User login
Risk Factors for Discharge to Rehabilitation Among Hip Fracture Patients
Length of stay (LOS) is a significant driver of costs after hip fracture surgery.1-3 Multiple studies have identified factors associated with increased LOS in hip fracture patients. These factors include admission time, delay to surgery, presence of comorbidities, and older age.4-9
One significant and potentially modifiable factor affecting LOS is delayed transfer to a rehabilitation center after surgery.8-11 Although patients after orthopedic surgeries require additional rehabilitation services or subacute care directly attributable to their injuries, specialized rehabilitation centers may not always have beds readily available.6-11 Studies have shown that delays in transfer to skilled nursing facilities or rehabilitation centers are highly common among orthopedic patients.8 It is therefore imperative that orthopedists have a mechanism for predicting and identifying which patients require rehabilitation services early in the postoperative period. Identifying risk factors and stratifying patients who are most likely to require rehabilitation would facilitate the early transfer of these patients and thereby directly decrease LOS and hospitalization-related costs.
In this article, we report results from prospective, national, multicenter data to identify commonly measured risk factors for discharge to rehabilitation facilities for hip fracture patients. Through multivariate analysis of ACS-NSQIP (American College of Surgeons National Surgical Quality Improvement Program) data, we determined which risk factors significantly predispose patients to discharge to rehabilitation centers versus discharge home. Knowledge of these risk factors allows the practicing orthopedist to be better equipped to identify patients who require additional rehabilitation early in the postoperative course. By mobilizing case managers and social workers to help avoid delays in the transfers of these identified patients, LOS-associated costs may ultimately decrease.
Materials and Methods
After obtaining institutional review board approval for this study from the Office of Research at Vanderbilt University, we prospectively collected 2011 discharge data from the ACS-NSQIP database (these data are unavailable for earlier years). All patients who underwent hip fracture surgery in 2011 were identified by CPT (Current Procedural Terminology) codes. Cases of patients with unknown discharge information and of those who died during their hospitalizations were excluded from analysis. For the remaining patients, discharge information as categorized by ACS-NSQIP included skilled care (eg, subacute hospital, skilled nursing home), unskilled facility (eg, nursing home, assisted facility), separate acute care, and rehabilitation. All other patients were discharged home without additional assistance or to the previous home where they received chronic care, assisted living, or unskilled aid. Patients were dichotomized according to whether they were discharged home or to one of the rehabilitation facilities mentioned.
To determine which risk factors significantly contributed to a patient’s discharge to rehabilitation, we ran univariate analyses using Fisher exact tests for categorical variables and Student t tests for continuous variables on multiple patient factors, including demographics, preoperative comorbidities, and operative factors. Demographics included age and sex. Preoperative comorbidities included 32 conditions: diabetes mellitus, active smoking status, current alcohol use, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, hypertension requiring medication, history of esophageal varices, history of myocardial infarction, current renal failure, current dialysis dependence, steroid use, recent weight loss, existing bleeding disorder, transfusion before discharge, presence of central nervous system tumor, recent chemotherapy, recent radiation therapy, previous percutaneous coronary intervention, previous percutaneous coronary stenting, history of angina, peripheral vascular disease, cerebrovascular accidents, recent surgery (within 30 days), rest pain, impaired sensorium, history of transient ischemic attacks, current hemiplegia status, current paraplegia status, current quadriplegia status, current ascites, hypertension, and disseminated cancer. Operative factors included wound infection, DNR (do not resuscitate) status, ventilator support, anesthesia type, wound class, ASA (American Society of Anesthesiologists) class, and operative time.
For the univariate analyses, significance was set at P < .05. Demographics, preoperative comorbidities, and operative factors that were significantly associated with discharge to a rehabilitation facility in the univariate analysis were selected as covariates for a multivariate analysis. We incorporated a binary logistic regression to analyze which of these significant risk factors are correlated with a patient’s discharge to a rehabilitation facility after hip fracture surgery.
Results
A total of 4974 patients undergoing surgery for hip fractures in 2011 were identified. Of these patients, 4815 had complete information on discharge location and were included in the analysis.
Table 1 lists the results of the univariate analysis comparing demographics, preoperative comorbidities, and operative factors between the home and rehabilitation groups. Both age (P < .001) and sex (P = .012) were significantly different between groups; the rehabilitation group was older by about 10 years and included significantly more females. In addition to demographic factors, 16 preoperative comorbidities, and 5 surgical factors were significantly associated with discharge to rehabilitation.
Surgery type significantly affected discharge to rehabilitation (Figure). Patients who were undergoing open plating of a femoral neck fracture or intramedullary nailing of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture constituted 30% of all patients discharged to rehabilitation centers. In contrast, patients undergoing percutaneous skeletal fixation of a proximal femoral fracture constituted only 5.5% of all patients discharged to rehabilitation. Based on surgery type, we broke down discharge location further, into categories of skilled nursing facility, unskilled facility (not patient’s previous home), separate acute-care facility, dedicated rehabilitation center, and home. Of all 4815 patients combined, 2102 (43.6%) were discharged to a skilled nursing facility, 31 (0.6%) to an unskilled facility (not home), 106 (2.2%) to separate acute care, 1312 (27.2%) to a dedicated rehabilitation center, and 950 (19.7%) home.
Table 2 lists the significant results from the multivariate logistical analysis comparing discharge to a rehabilitation center and discharge home after controlling for the significant risk factors (Table 1). Current diabetes, history of dyspnea, previous myocardial infarction, history of ischemic attacks, current bleeding disorder, transfusion during hospitalization, previous percutaneous cardiac stenting, chemotherapy, past cerebrovascular accident, presence of cancer, surgery type based on CPT code, history of chronic obstructive pulmonary disease or congestive heart failure, current smoking status, and operative time longer than 90 minutes were not significantly correlated with discharge to rehabilitation in the multivariate analysis. All significant factors were associated with higher odds of discharge to rehabilitation except for DNR status. DNR patients were 2.04 times more likely (95% CI, 1.49-2.78; P < .001) to be discharged home than to rehabilitation centers.
Applying these adjusted odds ratios, we see that an elderly woman (age, >65 years) who underwent general anesthesia with an ASA class higher than 2 was 17.63 times more likely than a patient without these risk factors to be discharged to rehabilitation. If this patient were also dialysis-dependent, she would be 61.52 times more likely than a similar patient without dialysis needs to be discharged to rehabilitation.
Even when controlling for all significant and nonsignificant variables in multivariate logistical analysis, age over 65 years (β = 1.05; P < .001), female sex (β = 1.76; P = .004), dialysis dependence (β = 12.98; P = .036), hypertension requiring medication (β = 1.53; P = .032), and ASA class higher than 2 (β = 1.98; P = .001) were found to be significant risk factors for discharge to rehabilitation.
Discussion
This study was the first to investigate the issue of which patient risk factors allow the practicing orthopedist to identify patients who require rehabilitation after hip fracture surgery. Through our multivariate analysis, which controlled for demographics, comorbidities, and operative factors, we found that older age, female sex, history of percutaneous coronary intervention, dialysis dependence, general anesthesia, and ASA class higher than 2 significantly increased the odds of discharge to a rehabilitation center versus home.
Using our study’s results, we can create a risk stratification model for patients and thereby a means of targeting patients who need rehabilitation and starting the process of finding a rehabilitation bed early in the postoperative course. Our study’s variables are easily measured metrics that may be collected in any hospital setting. Especially for hip fracture patients, early planning and discharge to the appropriate rehabilitation center are important in decreasing LOS and associated hospitalization costs. According to one report,3 about 85% of all hip fracture costs are directly related to LOS, given the unnecessarily long rehabilitation periods in hospitals. Hollingworth and colleagues2 compared costs for patients who remained in the hospital with costs for those discharged with rehabilitation services. Overall costs were significantly lower for patients discharged home with rehabilitation. The authors concluded that 40% of hip fracture patients may be suitable for early discharge.2 In an analysis of Medicare payments for hip fracture treatment, hospital costs including LOS accounted for 60% of all payments.12 The results of these 2 studies suggest that the overall driver of hip fracture costs is prolonged LOS and that, if patients are discharged to rehabilitation, then overall costs may be lowered through a direct reduction in hospital LOS. Given that hip fractures account for almost 350,000 hospital admissions in the United States each year, and using our institution’s average hospital charge per day ($4500), about $1.6 billion may be saved if each patient’s LOS decreased by 1 day.13 Although multiple factors affect LOS, discharge planning is under orthopedists’ direct control. Therefore, early identification of patients who will require rehabilitation may help reduce LOS-associated costs in our health care system.
The patient variables that were significantly associated with discharge to rehabilitation are also associated with increased morbidity and mortality in hip fracture patients, according to the literature,14-20 which provides some external validation of using these risk factors as predictors for rehabilitation. A patient with one of these risk factors may require rehabilitation, given that rehabilitation services are specifically linked to lower morbidity and mortality rates among hip fracture patients. For example, patients with dialysis needs were 3.49 times more likely to be discharged to a rehabilitation center in our study. In a 2000 study by Coco and Rush,16 hip fracture patients on dialysis had a 1-year mortality rate 2.5 times higher than that of patients who were not dialysis-dependent. In 2010, Cameron and colleagues17 found that cardiovascular disease was associated with a 2.68 times higher risk of mortality in hip fracture patients. Similarly in our study, both hypertension and history of percutaneous coronary intervention were associated with discharge to rehabilitation. We found higher odds of discharge to rehabilitation with higher ASA classes, which mirror results from a study by Michel and colleagues,15 who found that higher (vs lower) preoperative ASA classes were associated with higher 1-year mortality in hip fracture patients. Interestingly, DNR status was associated with higher odds of discharge home, which may reflect patients’ desires to forgo noninvasive or lifesaving procedures that may be performed at rehabilitation facilities. Although general anesthesia predisposed patients to discharge to a rehabilitation center, multiple studies have found no association between anesthesia type and postoperative mortality rates for hip fracture patients.18,19 Last, Marcantonio and colleagues20 found delirium specifically had a higher odds ratio for discharge, but our univariate analysis did not find a significant association between impaired sensorium and discharge location. Given the correlation of our risk factors with increased morbidity and mortality in the literature, our study’s results provide the initial groundwork for creating a risk calculator that orthopedists can use to predict discharge to rehabilitation.
Our study had some limitations. Although we analyzed a large number of demographics, preoperative comorbidities, and surgical factors, our univariate analysis was limited to information in the ACS-NSQIP database. We did not incorporate other clinically relevant factors (eg, social factors, including patients’ support networks) that may influence discharge decisions. Furthermore, ACS-NSQIP records patient data only up to 30 days after surgery. Discharge information for the time after that was missing for a subset of hip fracture patients, and these patients had to be excluded, potentially skewing our data. ACS-NSQIP also does not collect cost data for patients based on hospitalization or LOS, so we could not determine whether patients discharged to rehabilitation incurred higher costs because of longer hospitalizations.
Nevertheless, our study identified significant patient and operative variables that are associated with discharge to a rehabilitation center. By identifying hip fracture patients with these risk factors early and mobilizing the appropriate resources, practicing orthopedists should be better equipped to help facilitate the discharge of patients to the appropriate location after surgery. Validation of these risk factors should be prospectively determined with an analysis of LOS and cost implications. Use of a risk calculator may in fact result in decreased LOS and hospital-related costs. Furthermore, using these risk factors in a prospective patient cohort would help validate their use and determine whether there is clinical correlation. The orthopedists in our institution are becoming more aware of these risk factors, but validation is necessary.
1. Garcia AE, Bonnaig JV, Yoneda ZT, et al. Patient variables which may predict length of stay and hospital costs in elderly patients with hip fracture. J Orthop Trauma. 2012;26(11):620-623.
2. Hollingworth W, Todd C, Parker M, Roberts JA, Williams R. Cost analysis of early discharge after hip fracture. BMJ. 1993;307(6909):903-906.
3. Sund R, Riihimäki J, Mäkelä M, et al. Modeling the length of the care episode after hip fracture: does the type of fracture matter? Scand J Surg. 2009;98(3):169-174.
4. Fox KM, Magaziner J, Hebel JR, Kenzora JE, Kashner TM. Intertrochanteric versus femoral neck hip fractures: differential characteristics, treatment, and sequelae. J Gerontol A Biol Sci Med Sci. 1999;54(12):M635-M640.
5. Foss NB, Palm H, Krasheninnikoff M, Kehlet H, Gebuhr P. Impact of surgical complications on length of stay after hip fracture surgery. Injury. 2007;38(7):780-784.
6. Lefaivre KA, Macadam SA, Davidson DJ, Gandhi R, Chan H, Broekhuyse HM. Length of stay, mortality, morbidity and delay to surgery in hip fractures. J Bone Joint Surg Br. 2009;91(7):922-927.
7. Clague JE, Craddock E, Andrew G, Horan MA, Pendleton N. Predictors of outcome following hip fracture. Admission time predicts length of stay and in-hospital mortality. Injury. 2002;33(1):1-6.
8. Parker MJ, Todd CJ, Palmer CR, et al. Inter-hospital variations in length of hospital stay following hip fracture. Age Ageing. 1998;27(31):333-337.
9. Brasel KJ, Rasmussen J, Cauley C, Weigelt JA. Reasons for delayed discharge of trauma patients. J Surg Res. 2002;107(2):223-226.
10. Bonar SK, Tinetti ME, Speechley M, Cooney LM. Factors associated with short- versus long-term skilled nursing facility placement among community-living hip fracture patients. J Am Geriatr Soc. 1990;38(10):1139-1144.
11. Bentler SE, Liu L, Obrizan M, et al. The aftermath of hip fracture: discharge placement, functional status change, and mortality. Am J Epidemiol. 2009;170(10):1290-1299.
12. Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res. 2010;45(6 pt 1):1783-1795.
13. American Academy of Orthopaedic Surgeons. Burden of Musculoskeletal Diseases in the United States: Prevalence, Societal and Economic Cost. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2008.
14. Maciejewski ML, Radcliff A, Henderson WG, et al. Determinants of postsurgical discharge setting for male hip fracture patients. J Rehabil Res Dev. 2013;50(9):1267-1276.
15. Michel JP, Klopfenstein C, Hoffmeyer P, Stern R, Grab B. Hip fracture surgery: is the pre-operative American Society of Anesthesiologists (ASA) score a predictor of functional outcome? Aging Clin Exp Res. 2002;14(5):389-394.
16. Coco M, Rush H. Increased incidence of hip fractures in dialysis patients with low serum parathyroid hormone. Am J Kidney Dis. 2000;36(6):1115-1121.
17. Cameron ID, Chen JS, March LM, et al. Hip fracture causes excess mortality owing to cardiovascular and infectious disease in institutionalized older people: a prospective 5-year study. J Bone Miner Res. 2010;25(4):866-872.
18. White SM, Moppett IK, Griffiths R. Outcome by mode of anaesthesia for hip fracture surgery. An observational audit of 65 535 patients in a national dataset. Anaesthesia. 2014;69(3):224-230.
19. Le-Wendling L, Bihorac A, Baslanti TO, et al. Regional anesthesia as compared with general anesthesia for surgery in geriatric patients with hip fracture: does it decrease morbidity, mortality, and health care costs? Results of a single-centered study. Pain Med. 2012;13(7):948-956.
20. Marcantonio ER, Flacker JM, Michaels M, Resnick NM. Delirium is independently associated with poor functional recovery after hip fracture. J Am Geriatr Soc. 2000;48(6):618-624.
Length of stay (LOS) is a significant driver of costs after hip fracture surgery.1-3 Multiple studies have identified factors associated with increased LOS in hip fracture patients. These factors include admission time, delay to surgery, presence of comorbidities, and older age.4-9
One significant and potentially modifiable factor affecting LOS is delayed transfer to a rehabilitation center after surgery.8-11 Although patients after orthopedic surgeries require additional rehabilitation services or subacute care directly attributable to their injuries, specialized rehabilitation centers may not always have beds readily available.6-11 Studies have shown that delays in transfer to skilled nursing facilities or rehabilitation centers are highly common among orthopedic patients.8 It is therefore imperative that orthopedists have a mechanism for predicting and identifying which patients require rehabilitation services early in the postoperative period. Identifying risk factors and stratifying patients who are most likely to require rehabilitation would facilitate the early transfer of these patients and thereby directly decrease LOS and hospitalization-related costs.
In this article, we report results from prospective, national, multicenter data to identify commonly measured risk factors for discharge to rehabilitation facilities for hip fracture patients. Through multivariate analysis of ACS-NSQIP (American College of Surgeons National Surgical Quality Improvement Program) data, we determined which risk factors significantly predispose patients to discharge to rehabilitation centers versus discharge home. Knowledge of these risk factors allows the practicing orthopedist to be better equipped to identify patients who require additional rehabilitation early in the postoperative course. By mobilizing case managers and social workers to help avoid delays in the transfers of these identified patients, LOS-associated costs may ultimately decrease.
Materials and Methods
After obtaining institutional review board approval for this study from the Office of Research at Vanderbilt University, we prospectively collected 2011 discharge data from the ACS-NSQIP database (these data are unavailable for earlier years). All patients who underwent hip fracture surgery in 2011 were identified by CPT (Current Procedural Terminology) codes. Cases of patients with unknown discharge information and of those who died during their hospitalizations were excluded from analysis. For the remaining patients, discharge information as categorized by ACS-NSQIP included skilled care (eg, subacute hospital, skilled nursing home), unskilled facility (eg, nursing home, assisted facility), separate acute care, and rehabilitation. All other patients were discharged home without additional assistance or to the previous home where they received chronic care, assisted living, or unskilled aid. Patients were dichotomized according to whether they were discharged home or to one of the rehabilitation facilities mentioned.
To determine which risk factors significantly contributed to a patient’s discharge to rehabilitation, we ran univariate analyses using Fisher exact tests for categorical variables and Student t tests for continuous variables on multiple patient factors, including demographics, preoperative comorbidities, and operative factors. Demographics included age and sex. Preoperative comorbidities included 32 conditions: diabetes mellitus, active smoking status, current alcohol use, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, hypertension requiring medication, history of esophageal varices, history of myocardial infarction, current renal failure, current dialysis dependence, steroid use, recent weight loss, existing bleeding disorder, transfusion before discharge, presence of central nervous system tumor, recent chemotherapy, recent radiation therapy, previous percutaneous coronary intervention, previous percutaneous coronary stenting, history of angina, peripheral vascular disease, cerebrovascular accidents, recent surgery (within 30 days), rest pain, impaired sensorium, history of transient ischemic attacks, current hemiplegia status, current paraplegia status, current quadriplegia status, current ascites, hypertension, and disseminated cancer. Operative factors included wound infection, DNR (do not resuscitate) status, ventilator support, anesthesia type, wound class, ASA (American Society of Anesthesiologists) class, and operative time.
For the univariate analyses, significance was set at P < .05. Demographics, preoperative comorbidities, and operative factors that were significantly associated with discharge to a rehabilitation facility in the univariate analysis were selected as covariates for a multivariate analysis. We incorporated a binary logistic regression to analyze which of these significant risk factors are correlated with a patient’s discharge to a rehabilitation facility after hip fracture surgery.
Results
A total of 4974 patients undergoing surgery for hip fractures in 2011 were identified. Of these patients, 4815 had complete information on discharge location and were included in the analysis.
Table 1 lists the results of the univariate analysis comparing demographics, preoperative comorbidities, and operative factors between the home and rehabilitation groups. Both age (P < .001) and sex (P = .012) were significantly different between groups; the rehabilitation group was older by about 10 years and included significantly more females. In addition to demographic factors, 16 preoperative comorbidities, and 5 surgical factors were significantly associated with discharge to rehabilitation.
Surgery type significantly affected discharge to rehabilitation (Figure). Patients who were undergoing open plating of a femoral neck fracture or intramedullary nailing of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture constituted 30% of all patients discharged to rehabilitation centers. In contrast, patients undergoing percutaneous skeletal fixation of a proximal femoral fracture constituted only 5.5% of all patients discharged to rehabilitation. Based on surgery type, we broke down discharge location further, into categories of skilled nursing facility, unskilled facility (not patient’s previous home), separate acute-care facility, dedicated rehabilitation center, and home. Of all 4815 patients combined, 2102 (43.6%) were discharged to a skilled nursing facility, 31 (0.6%) to an unskilled facility (not home), 106 (2.2%) to separate acute care, 1312 (27.2%) to a dedicated rehabilitation center, and 950 (19.7%) home.
Table 2 lists the significant results from the multivariate logistical analysis comparing discharge to a rehabilitation center and discharge home after controlling for the significant risk factors (Table 1). Current diabetes, history of dyspnea, previous myocardial infarction, history of ischemic attacks, current bleeding disorder, transfusion during hospitalization, previous percutaneous cardiac stenting, chemotherapy, past cerebrovascular accident, presence of cancer, surgery type based on CPT code, history of chronic obstructive pulmonary disease or congestive heart failure, current smoking status, and operative time longer than 90 minutes were not significantly correlated with discharge to rehabilitation in the multivariate analysis. All significant factors were associated with higher odds of discharge to rehabilitation except for DNR status. DNR patients were 2.04 times more likely (95% CI, 1.49-2.78; P < .001) to be discharged home than to rehabilitation centers.
Applying these adjusted odds ratios, we see that an elderly woman (age, >65 years) who underwent general anesthesia with an ASA class higher than 2 was 17.63 times more likely than a patient without these risk factors to be discharged to rehabilitation. If this patient were also dialysis-dependent, she would be 61.52 times more likely than a similar patient without dialysis needs to be discharged to rehabilitation.
Even when controlling for all significant and nonsignificant variables in multivariate logistical analysis, age over 65 years (β = 1.05; P < .001), female sex (β = 1.76; P = .004), dialysis dependence (β = 12.98; P = .036), hypertension requiring medication (β = 1.53; P = .032), and ASA class higher than 2 (β = 1.98; P = .001) were found to be significant risk factors for discharge to rehabilitation.
Discussion
This study was the first to investigate the issue of which patient risk factors allow the practicing orthopedist to identify patients who require rehabilitation after hip fracture surgery. Through our multivariate analysis, which controlled for demographics, comorbidities, and operative factors, we found that older age, female sex, history of percutaneous coronary intervention, dialysis dependence, general anesthesia, and ASA class higher than 2 significantly increased the odds of discharge to a rehabilitation center versus home.
Using our study’s results, we can create a risk stratification model for patients and thereby a means of targeting patients who need rehabilitation and starting the process of finding a rehabilitation bed early in the postoperative course. Our study’s variables are easily measured metrics that may be collected in any hospital setting. Especially for hip fracture patients, early planning and discharge to the appropriate rehabilitation center are important in decreasing LOS and associated hospitalization costs. According to one report,3 about 85% of all hip fracture costs are directly related to LOS, given the unnecessarily long rehabilitation periods in hospitals. Hollingworth and colleagues2 compared costs for patients who remained in the hospital with costs for those discharged with rehabilitation services. Overall costs were significantly lower for patients discharged home with rehabilitation. The authors concluded that 40% of hip fracture patients may be suitable for early discharge.2 In an analysis of Medicare payments for hip fracture treatment, hospital costs including LOS accounted for 60% of all payments.12 The results of these 2 studies suggest that the overall driver of hip fracture costs is prolonged LOS and that, if patients are discharged to rehabilitation, then overall costs may be lowered through a direct reduction in hospital LOS. Given that hip fractures account for almost 350,000 hospital admissions in the United States each year, and using our institution’s average hospital charge per day ($4500), about $1.6 billion may be saved if each patient’s LOS decreased by 1 day.13 Although multiple factors affect LOS, discharge planning is under orthopedists’ direct control. Therefore, early identification of patients who will require rehabilitation may help reduce LOS-associated costs in our health care system.
The patient variables that were significantly associated with discharge to rehabilitation are also associated with increased morbidity and mortality in hip fracture patients, according to the literature,14-20 which provides some external validation of using these risk factors as predictors for rehabilitation. A patient with one of these risk factors may require rehabilitation, given that rehabilitation services are specifically linked to lower morbidity and mortality rates among hip fracture patients. For example, patients with dialysis needs were 3.49 times more likely to be discharged to a rehabilitation center in our study. In a 2000 study by Coco and Rush,16 hip fracture patients on dialysis had a 1-year mortality rate 2.5 times higher than that of patients who were not dialysis-dependent. In 2010, Cameron and colleagues17 found that cardiovascular disease was associated with a 2.68 times higher risk of mortality in hip fracture patients. Similarly in our study, both hypertension and history of percutaneous coronary intervention were associated with discharge to rehabilitation. We found higher odds of discharge to rehabilitation with higher ASA classes, which mirror results from a study by Michel and colleagues,15 who found that higher (vs lower) preoperative ASA classes were associated with higher 1-year mortality in hip fracture patients. Interestingly, DNR status was associated with higher odds of discharge home, which may reflect patients’ desires to forgo noninvasive or lifesaving procedures that may be performed at rehabilitation facilities. Although general anesthesia predisposed patients to discharge to a rehabilitation center, multiple studies have found no association between anesthesia type and postoperative mortality rates for hip fracture patients.18,19 Last, Marcantonio and colleagues20 found delirium specifically had a higher odds ratio for discharge, but our univariate analysis did not find a significant association between impaired sensorium and discharge location. Given the correlation of our risk factors with increased morbidity and mortality in the literature, our study’s results provide the initial groundwork for creating a risk calculator that orthopedists can use to predict discharge to rehabilitation.
Our study had some limitations. Although we analyzed a large number of demographics, preoperative comorbidities, and surgical factors, our univariate analysis was limited to information in the ACS-NSQIP database. We did not incorporate other clinically relevant factors (eg, social factors, including patients’ support networks) that may influence discharge decisions. Furthermore, ACS-NSQIP records patient data only up to 30 days after surgery. Discharge information for the time after that was missing for a subset of hip fracture patients, and these patients had to be excluded, potentially skewing our data. ACS-NSQIP also does not collect cost data for patients based on hospitalization or LOS, so we could not determine whether patients discharged to rehabilitation incurred higher costs because of longer hospitalizations.
Nevertheless, our study identified significant patient and operative variables that are associated with discharge to a rehabilitation center. By identifying hip fracture patients with these risk factors early and mobilizing the appropriate resources, practicing orthopedists should be better equipped to help facilitate the discharge of patients to the appropriate location after surgery. Validation of these risk factors should be prospectively determined with an analysis of LOS and cost implications. Use of a risk calculator may in fact result in decreased LOS and hospital-related costs. Furthermore, using these risk factors in a prospective patient cohort would help validate their use and determine whether there is clinical correlation. The orthopedists in our institution are becoming more aware of these risk factors, but validation is necessary.
Length of stay (LOS) is a significant driver of costs after hip fracture surgery.1-3 Multiple studies have identified factors associated with increased LOS in hip fracture patients. These factors include admission time, delay to surgery, presence of comorbidities, and older age.4-9
One significant and potentially modifiable factor affecting LOS is delayed transfer to a rehabilitation center after surgery.8-11 Although patients after orthopedic surgeries require additional rehabilitation services or subacute care directly attributable to their injuries, specialized rehabilitation centers may not always have beds readily available.6-11 Studies have shown that delays in transfer to skilled nursing facilities or rehabilitation centers are highly common among orthopedic patients.8 It is therefore imperative that orthopedists have a mechanism for predicting and identifying which patients require rehabilitation services early in the postoperative period. Identifying risk factors and stratifying patients who are most likely to require rehabilitation would facilitate the early transfer of these patients and thereby directly decrease LOS and hospitalization-related costs.
In this article, we report results from prospective, national, multicenter data to identify commonly measured risk factors for discharge to rehabilitation facilities for hip fracture patients. Through multivariate analysis of ACS-NSQIP (American College of Surgeons National Surgical Quality Improvement Program) data, we determined which risk factors significantly predispose patients to discharge to rehabilitation centers versus discharge home. Knowledge of these risk factors allows the practicing orthopedist to be better equipped to identify patients who require additional rehabilitation early in the postoperative course. By mobilizing case managers and social workers to help avoid delays in the transfers of these identified patients, LOS-associated costs may ultimately decrease.
Materials and Methods
After obtaining institutional review board approval for this study from the Office of Research at Vanderbilt University, we prospectively collected 2011 discharge data from the ACS-NSQIP database (these data are unavailable for earlier years). All patients who underwent hip fracture surgery in 2011 were identified by CPT (Current Procedural Terminology) codes. Cases of patients with unknown discharge information and of those who died during their hospitalizations were excluded from analysis. For the remaining patients, discharge information as categorized by ACS-NSQIP included skilled care (eg, subacute hospital, skilled nursing home), unskilled facility (eg, nursing home, assisted facility), separate acute care, and rehabilitation. All other patients were discharged home without additional assistance or to the previous home where they received chronic care, assisted living, or unskilled aid. Patients were dichotomized according to whether they were discharged home or to one of the rehabilitation facilities mentioned.
To determine which risk factors significantly contributed to a patient’s discharge to rehabilitation, we ran univariate analyses using Fisher exact tests for categorical variables and Student t tests for continuous variables on multiple patient factors, including demographics, preoperative comorbidities, and operative factors. Demographics included age and sex. Preoperative comorbidities included 32 conditions: diabetes mellitus, active smoking status, current alcohol use, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, hypertension requiring medication, history of esophageal varices, history of myocardial infarction, current renal failure, current dialysis dependence, steroid use, recent weight loss, existing bleeding disorder, transfusion before discharge, presence of central nervous system tumor, recent chemotherapy, recent radiation therapy, previous percutaneous coronary intervention, previous percutaneous coronary stenting, history of angina, peripheral vascular disease, cerebrovascular accidents, recent surgery (within 30 days), rest pain, impaired sensorium, history of transient ischemic attacks, current hemiplegia status, current paraplegia status, current quadriplegia status, current ascites, hypertension, and disseminated cancer. Operative factors included wound infection, DNR (do not resuscitate) status, ventilator support, anesthesia type, wound class, ASA (American Society of Anesthesiologists) class, and operative time.
For the univariate analyses, significance was set at P < .05. Demographics, preoperative comorbidities, and operative factors that were significantly associated with discharge to a rehabilitation facility in the univariate analysis were selected as covariates for a multivariate analysis. We incorporated a binary logistic regression to analyze which of these significant risk factors are correlated with a patient’s discharge to a rehabilitation facility after hip fracture surgery.
Results
A total of 4974 patients undergoing surgery for hip fractures in 2011 were identified. Of these patients, 4815 had complete information on discharge location and were included in the analysis.
Table 1 lists the results of the univariate analysis comparing demographics, preoperative comorbidities, and operative factors between the home and rehabilitation groups. Both age (P < .001) and sex (P = .012) were significantly different between groups; the rehabilitation group was older by about 10 years and included significantly more females. In addition to demographic factors, 16 preoperative comorbidities, and 5 surgical factors were significantly associated with discharge to rehabilitation.
Surgery type significantly affected discharge to rehabilitation (Figure). Patients who were undergoing open plating of a femoral neck fracture or intramedullary nailing of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture constituted 30% of all patients discharged to rehabilitation centers. In contrast, patients undergoing percutaneous skeletal fixation of a proximal femoral fracture constituted only 5.5% of all patients discharged to rehabilitation. Based on surgery type, we broke down discharge location further, into categories of skilled nursing facility, unskilled facility (not patient’s previous home), separate acute-care facility, dedicated rehabilitation center, and home. Of all 4815 patients combined, 2102 (43.6%) were discharged to a skilled nursing facility, 31 (0.6%) to an unskilled facility (not home), 106 (2.2%) to separate acute care, 1312 (27.2%) to a dedicated rehabilitation center, and 950 (19.7%) home.
Table 2 lists the significant results from the multivariate logistical analysis comparing discharge to a rehabilitation center and discharge home after controlling for the significant risk factors (Table 1). Current diabetes, history of dyspnea, previous myocardial infarction, history of ischemic attacks, current bleeding disorder, transfusion during hospitalization, previous percutaneous cardiac stenting, chemotherapy, past cerebrovascular accident, presence of cancer, surgery type based on CPT code, history of chronic obstructive pulmonary disease or congestive heart failure, current smoking status, and operative time longer than 90 minutes were not significantly correlated with discharge to rehabilitation in the multivariate analysis. All significant factors were associated with higher odds of discharge to rehabilitation except for DNR status. DNR patients were 2.04 times more likely (95% CI, 1.49-2.78; P < .001) to be discharged home than to rehabilitation centers.
Applying these adjusted odds ratios, we see that an elderly woman (age, >65 years) who underwent general anesthesia with an ASA class higher than 2 was 17.63 times more likely than a patient without these risk factors to be discharged to rehabilitation. If this patient were also dialysis-dependent, she would be 61.52 times more likely than a similar patient without dialysis needs to be discharged to rehabilitation.
Even when controlling for all significant and nonsignificant variables in multivariate logistical analysis, age over 65 years (β = 1.05; P < .001), female sex (β = 1.76; P = .004), dialysis dependence (β = 12.98; P = .036), hypertension requiring medication (β = 1.53; P = .032), and ASA class higher than 2 (β = 1.98; P = .001) were found to be significant risk factors for discharge to rehabilitation.
Discussion
This study was the first to investigate the issue of which patient risk factors allow the practicing orthopedist to identify patients who require rehabilitation after hip fracture surgery. Through our multivariate analysis, which controlled for demographics, comorbidities, and operative factors, we found that older age, female sex, history of percutaneous coronary intervention, dialysis dependence, general anesthesia, and ASA class higher than 2 significantly increased the odds of discharge to a rehabilitation center versus home.
Using our study’s results, we can create a risk stratification model for patients and thereby a means of targeting patients who need rehabilitation and starting the process of finding a rehabilitation bed early in the postoperative course. Our study’s variables are easily measured metrics that may be collected in any hospital setting. Especially for hip fracture patients, early planning and discharge to the appropriate rehabilitation center are important in decreasing LOS and associated hospitalization costs. According to one report,3 about 85% of all hip fracture costs are directly related to LOS, given the unnecessarily long rehabilitation periods in hospitals. Hollingworth and colleagues2 compared costs for patients who remained in the hospital with costs for those discharged with rehabilitation services. Overall costs were significantly lower for patients discharged home with rehabilitation. The authors concluded that 40% of hip fracture patients may be suitable for early discharge.2 In an analysis of Medicare payments for hip fracture treatment, hospital costs including LOS accounted for 60% of all payments.12 The results of these 2 studies suggest that the overall driver of hip fracture costs is prolonged LOS and that, if patients are discharged to rehabilitation, then overall costs may be lowered through a direct reduction in hospital LOS. Given that hip fractures account for almost 350,000 hospital admissions in the United States each year, and using our institution’s average hospital charge per day ($4500), about $1.6 billion may be saved if each patient’s LOS decreased by 1 day.13 Although multiple factors affect LOS, discharge planning is under orthopedists’ direct control. Therefore, early identification of patients who will require rehabilitation may help reduce LOS-associated costs in our health care system.
The patient variables that were significantly associated with discharge to rehabilitation are also associated with increased morbidity and mortality in hip fracture patients, according to the literature,14-20 which provides some external validation of using these risk factors as predictors for rehabilitation. A patient with one of these risk factors may require rehabilitation, given that rehabilitation services are specifically linked to lower morbidity and mortality rates among hip fracture patients. For example, patients with dialysis needs were 3.49 times more likely to be discharged to a rehabilitation center in our study. In a 2000 study by Coco and Rush,16 hip fracture patients on dialysis had a 1-year mortality rate 2.5 times higher than that of patients who were not dialysis-dependent. In 2010, Cameron and colleagues17 found that cardiovascular disease was associated with a 2.68 times higher risk of mortality in hip fracture patients. Similarly in our study, both hypertension and history of percutaneous coronary intervention were associated with discharge to rehabilitation. We found higher odds of discharge to rehabilitation with higher ASA classes, which mirror results from a study by Michel and colleagues,15 who found that higher (vs lower) preoperative ASA classes were associated with higher 1-year mortality in hip fracture patients. Interestingly, DNR status was associated with higher odds of discharge home, which may reflect patients’ desires to forgo noninvasive or lifesaving procedures that may be performed at rehabilitation facilities. Although general anesthesia predisposed patients to discharge to a rehabilitation center, multiple studies have found no association between anesthesia type and postoperative mortality rates for hip fracture patients.18,19 Last, Marcantonio and colleagues20 found delirium specifically had a higher odds ratio for discharge, but our univariate analysis did not find a significant association between impaired sensorium and discharge location. Given the correlation of our risk factors with increased morbidity and mortality in the literature, our study’s results provide the initial groundwork for creating a risk calculator that orthopedists can use to predict discharge to rehabilitation.
Our study had some limitations. Although we analyzed a large number of demographics, preoperative comorbidities, and surgical factors, our univariate analysis was limited to information in the ACS-NSQIP database. We did not incorporate other clinically relevant factors (eg, social factors, including patients’ support networks) that may influence discharge decisions. Furthermore, ACS-NSQIP records patient data only up to 30 days after surgery. Discharge information for the time after that was missing for a subset of hip fracture patients, and these patients had to be excluded, potentially skewing our data. ACS-NSQIP also does not collect cost data for patients based on hospitalization or LOS, so we could not determine whether patients discharged to rehabilitation incurred higher costs because of longer hospitalizations.
Nevertheless, our study identified significant patient and operative variables that are associated with discharge to a rehabilitation center. By identifying hip fracture patients with these risk factors early and mobilizing the appropriate resources, practicing orthopedists should be better equipped to help facilitate the discharge of patients to the appropriate location after surgery. Validation of these risk factors should be prospectively determined with an analysis of LOS and cost implications. Use of a risk calculator may in fact result in decreased LOS and hospital-related costs. Furthermore, using these risk factors in a prospective patient cohort would help validate their use and determine whether there is clinical correlation. The orthopedists in our institution are becoming more aware of these risk factors, but validation is necessary.
1. Garcia AE, Bonnaig JV, Yoneda ZT, et al. Patient variables which may predict length of stay and hospital costs in elderly patients with hip fracture. J Orthop Trauma. 2012;26(11):620-623.
2. Hollingworth W, Todd C, Parker M, Roberts JA, Williams R. Cost analysis of early discharge after hip fracture. BMJ. 1993;307(6909):903-906.
3. Sund R, Riihimäki J, Mäkelä M, et al. Modeling the length of the care episode after hip fracture: does the type of fracture matter? Scand J Surg. 2009;98(3):169-174.
4. Fox KM, Magaziner J, Hebel JR, Kenzora JE, Kashner TM. Intertrochanteric versus femoral neck hip fractures: differential characteristics, treatment, and sequelae. J Gerontol A Biol Sci Med Sci. 1999;54(12):M635-M640.
5. Foss NB, Palm H, Krasheninnikoff M, Kehlet H, Gebuhr P. Impact of surgical complications on length of stay after hip fracture surgery. Injury. 2007;38(7):780-784.
6. Lefaivre KA, Macadam SA, Davidson DJ, Gandhi R, Chan H, Broekhuyse HM. Length of stay, mortality, morbidity and delay to surgery in hip fractures. J Bone Joint Surg Br. 2009;91(7):922-927.
7. Clague JE, Craddock E, Andrew G, Horan MA, Pendleton N. Predictors of outcome following hip fracture. Admission time predicts length of stay and in-hospital mortality. Injury. 2002;33(1):1-6.
8. Parker MJ, Todd CJ, Palmer CR, et al. Inter-hospital variations in length of hospital stay following hip fracture. Age Ageing. 1998;27(31):333-337.
9. Brasel KJ, Rasmussen J, Cauley C, Weigelt JA. Reasons for delayed discharge of trauma patients. J Surg Res. 2002;107(2):223-226.
10. Bonar SK, Tinetti ME, Speechley M, Cooney LM. Factors associated with short- versus long-term skilled nursing facility placement among community-living hip fracture patients. J Am Geriatr Soc. 1990;38(10):1139-1144.
11. Bentler SE, Liu L, Obrizan M, et al. The aftermath of hip fracture: discharge placement, functional status change, and mortality. Am J Epidemiol. 2009;170(10):1290-1299.
12. Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res. 2010;45(6 pt 1):1783-1795.
13. American Academy of Orthopaedic Surgeons. Burden of Musculoskeletal Diseases in the United States: Prevalence, Societal and Economic Cost. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2008.
14. Maciejewski ML, Radcliff A, Henderson WG, et al. Determinants of postsurgical discharge setting for male hip fracture patients. J Rehabil Res Dev. 2013;50(9):1267-1276.
15. Michel JP, Klopfenstein C, Hoffmeyer P, Stern R, Grab B. Hip fracture surgery: is the pre-operative American Society of Anesthesiologists (ASA) score a predictor of functional outcome? Aging Clin Exp Res. 2002;14(5):389-394.
16. Coco M, Rush H. Increased incidence of hip fractures in dialysis patients with low serum parathyroid hormone. Am J Kidney Dis. 2000;36(6):1115-1121.
17. Cameron ID, Chen JS, March LM, et al. Hip fracture causes excess mortality owing to cardiovascular and infectious disease in institutionalized older people: a prospective 5-year study. J Bone Miner Res. 2010;25(4):866-872.
18. White SM, Moppett IK, Griffiths R. Outcome by mode of anaesthesia for hip fracture surgery. An observational audit of 65 535 patients in a national dataset. Anaesthesia. 2014;69(3):224-230.
19. Le-Wendling L, Bihorac A, Baslanti TO, et al. Regional anesthesia as compared with general anesthesia for surgery in geriatric patients with hip fracture: does it decrease morbidity, mortality, and health care costs? Results of a single-centered study. Pain Med. 2012;13(7):948-956.
20. Marcantonio ER, Flacker JM, Michaels M, Resnick NM. Delirium is independently associated with poor functional recovery after hip fracture. J Am Geriatr Soc. 2000;48(6):618-624.
1. Garcia AE, Bonnaig JV, Yoneda ZT, et al. Patient variables which may predict length of stay and hospital costs in elderly patients with hip fracture. J Orthop Trauma. 2012;26(11):620-623.
2. Hollingworth W, Todd C, Parker M, Roberts JA, Williams R. Cost analysis of early discharge after hip fracture. BMJ. 1993;307(6909):903-906.
3. Sund R, Riihimäki J, Mäkelä M, et al. Modeling the length of the care episode after hip fracture: does the type of fracture matter? Scand J Surg. 2009;98(3):169-174.
4. Fox KM, Magaziner J, Hebel JR, Kenzora JE, Kashner TM. Intertrochanteric versus femoral neck hip fractures: differential characteristics, treatment, and sequelae. J Gerontol A Biol Sci Med Sci. 1999;54(12):M635-M640.
5. Foss NB, Palm H, Krasheninnikoff M, Kehlet H, Gebuhr P. Impact of surgical complications on length of stay after hip fracture surgery. Injury. 2007;38(7):780-784.
6. Lefaivre KA, Macadam SA, Davidson DJ, Gandhi R, Chan H, Broekhuyse HM. Length of stay, mortality, morbidity and delay to surgery in hip fractures. J Bone Joint Surg Br. 2009;91(7):922-927.
7. Clague JE, Craddock E, Andrew G, Horan MA, Pendleton N. Predictors of outcome following hip fracture. Admission time predicts length of stay and in-hospital mortality. Injury. 2002;33(1):1-6.
8. Parker MJ, Todd CJ, Palmer CR, et al. Inter-hospital variations in length of hospital stay following hip fracture. Age Ageing. 1998;27(31):333-337.
9. Brasel KJ, Rasmussen J, Cauley C, Weigelt JA. Reasons for delayed discharge of trauma patients. J Surg Res. 2002;107(2):223-226.
10. Bonar SK, Tinetti ME, Speechley M, Cooney LM. Factors associated with short- versus long-term skilled nursing facility placement among community-living hip fracture patients. J Am Geriatr Soc. 1990;38(10):1139-1144.
11. Bentler SE, Liu L, Obrizan M, et al. The aftermath of hip fracture: discharge placement, functional status change, and mortality. Am J Epidemiol. 2009;170(10):1290-1299.
12. Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res. 2010;45(6 pt 1):1783-1795.
13. American Academy of Orthopaedic Surgeons. Burden of Musculoskeletal Diseases in the United States: Prevalence, Societal and Economic Cost. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2008.
14. Maciejewski ML, Radcliff A, Henderson WG, et al. Determinants of postsurgical discharge setting for male hip fracture patients. J Rehabil Res Dev. 2013;50(9):1267-1276.
15. Michel JP, Klopfenstein C, Hoffmeyer P, Stern R, Grab B. Hip fracture surgery: is the pre-operative American Society of Anesthesiologists (ASA) score a predictor of functional outcome? Aging Clin Exp Res. 2002;14(5):389-394.
16. Coco M, Rush H. Increased incidence of hip fractures in dialysis patients with low serum parathyroid hormone. Am J Kidney Dis. 2000;36(6):1115-1121.
17. Cameron ID, Chen JS, March LM, et al. Hip fracture causes excess mortality owing to cardiovascular and infectious disease in institutionalized older people: a prospective 5-year study. J Bone Miner Res. 2010;25(4):866-872.
18. White SM, Moppett IK, Griffiths R. Outcome by mode of anaesthesia for hip fracture surgery. An observational audit of 65 535 patients in a national dataset. Anaesthesia. 2014;69(3):224-230.
19. Le-Wendling L, Bihorac A, Baslanti TO, et al. Regional anesthesia as compared with general anesthesia for surgery in geriatric patients with hip fracture: does it decrease morbidity, mortality, and health care costs? Results of a single-centered study. Pain Med. 2012;13(7):948-956.
20. Marcantonio ER, Flacker JM, Michaels M, Resnick NM. Delirium is independently associated with poor functional recovery after hip fracture. J Am Geriatr Soc. 2000;48(6):618-624.
