherence score for all regularly scheduled medications. This quantitative measure focused on adherence to medications patients knew they should be taking and did not measure medication discrepancies (sometimes termed unintentional nonadherence).

Analysis

Patient characteristics were summarized and reported using simple descriptive statistics. Candidate predictors of postdischarge medication adherence were chosen a priori from patient characteristics assessed during hospital admission. These included patient age, gender, race, ethnicity, marital status, insurance, years of education, presence of primary care physician (PCP), study site, number of preadmission medications, medication understanding, baseline adherence, cognition, and health literacy. Unadjusted results were calculated using univariable linear regression, with each patient's adherence score as the dependent variable and each predictor as the independent variable. Adjusted results were then derived using multivariable linear regression with all the candidate predictors in the model.

Lastly, because of missing data for some predictors, in particular baseline adherence and medication understanding, multiple imputation techniques were used to impute missing data and increase statistical power.23 We used the Markov Chain Monte Carlo (MCMC) method for multiple imputation, which generally assumes that the data came from a normal distribution and that the missing data are missing at random. Because of the essentially normal distribution of the data, and because the amount of missing data was so small (<1% for almost all variables, 5% for baseline adherence, and 8% for medication understanding), we expected little bias and present the complete case analysis, which maximized statistical power.

Two‐sided P values <0.05 were considered significant, and SAS version 9.2 (Cary, NC) was used for all analyses.

RESULTS

Table 1 shows descriptive baseline patient characteristics of study sample (responders) as well as nonresponders at 30 days. For the responders, the mean age of the 646 patients was 61.2 years, 94.7% were insured, and 19.3% had inadequate or marginal health literacy. Patients were prescribed an average of 8 preadmission medications. Most patients (92.3%) had a regular PCP prior to admission. Nonresponders had nonsignificant trends towards having lower health literacy, medication understanding, and baseline medication adherence.

Baseline Characteristics
CharacteristicTotal N, 30‐Day RespondentsValueTotal N, NonrespondentsValue
  • Abbreviations: PCP, primary care physician; SD, standard deviation; s‐TOFHLA, short form of the Test of Functional Health Literacy in Adults. *03, with 3 indicating better understanding. 036, with higher scores indicating higher health literacy. 04, with 4 indicating higher baseline adherence. 05, with higher scores indicating better cognition; a score <3 indicates dementia.

Age, mean in yr (SD)64661.2 (13.5)4555.4 (14.3)
Gender, N (percentage)646 45 
Female272 (42.1)18 (40.0)
Male374 (57.9)27 (60.0)
Race, N (percentage)643 45 
White511 (79.5)32 (71.1)
Black104 (16.2)11 (24.4)
Other28 (4.4)2 (4.4)
Ethnicity, N (percentage)639 45 
Hispanic24 (3.8)1 (2.2)
Not Hispanic615 (96.2)44 (97.8)
Marital status, N (percentage)646 45 
Married/cohabitate382 (59.1)20 (44.4)
Separated/divorced118 (18.3)11 (24.4)
Widowed81 (12.5)5 (11.1)
Never married65 (10.1)9 (2.0)
Insurance type, N (percentage)646 45 
Medicaid53 (8.2)5 (11.1)
Medicare270 (41.8)13 (28.9)
Private289 (44.7)19 (42.2)
Self‐pay34 (5.3)8 (17.8)
Years of education, mean in yr (SD)64314.0 (3.1)4513.3 (2.7)
Presence of PCP prior to admission, N (percentage)646 45 
Yes596 (92.3)38 (84.4)
No50 (7.74)7 (15.6)
Site, N (percentage)646 45 
Site 1358 (55.4)8 (17.8)
Site 2288 (44.6)37 (82.2)
No. of preadmission medications, mean no. (SD)6417.8 (4.8)457.7 (5.4)
Medication understanding score, mean (SD)*5972.4 (0.5)402.2 (0.62)
Health literacy (s‐TOFHLA) score, mean (SD)64229.1 (8.9)4526.0 (12.0)
Baseline adherence (SD)6132.7 (1.1)452.4 (1.2)
MiniCog score, N (percentage)646 45 
Demented63 (9.8)5 (11.1)
Not demented583 (90.2)40 (88.9)

The average postdischarge adherence score was 95% (standard deviation [SD] = 10.2%), and less than 10% of patients had an adherence score of less than 85%; overall the distribution was left‐skewed. Table 2 illustrates crude and adjusted parameter estimates for variables in the model. Table 3 shows significant findings in the fully adjusted model, which used multiple imputation techniques to account for missing data.

Crude and Adjusted Measurements
PredictorCrude Parameter Estimate (Beta) With 95% Confidence IntervalsP ValueAdjusted Parameter Estimate (Beta) With 95% Confidence IntervalsP Value
  • NOTE: For crude estimates, value is category vs absence of parameter in univariable testing. For adjusted estimates of categorical variables, value is each category compared to referent category. Beta‐coefficient represents absolute change in adherence (eg, 0.010 for age means a 1% absolute increase in adherence for every 10 yr increase in patient age). Abbreviations: PCP, primary care physician; Ref, referent; s‐TOFHLA, short form of the Test of Functional Health Literacy in Adults.

Age per 10 yr0.010 (0.007, 0.020)<0.00010.010 (0.002, 0.020)0.018
Male gender0.012 (0.004, 0.028)0.1370.003 (0.014, 0.020)0.727
Race/ethnicity    
White0.011 (0.009, 0.031)0.266RefRef
Black0.017 (0.038, 0.005)0.130.006 (0.017, 0.030)0.598
Other0.010 (0.029, 0.049)0.5990.017 (0.027, 0.062)0.446
Hispanic/Latino0.005 (0.037, 0.047)0.8030.036 (0.013, 0.085)0.149
Marital status    
Married/cohabitate0.006 (0.011, 0.022)0.500RefRef
Separated/divorced0.005 (0.025, 0.016)0.6640.009 (0.014, 0.031)0.446
Widowed0.001 (0.023, 0.025)0.9220.013 (0.039, 0.013)0.338
Never married0.009 (0.035, 0.018)0.5150.004 (0.033, 0.025)0.784
Insurance type    
Private0.008 (0.008, 0.024)0.347RefRef
Medicaid0.046 (0.075, 0.018)0.0020.026 (0.058, 0.007)0.121
Medicare0.012 (0.004, 0.028)0.1380.002 (0.023, 0.018)0.844
Self‐pay0.027 (0.062, 0.008)0.1350.029 (0.073, 0.015)0.202
Years of education0.003 (0.0003, 0.005)0.0280.0001 (0.003, 0.003)0.949
Presence of PCP prior to admission0.007 (0.022, 0.037)0.6300.002 (0.032, 0.036)0.888
Site0.050 (0.065, 0.034)<0.00010.038 (0.056, 0.021)<0.0001
No. of preadmission medications0.0003 (0.002, 0.001)0.6840.0001 (0.002, 0.002)0.918
Medication understanding score per point0.007 (0.009, 0.023)0.3900.006 (0.011, 0.023)0.513
Health literacy (s‐TOFHLA) score per 10 points0.0006 (0.008, 0.01)0.8970.003 (0.008, 0.01)0.644
Baseline adherence per point0.023 (0.016, 0.031)<0.00010.017 (0.009, 0.024)<0.0001
Cognitive function0.004 (0.022, 0.031)0.7570.008 (0.019, 0.036)0.549
Significant Results in Adjusted Analyses With Multiple Imputation
PredictorParameter Estimate (Beta) With 95% Confidence IntervalsP Value
  • NOTE: Total observations, 646; 67 with missing values. All variables adjusted for gender, race, cognitive function, number of preadmission medications, marital status, health literacy score, medication understanding score, presence of primary care physician (PCP), years of school, Hispanic/Latino ethnicity. Abbreviations: Ref, referent.

Age per 10 yr0.010 (0.004, 0.020)0.004
Insurance type  
PrivateRefRef
Medicaid0.045 (0.076, 0.014)0.005
Medicare0.010 (0.030, 0.010)0.333
Self‐pay0.013 (0.050, 0.025)0.512
Site0.036 (0.053, 0.019)<0.0001
Baseline adherence per point0.016 (0.008, 0.024)<0.0001

Intervention arm was of borderline statistical significance in predicting postdischarge adherence (P = 0.052), and so was removed from the final model. Study site, age, insurance, and baseline adherence were the only significant independent predictors of postdischarge adherence in the fully adjusted model (Table 3). For example, for every 10‐year increase in age, patients had, on average, an adjusted 1% absolute increase in their adherence score (95% confidence interval [CI] 0.4% to 2.0%). For every 1‐point increase in baseline medication adherence (based on the Morisky scale), there was a 1.6% absolute increase in medication adherence (95% CI 0.8% to 2.4%). In unadjusted analyses, patients with Medicaid were less adherent with medications after discharge than were patients with private insurance. This difference became nonsignificant in adjusted analyses, but when analyses were repeated using multiple imputation techniques, the results again became statistically significantMedicaid insurance was associated with a 4.5% absolute decrease in postdischarge adherence compared with private insurance (95% CI 7.6% to 1.4%). Study site (specifically, Brigham and Women's Hospital) was also a significant predictor of greater postdischarge medication adherence. Years of education was a significant predictor of adherence in unadjusted analyses, but was not an independent predictor when adjusted for other factors. When baseline adherence was removed from the multiple imputation model, there were no changes in which factors were significant predictors of adherence.

DISCUSSION

In this study, we found that low baseline adherence, younger age, Medicaid insurance, and study site were significant predictors of lower 30‐day medication adherence. Of particular interest is our finding regarding baseline adherence, a simple measure to obtain on hospitalized patients. It is notable that in our study, education was not an independent significant predictor of postdischarge adherence, even when baseline adherence was removed from the model. The same is true for medication understanding, cognitive function, and health literacy.

Older patients appeared more adherent with medications in the month after hospital discharge, perhaps reflecting increased interaction with the healthcare system (appointments, number of physician interactions), a greater belief in the importance of chronic medication management, or a higher level of experience with managing medications. A similar relationship between age and adherence has been shown in outpatient studies of patients with hypertension, diabetes, and other chronic diseases.2427

Medicaid patients may be less likely to remain adherent because of the plan's limited coverage of medications relative to patients' ability to pay. For example, Medicaid in Tennessee covers the first 5 generic medications at no cost to the patient but has co‐payments for additional medications and for brand name drugs. Medicaid in Massachusetts has co‐payments of $1 to $3 for each medication. Alternatively, Medicaid insurance may be a marker for other patient characteristics associated with low adherence for which we were not fully able to adjust.

Site differences were also notable in this study; these differences could have been due to differences in insurance coverage in Tennessee versus Massachusetts (which has near‐universal coverage), differences in types of insurance (eg, fewer patients at Brigham and Women's Hospital had Medicaid than at Vanderbilt), cultural and geographic differences between the 2 locations, or other differences in transitional care between the 2 sites.

This study corroborates previous literature on medication adherence (specifically unintentional nonadherence) in the outpatient setting,4, 811 for example, on the association of younger age with low adherence in certain populations. On the other hand, it may contrast with previous literature which has sometimes shown a relationship between patient education or health literacy and medication adherence.14, 15, 2835 However, previous studies have not focused on the transition from inpatient to outpatient settings. Perhaps intensive medication education in the hospital, even under usual care, mitigates the effects of these factors on postdischarge adherence. Finally, baseline adherence seems to correlate with postdischarge adherence, a finding which makes intuitive sense and has been previously reported for specific medications.36

There are several limitations to this study. Although large, the study was performed at only 2 clinical sites where most patients were white and fairly well‐educated, perhaps because patients admitted to a tertiary care center with ACS or ADHF are more affluent than general medical inpatients as a whole; this may limit generalizability. Postdischarge medication adherence might have been higher than in other patient populations given the nature of the population, possible loss‐to‐follow‐up bias, and the fact that half of the subjects received an intervention designed to improve medication management after discharge; such low rates of nonadherence in our study may have reduced our ability to detect important predictors in our models. In addition, the period of follow‐up was 30 days, thus limiting our findings to short‐term postdischarge medication adherence. Postdischarge medication adherence was based on patient self‐report, which not only assumed that the patient was still managing his/her own medications after discharge, but may also be susceptible to both recall and social acceptability bias, which might overestimate our adherence scores, again limiting our ability to detect important predictors of nonadherence. However, other studies have shown a good correlation between self‐reported medication adherence and other more objective measures,37, 38 and recall was only for 7 days, a measure used previously in the literature39, 40 and one designed to reduce recall bias. Systematic underreporting in certain patient populations is less likely but possible.

In the future, research should focus on targeting patients who have low baseline adherence to evaluate the effects of various interventions on postdischarge medication outcomes. Repeating the study in a population with a high prevalence of low health literacy might be illuminating, given that previous studies have shown that patients with low health literacy have less ability to identify their medications and have less refill adherence.29, 30

In conclusion, in patients hospitalized with cardiovascular disease, predictors of lower postdischarge adherence include younger age, Medicaid insurance, and low baseline adherence. It may be prudent to assess baseline adherence and insurance type in hospitalized patients in order to identify those who may benefit from additional assistance to improve medication adherence and medication safety during transitions in care.

Acknowledgements

Meeting Presentations: SGIM New England Regional Meeting, oral presentation, Boston, MA, March 4, 2011; and SGIM National Meeting, poster presentation, Phoenix, AZ, May 6, 2011. Dr Schnipper had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Disclosures: Financial support was provided by R01 HL089755 (NHLBI, Kripalani), K23 HL077597 (NHLBI, Kripalani), K08 HL072806 (NHLBI, Schnipper), T32HP10251‐02 (Cohen), and by the Division of General Medicine, Massachusetts General Hospital and the Harvard Medical School Fellowship in General Medicine and Primary Care (Cohen). Dr Kripalani is a consultant to and holds equity in PictureRx, LLC, which makes patient education tools to improve medication management. PictureRx did not provide materials or funding for this study. All other authors disclose no relevant or financial conflicts of interest.

References
  1. Osterberg L,Blaschke T.Adherence to medication.N Engl J Med.2005;353(5):487497.
  2. Coleman EA,Smith JD,Raha D,Min SJ.Posthospital medication discrepancies: prevalence and contributing factors.Arch Intern Med.2005;165(16):18421847.
  3. Cua YM,Kripalani S.Medication use in the transition from hospital to home.Ann Acad Med Singapore.2008;37(2):136141.
  4. Moore C,Wisnivesky J,Williams S,McGinn T.Medical errors related to discontinuity of care from an inpatient to an outpatient setting.J Gen Intern Med.2003;18(8):646651.
  5. Forster AJ,Murff HJ,Peterson JF,Gandhi TK,Bates DW.The incidence and severity of adverse events affecting patients after discharge from the hospital.Ann Intern Med.2003;138(3):161167.
  6. Forster AJ,Murff HJ,Peterson JF,Gandhi TK,Bates DW.Adverse drug events occurring following hospital discharge.J Gen Intern Med.2005;20(4):317323.
  7. Schnipper JL,Kirwin JL,Cotugno MC, et al.Role of pharmacist counseling in preventing adverse drug events after hospitalization.Arch Intern Med.2006;166(5):565571.
  8. Vira T,Colquhoun M,Etchells E.Reconcilable differences: correcting medication errors at hospital admission and discharge.Qual Saf Health Care.2006;15(2):122126.
  9. Hassan M,Lage MJ.Risk of rehospitalization among bipolar disorder patients who are nonadherent to antipsychotic therapy after hospital discharge.Am J Health Syst Pharm.2009;66(4):358365.
  10. Mansur N,Weiss A,Hoffman A,Gruenewald T,Beloosesky Y.Continuity and adherence to long‐term drug treatment by geriatric patients after hospital discharge: a prospective cohort study.Drugs Aging.2008;25(10):861870.
  11. Kripalani S,Henderson LE,Jacobson TA,Vaccarino V.Medication use among inner‐city patients after hospital discharge: patient‐reported barriers and solutions.Mayo Clin Proc.2008;83(5):529535.
  12. Lindquist LA,Go L,Fleisher J,Jain N,Friesema E,Baker DW.Relationship of health literacy to intentional and unintentional non‐adherence of hospital discharge medications.J Gen Intern Med.2012;27(2):173178.
  13. Office of Disease Prevention and Health Promotion, US Department of Health and Human Services.Healthy People 2010. Available at: http://www.healthypeople.gov/Document/pdf/uih/2010uih.pdf. Accessed February 15,2012.
  14. Davis TC,Wolf MS,Bass PF, et al.Literacy and misunderstanding prescription drug labels.Ann Intern Med.2006;145(12):887894.
  15. Kripalani S,Henderson LE,Chiu EY,Robertson R,Kolm P,Jacobson TA.Predictors of medication self‐management skill in a low‐literacy population.J Gen Intern Med.2006;21(8):852856.
  16. Schnipper JL,Roumie CL,Cawthorn C, et al;for the PILL‐CVD Study Group.Rationale and design of the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL‐CVD) study.Circ Cardiovasc Qual Outcomes.2010;3(2):212219.
  17. Borson S,Scanlan JM,Watanabe J,Tu SP,Lessig M.Simplifying detection of cognitive impairment: comparison of the Mini‐Cog and Mini‐Mental State Examination in a multiethnic sample.J Am Geriatr Soc.2005;53(5):871874.
  18. Nurss JR.Short Test of Functional Health Literacy in Adults.Snow Camp, NC:Peppercorn Books and Press;1998.
  19. Morisky DE,Ang A,Krousel‐Wood M,Ward HJ.Predictive validity of a medication adherence measure in an outpatient setting.J Clin Hypertens (Greenwich).2008;10(5):348354.
  20. Marvanova M,Roumie CL,Eden SK,Cawthon C,Schnipper JL,Kripalani S.Health literacy and medication understanding among hospitalized adults.J Hosp Med. In press.
  21. Marvanova M,Roumie CL,Eden SK,Cawthon C,Schnipper JL,Kripalani S.Health literacy and medication understanding among hospitalized adults.J Hosp Med.2011;6(9):488493.
  22. Toobert DJ,Hampson SE,Glasgow RE.The summary of diabetes self‐care activities measure: results from 7 studies and a revised scale.Diabetes Care.2000;23(7):943950.
  23. Rubin DB.Multiple Imputation for Nonresponse in Surveys.New York, NY:John Wiley 1987.
  24. Hinkin CH,Hardy DJ,Mason KI, et al.Medication adherence in HIV‐infected adults: effect of patient age, cognitive status, and substance abuse.AIDS.2004;18(suppl 1):S19S25.
  25. Wong MC,Jiang JY,Griffiths SM.Factors associated with antihypertensive drug compliance in 83,884 Chinese patients: a cohort study.J Epidemiol Community Health.2010;64(10):895901.
  26. Wong MC,Kong AP,So WY,Jiang JY,Chan JC,Griffiths SM.Adherence to oral hypoglycemic agents in 26,782 Chinese patients: a cohort study.J Clin Pharmacol.2011;51(10):14741482.
  27. Gazmararian J,Jacobson KL,Pan Y,Schmotzer B,Kripalani S.Effect of a pharmacy‐based health literacy intervention and patient characteristics on medication refill adherence in an urban health system.Ann Pharmacother.2010;44(1):8087.
  28. Kalichman SC,Ramachandran B,Catz S.Adherence to combination antiretroviral therapies in HIV patients of low health literacy.J Gen Intern Med.1999;14(5):267273.
  29. Gazmararian JA,Kripalani S,Miller MJ,Echt KV,Ren J,Rask K.Factors associated with medication refill adherence in cardiovascular‐related diseases: a focus on health literacy.J Gen Intern Med.2006;21(12):12151221.
  30. Persell SD,Osborn CY,Richard R,Skripkauskas S,Wolf MS.Limited health literacy is a barrier to medication reconciliation in ambulatory care.J Gen Intern Med.2007;22(11):15231526.
  31. Chew LD,Bradley KA,Flum DR,Cornia PB,Koepsell TD.The impact of low health literacy on surgical practice.Am J Surg.2004;188(3):250253.
  32. Gatti ME,Jacobson KL,Gazmararian JA,Schmotzer B,Kripalani S.Relationships between beliefs about medications and adherence.Am J Health Syst Pharm.2009;66(7):657664.
  33. Fang MC,Machtinger EL,Wang F,Schillinger D.Health literacy and anticoagulation‐related outcomes among patients taking warfarin.J Gen Intern Med.2006;21(8):841846.
  34. Paasche‐Orlow MK,Cheng DM,Palepu A,Meli S,Faber V,Samet JH.Health literacy, antiretroviral adherence, and HIV‐RNA suppression: a longitudinal perspective.J Gen Intern Med.2006;21(8):835840.
  35. Platt AB,Localio AR,Brensinger CM, et al.Risk factors for nonadherence to warfarin: results from the IN‐RANGE study.Pharmacoepidemiol Drug Saf.2008;17(9):853860.
  36. Muntner P,Mann DM,Woodward M, et al.Predictors of low clopidogrel adherence following percutaneous coronary intervention.Am J Cardiol.2011;108(6):822827.
  37. Shi L,Liu J,Fonseca V,Walker P,Kalsekar A,Pawaskar M.Correlation between adherence rates measured by MEMS and self‐reported questionnaires: a meta‐analysis.Health Qual Life Outcomes.2010;8:99.
  38. Shi L,Liu J,Koleva Y,Fonseca V,Kalsekar A,Pawaskar M.Concordance of adherence measurement using self‐reported adherence questionnaires and medication monitoring devices.Pharmacoeconomics.2010;28(12):10971107.
  39. Grant RW,Devita NG,Singer DE,Meigs JB.Polypharmacy and medication adherence in patients with type 2 diabetes.Diabetes Care.2003;26(5):14081412.
  40. Grant RW,Devita NG,Singer DE,Meigs JB.Improving adherence and reducing medication discrepancies in patients with diabetes.Ann Pharmacother.2003;37(7–8):962969.
References
  1. Osterberg L,Blaschke T.Adherence to medication.N Engl J Med.2005;353(5):487497.
  2. Coleman EA,Smith JD,Raha D,Min SJ.Posthospital medication discrepancies: prevalence and contributing factors.Arch Intern Med.2005;165(16):18421847.
  3. Cua YM,Kripalani S.Medication use in the transition from hospital to home.Ann Acad Med Singapore.2008;37(2):136141.
  4. Moore C,Wisnivesky J,Williams S,McGinn T.Medical errors related to discontinuity of care from an inpatient to an outpatient setting.J Gen Intern Med.2003;18(8):646651.
  5. Forster AJ,Murff HJ,Peterson JF,Gandhi TK,Bates DW.The incidence and severity of adverse events affecting patients after discharge from the hospital.Ann Intern Med.2003;138(3):161167.
  6. Forster AJ,Murff HJ,Peterson JF,Gandhi TK,Bates DW.Adverse drug events occurring following hospital discharge.J Gen Intern Med.2005;20(4):317323.
  7. Schnipper JL,Kirwin JL,Cotugno MC, et al.Role of pharmacist counseling in preventing adverse drug events after hospitalization.Arch Intern Med.2006;166(5):565571.
  8. Vira T,Colquhoun M,Etchells E.Reconcilable differences: correcting medication errors at hospital admission and discharge.Qual Saf Health Care.2006;15(2):122126.
  9. Hassan M,Lage MJ.Risk of rehospitalization among bipolar disorder patients who are nonadherent to antipsychotic therapy after hospital discharge.Am J Health Syst Pharm.2009;66(4):358365.
  10. Mansur N,Weiss A,Hoffman A,Gruenewald T,Beloosesky Y.Continuity and adherence to long‐term drug treatment by geriatric patients after hospital discharge: a prospective cohort study.Drugs Aging.2008;25(10):861870.
  11. Kripalani S,Henderson LE,Jacobson TA,Vaccarino V.Medication use among inner‐city patients after hospital discharge: patient‐reported barriers and solutions.Mayo Clin Proc.2008;83(5):529535.
  12. Lindquist LA,Go L,Fleisher J,Jain N,Friesema E,Baker DW.Relationship of health literacy to intentional and unintentional non‐adherence of hospital discharge medications.J Gen Intern Med.2012;27(2):173178.
  13. Office of Disease Prevention and Health Promotion, US Department of Health and Human Services.Healthy People 2010. Available at: http://www.healthypeople.gov/Document/pdf/uih/2010uih.pdf. Accessed February 15,2012.
  14. Davis TC,Wolf MS,Bass PF, et al.Literacy and misunderstanding prescription drug labels.Ann Intern Med.2006;145(12):887894.
  15. Kripalani S,Henderson LE,Chiu EY,Robertson R,Kolm P,Jacobson TA.Predictors of medication self‐management skill in a low‐literacy population.J Gen Intern Med.2006;21(8):852856.
  16. Schnipper JL,Roumie CL,Cawthorn C, et al;for the PILL‐CVD Study Group.Rationale and design of the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL‐CVD) study.Circ Cardiovasc Qual Outcomes.2010;3(2):212219.
  17. Borson S,Scanlan JM,Watanabe J,Tu SP,Lessig M.Simplifying detection of cognitive impairment: comparison of the Mini‐Cog and Mini‐Mental State Examination in a multiethnic sample.J Am Geriatr Soc.2005;53(5):871874.
  18. Nurss JR.Short Test of Functional Health Literacy in Adults.Snow Camp, NC:Peppercorn Books and Press;1998.
  19. Morisky DE,Ang A,Krousel‐Wood M,Ward HJ.Predictive validity of a medication adherence measure in an outpatient setting.J Clin Hypertens (Greenwich).2008;10(5):348354.
  20. Marvanova M,Roumie CL,Eden SK,Cawthon C,Schnipper JL,Kripalani S.Health literacy and medication understanding among hospitalized adults.J Hosp Med. In press.
  21. Marvanova M,Roumie CL,Eden SK,Cawthon C,Schnipper JL,Kripalani S.Health literacy and medication understanding among hospitalized adults.J Hosp Med.2011;6(9):488493.
  22. Toobert DJ,Hampson SE,Glasgow RE.The summary of diabetes self‐care activities measure: results from 7 studies and a revised scale.Diabetes Care.2000;23(7):943950.
  23. Rubin DB.Multiple Imputation for Nonresponse in Surveys.New York, NY:John Wiley 1987.
  24. Hinkin CH,Hardy DJ,Mason KI, et al.Medication adherence in HIV‐infected adults: effect of patient age, cognitive status, and substance abuse.AIDS.2004;18(suppl 1):S19S25.
  25. Wong MC,Jiang JY,Griffiths SM.Factors associated with antihypertensive drug compliance in 83,884 Chinese patients: a cohort study.J Epidemiol Community Health.2010;64(10):895901.
  26. Wong MC,Kong AP,So WY,Jiang JY,Chan JC,Griffiths SM.Adherence to oral hypoglycemic agents in 26,782 Chinese patients: a cohort study.J Clin Pharmacol.2011;51(10):14741482.
  27. Gazmararian J,Jacobson KL,Pan Y,Schmotzer B,Kripalani S.Effect of a pharmacy‐based health literacy intervention and patient characteristics on medication refill adherence in an urban health system.Ann Pharmacother.2010;44(1):8087.
  28. Kalichman SC,Ramachandran B,Catz S.Adherence to combination antiretroviral therapies in HIV patients of low health literacy.J Gen Intern Med.1999;14(5):267273.
  29. Gazmararian JA,Kripalani S,Miller MJ,Echt KV,Ren J,Rask K.Factors associated with medication refill adherence in cardiovascular‐related diseases: a focus on health literacy.J Gen Intern Med.2006;21(12):12151221.
  30. Persell SD,Osborn CY,Richard R,Skripkauskas S,Wolf MS.Limited health literacy is a barrier to medication reconciliation in ambulatory care.J Gen Intern Med.2007;22(11):15231526.
  31. Chew LD,Bradley KA,Flum DR,Cornia PB,Koepsell TD.The impact of low health literacy on surgical practice.Am J Surg.2004;188(3):250253.
  32. Gatti ME,Jacobson KL,Gazmararian JA,Schmotzer B,Kripalani S.Relationships between beliefs about medications and adherence.Am J Health Syst Pharm.2009;66(7):657664.
  33. Fang MC,Machtinger EL,Wang F,Schillinger D.Health literacy and anticoagulation‐related outcomes among patients taking warfarin.J Gen Intern Med.2006;21(8):841846.
  34. Paasche‐Orlow MK,Cheng DM,Palepu A,Meli S,Faber V,Samet JH.Health literacy, antiretroviral adherence, and HIV‐RNA suppression: a longitudinal perspective.J Gen Intern Med.2006;21(8):835840.
  35. Platt AB,Localio AR,Brensinger CM, et al.Risk factors for nonadherence to warfarin: results from the IN‐RANGE study.Pharmacoepidemiol Drug Saf.2008;17(9):853860.
  36. Muntner P,Mann DM,Woodward M, et al.Predictors of low clopidogrel adherence following percutaneous coronary intervention.Am J Cardiol.2011;108(6):822827.
  37. Shi L,Liu J,Fonseca V,Walker P,Kalsekar A,Pawaskar M.Correlation between adherence rates measured by MEMS and self‐reported questionnaires: a meta‐analysis.Health Qual Life Outcomes.2010;8:99.
  38. Shi L,Liu J,Koleva Y,Fonseca V,Kalsekar A,Pawaskar M.Concordance of adherence measurement using self‐reported adherence questionnaires and medication monitoring devices.Pharmacoeconomics.2010;28(12):10971107.
  39. Grant RW,Devita NG,Singer DE,Meigs JB.Polypharmacy and medication adherence in patients with type 2 diabetes.Diabetes Care.2003;26(5):14081412.
  40. Grant RW,Devita NG,Singer DE,Meigs JB.Improving adherence and reducing medication discrepancies in patients with diabetes.Ann Pharmacother.2003;37(7–8):962969.
Issue
Journal of Hospital Medicine - 7(6)
Issue
Journal of Hospital Medicine - 7(6)
Page Number
470-475
Page Number
470-475
Publications
Publications
Article Type
Display Headline
Predictors of medication adherence postdischarge: The impact of patient age, insurance status, and prior adherence
Display Headline
Predictors of medication adherence postdischarge: The impact of patient age, insurance status, and prior adherence
Sections
Article Source

Copyright © 2012 Society of Hospital Medicine

Disallow All Ads
Correspondence Location
General Medicine Unit, Massachusetts General Hospital, 50 Staniford St, 9th Floor, Boston, MA 02114
Content Gating
No Gating (article Unlocked/Free)
Alternative CME
Article PDF Media
Media Files

Post‐Discharge Inpatients With Depressive Symptoms

Article Type
Changed
Sun, 05/28/2017 - 20:14
Display Headline
Post‐discharge hospital utilization among adult medical inpatients with depressive symptoms

Fully 19% of Medicare patients are readmitted to the hospital within 30 days of discharge.1 This represents a large amount of potentially avoidable morbidity and cost. Indeed, projects to improve the discharge process and post‐hospital care have shown that as much as one‐third of hospital utilization in the month after discharge can be avoided.2 Consequently, the rate of early, unplanned hospital utilization after discharge has emerged as an important indicator of hospital quality and the Centers for Medicare and Medicaid Services (CMS) has proposed a policy to decrease payments to hospitals with high rates of early unplanned hospital utilization. Thus, there is great interest in identifying modifiable risk factors for rehospitalization that could be used to refine intervention models and lead to improvements in quality of care, patient outcomes, and cost savings.

To date, known predictors of readmission include: lower socioeconomic status,3 history of prior hospitalization4 and advanced age,5 length of stay greater than 7 days,6 a high burden of comorbid illnesses (based on Charlson score),7 poor social support,8 and specific diagnoses (eg, congestive heart failure, chronic obstructive pulmonary disease [COPD] and myocardial infarction).5, 9, 10 In addition, unplanned readmissions and emergency department (ED) visits have been linked to polypharmacy and adverse drug events related to treatment with medications such as warfarin, digoxin and narcotics.11, 12 Another characteristic that has also been linked to readmission is depression;13 however5 reports supporting this association are from studies of elderly patients or with patients who have specific diagnoses (eg, congestive heart failure [CHF], COPD, myocardial infarction).1416

Depression is common, affecting 13% to 16% of people in the US, and is recognized as an important risk factor for poor outcomes among patients with various chronic illnesses.1719 The mechanisms by which depression can be linked to health outcomes and health service utilization have been studied in age‐specific or disease‐specific cohorts such as cardiac patients or frail elders and include both physiologic factors such as hypercoagulability and hyperinflammatory conditions, as well as behavioral factors such as poor self‐care behaviors and heightened sensitivity to somatic symptoms. How these mechanisms link depression to health outcomes and hospital utilization in a general medical population is not clearly understood. Kartha et al.13 reported findings indicating that depression is a risk factor for rehospitalization in general medical inpatients, but the study sample was relatively small and the study design methodology significantly limited its generalizability.12 It would be useful to provide supporting evidence showing depression as an important risk factor for readmission in the general medical in‐patient population using more rigorous study methods and a larger cohort.

We hypothesized that depressive symptoms would be an independent risk factor for early unplanned hospital utilization after discharge for all medical patients. Therefore, we conducted a secondary analysis of the Project RED clinical trial dataset to assess the association between a positive depression screen during inpatient hospitalization and the rate of subsequent hospital utilization.

Methods

Data from the Project RED clinical trial were reviewed for inclusion in a secondary analysis. Complete data were available for 738 of the 749 subjects recruited for Project RED.

Project RED Setting and Participants

Project RED was a two‐armed randomized controlled trial of English‐speaking adult patients, 18 years or older, admitted to the teaching service of Boston Medical Center, a large urban safety‐net hospital with an ethnically diverse patient population. A total of 749 subjects were enrolled and randomized between January 3, 2006 and October 18, 2007. Patients were required to have a telephone, be able to comprehend study details and the consent process in English, and have plans to be discharged to a US community. Patients were not enrolled if they were admitted from a skilled nursing facility or other hospital, transferred to a different hospital service prior to enrollment, admitted for a planned hospitalization, on hospital precautions, on suicide watch, deaf or blind. The Institutional Review Board of Boston University approved all study activities. A full description of the methods for the Project RED trial has been described previously.2

Outcome Variable

The primary endpoint was rate of hospital utilization within 30 days of discharge from the index admission, defined as the total number of ED visits and readmissions per subject within 30 days of the index discharge. Hospital utilization rates within 60 and 90 days of the index hospitalization discharge were also analyzed as secondary outcomes. Any ED visit in which a subject was subsequently admitted to the hospital was only counted as a readmission. Outcome data were collected by reviewing the hospital's electronic medical records (EMRs) and by contacting subjects by telephone 30 days after discharge. Dates of hospital utilization occurring at Boston Medical Center were obtained from the EMR, while those at other hospitals were collected through subject report. Subjects who could not be reached within 60 days of discharge were assumed alive.

Primary Independent Variable

The primary independent variable of interest was depressive symptoms defined as a positive score for minor or major depression on the nine‐item Patient Health Questionnaire (PHQ‐9) depression screening tool.20 A dichotomized variable was created using a standardized scoring system to determine the screening cut‐off for major or minor depressive symptoms.19

Statistical Analysis

Demographic and other characteristics of the subjects were compared by depression status (Table 1). Potential confounders were identified a priori from the available literature on factors associated with rehospitalization. These included age, gender, marital status, health literacy score (rapid estimate of health literacy in adult medicine tool [REALM]),21 Charlson score,22 insurance type, employment status, income level, homelessness status within past three months, hospital utilization within the 6 months prior to the index hospitalization, educational attainment, length of hospital stay and Project RED study group assignment. Bivariate analyses were conducted to determine which covariates were significant confounders of the relationship between depression and hospital utilization within 30 days of discharge. Chi‐square tests were used for categorical variables and t‐tests for continuous variables.

Baseline Characteristics of Study Subjects by Depression Screen Status
CharacteristicDepression Screen*
Negative (n = 500)Positive (n = 238)P Value
  • NOTE: Some columns may not add up to 100% due to omission of Other categories.

  • Abbreviations: PCP, primary care provider; PHQ9, Patient Health Questionnaire‐9; REALM, Rapid Estimate of Health Literacy in Adult Medicine tool; SD, standard deviation.

  • Positive depressive symptom screen determined by PHQ9 screen tool, a nine‐item 4‐point Likert scale, standard scoring algorithm to screen for major and minor depression. A score of 5 or higher indicates a positive depression symptom screen.17

  • Free Care is a Massachusetts state program for uninsured patients.

  • Health literacy categories correspond to total score as determined by REALM.18

  • Have PCP refers to subject self‐identifying PCP at time of Project RED study enrollment.

  • Unmarried marital status includes subjects identified as divorced, widow, single, partnered.

  • Charlson Comorbity Index Score reflects the cumulative increased likelihood of 1‐year mortality. The higher the score the more severe the comorbid condition. A 33% increase in risk for death is reflected in a 1‐point increase in weights. The minimum score is 0. There is no maximum score.19

  • Project RED study intervention group refers to subjects who received the 3‐armed discharge intervention.

  • Frequent Utilizer is defined as a subject with 2 or more hospital utilizations in 6 months prior to Project RED clinical trial index admission.

Race, No. (%)   
White140 (30)66 (30) 
Black268 (58)117 (54) 
Hispanic47 (10)29 (13)0.760
Insurance, No. (%)   
Private95 (19)22 (9) 
Medicare69 (14)30 (13) 
Medicaid214 (43)143 (61) 
Free care118 (24)40 (17)<0.001
Education, No. (%)   
<8th grade33 (7)21 (9) 
Some high school82 (17)52 (22) 
High school grad192 (38)90 (38) 
Some college126 (25)51 (22) 
College grad67 (13)22 (9)0.135
Health Literacy   
Grade 3 and below64 (13)44 (19) 
Grade 4654 (11)22 (10) 
Grade 78156 (32)73 (32) 
Grade 9 and above213 (44)89 (39)0.170
Income, $, No. (%)   
No income61 (12)37 (16) 
<10K77 (15)61 (26) 
1020K96 (19)35 (15) 
2050K97 (19)34 (14) 
50100K35 (8)7 (2) 
No answer132 (27)64 (27)0.002
Employment status, No. (%)   
Full time142 (28)34 (14) 
Part time57 (11)30 (13) 
Not Working297 (59)171 (72)<0.001
Age, mean (SD), years49.9 (16.0)49.6 (13.3)0.802
Gender: No. (%) Female239 (48)133 (56)0.040
Have PCP, No. (%) Yes399 (80)197 (83)0.340
Marital status, No. (%) unmarried365 (73)201 (85)<0.001
Charlson score, mean (SD)1.058 (1.6)1.56 (2.39)0.001
RED study group,# No. (%)   
Intervention243 (49)127 (53)0.22
Length of stay, days, mean (SD)2.5 (2.8)3.1 (3.8)0.016
Homeless in last 3 months, No. (%)45 (9)30 (13)0.130
Frequent utilizer,** No. (%)159 (32)104 (44)0.002

Age, length of stay, and Charlson score were used as continuous variables. Gender, marital status, frequent prior utilization (01 vs. 2 or more), and homelessness were treated as dichotomous variables. Categorical variables were created for, educational attainment (less than eighth grade, some high school, high school graduate, some college, college graduate), insurance type (Medicare, Medicaid, private insurance or free care), income level (no income, less than $10,000 per year, $10,00020,000, $20,00050,000, $50,000100,000, no answer), level of health literacy (grade 3 and below, grade 46, grade 78, grade 9 or above) and employment status(working full‐time, working part‐time, not working, no answer).

The 30‐day hospital utilization rate reflects the number of hospital utilization events within 30 days of discharge per subject. The same method was used to calculate hospital utilization rates within 60 and 90 days of discharge respectively. The unadjusted incident rate ratio (IRR) was calculated as the ratio of the rate of hospital utilizations among patients with depressive symptoms versus patients without depressive symptoms. Data for hospital utilization at 30, 60, and 90 days are cumulative.

Poisson models were used to test for significant differences between the predicted and observed number of hospitalization events at 30 days. A backward stepwise regression was conducted to identify and control for relevant confounders and construct the final, best‐fit model for the association between depression and hospital reutilization. A statistical significance level of P = 0.10 was used for the stepwise regression. To evaluate potential interactions between depression and the Project RED intervention, interaction terms were included. Two‐sided significance tests were used. P values of less than 0.05 were considered to indicate statistical significance. All data were analyzed with S‐Plus 8.0 (Seattle, WA).

In addition, a Kaplan‐Meier hazard curve was generated for the first hospital utilization event, ED visit or readmission, for the 30‐day period following discharge and compared with a log‐rank test.

Results

A total of 28% of subjects were categorized as having a positive depression screen. More women (36%) had positive depression screens than men (28%). Among patients with a positive depression screen, 58% had a history of depression and 53% were currently taking medications at the time of enrollment, compared with 25% and 22% respectively for subjects with a negative depression screen. Table 1 presents the means or percentages for baseline characteristics by depression status in the analytic cohort. Subjects with Medicaid for insurance had a higher rate of depression (61%) than subjects with Medicare (13%), private insurance (9%), or those who qualified for the Free Care pool (17%) which is the Massachusetts state funding for healthcare to uninsured persons. Subjects who were unemployed, unmarried, or who reported earnings less than $10,000 per year were also more likely to screen positive for depression. In addition, depressed subjects had a higher severity of co‐morbid disease and longer length of stay for the index hospitalization. Patients categorized as frequent utilizers (2 or more prior admissions) for the 6 months prior to the index hospitalization were also more likely to be depressed. Of further note, is the relatively younger average age among both depressive patients (49.6 years) and non‐depressive patients (49.9) of these study subjects.

The unadjusted hospital utilization rate at 30, 60, and 90 days post‐discharge by depression status is shown in Table 2. At 30 days post‐discharge, those with depressive symptoms had a higher rate of hospital utilization than those without depressive symptoms (0.563 vs. 0.296). In other words, 56 utilization events occurred per 100 patients with depressive symptoms, compared with 30 utilization events per 100 patients without depressive symptoms. The unadjusted 30‐day post‐discharge hospital utilization rate among those with depressive symptoms was higher compared with those without symptoms (IRR, 1.90, 95% confidence interval [CI], 1.242.71). A similar trend was found among subjects at 60 and 90 days post‐discharge.

Number of Hospital Utilizations, Hospital Utilization Rate, Unadjusted IRR at 30, 60 and 90 Days by Depression Screen Status*
Hospital UtilizationDepression Screen*P ValueIRR (CI)
Negative, n = 500 (68%)Positive, n = 238 (32%)
  • Abbreviations: CI, confidence interval; ED, emergency department; IRR, incident rate ratio.

  • Depression screen determined by scoring of Patient Health Questionnaire‐9 (PHQ9). Depressive symptom score of 5 points or higher is designated as positive.17

  • Number of hospital utilizations include all ED visits and hospital readmissions following discharge from Project RED index admission. ED visits leading to hospital admission are counted as one event. Sum reflects cumulative number of events over 30, 60 and 90 days.

No. of hospital utilizations140134 1.90 (1.51,2.40)
30‐day hospital utilization rate0.2960.563<0.001 
No. of hospital utilizations231205 1.87 (1.55,2.26)
60‐day hospital utilization rate0.4630.868<0.001 
No. of hospital utilizations324275 1.79 (1.53,2.10)
90‐day hospital utilization rate0.6481.165<0.001 

Poisson regression analyses were conducted to control for potential confounding in the relationship between depressive symptoms and hospital utilization rate within 30 days after discharge (Table 3). After controlling for relevant confounders, including age, gender, employment status, frequent prior hospitalization status, marital status, Charlson score, Project RED study group assignment and the interaction variable for RED study group assignment and depression, the association between symptoms of depression, and hospital utilization rate remained significant (IRR, 1.73; 95% CI, 1.272.36).

Adjusted Incident Rate Ratio of Hospital Utilization Within 30 Days of Discharge
CharacteristicsIRRCIP Value
  • Abbreviations: CI, confidence interval; ED, emergency department; IRR, incident rate ratio, PHQ9, Patient Health Questionnaire‐9.

  • Positive depressive symptom screen determined by PHQ9 screen tool, a nine‐item 4‐point Likert scale, standard scoring algorithm to screen for major and minor depression. A score of 5 or higher indicates a positive depression symptom screen.17

  • Unmarried refers to subjects whose self‐reported marital status includes divorced, single, partnered or widowed.

  • Frequent utilizer: 2 or more ED or hospital admissions visits in prior 6 months from index admission.

  • Refers to Project RED study group assignment.

  • Charlson Comorbity Index Score reflects the cumulative increased likelihood of 1‐year mortality. The higher the score the more severe the comorbid condition. A 33% increase in risk for death is reflected in a 1‐point increase in weights. The minimum score is 0. There is no maximum score.19

  • Interaction term of Project RED study group assignment and depressive symptom category (positive or negative).

Depression symptoms*  <0.001
Positive1.731.272.36 
NegativeREF1.0 
Gender  <0.001
Male1.871.472.40 
FemaleREF1.0 
Marital status  0.005
Married0.6250.440.89 
Unmarried1.0REF 
Frequent utilizer  <0.001
2+ prior visits2.451.923.15 
<2 prior visits1.0REF 
Study group  0.054
Intervention0.760.551.06 
Control1.0REF 
Employment   
Part time1.400.852.300.095
Not working1.671.152.440.003
Other0.520.073.850.262
Full time1.0REF 
Charlson Score0.980.921.040.250
Group* depression0.840.521.360.236
Age1.000.991.010.375

Figure 1 depicts the Kaplan‐Meier hazard curve generated for time to first hospital utilization, stratified by depression status. While 21% of participants without symptoms of depression had a hospital utilization within 30 days, fully 29% of participants with symptoms of depression had a hospital utilization within 30 days (P = 0.011).

Figure 1
Hazard for hospital utilization among subjects with and without depressive symptoms in 30 days following hospital discharge.

Discussion

Our study shows hospitalized patients who screen positive for depressive symptoms are significantly more likely to have a hospital visit (emergency room or rehospitalization) within 30 days of discharge than those who do not screen positive for depressive symptoms among medical patients admitted to an urban, academic, safety‐net hospital. These findings are consistent with, and extend, prior reports regarding depression and rehospitalization in specific populations (ie, geriatrics) and specific diagnoses (ie, cardiovascular disease [CVD] and COPD).1012 We observed a 73% higher incidence rate for hospital utilization within 30 days of discharge for those with symptoms of depression. This puts symptoms of depression on par with frequent prior rehospitalization, advanced age and low social support, as known risk factors for rehospitalization.4, 5, 23

Also of significance is the relatively young age of this study population (49.9 years non‐depressive patients and 49.6 years for depressive patients) compared with the study cohorts used for research in the majority of the existing literature. The chief reason for the young age of our cohort is that potential subjects were excluded if they came from a skilled nursing facility or other hospital. This may limit the generalizability of our findings; however, it seems likely that interventions relating to depression and transitions of care will need to be quite different for patients that reside in long‐term care facilities vs. patients that live in the community. For example, patients living in the community may have significant barriers to access post‐discharge services due to insurance status and are more likely to be sensitive to variations in social support.

Early rehospitalization is associated with significant morbidity, mortality, and expense. It is also a potential marker for poor quality of care.24 Concerns for patient safety, escalating healthcare costs, and possible change in hospital reimbursement mechanisms are fueling the search for modifiable risk factors associated with early rehospitalization. Our data provide evidence that symptoms of depression may be an important focus of attention. We do not know, however whether treating hospitalized patients who screen positive for depression will decrease early rehospitalization and emergency room utilization rates.

Various physiologic and behavioral mechanisms may link symptoms of depression to hospital utilization after discharge. For example, depressed patients with features of somatization may be more likely to experience worrisome physical symptoms after discharge and present prematurely for reevaluation. Patients who are sicker in some fashion not captured by our measured confounders may have symptoms of depression related to chronic, debilitating disease warranting early return to the hospital. Depression may also yield nonadherence to aspects of the discharge treatment plan leading to rehospitalization as a result of poor post‐discharge disease management. For example, research shows that patients with depression following coronary artery bypass surgery are less likely to adhere with cardiac rehabilitation programs.25 Likewise, depression among chronically ill patients such as diabetics, asthmatics, or human immunodeficiency virus (HIV)‐positive patients impairs medication adherence and self‐care behavior which may lead to disease relapse or recurrence.2628 One study examining depression effects on hypertensive medicine adherence in African Americans identified self‐efficacy as a mediating factor between depression and nonadherence.29 This implies that interventions such as self‐management education, a program through which chronically‐ill patients learn to better manage their illnesses through enhanced self‐confidence and problem‐solving strategies (including mood disorder challenges) may reduce early rehospitalization among depressed patients.30

There is also evidence that depression may have direct physiologic consequences. In patients with CVD, depression is associated with poor outcomes possibly related to decreased heart rate variability, hypercoagulability, high burdens of inflammatory markers, and severity of left ventricular dysfunction.3134 Similarly, depression among HIV/acquired immune deficiency syndrome (AIDS), diabetics and multiple sclerosis (MS) patients is linked to heightened levels of proinflammatory markers and less favorable outcomes that may signal a more severe form of the disease or an impaired response to treatment.3538 Indeed, MS investigators now hypothesize that the proinflammatory environment associated with the neurologic manifestations of MS are also causing depression symptoms among MS patients.34 This theory contrasts the common belief that depression in the chronically ill manifests independent of the chronic illness or in response to living with chronic disease.

A major strength of the current study is the large dataset and the broad range of covariates available for analyses. However, several limitations should be noted. First, data on hospital utilization outside Boston Medical Center were determined by patient self‐report and were not confirmed by document review. Second, we do not know the direction of the associations we report. If symptoms of depression are merely the consequence of having a higher disease burden, treatment of the underlying disease may be the most important response. While this is possible, our model does include several variables (eg, Charlson score and length of stay) that are likely to adjust for disease severity, pointing to the likelihood that symptoms of depression truly predict hospital utilization in a fashion that is independent of disease severity. Third, our results may not be generalizable to populations other than those served by urban safety‐net hospitals or other populations excluded from the Project RED trial (eg, non‐English speaking patients and patients from nursing homes). Finally, social factors such as substance use and social support system variables may residually confound the relationship between depression and hospital reutilization demonstrated in this study. While this dataset does not include a measure of social support other than marital status and housing status, data is available on substance use. Analyses conducted by our colleagues using Project RED data found that in this study population depression was significantly more prevalent among substance users (29% vs. 14%) compared with non‐users and that substance use is an independent risk factor for hospital reutilization (unpublished data).

Our findings linking depression to increased hospital utilization also warrant further consideration from healthcare policymakers. Central to the Obama Administration's February 2009 healthcare reform proposal is the pursuit of cost savings through reductions in unplanned hospital readmissions.39 Thus, identifying potentially modifiable risk factors for readmission, such as depression, is of great concern to healthcare providers and policymakers across the nation. If, through testing of interventions, depression proves to be a modifiable risk for readmission, policymakers, while negotiating healthcare reform measures, must provide for the services required to address this comorbidity at the time of discharge. For example, if a patient screens positive for depressive symptoms during a hospitalization for COPD exacerbation, will the proposed payment reforms allow for mental health services during the immediate post‐discharge period in order to reduce the likelihood of hospital readmission? Will those mental health services be readily available? Payment reforms that account for all necessary transitional care services will indeed help reduce readmission costs with less risk for untoward consequences.

In conclusion, our results indicate that a positive depression screen is a significant risk factor for early post‐discharge hospital utilization among hospitalized adults on a general medical service, even after controlling for relevant confounders. Screening for depression during acute hospitalizations may be an important step in identifying patients at increased risk for readmission. Future research should focus on further characterizing and stratifying populations at highest risk for depression. Efforts should also include developing and evaluating targeted interventions for patients with symptoms of depression among hospitalized patients as part of discharge planning. Timely depression therapy during the hospitalization or following hospital discharge might reduce costly readmissions and enhance patient safety.

References
  1. Jenks SF,Williams MV,Coleman EA.Rehospitalizations among patients in the Medicare fee‐for‐service program.N Engl J Med.2009;360(14):14571459.
  2. Jack BW,Chetty VK,Anthony D, et al.The reengineered hospital discharge program to decrease rehospitalization.Ann Intern Med.2009;150(3):178187.
  3. Weissman JS,Stern RS,Epstein AM.The impact of patient socioeconomic status and other social factors on readmission: a prospective study in four Massachusetts hospitals.Inquiry.1994;31(2):163172.
  4. van Walraven C,Mamdani M,Fang J,Austin PC.Continuity of care and patient outcomes after hospital discharge.J Gen Intern Med.2004;19:624631. [PMID: 15209600]
  5. Marcantonio ER,McKean S,Goldfinger M,Kleenfield S,Yurkofsky M,Brennan TA.Factors associated with unplanned hospital readmission among patients 65 years of age and older in a Medicare managed care plan.Am J Med.1999;107(1):1317.
  6. Krumholz HM,Parent EM,Tu N, et al.Readmission after hospitalization for congestive heart failure among Medicare beneficiaries.Arch Intern Med.1997;157(1):99104.
  7. Librero J,Peiro S,Ordinana R.Chronic comorbidity and outcomes of hospital care: length of stay, mortality and readmission at 30 and 365 days.J Clin Epidemiol.1999;52(3):171179.
  8. Rodríguez‐Artalejo F,Guallar‐Castillón P,Herrera MC, et al.Social network as a predictor of hospital readmission and mortality among older patients with heart failure.J Card Fail.2006;12:621627.
  9. Wong AW,Gan WQ,Burns J,Sin DD,van Eeden SF.Acute exacerbation of chronic obstructive pulmonary disease: influence of social factors in determining length of stay and readmission rates.Can Respir J.2008;15(7):361364.
  10. Parashar S,Rumsfeld JS,Spertus JA, et al.Time course of depression and outcome of myocardial infarction.Arch Intern Med.2006;166(18):20352043.
  11. Budpitz DS,Shebab N,Kegler SR, et al.Medication use leading to emergency department visits for adverse drug events in older adults.Ann Intern Med.2007;147(11):755765.
  12. Campbell SE,Seymour DG,Primrose WR.A systematic literature review of factors affecting outcomes in older medical patients admitted to hospital.Age Ageing.2004;33(2):110115.
  13. Kartha A,Anthony D,Manasseh CS, et al.Depression is a risk factor for rehospitalization in medical inpatients.Prim Care Companion J Clin Psychiatry.2007;9(4):256262.
  14. Almagro P,Barreiro Bienvenido ,Ochoa de Echaguen A, et al.Risk factors for hospital readmission in patients with chronic obstructive pulmonary disease.Respiration.2006;73:311317.
  15. Frasure‐Smith N,Lesperance F,Gravel G, et al.Depression and healthcare costs during the first year following myocardial infarction.J Psychosom Res.2000;48(4–5):471478.
  16. Jiang W,Alexander J,Christopher E, et al.Relationship of depression to increased risk of mortality and rehospitalization.Arch Intern Med.2001;161(15):18491856.
  17. Parashar S,Rumsfeld JS,Spertus JA, et al.Time course of depression and outcome of myocardial infarction.Arch Intern Med.2006;166:20352043.
  18. Scherer M,Herrmann‐Lingen C.Single item on positive affect is associated with 1‐year survival in consecutive medical inpatients.J Gen Hosp Psych.2009;31:813.
  19. Hasin DS,Goodwin RD,Stinson FS,Grant BF.Epidemiology of major depressive disorder: results from the National Epidemiologic Survey on Alcoholism and Related Conditions.Arch Gen Psychiatry.2005;62(10):1097106.
  20. Kroenke K,Spitzer RL,Williams JB.The PHQ‐9: Validity of a brief depression severity measure.J Gen Intern Med.2001;16:606613. [PMID:11556941]
  21. Davis TC,Long SW,Jackson RH, et al.Rapid estimate of adult literacy in medicine: a shortened screening instrument.Fam Med.1993;25:391395. [PMID:8349060]
  22. Charlson ME,Pompei P,Ales KL,MacKenzie CR.A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis.1987;40:373383. [PMID: 3558716]
  23. Rodriguez‐Artalejo F,Guallar‐Castillon P,Herrera MC, et al.Social network as a predictor of hospital readmission and mortality among older patients with heart failure.J Card Fail.2006;12(8):621627.
  24. Ashton CM,Del Junco DJ,Souchek J,Wray NP,Mansyr CL.The association between the quality of inpatient care and early readmission: a meta‐analysis of the evidence.Med Care.1997;35(10):10441059.
  25. Kronish IM,Rieckmann N,Halm FA, et al.Persistent depression affects adherence to secondary prevention behaviors after acute coronary syndromes.J Gen Intern Med.2006;21(11):11781183.
  26. Cukor D,Rosenthal DS,Jindal RM,Brown CD,Kimmel PL.Depression is an important contributor to low medication adherence in hemodialyzed patients and transplant recipients.Kidney Int.2009;75(11):12231229.
  27. Gonzalez JS,Safren SA,Delahanty LM, et al.Symptoms of depression prospectively predict poorer self‐care in patients with Type 2 diabetes.Diabet Med.2008;25(9):11021107.
  28. Lima VD,Geller J,Bangsberg DR, et al.The effect of adherence on the association between depressive symptoms and mortality among HIV‐infected individuals first initiating HAART.AIDS.2007;21(9):11751183.
  29. Schoenthaler A,Ogedegbe G,Allegrante JP.Self‐efficacy mediates the relationship between depressive symptoms and medication adherence.Health Educ Behav.2009;36(1):127137.
  30. Bodenheimer T,Lorig K,Holman H,Grumbach K.Patient self‐management of chronic disease in primary care.JAMA.2002;288(19):24692475.
  31. McFarlane AM.Effects of sertraline on the recovery rate of cardiac autonomic function in depressed patients after acute myocardial infarction.Am Heart J.2001;142:617623.
  32. van Melle JP,de Jonge P,Ormel J, et al.Relationship between left ventricular dysfunction and depression following myocardial infarction: data from the MIND‐IT.Eur Heart J.2005;26:26502656.
  33. Serebruany VL,Glassman AH,Malinin AI, et al.Platelet/endothelial biomarkers in depressed patients treated with the selective serotonin reuptake inhibitor sertraline after acugte coronary events: the Sertraline AntiDepressant Heart Attack Randomized Trial (SADHART) Platelet SubStudy.Circulation.2003;108:939944.
  34. Mulvihill NT,Foley JB.Inflammation in acute coronary syndromes.Cleve Clin J Med.2002;69(Suppl2):SII130SII142.
  35. Gold SM,Irwin MR.Depression and immunity: inflammation and depressive symptoms in multiple sclerosis.Neurol Clin.2006;24(3):507519.
  36. Brydon L,Walker C,Wawrzyniak A, et al.Synergistic effects of psychological and immune stressors on inflammatory cytokines and sickness responses in humans.Brain Behav Immun.2009;23(2):217224.
  37. Gresson JM,Hurwitz BE,Llabre MM,Schneiderman N,Penedo FJ,Klimas NG.Psychological distress, killer lymphocytes and disease severity in HIV/AIDS.Brain Behav Immun.2008;22(6):901911.
  38. Pizzi C,Manzoli L,Mancini S,Costa GM.Analysis of potential predictors of depression among coronary heart disease risk factors including heart rate variability, markers of inflammation, and endothelial function.Eur Heart J.2008;29(9):11101117.
  39. Connolly C.Obama proposes $634 billion fund for health care.Washington Post. February 26,2009:A1.
Article PDF
Issue
Journal of Hospital Medicine - 5(7)
Publications
Page Number
378-384
Legacy Keywords
depression, hospital discharge, patient safety, readmission, rehospitalization
Sections
Article PDF
Article PDF

Fully 19% of Medicare patients are readmitted to the hospital within 30 days of discharge.1 This represents a large amount of potentially avoidable morbidity and cost. Indeed, projects to improve the discharge process and post‐hospital care have shown that as much as one‐third of hospital utilization in the month after discharge can be avoided.2 Consequently, the rate of early, unplanned hospital utilization after discharge has emerged as an important indicator of hospital quality and the Centers for Medicare and Medicaid Services (CMS) has proposed a policy to decrease payments to hospitals with high rates of early unplanned hospital utilization. Thus, there is great interest in identifying modifiable risk factors for rehospitalization that could be used to refine intervention models and lead to improvements in quality of care, patient outcomes, and cost savings.

To date, known predictors of readmission include: lower socioeconomic status,3 history of prior hospitalization4 and advanced age,5 length of stay greater than 7 days,6 a high burden of comorbid illnesses (based on Charlson score),7 poor social support,8 and specific diagnoses (eg, congestive heart failure, chronic obstructive pulmonary disease [COPD] and myocardial infarction).5, 9, 10 In addition, unplanned readmissions and emergency department (ED) visits have been linked to polypharmacy and adverse drug events related to treatment with medications such as warfarin, digoxin and narcotics.11, 12 Another characteristic that has also been linked to readmission is depression;13 however5 reports supporting this association are from studies of elderly patients or with patients who have specific diagnoses (eg, congestive heart failure [CHF], COPD, myocardial infarction).1416

Depression is common, affecting 13% to 16% of people in the US, and is recognized as an important risk factor for poor outcomes among patients with various chronic illnesses.1719 The mechanisms by which depression can be linked to health outcomes and health service utilization have been studied in age‐specific or disease‐specific cohorts such as cardiac patients or frail elders and include both physiologic factors such as hypercoagulability and hyperinflammatory conditions, as well as behavioral factors such as poor self‐care behaviors and heightened sensitivity to somatic symptoms. How these mechanisms link depression to health outcomes and hospital utilization in a general medical population is not clearly understood. Kartha et al.13 reported findings indicating that depression is a risk factor for rehospitalization in general medical inpatients, but the study sample was relatively small and the study design methodology significantly limited its generalizability.12 It would be useful to provide supporting evidence showing depression as an important risk factor for readmission in the general medical in‐patient population using more rigorous study methods and a larger cohort.

We hypothesized that depressive symptoms would be an independent risk factor for early unplanned hospital utilization after discharge for all medical patients. Therefore, we conducted a secondary analysis of the Project RED clinical trial dataset to assess the association between a positive depression screen during inpatient hospitalization and the rate of subsequent hospital utilization.

Methods

Data from the Project RED clinical trial were reviewed for inclusion in a secondary analysis. Complete data were available for 738 of the 749 subjects recruited for Project RED.

Project RED Setting and Participants

Project RED was a two‐armed randomized controlled trial of English‐speaking adult patients, 18 years or older, admitted to the teaching service of Boston Medical Center, a large urban safety‐net hospital with an ethnically diverse patient population. A total of 749 subjects were enrolled and randomized between January 3, 2006 and October 18, 2007. Patients were required to have a telephone, be able to comprehend study details and the consent process in English, and have plans to be discharged to a US community. Patients were not enrolled if they were admitted from a skilled nursing facility or other hospital, transferred to a different hospital service prior to enrollment, admitted for a planned hospitalization, on hospital precautions, on suicide watch, deaf or blind. The Institutional Review Board of Boston University approved all study activities. A full description of the methods for the Project RED trial has been described previously.2

Outcome Variable

The primary endpoint was rate of hospital utilization within 30 days of discharge from the index admission, defined as the total number of ED visits and readmissions per subject within 30 days of the index discharge. Hospital utilization rates within 60 and 90 days of the index hospitalization discharge were also analyzed as secondary outcomes. Any ED visit in which a subject was subsequently admitted to the hospital was only counted as a readmission. Outcome data were collected by reviewing the hospital's electronic medical records (EMRs) and by contacting subjects by telephone 30 days after discharge. Dates of hospital utilization occurring at Boston Medical Center were obtained from the EMR, while those at other hospitals were collected through subject report. Subjects who could not be reached within 60 days of discharge were assumed alive.

Primary Independent Variable

The primary independent variable of interest was depressive symptoms defined as a positive score for minor or major depression on the nine‐item Patient Health Questionnaire (PHQ‐9) depression screening tool.20 A dichotomized variable was created using a standardized scoring system to determine the screening cut‐off for major or minor depressive symptoms.19

Statistical Analysis

Demographic and other characteristics of the subjects were compared by depression status (Table 1). Potential confounders were identified a priori from the available literature on factors associated with rehospitalization. These included age, gender, marital status, health literacy score (rapid estimate of health literacy in adult medicine tool [REALM]),21 Charlson score,22 insurance type, employment status, income level, homelessness status within past three months, hospital utilization within the 6 months prior to the index hospitalization, educational attainment, length of hospital stay and Project RED study group assignment. Bivariate analyses were conducted to determine which covariates were significant confounders of the relationship between depression and hospital utilization within 30 days of discharge. Chi‐square tests were used for categorical variables and t‐tests for continuous variables.

Baseline Characteristics of Study Subjects by Depression Screen Status
CharacteristicDepression Screen*
Negative (n = 500)Positive (n = 238)P Value
  • NOTE: Some columns may not add up to 100% due to omission of Other categories.

  • Abbreviations: PCP, primary care provider; PHQ9, Patient Health Questionnaire‐9; REALM, Rapid Estimate of Health Literacy in Adult Medicine tool; SD, standard deviation.

  • Positive depressive symptom screen determined by PHQ9 screen tool, a nine‐item 4‐point Likert scale, standard scoring algorithm to screen for major and minor depression. A score of 5 or higher indicates a positive depression symptom screen.17

  • Free Care is a Massachusetts state program for uninsured patients.

  • Health literacy categories correspond to total score as determined by REALM.18

  • Have PCP refers to subject self‐identifying PCP at time of Project RED study enrollment.

  • Unmarried marital status includes subjects identified as divorced, widow, single, partnered.

  • Charlson Comorbity Index Score reflects the cumulative increased likelihood of 1‐year mortality. The higher the score the more severe the comorbid condition. A 33% increase in risk for death is reflected in a 1‐point increase in weights. The minimum score is 0. There is no maximum score.19

  • Project RED study intervention group refers to subjects who received the 3‐armed discharge intervention.

  • Frequent Utilizer is defined as a subject with 2 or more hospital utilizations in 6 months prior to Project RED clinical trial index admission.

Race, No. (%)   
White140 (30)66 (30) 
Black268 (58)117 (54) 
Hispanic47 (10)29 (13)0.760
Insurance, No. (%)   
Private95 (19)22 (9) 
Medicare69 (14)30 (13) 
Medicaid214 (43)143 (61) 
Free care118 (24)40 (17)<0.001
Education, No. (%)   
<8th grade33 (7)21 (9) 
Some high school82 (17)52 (22) 
High school grad192 (38)90 (38) 
Some college126 (25)51 (22) 
College grad67 (13)22 (9)0.135
Health Literacy   
Grade 3 and below64 (13)44 (19) 
Grade 4654 (11)22 (10) 
Grade 78156 (32)73 (32) 
Grade 9 and above213 (44)89 (39)0.170
Income, $, No. (%)   
No income61 (12)37 (16) 
<10K77 (15)61 (26) 
1020K96 (19)35 (15) 
2050K97 (19)34 (14) 
50100K35 (8)7 (2) 
No answer132 (27)64 (27)0.002
Employment status, No. (%)   
Full time142 (28)34 (14) 
Part time57 (11)30 (13) 
Not Working297 (59)171 (72)<0.001
Age, mean (SD), years49.9 (16.0)49.6 (13.3)0.802
Gender: No. (%) Female239 (48)133 (56)0.040
Have PCP, No. (%) Yes399 (80)197 (83)0.340
Marital status, No. (%) unmarried365 (73)201 (85)<0.001
Charlson score, mean (SD)1.058 (1.6)1.56 (2.39)0.001
RED study group,# No. (%)   
Intervention243 (49)127 (53)0.22
Length of stay, days, mean (SD)2.5 (2.8)3.1 (3.8)0.016
Homeless in last 3 months, No. (%)45 (9)30 (13)0.130
Frequent utilizer,** No. (%)159 (32)104 (44)0.002

Age, length of stay, and Charlson score were used as continuous variables. Gender, marital status, frequent prior utilization (01 vs. 2 or more), and homelessness were treated as dichotomous variables. Categorical variables were created for, educational attainment (less than eighth grade, some high school, high school graduate, some college, college graduate), insurance type (Medicare, Medicaid, private insurance or free care), income level (no income, less than $10,000 per year, $10,00020,000, $20,00050,000, $50,000100,000, no answer), level of health literacy (grade 3 and below, grade 46, grade 78, grade 9 or above) and employment status(working full‐time, working part‐time, not working, no answer).

The 30‐day hospital utilization rate reflects the number of hospital utilization events within 30 days of discharge per subject. The same method was used to calculate hospital utilization rates within 60 and 90 days of discharge respectively. The unadjusted incident rate ratio (IRR) was calculated as the ratio of the rate of hospital utilizations among patients with depressive symptoms versus patients without depressive symptoms. Data for hospital utilization at 30, 60, and 90 days are cumulative.

Poisson models were used to test for significant differences between the predicted and observed number of hospitalization events at 30 days. A backward stepwise regression was conducted to identify and control for relevant confounders and construct the final, best‐fit model for the association between depression and hospital reutilization. A statistical significance level of P = 0.10 was used for the stepwise regression. To evaluate potential interactions between depression and the Project RED intervention, interaction terms were included. Two‐sided significance tests were used. P values of less than 0.05 were considered to indicate statistical significance. All data were analyzed with S‐Plus 8.0 (Seattle, WA).

In addition, a Kaplan‐Meier hazard curve was generated for the first hospital utilization event, ED visit or readmission, for the 30‐day period following discharge and compared with a log‐rank test.

Results

A total of 28% of subjects were categorized as having a positive depression screen. More women (36%) had positive depression screens than men (28%). Among patients with a positive depression screen, 58% had a history of depression and 53% were currently taking medications at the time of enrollment, compared with 25% and 22% respectively for subjects with a negative depression screen. Table 1 presents the means or percentages for baseline characteristics by depression status in the analytic cohort. Subjects with Medicaid for insurance had a higher rate of depression (61%) than subjects with Medicare (13%), private insurance (9%), or those who qualified for the Free Care pool (17%) which is the Massachusetts state funding for healthcare to uninsured persons. Subjects who were unemployed, unmarried, or who reported earnings less than $10,000 per year were also more likely to screen positive for depression. In addition, depressed subjects had a higher severity of co‐morbid disease and longer length of stay for the index hospitalization. Patients categorized as frequent utilizers (2 or more prior admissions) for the 6 months prior to the index hospitalization were also more likely to be depressed. Of further note, is the relatively younger average age among both depressive patients (49.6 years) and non‐depressive patients (49.9) of these study subjects.

The unadjusted hospital utilization rate at 30, 60, and 90 days post‐discharge by depression status is shown in Table 2. At 30 days post‐discharge, those with depressive symptoms had a higher rate of hospital utilization than those without depressive symptoms (0.563 vs. 0.296). In other words, 56 utilization events occurred per 100 patients with depressive symptoms, compared with 30 utilization events per 100 patients without depressive symptoms. The unadjusted 30‐day post‐discharge hospital utilization rate among those with depressive symptoms was higher compared with those without symptoms (IRR, 1.90, 95% confidence interval [CI], 1.242.71). A similar trend was found among subjects at 60 and 90 days post‐discharge.

Number of Hospital Utilizations, Hospital Utilization Rate, Unadjusted IRR at 30, 60 and 90 Days by Depression Screen Status*
Hospital UtilizationDepression Screen*P ValueIRR (CI)
Negative, n = 500 (68%)Positive, n = 238 (32%)
  • Abbreviations: CI, confidence interval; ED, emergency department; IRR, incident rate ratio.

  • Depression screen determined by scoring of Patient Health Questionnaire‐9 (PHQ9). Depressive symptom score of 5 points or higher is designated as positive.17

  • Number of hospital utilizations include all ED visits and hospital readmissions following discharge from Project RED index admission. ED visits leading to hospital admission are counted as one event. Sum reflects cumulative number of events over 30, 60 and 90 days.

No. of hospital utilizations140134 1.90 (1.51,2.40)
30‐day hospital utilization rate0.2960.563<0.001 
No. of hospital utilizations231205 1.87 (1.55,2.26)
60‐day hospital utilization rate0.4630.868<0.001 
No. of hospital utilizations324275 1.79 (1.53,2.10)
90‐day hospital utilization rate0.6481.165<0.001 

Poisson regression analyses were conducted to control for potential confounding in the relationship between depressive symptoms and hospital utilization rate within 30 days after discharge (Table 3). After controlling for relevant confounders, including age, gender, employment status, frequent prior hospitalization status, marital status, Charlson score, Project RED study group assignment and the interaction variable for RED study group assignment and depression, the association between symptoms of depression, and hospital utilization rate remained significant (IRR, 1.73; 95% CI, 1.272.36).

Adjusted Incident Rate Ratio of Hospital Utilization Within 30 Days of Discharge
CharacteristicsIRRCIP Value
  • Abbreviations: CI, confidence interval; ED, emergency department; IRR, incident rate ratio, PHQ9, Patient Health Questionnaire‐9.

  • Positive depressive symptom screen determined by PHQ9 screen tool, a nine‐item 4‐point Likert scale, standard scoring algorithm to screen for major and minor depression. A score of 5 or higher indicates a positive depression symptom screen.17

  • Unmarried refers to subjects whose self‐reported marital status includes divorced, single, partnered or widowed.

  • Frequent utilizer: 2 or more ED or hospital admissions visits in prior 6 months from index admission.

  • Refers to Project RED study group assignment.

  • Charlson Comorbity Index Score reflects the cumulative increased likelihood of 1‐year mortality. The higher the score the more severe the comorbid condition. A 33% increase in risk for death is reflected in a 1‐point increase in weights. The minimum score is 0. There is no maximum score.19

  • Interaction term of Project RED study group assignment and depressive symptom category (positive or negative).

Depression symptoms*  <0.001
Positive1.731.272.36 
NegativeREF1.0 
Gender  <0.001
Male1.871.472.40 
FemaleREF1.0 
Marital status  0.005
Married0.6250.440.89 
Unmarried1.0REF 
Frequent utilizer  <0.001
2+ prior visits2.451.923.15 
<2 prior visits1.0REF 
Study group  0.054
Intervention0.760.551.06 
Control1.0REF 
Employment   
Part time1.400.852.300.095
Not working1.671.152.440.003
Other0.520.073.850.262
Full time1.0REF 
Charlson Score0.980.921.040.250
Group* depression0.840.521.360.236
Age1.000.991.010.375

Figure 1 depicts the Kaplan‐Meier hazard curve generated for time to first hospital utilization, stratified by depression status. While 21% of participants without symptoms of depression had a hospital utilization within 30 days, fully 29% of participants with symptoms of depression had a hospital utilization within 30 days (P = 0.011).

Figure 1
Hazard for hospital utilization among subjects with and without depressive symptoms in 30 days following hospital discharge.

Discussion

Our study shows hospitalized patients who screen positive for depressive symptoms are significantly more likely to have a hospital visit (emergency room or rehospitalization) within 30 days of discharge than those who do not screen positive for depressive symptoms among medical patients admitted to an urban, academic, safety‐net hospital. These findings are consistent with, and extend, prior reports regarding depression and rehospitalization in specific populations (ie, geriatrics) and specific diagnoses (ie, cardiovascular disease [CVD] and COPD).1012 We observed a 73% higher incidence rate for hospital utilization within 30 days of discharge for those with symptoms of depression. This puts symptoms of depression on par with frequent prior rehospitalization, advanced age and low social support, as known risk factors for rehospitalization.4, 5, 23

Also of significance is the relatively young age of this study population (49.9 years non‐depressive patients and 49.6 years for depressive patients) compared with the study cohorts used for research in the majority of the existing literature. The chief reason for the young age of our cohort is that potential subjects were excluded if they came from a skilled nursing facility or other hospital. This may limit the generalizability of our findings; however, it seems likely that interventions relating to depression and transitions of care will need to be quite different for patients that reside in long‐term care facilities vs. patients that live in the community. For example, patients living in the community may have significant barriers to access post‐discharge services due to insurance status and are more likely to be sensitive to variations in social support.

Early rehospitalization is associated with significant morbidity, mortality, and expense. It is also a potential marker for poor quality of care.24 Concerns for patient safety, escalating healthcare costs, and possible change in hospital reimbursement mechanisms are fueling the search for modifiable risk factors associated with early rehospitalization. Our data provide evidence that symptoms of depression may be an important focus of attention. We do not know, however whether treating hospitalized patients who screen positive for depression will decrease early rehospitalization and emergency room utilization rates.

Various physiologic and behavioral mechanisms may link symptoms of depression to hospital utilization after discharge. For example, depressed patients with features of somatization may be more likely to experience worrisome physical symptoms after discharge and present prematurely for reevaluation. Patients who are sicker in some fashion not captured by our measured confounders may have symptoms of depression related to chronic, debilitating disease warranting early return to the hospital. Depression may also yield nonadherence to aspects of the discharge treatment plan leading to rehospitalization as a result of poor post‐discharge disease management. For example, research shows that patients with depression following coronary artery bypass surgery are less likely to adhere with cardiac rehabilitation programs.25 Likewise, depression among chronically ill patients such as diabetics, asthmatics, or human immunodeficiency virus (HIV)‐positive patients impairs medication adherence and self‐care behavior which may lead to disease relapse or recurrence.2628 One study examining depression effects on hypertensive medicine adherence in African Americans identified self‐efficacy as a mediating factor between depression and nonadherence.29 This implies that interventions such as self‐management education, a program through which chronically‐ill patients learn to better manage their illnesses through enhanced self‐confidence and problem‐solving strategies (including mood disorder challenges) may reduce early rehospitalization among depressed patients.30

There is also evidence that depression may have direct physiologic consequences. In patients with CVD, depression is associated with poor outcomes possibly related to decreased heart rate variability, hypercoagulability, high burdens of inflammatory markers, and severity of left ventricular dysfunction.3134 Similarly, depression among HIV/acquired immune deficiency syndrome (AIDS), diabetics and multiple sclerosis (MS) patients is linked to heightened levels of proinflammatory markers and less favorable outcomes that may signal a more severe form of the disease or an impaired response to treatment.3538 Indeed, MS investigators now hypothesize that the proinflammatory environment associated with the neurologic manifestations of MS are also causing depression symptoms among MS patients.34 This theory contrasts the common belief that depression in the chronically ill manifests independent of the chronic illness or in response to living with chronic disease.

A major strength of the current study is the large dataset and the broad range of covariates available for analyses. However, several limitations should be noted. First, data on hospital utilization outside Boston Medical Center were determined by patient self‐report and were not confirmed by document review. Second, we do not know the direction of the associations we report. If symptoms of depression are merely the consequence of having a higher disease burden, treatment of the underlying disease may be the most important response. While this is possible, our model does include several variables (eg, Charlson score and length of stay) that are likely to adjust for disease severity, pointing to the likelihood that symptoms of depression truly predict hospital utilization in a fashion that is independent of disease severity. Third, our results may not be generalizable to populations other than those served by urban safety‐net hospitals or other populations excluded from the Project RED trial (eg, non‐English speaking patients and patients from nursing homes). Finally, social factors such as substance use and social support system variables may residually confound the relationship between depression and hospital reutilization demonstrated in this study. While this dataset does not include a measure of social support other than marital status and housing status, data is available on substance use. Analyses conducted by our colleagues using Project RED data found that in this study population depression was significantly more prevalent among substance users (29% vs. 14%) compared with non‐users and that substance use is an independent risk factor for hospital reutilization (unpublished data).

Our findings linking depression to increased hospital utilization also warrant further consideration from healthcare policymakers. Central to the Obama Administration's February 2009 healthcare reform proposal is the pursuit of cost savings through reductions in unplanned hospital readmissions.39 Thus, identifying potentially modifiable risk factors for readmission, such as depression, is of great concern to healthcare providers and policymakers across the nation. If, through testing of interventions, depression proves to be a modifiable risk for readmission, policymakers, while negotiating healthcare reform measures, must provide for the services required to address this comorbidity at the time of discharge. For example, if a patient screens positive for depressive symptoms during a hospitalization for COPD exacerbation, will the proposed payment reforms allow for mental health services during the immediate post‐discharge period in order to reduce the likelihood of hospital readmission? Will those mental health services be readily available? Payment reforms that account for all necessary transitional care services will indeed help reduce readmission costs with less risk for untoward consequences.

In conclusion, our results indicate that a positive depression screen is a significant risk factor for early post‐discharge hospital utilization among hospitalized adults on a general medical service, even after controlling for relevant confounders. Screening for depression during acute hospitalizations may be an important step in identifying patients at increased risk for readmission. Future research should focus on further characterizing and stratifying populations at highest risk for depression. Efforts should also include developing and evaluating targeted interventions for patients with symptoms of depression among hospitalized patients as part of discharge planning. Timely depression therapy during the hospitalization or following hospital discharge might reduce costly readmissions and enhance patient safety.

Fully 19% of Medicare patients are readmitted to the hospital within 30 days of discharge.1 This represents a large amount of potentially avoidable morbidity and cost. Indeed, projects to improve the discharge process and post‐hospital care have shown that as much as one‐third of hospital utilization in the month after discharge can be avoided.2 Consequently, the rate of early, unplanned hospital utilization after discharge has emerged as an important indicator of hospital quality and the Centers for Medicare and Medicaid Services (CMS) has proposed a policy to decrease payments to hospitals with high rates of early unplanned hospital utilization. Thus, there is great interest in identifying modifiable risk factors for rehospitalization that could be used to refine intervention models and lead to improvements in quality of care, patient outcomes, and cost savings.

To date, known predictors of readmission include: lower socioeconomic status,3 history of prior hospitalization4 and advanced age,5 length of stay greater than 7 days,6 a high burden of comorbid illnesses (based on Charlson score),7 poor social support,8 and specific diagnoses (eg, congestive heart failure, chronic obstructive pulmonary disease [COPD] and myocardial infarction).5, 9, 10 In addition, unplanned readmissions and emergency department (ED) visits have been linked to polypharmacy and adverse drug events related to treatment with medications such as warfarin, digoxin and narcotics.11, 12 Another characteristic that has also been linked to readmission is depression;13 however5 reports supporting this association are from studies of elderly patients or with patients who have specific diagnoses (eg, congestive heart failure [CHF], COPD, myocardial infarction).1416

Depression is common, affecting 13% to 16% of people in the US, and is recognized as an important risk factor for poor outcomes among patients with various chronic illnesses.1719 The mechanisms by which depression can be linked to health outcomes and health service utilization have been studied in age‐specific or disease‐specific cohorts such as cardiac patients or frail elders and include both physiologic factors such as hypercoagulability and hyperinflammatory conditions, as well as behavioral factors such as poor self‐care behaviors and heightened sensitivity to somatic symptoms. How these mechanisms link depression to health outcomes and hospital utilization in a general medical population is not clearly understood. Kartha et al.13 reported findings indicating that depression is a risk factor for rehospitalization in general medical inpatients, but the study sample was relatively small and the study design methodology significantly limited its generalizability.12 It would be useful to provide supporting evidence showing depression as an important risk factor for readmission in the general medical in‐patient population using more rigorous study methods and a larger cohort.

We hypothesized that depressive symptoms would be an independent risk factor for early unplanned hospital utilization after discharge for all medical patients. Therefore, we conducted a secondary analysis of the Project RED clinical trial dataset to assess the association between a positive depression screen during inpatient hospitalization and the rate of subsequent hospital utilization.

Methods

Data from the Project RED clinical trial were reviewed for inclusion in a secondary analysis. Complete data were available for 738 of the 749 subjects recruited for Project RED.

Project RED Setting and Participants

Project RED was a two‐armed randomized controlled trial of English‐speaking adult patients, 18 years or older, admitted to the teaching service of Boston Medical Center, a large urban safety‐net hospital with an ethnically diverse patient population. A total of 749 subjects were enrolled and randomized between January 3, 2006 and October 18, 2007. Patients were required to have a telephone, be able to comprehend study details and the consent process in English, and have plans to be discharged to a US community. Patients were not enrolled if they were admitted from a skilled nursing facility or other hospital, transferred to a different hospital service prior to enrollment, admitted for a planned hospitalization, on hospital precautions, on suicide watch, deaf or blind. The Institutional Review Board of Boston University approved all study activities. A full description of the methods for the Project RED trial has been described previously.2

Outcome Variable

The primary endpoint was rate of hospital utilization within 30 days of discharge from the index admission, defined as the total number of ED visits and readmissions per subject within 30 days of the index discharge. Hospital utilization rates within 60 and 90 days of the index hospitalization discharge were also analyzed as secondary outcomes. Any ED visit in which a subject was subsequently admitted to the hospital was only counted as a readmission. Outcome data were collected by reviewing the hospital's electronic medical records (EMRs) and by contacting subjects by telephone 30 days after discharge. Dates of hospital utilization occurring at Boston Medical Center were obtained from the EMR, while those at other hospitals were collected through subject report. Subjects who could not be reached within 60 days of discharge were assumed alive.

Primary Independent Variable

The primary independent variable of interest was depressive symptoms defined as a positive score for minor or major depression on the nine‐item Patient Health Questionnaire (PHQ‐9) depression screening tool.20 A dichotomized variable was created using a standardized scoring system to determine the screening cut‐off for major or minor depressive symptoms.19

Statistical Analysis

Demographic and other characteristics of the subjects were compared by depression status (Table 1). Potential confounders were identified a priori from the available literature on factors associated with rehospitalization. These included age, gender, marital status, health literacy score (rapid estimate of health literacy in adult medicine tool [REALM]),21 Charlson score,22 insurance type, employment status, income level, homelessness status within past three months, hospital utilization within the 6 months prior to the index hospitalization, educational attainment, length of hospital stay and Project RED study group assignment. Bivariate analyses were conducted to determine which covariates were significant confounders of the relationship between depression and hospital utilization within 30 days of discharge. Chi‐square tests were used for categorical variables and t‐tests for continuous variables.

Baseline Characteristics of Study Subjects by Depression Screen Status
CharacteristicDepression Screen*
Negative (n = 500)Positive (n = 238)P Value
  • NOTE: Some columns may not add up to 100% due to omission of Other categories.

  • Abbreviations: PCP, primary care provider; PHQ9, Patient Health Questionnaire‐9; REALM, Rapid Estimate of Health Literacy in Adult Medicine tool; SD, standard deviation.

  • Positive depressive symptom screen determined by PHQ9 screen tool, a nine‐item 4‐point Likert scale, standard scoring algorithm to screen for major and minor depression. A score of 5 or higher indicates a positive depression symptom screen.17

  • Free Care is a Massachusetts state program for uninsured patients.

  • Health literacy categories correspond to total score as determined by REALM.18

  • Have PCP refers to subject self‐identifying PCP at time of Project RED study enrollment.

  • Unmarried marital status includes subjects identified as divorced, widow, single, partnered.

  • Charlson Comorbity Index Score reflects the cumulative increased likelihood of 1‐year mortality. The higher the score the more severe the comorbid condition. A 33% increase in risk for death is reflected in a 1‐point increase in weights. The minimum score is 0. There is no maximum score.19

  • Project RED study intervention group refers to subjects who received the 3‐armed discharge intervention.

  • Frequent Utilizer is defined as a subject with 2 or more hospital utilizations in 6 months prior to Project RED clinical trial index admission.

Race, No. (%)   
White140 (30)66 (30) 
Black268 (58)117 (54) 
Hispanic47 (10)29 (13)0.760
Insurance, No. (%)   
Private95 (19)22 (9) 
Medicare69 (14)30 (13) 
Medicaid214 (43)143 (61) 
Free care118 (24)40 (17)<0.001
Education, No. (%)   
<8th grade33 (7)21 (9) 
Some high school82 (17)52 (22) 
High school grad192 (38)90 (38) 
Some college126 (25)51 (22) 
College grad67 (13)22 (9)0.135
Health Literacy   
Grade 3 and below64 (13)44 (19) 
Grade 4654 (11)22 (10) 
Grade 78156 (32)73 (32) 
Grade 9 and above213 (44)89 (39)0.170
Income, $, No. (%)   
No income61 (12)37 (16) 
<10K77 (15)61 (26) 
1020K96 (19)35 (15) 
2050K97 (19)34 (14) 
50100K35 (8)7 (2) 
No answer132 (27)64 (27)0.002
Employment status, No. (%)   
Full time142 (28)34 (14) 
Part time57 (11)30 (13) 
Not Working297 (59)171 (72)<0.001
Age, mean (SD), years49.9 (16.0)49.6 (13.3)0.802
Gender: No. (%) Female239 (48)133 (56)0.040
Have PCP, No. (%) Yes399 (80)197 (83)0.340
Marital status, No. (%) unmarried365 (73)201 (85)<0.001
Charlson score, mean (SD)1.058 (1.6)1.56 (2.39)0.001
RED study group,# No. (%)   
Intervention243 (49)127 (53)0.22
Length of stay, days, mean (SD)2.5 (2.8)3.1 (3.8)0.016
Homeless in last 3 months, No. (%)45 (9)30 (13)0.130
Frequent utilizer,** No. (%)159 (32)104 (44)0.002

Age, length of stay, and Charlson score were used as continuous variables. Gender, marital status, frequent prior utilization (01 vs. 2 or more), and homelessness were treated as dichotomous variables. Categorical variables were created for, educational attainment (less than eighth grade, some high school, high school graduate, some college, college graduate), insurance type (Medicare, Medicaid, private insurance or free care), income level (no income, less than $10,000 per year, $10,00020,000, $20,00050,000, $50,000100,000, no answer), level of health literacy (grade 3 and below, grade 46, grade 78, grade 9 or above) and employment status(working full‐time, working part‐time, not working, no answer).

The 30‐day hospital utilization rate reflects the number of hospital utilization events within 30 days of discharge per subject. The same method was used to calculate hospital utilization rates within 60 and 90 days of discharge respectively. The unadjusted incident rate ratio (IRR) was calculated as the ratio of the rate of hospital utilizations among patients with depressive symptoms versus patients without depressive symptoms. Data for hospital utilization at 30, 60, and 90 days are cumulative.

Poisson models were used to test for significant differences between the predicted and observed number of hospitalization events at 30 days. A backward stepwise regression was conducted to identify and control for relevant confounders and construct the final, best‐fit model for the association between depression and hospital reutilization. A statistical significance level of P = 0.10 was used for the stepwise regression. To evaluate potential interactions between depression and the Project RED intervention, interaction terms were included. Two‐sided significance tests were used. P values of less than 0.05 were considered to indicate statistical significance. All data were analyzed with S‐Plus 8.0 (Seattle, WA).

In addition, a Kaplan‐Meier hazard curve was generated for the first hospital utilization event, ED visit or readmission, for the 30‐day period following discharge and compared with a log‐rank test.

Results

A total of 28% of subjects were categorized as having a positive depression screen. More women (36%) had positive depression screens than men (28%). Among patients with a positive depression screen, 58% had a history of depression and 53% were currently taking medications at the time of enrollment, compared with 25% and 22% respectively for subjects with a negative depression screen. Table 1 presents the means or percentages for baseline characteristics by depression status in the analytic cohort. Subjects with Medicaid for insurance had a higher rate of depression (61%) than subjects with Medicare (13%), private insurance (9%), or those who qualified for the Free Care pool (17%) which is the Massachusetts state funding for healthcare to uninsured persons. Subjects who were unemployed, unmarried, or who reported earnings less than $10,000 per year were also more likely to screen positive for depression. In addition, depressed subjects had a higher severity of co‐morbid disease and longer length of stay for the index hospitalization. Patients categorized as frequent utilizers (2 or more prior admissions) for the 6 months prior to the index hospitalization were also more likely to be depressed. Of further note, is the relatively younger average age among both depressive patients (49.6 years) and non‐depressive patients (49.9) of these study subjects.

The unadjusted hospital utilization rate at 30, 60, and 90 days post‐discharge by depression status is shown in Table 2. At 30 days post‐discharge, those with depressive symptoms had a higher rate of hospital utilization than those without depressive symptoms (0.563 vs. 0.296). In other words, 56 utilization events occurred per 100 patients with depressive symptoms, compared with 30 utilization events per 100 patients without depressive symptoms. The unadjusted 30‐day post‐discharge hospital utilization rate among those with depressive symptoms was higher compared with those without symptoms (IRR, 1.90, 95% confidence interval [CI], 1.242.71). A similar trend was found among subjects at 60 and 90 days post‐discharge.

Number of Hospital Utilizations, Hospital Utilization Rate, Unadjusted IRR at 30, 60 and 90 Days by Depression Screen Status*
Hospital UtilizationDepression Screen*P ValueIRR (CI)
Negative, n = 500 (68%)Positive, n = 238 (32%)
  • Abbreviations: CI, confidence interval; ED, emergency department; IRR, incident rate ratio.

  • Depression screen determined by scoring of Patient Health Questionnaire‐9 (PHQ9). Depressive symptom score of 5 points or higher is designated as positive.17

  • Number of hospital utilizations include all ED visits and hospital readmissions following discharge from Project RED index admission. ED visits leading to hospital admission are counted as one event. Sum reflects cumulative number of events over 30, 60 and 90 days.

No. of hospital utilizations140134 1.90 (1.51,2.40)
30‐day hospital utilization rate0.2960.563<0.001 
No. of hospital utilizations231205 1.87 (1.55,2.26)
60‐day hospital utilization rate0.4630.868<0.001 
No. of hospital utilizations324275 1.79 (1.53,2.10)
90‐day hospital utilization rate0.6481.165<0.001 

Poisson regression analyses were conducted to control for potential confounding in the relationship between depressive symptoms and hospital utilization rate within 30 days after discharge (Table 3). After controlling for relevant confounders, including age, gender, employment status, frequent prior hospitalization status, marital status, Charlson score, Project RED study group assignment and the interaction variable for RED study group assignment and depression, the association between symptoms of depression, and hospital utilization rate remained significant (IRR, 1.73; 95% CI, 1.272.36).

Adjusted Incident Rate Ratio of Hospital Utilization Within 30 Days of Discharge
CharacteristicsIRRCIP Value
  • Abbreviations: CI, confidence interval; ED, emergency department; IRR, incident rate ratio, PHQ9, Patient Health Questionnaire‐9.

  • Positive depressive symptom screen determined by PHQ9 screen tool, a nine‐item 4‐point Likert scale, standard scoring algorithm to screen for major and minor depression. A score of 5 or higher indicates a positive depression symptom screen.17

  • Unmarried refers to subjects whose self‐reported marital status includes divorced, single, partnered or widowed.

  • Frequent utilizer: 2 or more ED or hospital admissions visits in prior 6 months from index admission.

  • Refers to Project RED study group assignment.

  • Charlson Comorbity Index Score reflects the cumulative increased likelihood of 1‐year mortality. The higher the score the more severe the comorbid condition. A 33% increase in risk for death is reflected in a 1‐point increase in weights. The minimum score is 0. There is no maximum score.19

  • Interaction term of Project RED study group assignment and depressive symptom category (positive or negative).

Depression symptoms*  <0.001
Positive1.731.272.36 
NegativeREF1.0 
Gender  <0.001
Male1.871.472.40 
FemaleREF1.0 
Marital status  0.005
Married0.6250.440.89 
Unmarried1.0REF 
Frequent utilizer  <0.001
2+ prior visits2.451.923.15 
<2 prior visits1.0REF 
Study group  0.054
Intervention0.760.551.06 
Control1.0REF 
Employment   
Part time1.400.852.300.095
Not working1.671.152.440.003
Other0.520.073.850.262
Full time1.0REF 
Charlson Score0.980.921.040.250
Group* depression0.840.521.360.236
Age1.000.991.010.375

Figure 1 depicts the Kaplan‐Meier hazard curve generated for time to first hospital utilization, stratified by depression status. While 21% of participants without symptoms of depression had a hospital utilization within 30 days, fully 29% of participants with symptoms of depression had a hospital utilization within 30 days (P = 0.011).

Figure 1
Hazard for hospital utilization among subjects with and without depressive symptoms in 30 days following hospital discharge.

Discussion

Our study shows hospitalized patients who screen positive for depressive symptoms are significantly more likely to have a hospital visit (emergency room or rehospitalization) within 30 days of discharge than those who do not screen positive for depressive symptoms among medical patients admitted to an urban, academic, safety‐net hospital. These findings are consistent with, and extend, prior reports regarding depression and rehospitalization in specific populations (ie, geriatrics) and specific diagnoses (ie, cardiovascular disease [CVD] and COPD).1012 We observed a 73% higher incidence rate for hospital utilization within 30 days of discharge for those with symptoms of depression. This puts symptoms of depression on par with frequent prior rehospitalization, advanced age and low social support, as known risk factors for rehospitalization.4, 5, 23

Also of significance is the relatively young age of this study population (49.9 years non‐depressive patients and 49.6 years for depressive patients) compared with the study cohorts used for research in the majority of the existing literature. The chief reason for the young age of our cohort is that potential subjects were excluded if they came from a skilled nursing facility or other hospital. This may limit the generalizability of our findings; however, it seems likely that interventions relating to depression and transitions of care will need to be quite different for patients that reside in long‐term care facilities vs. patients that live in the community. For example, patients living in the community may have significant barriers to access post‐discharge services due to insurance status and are more likely to be sensitive to variations in social support.

Early rehospitalization is associated with significant morbidity, mortality, and expense. It is also a potential marker for poor quality of care.24 Concerns for patient safety, escalating healthcare costs, and possible change in hospital reimbursement mechanisms are fueling the search for modifiable risk factors associated with early rehospitalization. Our data provide evidence that symptoms of depression may be an important focus of attention. We do not know, however whether treating hospitalized patients who screen positive for depression will decrease early rehospitalization and emergency room utilization rates.

Various physiologic and behavioral mechanisms may link symptoms of depression to hospital utilization after discharge. For example, depressed patients with features of somatization may be more likely to experience worrisome physical symptoms after discharge and present prematurely for reevaluation. Patients who are sicker in some fashion not captured by our measured confounders may have symptoms of depression related to chronic, debilitating disease warranting early return to the hospital. Depression may also yield nonadherence to aspects of the discharge treatment plan leading to rehospitalization as a result of poor post‐discharge disease management. For example, research shows that patients with depression following coronary artery bypass surgery are less likely to adhere with cardiac rehabilitation programs.25 Likewise, depression among chronically ill patients such as diabetics, asthmatics, or human immunodeficiency virus (HIV)‐positive patients impairs medication adherence and self‐care behavior which may lead to disease relapse or recurrence.2628 One study examining depression effects on hypertensive medicine adherence in African Americans identified self‐efficacy as a mediating factor between depression and nonadherence.29 This implies that interventions such as self‐management education, a program through which chronically‐ill patients learn to better manage their illnesses through enhanced self‐confidence and problem‐solving strategies (including mood disorder challenges) may reduce early rehospitalization among depressed patients.30

There is also evidence that depression may have direct physiologic consequences. In patients with CVD, depression is associated with poor outcomes possibly related to decreased heart rate variability, hypercoagulability, high burdens of inflammatory markers, and severity of left ventricular dysfunction.3134 Similarly, depression among HIV/acquired immune deficiency syndrome (AIDS), diabetics and multiple sclerosis (MS) patients is linked to heightened levels of proinflammatory markers and less favorable outcomes that may signal a more severe form of the disease or an impaired response to treatment.3538 Indeed, MS investigators now hypothesize that the proinflammatory environment associated with the neurologic manifestations of MS are also causing depression symptoms among MS patients.34 This theory contrasts the common belief that depression in the chronically ill manifests independent of the chronic illness or in response to living with chronic disease.

A major strength of the current study is the large dataset and the broad range of covariates available for analyses. However, several limitations should be noted. First, data on hospital utilization outside Boston Medical Center were determined by patient self‐report and were not confirmed by document review. Second, we do not know the direction of the associations we report. If symptoms of depression are merely the consequence of having a higher disease burden, treatment of the underlying disease may be the most important response. While this is possible, our model does include several variables (eg, Charlson score and length of stay) that are likely to adjust for disease severity, pointing to the likelihood that symptoms of depression truly predict hospital utilization in a fashion that is independent of disease severity. Third, our results may not be generalizable to populations other than those served by urban safety‐net hospitals or other populations excluded from the Project RED trial (eg, non‐English speaking patients and patients from nursing homes). Finally, social factors such as substance use and social support system variables may residually confound the relationship between depression and hospital reutilization demonstrated in this study. While this dataset does not include a measure of social support other than marital status and housing status, data is available on substance use. Analyses conducted by our colleagues using Project RED data found that in this study population depression was significantly more prevalent among substance users (29% vs. 14%) compared with non‐users and that substance use is an independent risk factor for hospital reutilization (unpublished data).

Our findings linking depression to increased hospital utilization also warrant further consideration from healthcare policymakers. Central to the Obama Administration's February 2009 healthcare reform proposal is the pursuit of cost savings through reductions in unplanned hospital readmissions.39 Thus, identifying potentially modifiable risk factors for readmission, such as depression, is of great concern to healthcare providers and policymakers across the nation. If, through testing of interventions, depression proves to be a modifiable risk for readmission, policymakers, while negotiating healthcare reform measures, must provide for the services required to address this comorbidity at the time of discharge. For example, if a patient screens positive for depressive symptoms during a hospitalization for COPD exacerbation, will the proposed payment reforms allow for mental health services during the immediate post‐discharge period in order to reduce the likelihood of hospital readmission? Will those mental health services be readily available? Payment reforms that account for all necessary transitional care services will indeed help reduce readmission costs with less risk for untoward consequences.

In conclusion, our results indicate that a positive depression screen is a significant risk factor for early post‐discharge hospital utilization among hospitalized adults on a general medical service, even after controlling for relevant confounders. Screening for depression during acute hospitalizations may be an important step in identifying patients at increased risk for readmission. Future research should focus on further characterizing and stratifying populations at highest risk for depression. Efforts should also include developing and evaluating targeted interventions for patients with symptoms of depression among hospitalized patients as part of discharge planning. Timely depression therapy during the hospitalization or following hospital discharge might reduce costly readmissions and enhance patient safety.

References
  1. Jenks SF,Williams MV,Coleman EA.Rehospitalizations among patients in the Medicare fee‐for‐service program.N Engl J Med.2009;360(14):14571459.
  2. Jack BW,Chetty VK,Anthony D, et al.The reengineered hospital discharge program to decrease rehospitalization.Ann Intern Med.2009;150(3):178187.
  3. Weissman JS,Stern RS,Epstein AM.The impact of patient socioeconomic status and other social factors on readmission: a prospective study in four Massachusetts hospitals.Inquiry.1994;31(2):163172.
  4. van Walraven C,Mamdani M,Fang J,Austin PC.Continuity of care and patient outcomes after hospital discharge.J Gen Intern Med.2004;19:624631. [PMID: 15209600]
  5. Marcantonio ER,McKean S,Goldfinger M,Kleenfield S,Yurkofsky M,Brennan TA.Factors associated with unplanned hospital readmission among patients 65 years of age and older in a Medicare managed care plan.Am J Med.1999;107(1):1317.
  6. Krumholz HM,Parent EM,Tu N, et al.Readmission after hospitalization for congestive heart failure among Medicare beneficiaries.Arch Intern Med.1997;157(1):99104.
  7. Librero J,Peiro S,Ordinana R.Chronic comorbidity and outcomes of hospital care: length of stay, mortality and readmission at 30 and 365 days.J Clin Epidemiol.1999;52(3):171179.
  8. Rodríguez‐Artalejo F,Guallar‐Castillón P,Herrera MC, et al.Social network as a predictor of hospital readmission and mortality among older patients with heart failure.J Card Fail.2006;12:621627.
  9. Wong AW,Gan WQ,Burns J,Sin DD,van Eeden SF.Acute exacerbation of chronic obstructive pulmonary disease: influence of social factors in determining length of stay and readmission rates.Can Respir J.2008;15(7):361364.
  10. Parashar S,Rumsfeld JS,Spertus JA, et al.Time course of depression and outcome of myocardial infarction.Arch Intern Med.2006;166(18):20352043.
  11. Budpitz DS,Shebab N,Kegler SR, et al.Medication use leading to emergency department visits for adverse drug events in older adults.Ann Intern Med.2007;147(11):755765.
  12. Campbell SE,Seymour DG,Primrose WR.A systematic literature review of factors affecting outcomes in older medical patients admitted to hospital.Age Ageing.2004;33(2):110115.
  13. Kartha A,Anthony D,Manasseh CS, et al.Depression is a risk factor for rehospitalization in medical inpatients.Prim Care Companion J Clin Psychiatry.2007;9(4):256262.
  14. Almagro P,Barreiro Bienvenido ,Ochoa de Echaguen A, et al.Risk factors for hospital readmission in patients with chronic obstructive pulmonary disease.Respiration.2006;73:311317.
  15. Frasure‐Smith N,Lesperance F,Gravel G, et al.Depression and healthcare costs during the first year following myocardial infarction.J Psychosom Res.2000;48(4–5):471478.
  16. Jiang W,Alexander J,Christopher E, et al.Relationship of depression to increased risk of mortality and rehospitalization.Arch Intern Med.2001;161(15):18491856.
  17. Parashar S,Rumsfeld JS,Spertus JA, et al.Time course of depression and outcome of myocardial infarction.Arch Intern Med.2006;166:20352043.
  18. Scherer M,Herrmann‐Lingen C.Single item on positive affect is associated with 1‐year survival in consecutive medical inpatients.J Gen Hosp Psych.2009;31:813.
  19. Hasin DS,Goodwin RD,Stinson FS,Grant BF.Epidemiology of major depressive disorder: results from the National Epidemiologic Survey on Alcoholism and Related Conditions.Arch Gen Psychiatry.2005;62(10):1097106.
  20. Kroenke K,Spitzer RL,Williams JB.The PHQ‐9: Validity of a brief depression severity measure.J Gen Intern Med.2001;16:606613. [PMID:11556941]
  21. Davis TC,Long SW,Jackson RH, et al.Rapid estimate of adult literacy in medicine: a shortened screening instrument.Fam Med.1993;25:391395. [PMID:8349060]
  22. Charlson ME,Pompei P,Ales KL,MacKenzie CR.A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis.1987;40:373383. [PMID: 3558716]
  23. Rodriguez‐Artalejo F,Guallar‐Castillon P,Herrera MC, et al.Social network as a predictor of hospital readmission and mortality among older patients with heart failure.J Card Fail.2006;12(8):621627.
  24. Ashton CM,Del Junco DJ,Souchek J,Wray NP,Mansyr CL.The association between the quality of inpatient care and early readmission: a meta‐analysis of the evidence.Med Care.1997;35(10):10441059.
  25. Kronish IM,Rieckmann N,Halm FA, et al.Persistent depression affects adherence to secondary prevention behaviors after acute coronary syndromes.J Gen Intern Med.2006;21(11):11781183.
  26. Cukor D,Rosenthal DS,Jindal RM,Brown CD,Kimmel PL.Depression is an important contributor to low medication adherence in hemodialyzed patients and transplant recipients.Kidney Int.2009;75(11):12231229.
  27. Gonzalez JS,Safren SA,Delahanty LM, et al.Symptoms of depression prospectively predict poorer self‐care in patients with Type 2 diabetes.Diabet Med.2008;25(9):11021107.
  28. Lima VD,Geller J,Bangsberg DR, et al.The effect of adherence on the association between depressive symptoms and mortality among HIV‐infected individuals first initiating HAART.AIDS.2007;21(9):11751183.
  29. Schoenthaler A,Ogedegbe G,Allegrante JP.Self‐efficacy mediates the relationship between depressive symptoms and medication adherence.Health Educ Behav.2009;36(1):127137.
  30. Bodenheimer T,Lorig K,Holman H,Grumbach K.Patient self‐management of chronic disease in primary care.JAMA.2002;288(19):24692475.
  31. McFarlane AM.Effects of sertraline on the recovery rate of cardiac autonomic function in depressed patients after acute myocardial infarction.Am Heart J.2001;142:617623.
  32. van Melle JP,de Jonge P,Ormel J, et al.Relationship between left ventricular dysfunction and depression following myocardial infarction: data from the MIND‐IT.Eur Heart J.2005;26:26502656.
  33. Serebruany VL,Glassman AH,Malinin AI, et al.Platelet/endothelial biomarkers in depressed patients treated with the selective serotonin reuptake inhibitor sertraline after acugte coronary events: the Sertraline AntiDepressant Heart Attack Randomized Trial (SADHART) Platelet SubStudy.Circulation.2003;108:939944.
  34. Mulvihill NT,Foley JB.Inflammation in acute coronary syndromes.Cleve Clin J Med.2002;69(Suppl2):SII130SII142.
  35. Gold SM,Irwin MR.Depression and immunity: inflammation and depressive symptoms in multiple sclerosis.Neurol Clin.2006;24(3):507519.
  36. Brydon L,Walker C,Wawrzyniak A, et al.Synergistic effects of psychological and immune stressors on inflammatory cytokines and sickness responses in humans.Brain Behav Immun.2009;23(2):217224.
  37. Gresson JM,Hurwitz BE,Llabre MM,Schneiderman N,Penedo FJ,Klimas NG.Psychological distress, killer lymphocytes and disease severity in HIV/AIDS.Brain Behav Immun.2008;22(6):901911.
  38. Pizzi C,Manzoli L,Mancini S,Costa GM.Analysis of potential predictors of depression among coronary heart disease risk factors including heart rate variability, markers of inflammation, and endothelial function.Eur Heart J.2008;29(9):11101117.
  39. Connolly C.Obama proposes $634 billion fund for health care.Washington Post. February 26,2009:A1.
References
  1. Jenks SF,Williams MV,Coleman EA.Rehospitalizations among patients in the Medicare fee‐for‐service program.N Engl J Med.2009;360(14):14571459.
  2. Jack BW,Chetty VK,Anthony D, et al.The reengineered hospital discharge program to decrease rehospitalization.Ann Intern Med.2009;150(3):178187.
  3. Weissman JS,Stern RS,Epstein AM.The impact of patient socioeconomic status and other social factors on readmission: a prospective study in four Massachusetts hospitals.Inquiry.1994;31(2):163172.
  4. van Walraven C,Mamdani M,Fang J,Austin PC.Continuity of care and patient outcomes after hospital discharge.J Gen Intern Med.2004;19:624631. [PMID: 15209600]
  5. Marcantonio ER,McKean S,Goldfinger M,Kleenfield S,Yurkofsky M,Brennan TA.Factors associated with unplanned hospital readmission among patients 65 years of age and older in a Medicare managed care plan.Am J Med.1999;107(1):1317.
  6. Krumholz HM,Parent EM,Tu N, et al.Readmission after hospitalization for congestive heart failure among Medicare beneficiaries.Arch Intern Med.1997;157(1):99104.
  7. Librero J,Peiro S,Ordinana R.Chronic comorbidity and outcomes of hospital care: length of stay, mortality and readmission at 30 and 365 days.J Clin Epidemiol.1999;52(3):171179.
  8. Rodríguez‐Artalejo F,Guallar‐Castillón P,Herrera MC, et al.Social network as a predictor of hospital readmission and mortality among older patients with heart failure.J Card Fail.2006;12:621627.
  9. Wong AW,Gan WQ,Burns J,Sin DD,van Eeden SF.Acute exacerbation of chronic obstructive pulmonary disease: influence of social factors in determining length of stay and readmission rates.Can Respir J.2008;15(7):361364.
  10. Parashar S,Rumsfeld JS,Spertus JA, et al.Time course of depression and outcome of myocardial infarction.Arch Intern Med.2006;166(18):20352043.
  11. Budpitz DS,Shebab N,Kegler SR, et al.Medication use leading to emergency department visits for adverse drug events in older adults.Ann Intern Med.2007;147(11):755765.
  12. Campbell SE,Seymour DG,Primrose WR.A systematic literature review of factors affecting outcomes in older medical patients admitted to hospital.Age Ageing.2004;33(2):110115.
  13. Kartha A,Anthony D,Manasseh CS, et al.Depression is a risk factor for rehospitalization in medical inpatients.Prim Care Companion J Clin Psychiatry.2007;9(4):256262.
  14. Almagro P,Barreiro Bienvenido ,Ochoa de Echaguen A, et al.Risk factors for hospital readmission in patients with chronic obstructive pulmonary disease.Respiration.2006;73:311317.
  15. Frasure‐Smith N,Lesperance F,Gravel G, et al.Depression and healthcare costs during the first year following myocardial infarction.J Psychosom Res.2000;48(4–5):471478.
  16. Jiang W,Alexander J,Christopher E, et al.Relationship of depression to increased risk of mortality and rehospitalization.Arch Intern Med.2001;161(15):18491856.
  17. Parashar S,Rumsfeld JS,Spertus JA, et al.Time course of depression and outcome of myocardial infarction.Arch Intern Med.2006;166:20352043.
  18. Scherer M,Herrmann‐Lingen C.Single item on positive affect is associated with 1‐year survival in consecutive medical inpatients.J Gen Hosp Psych.2009;31:813.
  19. Hasin DS,Goodwin RD,Stinson FS,Grant BF.Epidemiology of major depressive disorder: results from the National Epidemiologic Survey on Alcoholism and Related Conditions.Arch Gen Psychiatry.2005;62(10):1097106.
  20. Kroenke K,Spitzer RL,Williams JB.The PHQ‐9: Validity of a brief depression severity measure.J Gen Intern Med.2001;16:606613. [PMID:11556941]
  21. Davis TC,Long SW,Jackson RH, et al.Rapid estimate of adult literacy in medicine: a shortened screening instrument.Fam Med.1993;25:391395. [PMID:8349060]
  22. Charlson ME,Pompei P,Ales KL,MacKenzie CR.A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis.1987;40:373383. [PMID: 3558716]
  23. Rodriguez‐Artalejo F,Guallar‐Castillon P,Herrera MC, et al.Social network as a predictor of hospital readmission and mortality among older patients with heart failure.J Card Fail.2006;12(8):621627.
  24. Ashton CM,Del Junco DJ,Souchek J,Wray NP,Mansyr CL.The association between the quality of inpatient care and early readmission: a meta‐analysis of the evidence.Med Care.1997;35(10):10441059.
  25. Kronish IM,Rieckmann N,Halm FA, et al.Persistent depression affects adherence to secondary prevention behaviors after acute coronary syndromes.J Gen Intern Med.2006;21(11):11781183.
  26. Cukor D,Rosenthal DS,Jindal RM,Brown CD,Kimmel PL.Depression is an important contributor to low medication adherence in hemodialyzed patients and transplant recipients.Kidney Int.2009;75(11):12231229.
  27. Gonzalez JS,Safren SA,Delahanty LM, et al.Symptoms of depression prospectively predict poorer self‐care in patients with Type 2 diabetes.Diabet Med.2008;25(9):11021107.
  28. Lima VD,Geller J,Bangsberg DR, et al.The effect of adherence on the association between depressive symptoms and mortality among HIV‐infected individuals first initiating HAART.AIDS.2007;21(9):11751183.
  29. Schoenthaler A,Ogedegbe G,Allegrante JP.Self‐efficacy mediates the relationship between depressive symptoms and medication adherence.Health Educ Behav.2009;36(1):127137.
  30. Bodenheimer T,Lorig K,Holman H,Grumbach K.Patient self‐management of chronic disease in primary care.JAMA.2002;288(19):24692475.
  31. McFarlane AM.Effects of sertraline on the recovery rate of cardiac autonomic function in depressed patients after acute myocardial infarction.Am Heart J.2001;142:617623.
  32. van Melle JP,de Jonge P,Ormel J, et al.Relationship between left ventricular dysfunction and depression following myocardial infarction: data from the MIND‐IT.Eur Heart J.2005;26:26502656.
  33. Serebruany VL,Glassman AH,Malinin AI, et al.Platelet/endothelial biomarkers in depressed patients treated with the selective serotonin reuptake inhibitor sertraline after acugte coronary events: the Sertraline AntiDepressant Heart Attack Randomized Trial (SADHART) Platelet SubStudy.Circulation.2003;108:939944.
  34. Mulvihill NT,Foley JB.Inflammation in acute coronary syndromes.Cleve Clin J Med.2002;69(Suppl2):SII130SII142.
  35. Gold SM,Irwin MR.Depression and immunity: inflammation and depressive symptoms in multiple sclerosis.Neurol Clin.2006;24(3):507519.
  36. Brydon L,Walker C,Wawrzyniak A, et al.Synergistic effects of psychological and immune stressors on inflammatory cytokines and sickness responses in humans.Brain Behav Immun.2009;23(2):217224.
  37. Gresson JM,Hurwitz BE,Llabre MM,Schneiderman N,Penedo FJ,Klimas NG.Psychological distress, killer lymphocytes and disease severity in HIV/AIDS.Brain Behav Immun.2008;22(6):901911.
  38. Pizzi C,Manzoli L,Mancini S,Costa GM.Analysis of potential predictors of depression among coronary heart disease risk factors including heart rate variability, markers of inflammation, and endothelial function.Eur Heart J.2008;29(9):11101117.
  39. Connolly C.Obama proposes $634 billion fund for health care.Washington Post. February 26,2009:A1.
Issue
Journal of Hospital Medicine - 5(7)
Issue
Journal of Hospital Medicine - 5(7)
Page Number
378-384
Page Number
378-384
Publications
Publications
Article Type
Display Headline
Post‐discharge hospital utilization among adult medical inpatients with depressive symptoms
Display Headline
Post‐discharge hospital utilization among adult medical inpatients with depressive symptoms
Legacy Keywords
depression, hospital discharge, patient safety, readmission, rehospitalization
Legacy Keywords
depression, hospital discharge, patient safety, readmission, rehospitalization
Sections
Article Source

Copyright © 2010 Society of Hospital Medicine

Disallow All Ads
Correspondence Location
Department of Family Medicine, Dowling 5, Boston Medical Center, 1 BMC Place, Boston, MA 02118
Content Gating
No Gating (article Unlocked/Free)
Alternative CME
Article PDF Media