David L. Hahn, MD, MS

Like many of my colleagues, I support President Obama’s call to demonstrate value as part of health care reform. One way to do that is through public reporting. The rationale is that public scrutiny of outcomes will motivate the health care “industry” to improve the “product” (outcomes), rather than accelerating valueless economic activity (process) that often benefits providers more than patients.

Fair enough. But does the existing system of quality indicators support the goals of reform identified by the Institute of Medicine (IOM)?¹ That is, does it make the system safer, more effective and efficient, timely, equitable, and patient-centered?

Not necessarily.

The reason is 2-fold. First, the best quality indicators are patient-oriented outcomes (eg, quality of life, morbidity, mortality), but that’s not what’s being reported. Second, many publicly reported surrogate measures are more harmful than helpful, and in need of serious reform themselves.

My experience

The Wisconsin Collaborative for Healthcare Quality (WCHQ), which I’ve been involved with for nearly 10 years, is composed of health care organizations, mine included, committed to voluntary reporting of quality metrics. The TABLE features a list of the metrics, chosen by the WCHQ, that are reported.

I’ve rated each metric on 2 criteria:

1. How good is the evidence for the screening tool or intervention? (There is good evidence for colorectal cancer screening, for example, but evidence for low-density lipoprotein [LDL] testing is poor.)
2. How good is the quality indicator itself, including the frequency? (There’s good evidence for Pap testing within 3 years, whereas twice-yearly HbA1c testing is opinion-based.)

TABLE 1
How do the “quality indicators” rate?

Some worrisome examples

While the ratings are partly subjective, they’re meant to illustrate that not all publicly reported metrics are supported by good evidence.

This is particularly troubling, given the fact that acting on fair or poor evidence may cause more harm than good. Consider these worrisome examples:

LDL control <100 mg/dL. I’ve known patients who had their first myocardial infarction when their LDL cholesterol was <100 mg/dL. After the event, these patients weren’t given a statin because they were already “at goal”; they subsequently had a reinfarction.

LDL should not be used as a quality indicator in secondary prevention for (at least) 2 reasons: First, some LDL-lowering drugs are harmful or have no net benefit (eg, estrogen in women, fibrates).² Second, statin benefit may or may not be related to lipid lowering, and the magnitude of benefit is not related to any arbitrary LDL goal.³

There is clear, compelling evidence supporting near-universal statin therapy for patients at high cardiovascular risk regardless of their LDL cholesterol values—but a lack of evidence that titrating lipid therapy to achieve proposed low LDL levels is beneficial or safe.⁴ Receiving the maximum tolerated dose of statin is therefore the appropriate evidence-based surrogate quality indicator, not LDL.

Mammography. The Cochrane collaboration has concluded that “for every 2000 women invited for [mammography] screening throughout 10 years, 1 will have her life prolonged,” and 10 healthy women who would not have been diagnosed without the screening will be treated unnecessarily.⁵ The Cochrane review thus concluded that it’s not clear whether mammography screening does more good than harm.⁵

The US Preventive Services Task Force (USPSTF) recently downgraded mammography screening for women over age 50 from an A- to a B-rated recommendation. The fine balance between benefit and harm in this and other USPSTF B-rated preventive measures requires that clinicians educate patients and elicit their preferences. But this doesn’t occur when health plans strive to outdo one another in achieving higher publicly reported screening goals.⁶ Documentation of valid shared decision-making, not screening rates, is the appropriate quality indicator.

The evidence vs the “business” of medicine

I have no illusions that my recommendations will be adopted easily—or soon. After all, we practice in an environment in which evidence-based practice recommendations can conflict with financial and operational goals perceived as necessary to survive. However, I believe that evidence trumps business in achieving the IOM goals.

It remains to be seen whether we can simultaneously move toward valid evidence-based public reporting and health-care financial reform. But one thing is clear: To insist that evidence-based patient-oriented quality indicators are too difficult to measure, or to ignore or deny the evidence, puts the lie to claims of patient-centered care and, ultimately, to long-needed health care reform looming on the horizon.

Like many of my colleagues, I support President Obama’s call to demonstrate value as part of health care reform. One way to do that is through public reporting. The rationale is that public scrutiny of outcomes will motivate the health care “industry” to improve the “product” (outcomes), rather than accelerating valueless economic activity (process) that often benefits providers more than patients.

Fair enough. But does the existing system of quality indicators support the goals of reform identified by the Institute of Medicine (IOM)?¹ That is, does it make the system safer, more effective and efficient, timely, equitable, and patient-centered?

Not necessarily.

The reason is 2-fold. First, the best quality indicators are patient-oriented outcomes (eg, quality of life, morbidity, mortality), but that’s not what’s being reported. Second, many publicly reported surrogate measures are more harmful than helpful, and in need of serious reform themselves.

My experience

The Wisconsin Collaborative for Healthcare Quality (WCHQ), which I’ve been involved with for nearly 10 years, is composed of health care organizations, mine included, committed to voluntary reporting of quality metrics. The TABLE features a list of the metrics, chosen by the WCHQ, that are reported.

I’ve rated each metric on 2 criteria:

1. How good is the evidence for the screening tool or intervention? (There is good evidence for colorectal cancer screening, for example, but evidence for low-density lipoprotein [LDL] testing is poor.)
2. How good is the quality indicator itself, including the frequency? (There’s good evidence for Pap testing within 3 years, whereas twice-yearly HbA1c testing is opinion-based.)

TABLE 1
How do the “quality indicators” rate?

Some worrisome examples

While the ratings are partly subjective, they’re meant to illustrate that not all publicly reported metrics are supported by good evidence.

This is particularly troubling, given the fact that acting on fair or poor evidence may cause more harm than good. Consider these worrisome examples:

LDL control <100 mg/dL. I’ve known patients who had their first myocardial infarction when their LDL cholesterol was <100 mg/dL. After the event, these patients weren’t given a statin because they were already “at goal”; they subsequently had a reinfarction.

LDL should not be used as a quality indicator in secondary prevention for (at least) 2 reasons: First, some LDL-lowering drugs are harmful or have no net benefit (eg, estrogen in women, fibrates).² Second, statin benefit may or may not be related to lipid lowering, and the magnitude of benefit is not related to any arbitrary LDL goal.³

There is clear, compelling evidence supporting near-universal statin therapy for patients at high cardiovascular risk regardless of their LDL cholesterol values—but a lack of evidence that titrating lipid therapy to achieve proposed low LDL levels is beneficial or safe.⁴ Receiving the maximum tolerated dose of statin is therefore the appropriate evidence-based surrogate quality indicator, not LDL.

Mammography. The Cochrane collaboration has concluded that “for every 2000 women invited for [mammography] screening throughout 10 years, 1 will have her life prolonged,” and 10 healthy women who would not have been diagnosed without the screening will be treated unnecessarily.⁵ The Cochrane review thus concluded that it’s not clear whether mammography screening does more good than harm.⁵

The US Preventive Services Task Force (USPSTF) recently downgraded mammography screening for women over age 50 from an A- to a B-rated recommendation. The fine balance between benefit and harm in this and other USPSTF B-rated preventive measures requires that clinicians educate patients and elicit their preferences. But this doesn’t occur when health plans strive to outdo one another in achieving higher publicly reported screening goals.⁶ Documentation of valid shared decision-making, not screening rates, is the appropriate quality indicator.

The evidence vs the “business” of medicine

I have no illusions that my recommendations will be adopted easily—or soon. After all, we practice in an environment in which evidence-based practice recommendations can conflict with financial and operational goals perceived as necessary to survive. However, I believe that evidence trumps business in achieving the IOM goals.

It remains to be seen whether we can simultaneously move toward valid evidence-based public reporting and health-care financial reform. But one thing is clear: To insist that evidence-based patient-oriented quality indicators are too difficult to measure, or to ignore or deny the evidence, puts the lie to claims of patient-centered care and, ultimately, to long-needed health care reform looming on the horizon.

Like many of my colleagues, I support President Obama’s call to demonstrate value as part of health care reform. One way to do that is through public reporting. The rationale is that public scrutiny of outcomes will motivate the health care “industry” to improve the “product” (outcomes), rather than accelerating valueless economic activity (process) that often benefits providers more than patients.

Fair enough. But does the existing system of quality indicators support the goals of reform identified by the Institute of Medicine (IOM)?¹ That is, does it make the system safer, more effective and efficient, timely, equitable, and patient-centered?

Not necessarily.

The reason is 2-fold. First, the best quality indicators are patient-oriented outcomes (eg, quality of life, morbidity, mortality), but that’s not what’s being reported. Second, many publicly reported surrogate measures are more harmful than helpful, and in need of serious reform themselves.

My experience

The Wisconsin Collaborative for Healthcare Quality (WCHQ), which I’ve been involved with for nearly 10 years, is composed of health care organizations, mine included, committed to voluntary reporting of quality metrics. The TABLE features a list of the metrics, chosen by the WCHQ, that are reported.

I’ve rated each metric on 2 criteria:

1. How good is the evidence for the screening tool or intervention? (There is good evidence for colorectal cancer screening, for example, but evidence for low-density lipoprotein [LDL] testing is poor.)
2. How good is the quality indicator itself, including the frequency? (There’s good evidence for Pap testing within 3 years, whereas twice-yearly HbA1c testing is opinion-based.)

TABLE 1
How do the “quality indicators” rate?

Some worrisome examples

While the ratings are partly subjective, they’re meant to illustrate that not all publicly reported metrics are supported by good evidence.

This is particularly troubling, given the fact that acting on fair or poor evidence may cause more harm than good. Consider these worrisome examples:

LDL control <100 mg/dL. I’ve known patients who had their first myocardial infarction when their LDL cholesterol was <100 mg/dL. After the event, these patients weren’t given a statin because they were already “at goal”; they subsequently had a reinfarction.

LDL should not be used as a quality indicator in secondary prevention for (at least) 2 reasons: First, some LDL-lowering drugs are harmful or have no net benefit (eg, estrogen in women, fibrates).² Second, statin benefit may or may not be related to lipid lowering, and the magnitude of benefit is not related to any arbitrary LDL goal.³

There is clear, compelling evidence supporting near-universal statin therapy for patients at high cardiovascular risk regardless of their LDL cholesterol values—but a lack of evidence that titrating lipid therapy to achieve proposed low LDL levels is beneficial or safe.⁴ Receiving the maximum tolerated dose of statin is therefore the appropriate evidence-based surrogate quality indicator, not LDL.

Mammography. The Cochrane collaboration has concluded that “for every 2000 women invited for [mammography] screening throughout 10 years, 1 will have her life prolonged,” and 10 healthy women who would not have been diagnosed without the screening will be treated unnecessarily.⁵ The Cochrane review thus concluded that it’s not clear whether mammography screening does more good than harm.⁵

The US Preventive Services Task Force (USPSTF) recently downgraded mammography screening for women over age 50 from an A- to a B-rated recommendation. The fine balance between benefit and harm in this and other USPSTF B-rated preventive measures requires that clinicians educate patients and elicit their preferences. But this doesn’t occur when health plans strive to outdo one another in achieving higher publicly reported screening goals.⁶ Documentation of valid shared decision-making, not screening rates, is the appropriate quality indicator.

The evidence vs the “business” of medicine

I have no illusions that my recommendations will be adopted easily—or soon. After all, we practice in an environment in which evidence-based practice recommendations can conflict with financial and operational goals perceived as necessary to survive. However, I believe that evidence trumps business in achieving the IOM goals.

It remains to be seen whether we can simultaneously move toward valid evidence-based public reporting and health-care financial reform. But one thing is clear: To insist that evidence-based patient-oriented quality indicators are too difficult to measure, or to ignore or deny the evidence, puts the lie to claims of patient-centered care and, ultimately, to long-needed health care reform looming on the horizon.

I was disappointed, recently, when I read a handout from the American College of Physicians (ACP) that was designed to help physicians talk to patients about the benefits and harms of screening mammography in women ages 40 to 49. The problem: The summary simply didn’t provide enough relevant information to help patients make an informed choice.

The patient handout was part of a collection of ACP articles that appeared in the Annals of Internal Medicine. (See related POEM.) The first article was a systematic review of screening mammography in women 40 to 49 years of age documenting current data on benefit and harms.¹

The ACP also published a clinical practice guideline that included the recommendation that “clinicians should inform women 40 to 49 years of age about the potential benefits and harms of screening mammography.”² Accompanying that article was a summary for patients.³ Unfortunately, though, this summary failed to provide answers to the following questions: (1) What is my risk of dying of breast cancer if I am not screened? (2) What is the quantitative benefit if I am screened? (3) What is the quantitative harm if I am screened?

Filling the gap between recommendations and practice

To provide patients with the information they need to make an informed choice, I developed a one-page patient information sheet (PATIENT HANDOUT). My handout outlines the benefit and harms as numbers per 1000 women over 10 years, as was done in a recently validated patient education pamphlet.⁴

I used a readily available epidemiological source⁵ for the base-case (without screening) breast cancer death risks, and calculated the putative decreases in breast cancer deaths due to screening using the 15% relative risk reduction adopted by the ACP,² the USPSTF,⁶ and the Cochrane Collaboration editors.⁷

PATIENT HANDOUT

Weighing benefit and harms: Mammography for women, ages 40–49
For every 1000 women in their 40s, what are the benefits and harms of mammogram screening over a period of 10 years?

The science behind the bullet points

I kept the grid simple, but recognize that my colleagues would appreciate knowing where the numbers came from, and what was the basis for certain explanatory statements. So here is some background on the first three bullets.

• “Benefit: Less likely to die from breast cancer.” The explanation (#1) that is tagged to this statement notes that the benefit of mammograms may be larger if a woman is at higher risk. This statement was taken from the ACP meta-analysis and guideline.^1,2 The second part of this explanatory sentence, which indicates that “the benefit of mammograms may be smaller or even zero if the best quality research studies are correct” is based on the results of the Cochrane Review⁷ and is acknowledged in the ACP guideline.²
• “Harm: More likely to have false alarms (false positive mammograms).” I presented the percentage of false positive mammograms as 45% based on the 5-year (30%) and 10-year (56%) false-positive percentages reported for the Harvard Pilgrim Health Care study by the ACP.²
• “Harm: More likely to have unnecessary diagnosis and treatment for breast cancer.” I listed the number of women likely to suffer this harm as 5, based on the Cochrane Collaboration Review on screening for breast cancer with mammography.⁷

Is it easier to just order the test?

In talking with my colleagues, I gather that many are aware that controversies surround the “B”-rated USPSTF recommendation for mammography for women ages 40 to 49. However, my colleagues often lack the time to discuss the details with each patient, so they tell me that it’s “just easier to order the test.”

I disagree. Physicians can use the patient information sheet I’ve developed to provide basic relevant information to women who are trying to decide whether or not to get a mammogram. If my sheet doesn’t meet your needs, consider others. (See Fast Track, below right.)

The ACP may have fallen short with its mammography screening summary for patients, but your discussion with patients need not.

Acknowledgments

The author would like to acknowledge Phil Colmenares, MD, MPH, for his valuable comments and input into the design of the patient information handout.

I was disappointed, recently, when I read a handout from the American College of Physicians (ACP) that was designed to help physicians talk to patients about the benefits and harms of screening mammography in women ages 40 to 49. The problem: The summary simply didn’t provide enough relevant information to help patients make an informed choice.

The patient handout was part of a collection of ACP articles that appeared in the Annals of Internal Medicine. (See related POEM.) The first article was a systematic review of screening mammography in women 40 to 49 years of age documenting current data on benefit and harms.¹

The ACP also published a clinical practice guideline that included the recommendation that “clinicians should inform women 40 to 49 years of age about the potential benefits and harms of screening mammography.”² Accompanying that article was a summary for patients.³ Unfortunately, though, this summary failed to provide answers to the following questions: (1) What is my risk of dying of breast cancer if I am not screened? (2) What is the quantitative benefit if I am screened? (3) What is the quantitative harm if I am screened?

Filling the gap between recommendations and practice

To provide patients with the information they need to make an informed choice, I developed a one-page patient information sheet (PATIENT HANDOUT). My handout outlines the benefit and harms as numbers per 1000 women over 10 years, as was done in a recently validated patient education pamphlet.⁴

I used a readily available epidemiological source⁵ for the base-case (without screening) breast cancer death risks, and calculated the putative decreases in breast cancer deaths due to screening using the 15% relative risk reduction adopted by the ACP,² the USPSTF,⁶ and the Cochrane Collaboration editors.⁷

PATIENT HANDOUT

Weighing benefit and harms: Mammography for women, ages 40–49
For every 1000 women in their 40s, what are the benefits and harms of mammogram screening over a period of 10 years?

The science behind the bullet points

I kept the grid simple, but recognize that my colleagues would appreciate knowing where the numbers came from, and what was the basis for certain explanatory statements. So here is some background on the first three bullets.

• “Benefit: Less likely to die from breast cancer.” The explanation (#1) that is tagged to this statement notes that the benefit of mammograms may be larger if a woman is at higher risk. This statement was taken from the ACP meta-analysis and guideline.^1,2 The second part of this explanatory sentence, which indicates that “the benefit of mammograms may be smaller or even zero if the best quality research studies are correct” is based on the results of the Cochrane Review⁷ and is acknowledged in the ACP guideline.²
• “Harm: More likely to have false alarms (false positive mammograms).” I presented the percentage of false positive mammograms as 45% based on the 5-year (30%) and 10-year (56%) false-positive percentages reported for the Harvard Pilgrim Health Care study by the ACP.²
• “Harm: More likely to have unnecessary diagnosis and treatment for breast cancer.” I listed the number of women likely to suffer this harm as 5, based on the Cochrane Collaboration Review on screening for breast cancer with mammography.⁷

Is it easier to just order the test?

In talking with my colleagues, I gather that many are aware that controversies surround the “B”-rated USPSTF recommendation for mammography for women ages 40 to 49. However, my colleagues often lack the time to discuss the details with each patient, so they tell me that it’s “just easier to order the test.”

I disagree. Physicians can use the patient information sheet I’ve developed to provide basic relevant information to women who are trying to decide whether or not to get a mammogram. If my sheet doesn’t meet your needs, consider others. (See Fast Track, below right.)

The ACP may have fallen short with its mammography screening summary for patients, but your discussion with patients need not.

Acknowledgments

The author would like to acknowledge Phil Colmenares, MD, MPH, for his valuable comments and input into the design of the patient information handout.

I was disappointed, recently, when I read a handout from the American College of Physicians (ACP) that was designed to help physicians talk to patients about the benefits and harms of screening mammography in women ages 40 to 49. The problem: The summary simply didn’t provide enough relevant information to help patients make an informed choice.

The patient handout was part of a collection of ACP articles that appeared in the Annals of Internal Medicine. (See related POEM.) The first article was a systematic review of screening mammography in women 40 to 49 years of age documenting current data on benefit and harms.¹

The ACP also published a clinical practice guideline that included the recommendation that “clinicians should inform women 40 to 49 years of age about the potential benefits and harms of screening mammography.”² Accompanying that article was a summary for patients.³ Unfortunately, though, this summary failed to provide answers to the following questions: (1) What is my risk of dying of breast cancer if I am not screened? (2) What is the quantitative benefit if I am screened? (3) What is the quantitative harm if I am screened?

Filling the gap between recommendations and practice

To provide patients with the information they need to make an informed choice, I developed a one-page patient information sheet (PATIENT HANDOUT). My handout outlines the benefit and harms as numbers per 1000 women over 10 years, as was done in a recently validated patient education pamphlet.⁴

I used a readily available epidemiological source⁵ for the base-case (without screening) breast cancer death risks, and calculated the putative decreases in breast cancer deaths due to screening using the 15% relative risk reduction adopted by the ACP,² the USPSTF,⁶ and the Cochrane Collaboration editors.⁷

PATIENT HANDOUT

Weighing benefit and harms: Mammography for women, ages 40–49
For every 1000 women in their 40s, what are the benefits and harms of mammogram screening over a period of 10 years?

The science behind the bullet points

I kept the grid simple, but recognize that my colleagues would appreciate knowing where the numbers came from, and what was the basis for certain explanatory statements. So here is some background on the first three bullets.

• “Benefit: Less likely to die from breast cancer.” The explanation (#1) that is tagged to this statement notes that the benefit of mammograms may be larger if a woman is at higher risk. This statement was taken from the ACP meta-analysis and guideline.^1,2 The second part of this explanatory sentence, which indicates that “the benefit of mammograms may be smaller or even zero if the best quality research studies are correct” is based on the results of the Cochrane Review⁷ and is acknowledged in the ACP guideline.²
• “Harm: More likely to have false alarms (false positive mammograms).” I presented the percentage of false positive mammograms as 45% based on the 5-year (30%) and 10-year (56%) false-positive percentages reported for the Harvard Pilgrim Health Care study by the ACP.²
• “Harm: More likely to have unnecessary diagnosis and treatment for breast cancer.” I listed the number of women likely to suffer this harm as 5, based on the Cochrane Collaboration Review on screening for breast cancer with mammography.⁷

Is it easier to just order the test?

In talking with my colleagues, I gather that many are aware that controversies surround the “B”-rated USPSTF recommendation for mammography for women ages 40 to 49. However, my colleagues often lack the time to discuss the details with each patient, so they tell me that it’s “just easier to order the test.”

I disagree. Physicians can use the patient information sheet I’ve developed to provide basic relevant information to women who are trying to decide whether or not to get a mammogram. If my sheet doesn’t meet your needs, consider others. (See Fast Track, below right.)

The ACP may have fallen short with its mammography screening summary for patients, but your discussion with patients need not.

Acknowledgments

The author would like to acknowledge Phil Colmenares, MD, MPH, for his valuable comments and input into the design of the patient information handout.

A psychiatrist colleague once remarked that I did a better-than-average job diagnosing depression. Maybe so, but I was never quite sure if my patients really met established diagnostic criteria, since I usually “gestalted” the diagnosis. I would talk to my patients about depression so they would understand and accept it as a treatable medical condition, not a moral failing.

Recommending appropriate therapy was fairly easy, since information on pharmacotherapy is so prevalent. But I rarely learned how well my depressed patients were doing, since I practiced the “no news is good news” model of chronic disease management.

Then I was invited to participate in a practice-based study of depression.¹ I learned a better way to manage depression, which was more effective, satisfying, and—believe it or not—more efficient than the old way. I learned that successful management also saved time and money by reducing fruitless diagnostic testing and return visits for unmet needs.

The 5 Linked Tasks Of Successful Management

Successful management of depression requires completion of 5 linked tasks: recognition, diagnosis, treatment, systematic follow-up, and outcome assessment. Most important for the busy practitioner, these tasks can be completed without disrupting “usual” patient care.

What specific tools did I acquire? I’ll describe them according to tasks; the first task—recognition—deserves special emphasis and is discussed last.

Diagnosis: 9 simple questions

Whenever I suspect the possibility of depression in a patient, I ask the 9 simple questions in the PRIME-MD Patient Health Questionnaire (PHQ-9).^2,3 The PHQ-9 instrument is the best available depression screening tool for primary care.⁴ An affirmative answer to either question A or B plus a total of 5 positive “nearly every day” replies indicates a diagnosis of major depression. Fewer symptoms indicate minor depression, for which therapy is individualized (Figure 1).

It takes about 1 or 2 minutes to make a PHQ-9 diagnosis of depression; listening to the patient’s story fleshes out the diagnosis and may reveal underlying comorbidities (such as alcohol use or physical abuse) that guide the next task.

I minimize diagnostic testing. In 16 months, 1 of 95 patients in my practice with major depression turned out to have a significant underlying medical illness responsible for his symptoms. I diagnosed a lymphoma during systematic follow-up when treatment failed.

FIGURE 1
PRIME-MD Patient Health Questionnaire (PHQ-9)

Treatment: What to discuss

I use most of the office visit to discuss the diagnosis, patient preferences for treatment (or no treatment), and pharmacotherapy indications, expectations, and side effects. In my experience, this discussion is usually completed within the allotted time of the appointment.

Follow-up: Make 3 appointments

I ask patients to make 3 follow-up appointments: at 1 week (to assess medication side effects, if medication is prescribed), at 4 weeks (to assess early response), and at 8 weeks (to assess maintenance of response). After the week-8 visit, I ask patients to return every 3 months, or sooner if not at goal.

If the patient declines pharmacologic treatment, or if referral for counseling is the sole treatment decision, the schedule of follow-up visits may vary. Visits would then be spent assessing whether the patient is improving or in need of additional treatment.

Outcome assessment: Take “emotional temperature”

I obtain a quantitative “depression score” (Figure 2) at the initial diagnostic visit and at each follow-up visit. This ad hoc scoring system, developed by my psychiatrist partners, complements a patient’s report of progress. Scores collected from individual patients can also be used to assess overall improvement in the practice population of depressed patients.

I have found that patients understand and support having their “emotional temperature” taken at each visit. The PHQ-9 can be used to quantitate improvement during follow-up visits.³

FIGURE 2
Depression Scoring System

Recognition: Pandora’s box?

No matter how efficient, practical, and “userfriendly,” a protocol is of no value if unused. I suspect that, consciously or unconsciously, many physicians avoid opening the “Pandora’s box” of depression, because they fear disrupting the office schedule.

In my practice, recognizing and managing depression—and other unexpected patient issues—necessitated, I now realize, a complete re-engineering of the way I practiced medicine. Suppose you went to a shoe store on Friday to buy a pair of hiking boots for a weekend outing and were told, “Sorry, you’ll have to come back next week: we sell hiking boots only on Tuesdays and Thursdays.” Think how absurd that would be, and then realize it is exactly how most doctors schedule patient visits—on the “quota system,” despite good evidence that patients’ primary care needs cannot be conveniently categorized and scheduled.

All I know about my schedule at the start of a day is that I will encounter 25 to 30 patients, and that I will leave the office when it closes. I will not know in advance the complaints, treatments, and time I should spend with each patient. This flexible scheduling works well and liberates me to address patients’ spontaneous needs as they arise during the office encounter.

Case Finding: Don’t Trust Appearances

For depression, the case-finding approach is preferable to screening the entire practice.⁵ I no longer trust appearances to recognize depression. While participating in the practice-based study of depression,¹ I encountered a hardworking, cheerful female executive who had undergone many diagnostic procedures and who did not at all appear depressed—but she was. She got better after treatment and expressed profound gratitude. Not all severely affected patients present saying, “I’m depressed, please treat me.”

Unexplained somatic complaints

In my practice, one third of depressed patients complain only of a variety of somatic symptoms. I tell medical students, “If you can’t figure out what’s going on in an office encounter within minutes (1 to 3 minutes, depending on the level of training), think depression or alcohol abuse.” This clinical pearl is memorable and mostly accurate; the differential diagnosis can be expanded throughout training.

Stress? Suspect depression

Two thirds of my patients with diagnosed depression allude to “stress,” but often only in a very roundabout manner. So, whenever a patient utters the word “stress” during an office encounter, out comes the PHQ-9. I cannot think of a simpler technique for recognizing depression. In general, it is also well to remember that patients suffering from chronic diseases will have a higher-than-average likelihood of associated depression.

Managing depression should be simple

The management of depression should resemble the management of diabetes: astute case finding (mass screening for diabetes is not recommended), clearly defined diagnosis (blood sugar level), effective treatment (both behavioral and pharmacologic), systematic follow-up, and a quantitative outcome (glycosylated hemoglobin level). Diabetes was formerly managed as an acute illness, since once upon a time it was acceptable just to keep patients out of ketoacidosis. Today we treat diabetes better, and patients do better. Now it is time to do better with depression.

Correspondence
David L. Hahn, MD, MS, Arcand Park Clinic, 3434 East Washington Ave, Madison WI 53704. E-mail: dlhahn@wiscmail.wisc.edu

A psychiatrist colleague once remarked that I did a better-than-average job diagnosing depression. Maybe so, but I was never quite sure if my patients really met established diagnostic criteria, since I usually “gestalted” the diagnosis. I would talk to my patients about depression so they would understand and accept it as a treatable medical condition, not a moral failing.

Recommending appropriate therapy was fairly easy, since information on pharmacotherapy is so prevalent. But I rarely learned how well my depressed patients were doing, since I practiced the “no news is good news” model of chronic disease management.

Then I was invited to participate in a practice-based study of depression.¹ I learned a better way to manage depression, which was more effective, satisfying, and—believe it or not—more efficient than the old way. I learned that successful management also saved time and money by reducing fruitless diagnostic testing and return visits for unmet needs.

The 5 Linked Tasks Of Successful Management

Successful management of depression requires completion of 5 linked tasks: recognition, diagnosis, treatment, systematic follow-up, and outcome assessment. Most important for the busy practitioner, these tasks can be completed without disrupting “usual” patient care.

What specific tools did I acquire? I’ll describe them according to tasks; the first task—recognition—deserves special emphasis and is discussed last.

Diagnosis: 9 simple questions

Whenever I suspect the possibility of depression in a patient, I ask the 9 simple questions in the PRIME-MD Patient Health Questionnaire (PHQ-9).^2,3 The PHQ-9 instrument is the best available depression screening tool for primary care.⁴ An affirmative answer to either question A or B plus a total of 5 positive “nearly every day” replies indicates a diagnosis of major depression. Fewer symptoms indicate minor depression, for which therapy is individualized (Figure 1).

It takes about 1 or 2 minutes to make a PHQ-9 diagnosis of depression; listening to the patient’s story fleshes out the diagnosis and may reveal underlying comorbidities (such as alcohol use or physical abuse) that guide the next task.

I minimize diagnostic testing. In 16 months, 1 of 95 patients in my practice with major depression turned out to have a significant underlying medical illness responsible for his symptoms. I diagnosed a lymphoma during systematic follow-up when treatment failed.

FIGURE 1
PRIME-MD Patient Health Questionnaire (PHQ-9)

Treatment: What to discuss

I use most of the office visit to discuss the diagnosis, patient preferences for treatment (or no treatment), and pharmacotherapy indications, expectations, and side effects. In my experience, this discussion is usually completed within the allotted time of the appointment.

Follow-up: Make 3 appointments

I ask patients to make 3 follow-up appointments: at 1 week (to assess medication side effects, if medication is prescribed), at 4 weeks (to assess early response), and at 8 weeks (to assess maintenance of response). After the week-8 visit, I ask patients to return every 3 months, or sooner if not at goal.

If the patient declines pharmacologic treatment, or if referral for counseling is the sole treatment decision, the schedule of follow-up visits may vary. Visits would then be spent assessing whether the patient is improving or in need of additional treatment.

Outcome assessment: Take “emotional temperature”

I obtain a quantitative “depression score” (Figure 2) at the initial diagnostic visit and at each follow-up visit. This ad hoc scoring system, developed by my psychiatrist partners, complements a patient’s report of progress. Scores collected from individual patients can also be used to assess overall improvement in the practice population of depressed patients.

I have found that patients understand and support having their “emotional temperature” taken at each visit. The PHQ-9 can be used to quantitate improvement during follow-up visits.³

FIGURE 2
Depression Scoring System

Recognition: Pandora’s box?

No matter how efficient, practical, and “userfriendly,” a protocol is of no value if unused. I suspect that, consciously or unconsciously, many physicians avoid opening the “Pandora’s box” of depression, because they fear disrupting the office schedule.

In my practice, recognizing and managing depression—and other unexpected patient issues—necessitated, I now realize, a complete re-engineering of the way I practiced medicine. Suppose you went to a shoe store on Friday to buy a pair of hiking boots for a weekend outing and were told, “Sorry, you’ll have to come back next week: we sell hiking boots only on Tuesdays and Thursdays.” Think how absurd that would be, and then realize it is exactly how most doctors schedule patient visits—on the “quota system,” despite good evidence that patients’ primary care needs cannot be conveniently categorized and scheduled.

All I know about my schedule at the start of a day is that I will encounter 25 to 30 patients, and that I will leave the office when it closes. I will not know in advance the complaints, treatments, and time I should spend with each patient. This flexible scheduling works well and liberates me to address patients’ spontaneous needs as they arise during the office encounter.

Case Finding: Don’t Trust Appearances

For depression, the case-finding approach is preferable to screening the entire practice.⁵ I no longer trust appearances to recognize depression. While participating in the practice-based study of depression,¹ I encountered a hardworking, cheerful female executive who had undergone many diagnostic procedures and who did not at all appear depressed—but she was. She got better after treatment and expressed profound gratitude. Not all severely affected patients present saying, “I’m depressed, please treat me.”

Unexplained somatic complaints

In my practice, one third of depressed patients complain only of a variety of somatic symptoms. I tell medical students, “If you can’t figure out what’s going on in an office encounter within minutes (1 to 3 minutes, depending on the level of training), think depression or alcohol abuse.” This clinical pearl is memorable and mostly accurate; the differential diagnosis can be expanded throughout training.

Stress? Suspect depression

Two thirds of my patients with diagnosed depression allude to “stress,” but often only in a very roundabout manner. So, whenever a patient utters the word “stress” during an office encounter, out comes the PHQ-9. I cannot think of a simpler technique for recognizing depression. In general, it is also well to remember that patients suffering from chronic diseases will have a higher-than-average likelihood of associated depression.

Managing depression should be simple

The management of depression should resemble the management of diabetes: astute case finding (mass screening for diabetes is not recommended), clearly defined diagnosis (blood sugar level), effective treatment (both behavioral and pharmacologic), systematic follow-up, and a quantitative outcome (glycosylated hemoglobin level). Diabetes was formerly managed as an acute illness, since once upon a time it was acceptable just to keep patients out of ketoacidosis. Today we treat diabetes better, and patients do better. Now it is time to do better with depression.

Correspondence
David L. Hahn, MD, MS, Arcand Park Clinic, 3434 East Washington Ave, Madison WI 53704. E-mail: dlhahn@wiscmail.wisc.edu

A psychiatrist colleague once remarked that I did a better-than-average job diagnosing depression. Maybe so, but I was never quite sure if my patients really met established diagnostic criteria, since I usually “gestalted” the diagnosis. I would talk to my patients about depression so they would understand and accept it as a treatable medical condition, not a moral failing.

Recommending appropriate therapy was fairly easy, since information on pharmacotherapy is so prevalent. But I rarely learned how well my depressed patients were doing, since I practiced the “no news is good news” model of chronic disease management.

Then I was invited to participate in a practice-based study of depression.¹ I learned a better way to manage depression, which was more effective, satisfying, and—believe it or not—more efficient than the old way. I learned that successful management also saved time and money by reducing fruitless diagnostic testing and return visits for unmet needs.

The 5 Linked Tasks Of Successful Management

Successful management of depression requires completion of 5 linked tasks: recognition, diagnosis, treatment, systematic follow-up, and outcome assessment. Most important for the busy practitioner, these tasks can be completed without disrupting “usual” patient care.

What specific tools did I acquire? I’ll describe them according to tasks; the first task—recognition—deserves special emphasis and is discussed last.

Diagnosis: 9 simple questions

Whenever I suspect the possibility of depression in a patient, I ask the 9 simple questions in the PRIME-MD Patient Health Questionnaire (PHQ-9).^2,3 The PHQ-9 instrument is the best available depression screening tool for primary care.⁴ An affirmative answer to either question A or B plus a total of 5 positive “nearly every day” replies indicates a diagnosis of major depression. Fewer symptoms indicate minor depression, for which therapy is individualized (Figure 1).

It takes about 1 or 2 minutes to make a PHQ-9 diagnosis of depression; listening to the patient’s story fleshes out the diagnosis and may reveal underlying comorbidities (such as alcohol use or physical abuse) that guide the next task.

I minimize diagnostic testing. In 16 months, 1 of 95 patients in my practice with major depression turned out to have a significant underlying medical illness responsible for his symptoms. I diagnosed a lymphoma during systematic follow-up when treatment failed.

FIGURE 1
PRIME-MD Patient Health Questionnaire (PHQ-9)

Treatment: What to discuss

I use most of the office visit to discuss the diagnosis, patient preferences for treatment (or no treatment), and pharmacotherapy indications, expectations, and side effects. In my experience, this discussion is usually completed within the allotted time of the appointment.

Follow-up: Make 3 appointments

I ask patients to make 3 follow-up appointments: at 1 week (to assess medication side effects, if medication is prescribed), at 4 weeks (to assess early response), and at 8 weeks (to assess maintenance of response). After the week-8 visit, I ask patients to return every 3 months, or sooner if not at goal.

If the patient declines pharmacologic treatment, or if referral for counseling is the sole treatment decision, the schedule of follow-up visits may vary. Visits would then be spent assessing whether the patient is improving or in need of additional treatment.

Outcome assessment: Take “emotional temperature”

I obtain a quantitative “depression score” (Figure 2) at the initial diagnostic visit and at each follow-up visit. This ad hoc scoring system, developed by my psychiatrist partners, complements a patient’s report of progress. Scores collected from individual patients can also be used to assess overall improvement in the practice population of depressed patients.

I have found that patients understand and support having their “emotional temperature” taken at each visit. The PHQ-9 can be used to quantitate improvement during follow-up visits.³

FIGURE 2
Depression Scoring System

Recognition: Pandora’s box?

No matter how efficient, practical, and “userfriendly,” a protocol is of no value if unused. I suspect that, consciously or unconsciously, many physicians avoid opening the “Pandora’s box” of depression, because they fear disrupting the office schedule.

In my practice, recognizing and managing depression—and other unexpected patient issues—necessitated, I now realize, a complete re-engineering of the way I practiced medicine. Suppose you went to a shoe store on Friday to buy a pair of hiking boots for a weekend outing and were told, “Sorry, you’ll have to come back next week: we sell hiking boots only on Tuesdays and Thursdays.” Think how absurd that would be, and then realize it is exactly how most doctors schedule patient visits—on the “quota system,” despite good evidence that patients’ primary care needs cannot be conveniently categorized and scheduled.

All I know about my schedule at the start of a day is that I will encounter 25 to 30 patients, and that I will leave the office when it closes. I will not know in advance the complaints, treatments, and time I should spend with each patient. This flexible scheduling works well and liberates me to address patients’ spontaneous needs as they arise during the office encounter.

Case Finding: Don’t Trust Appearances

For depression, the case-finding approach is preferable to screening the entire practice.⁵ I no longer trust appearances to recognize depression. While participating in the practice-based study of depression,¹ I encountered a hardworking, cheerful female executive who had undergone many diagnostic procedures and who did not at all appear depressed—but she was. She got better after treatment and expressed profound gratitude. Not all severely affected patients present saying, “I’m depressed, please treat me.”

Unexplained somatic complaints

In my practice, one third of depressed patients complain only of a variety of somatic symptoms. I tell medical students, “If you can’t figure out what’s going on in an office encounter within minutes (1 to 3 minutes, depending on the level of training), think depression or alcohol abuse.” This clinical pearl is memorable and mostly accurate; the differential diagnosis can be expanded throughout training.

Stress? Suspect depression

Two thirds of my patients with diagnosed depression allude to “stress,” but often only in a very roundabout manner. So, whenever a patient utters the word “stress” during an office encounter, out comes the PHQ-9. I cannot think of a simpler technique for recognizing depression. In general, it is also well to remember that patients suffering from chronic diseases will have a higher-than-average likelihood of associated depression.

Managing depression should be simple

The management of depression should resemble the management of diabetes: astute case finding (mass screening for diabetes is not recommended), clearly defined diagnosis (blood sugar level), effective treatment (both behavioral and pharmacologic), systematic follow-up, and a quantitative outcome (glycosylated hemoglobin level). Diabetes was formerly managed as an acute illness, since once upon a time it was acceptable just to keep patients out of ketoacidosis. Today we treat diabetes better, and patients do better. Now it is time to do better with depression.

Correspondence
David L. Hahn, MD, MS, Arcand Park Clinic, 3434 East Washington Ave, Madison WI 53704. E-mail: dlhahn@wiscmail.wisc.edu

Preventive care to the well patient has become an accepted activity in primary care.¹ Clinical preventive services require a high standard of proof of effectiveness.² Although evidence-based guidelines¹ and goals³ have been published, optimal delivery has been achieved only infrequently.⁴ The formidable list of its potential barriers⁴ includes physician factors (lack of consensus, motivation, compensation, time), patient factors (age, race, sex, patient acceptance, insurance status, visit frequency, not scheduling preventive visits), and system factors (absence of paramedical assistance, disorganized medical records, fragmentation of care, lack of a systematic delivery program). Strategies to improve the delivery of preventive services must recognize the complexity and uniqueness of each office practice^5-7 and advocate the use of quality improvement methodologies.⁸

Previous studies have shown that insurance coverage significantly affects preventive service delivery,⁹ but the fear that health maintenance organization (HMO) patients would receive fewer preventive services than fee-for-service (FFS) patients¹⁰ is unfounded.¹¹ The favorable effect of HMO enrollment on preventive service delivery is probably because of insurance coverage rather than self-selection.¹¹ Many studies of FFS and HMO care delivery have compared one physician group seeing FFS patients exclusively with another group seeing HMO patients.^12-18 In that study design, differences may be related to confounding physician and system factors rather than the type of insurance coverage.¹⁹

Physicians have been trained to deliver preventive care during the annual complete physical examination. The effectiveness of the traditional complete physical as a vehicle for preventive service delivery has been questioned recently,²⁰ however, and delivery of preventive services during acute care visits has been advocated as a more effective alternative.^21,22 Indeed, systematic delivery strategies that include the offer of these services during acute care visits have achieved the highest levels of preventive service delivery in primary care settings.^4,23 But these studies were performed in a small number of private practice settings, and it is unclear whether the same results apply in the larger organized systems of care.²⁴

As part of a program to increase delivery of adult clinical preventive services, we determined the major factors associated with the performance of specified preventive services in a large physician-owned health delivery system that included both HMO and FFS patients. We particularly sought to measure the quantity of preventive services associated with insurance status (HMO vs FFS) and visit type (preventive visits vs acute care only).

Methods

Practice Setting

We studied a multispecialty group practice with 15 offices located in Dane County, Wisconsin. The service area is representative of the county and contains one midsize city (190,000 population), several smaller cities and towns, and a surrounding rural area. During the audit period there was a county population of 351,362, of whom 149,225 (42%) were enrolled in HMOs, 7027 (2%) were covered by Medicaid, and 42,193 (12%) were uninsured. The group practice owned and operated its own prepaid health plan (DeanCare HMO), which was clinically managed by the physician-owners who were compensated for their services on the basis of a discounted FFS formula that did not provide financial incentives to manage HMO and FFS patients differently. The HMO paid for all preventive services deemed appropriate by the physician, while many FFS insurance carriers did not.

Audit of Computerized Billing Records

The group practice computerized billing system contained a record for every encounter. The records include provider and patient codes, the site of service, diagnosis codes, and any billable tests obtained within the group’s clinical laboratory and radiology facilities. Only group practice sites that used these facilities exclusively were audited. Patient demographic information was obtained by cross-referencing the patient code with a master enrollment file. Computerized billing data for 1985 to 1995 were audited. For each patient, yearly data were abstracted that included a flag for the presence/absence (1/0) of a billable preventive procedure. From this data file, a set of positive criteria was constructed to determine the prevalence of preventive service delivery on the basis of age, sex, and frequency criteria Table 1. Positive criteria were modeled after those developed by the National Committee on Quality Assurance.²⁵

The audited population included all adult outpatients (aged 18 years and older) who visited a defined panel of primary care providers at 10 sites at least once in 1995. In addition to age and sex, the following data were abstracted: billable preventive procedures (Papanicolaou test, screening mammogram, cholesterol test, tetanus immunization, fecal occult blood test, and sigmoidoscopy), total visits to primary care providers, insurance type (HMO, FFS, Medicare, Medicaid, or workers’ compensation), and visit type (acute care–only or scheduled preventive care visit). Cholesterol testing as part of a chemistry panel was counted separately from single cholesterol determinations or lipid profiles, which were classified as screening.

Computer Audit Validation Using Medical Record Reviews

At one study site, the computer audit results were validated by manual review of 200 randomly selected medical records. The validation audit found good agreement (±2.5%) for documentation of preventive procedures; the proportion of positive criteria for the computerized audit was always within 2.5% of that for the manual chart review of the same patient population. Computerized visit frequency counts were highly correlated with medical record review counts (R = 0.93; P <.0001). Medical record review also showed that, except for sigmoidoscopy, approximately 98% of audited preventive procedures had been performed for screening, not for symptoms. (Sixty-seven percent of sigmoidoscopies had been done for screening.)

A separate review of a stratified sample of 245 medical records⁴ found evidence that screening was performed outside the system of care (and not recorded in the transaction files) in 0.5% or fewer of preventive procedures; the only exception was Papanicolaou tests, which had been obtained outside the system in 13.8% of women aged 20 to 29 years and 4.5% of those aged 30 to 39 years.

Statistical Testing

We explored univariate associations between age, sex, visit frequency, insurance type, and visit type using analysis of variance for continuous dependent variables and the Fisher exact test for binary categorical variables. We used logistic regression to test whether preventive services were independently associated with those same variables. For the statistical testing, visit frequency was defined as all visits to primary care providers in 1995, and presence of a visit scheduled specifically for preventive care was coded as positive if one occurred during 1994 or 1995 (P <.05 was reported as significant).

Results

Thirty-five family practice physicians, 33 general internists, 15 obstetrician/gynecologists, and 12 physician assistants supervised by the primary care physicians made up the primary care provider panel. This panel encountered 75,621 outpatients at least once in 1995. Characteristics of this patient group are presented in Table 2.

Age was positively associated with men (mean = 47 years vs 45 years for women), visit frequency (R = .073), HMO insurance (mean = 45 years vs 41 years for FFS) and acute care–visits (mean = 48 years vs 43 years for preventive care visit) (P <.001 for all comparisons). Women were positively associated with visit frequency (mean = 3.4 visits per year vs 2.6 visits for men) and preventive care visits (52% of all women scheduled such visits vs 22% of men ).

HMO membership was equal for both sexes. Visit frequency was positively associated with HMO membership (mean = 3.2 visits vs 2.6 for FFS) and preventive care visits (mean = 3.1 visits vs 3.0 for acute care only) (P <.001 for both). HMO members scheduled more preventive visits than FFS patients in 1995 (36% and 31%, respectively; P <.001).

Table 3 presents the results for the Papanicolaou test, mammography, cholesterol, tetanus immunization, fecal occult blood testing, and sigmoidoscopy. In general, older patients received more preventive services than younger patients; differences by sex were inconsistent. Patients with 3 or more visits had consistently more positive criteria (7% to 12% more) than patients encountered only once or twice in 1995. HMO patients had 6% to 13% more screening than FFS patients. Associations with visit type were greatest: Patients who scheduled at least one preventive visit had 9% to 45% more positive criteria than patients seen only for acute care. Because of the large sample size, all these differences were statistically significant. Even a trivial positive association of age and mammography use achieved statistical significance (P = .01) in the multivariate model. Logistic regression analyses showed that, for every preventive service audited, the associations of visit frequency and type were statistically independent (P <.001 in all cases).

Table 4 illustrates the profound interaction between insurance type and visit type. HMO patients who scheduled a preventive visit had the highest rates of screening, while FFS patients seen only for acute care had rates that were 14% to 50% lower. FFS patients who scheduled a preventive visit had rates approaching the high rates of screening of the HMO/preventive visit group. Interestingly, HMO patients seen only for acute care had low rates of screening, closer to those of FFS/acute care–only than to the HMO patients who scheduled a preventive visit.

Discussion

The audit methodology we used in this study measured the delivery of adult clinical preventive services in a multispecialty, multisite group practice health system that treats both FFS and HMO patients. The audited group included all patients having at least one face-to-face encounter with a primary care provider in 1995. This methodology excluded patients who were not seen during the audited year. Peripheral sites where billable services were not recorded were also excluded. Positive audit criteria were determined according to nationally recognized norms.²⁵

Limitations

Two random medical record reviews validated the computerized billing file audit showing that most recorded preventive procedures were performed for screening, and few patients had received any preventive services elsewhere. (The exceptions were the approximately 33% of sigmoidoscopies that were done for symptoms and a small percentage of Papanicolaou tests that were performed outside the system.) The magnitude of these discrepancies was insufficient to alter the conclusions of our study. It is possible that physicians did not code for some preventive visits scheduled by FFS patients whose insurance did not cover prevention. If so, our audit could have underestimated the association of scheduling preventive visits with the delivery of such services for FFS patients. There was no incentive for physicians to undercode preventive visits for HMO patients. The system of care did not record assignment of patients to individual physicians. Therefore, it was not possible to audit individual providers. The system’s database did not record demographic information, such as education, income, ethnic origin, or race. Thus, we could not analyze health system performance in regard to socioeconomic status, which is an important additional predictor of preventive service use.²⁶ Nonbillable services, such as smoking cessation counseling and blood pressure testing, were not measurable using our methodology.

FFS Versus HMO Insurance

With one exception,²⁷ studies of HMO and FFS care have compared one group treating HMO patients with another group providing care for FFS patients. That type of study design raises the concern that system factors (different provider group attitudes, training, system access, protocols, and so forth) were responsible for the differences reported between HMO and FFS care. It is unlikely that this happened in our study, because both HMO and FFS patients were seen within the same system of care by the same physicians whose compensation formula did not discriminate between insurance types. In this system of care, HMO members had insurance to pay for any preventive service offered by the physician, while coverage for prevention was not uniformly available to FFS policy holders.

Our results that show HMO patients received more preventive services than FFS patients agree with the results of the National Health Interview Survey (NHIS),¹¹ which included a representative sample of the US population. It is unlikely that HMO self-selection accounted for the higher preventive services delivery rates for HMO patients in these studies. After controlling for factors correlated with selection into HMOs, health status, and use of medical services, results of the NHIS were not altered. In our study, only half of the HMO patients scheduled a preventive visit; the other half had lower screening rates Table 4.

The most important determinant of access to health care is having health insurance.²⁸ We believe the most simple explanation for the superiority of preventive service delivery to HMO patients in our system of care is that FFS patients without preventive coverage are reluctant to pay out-of-pocket expenses for these services. This belief is supported by the results of a recent study in managed care settings showing that physician compensation method was not significantly related to use, while plan benefit level was positively related to increased service delivery.⁹ Thus, providing insurance coverage to pay for preventive care is one potential strategy for increasing delivery.

Preventive Visits Versus Acute Care

Visit frequency has previously been shown to have a positive association with the delivery of preventive services.^4,29 Our study confirms the association with visit frequency, but found a statistically independent and greater association with visit type. Similar to results in other primary care studies,³⁰ we found that a scheduled preventive visit was strongly predictive of the delivery of preventive services.

One strategy, then, for rectifying this discrepancy is to insist that all patients schedule a yearly preventive visit. Currently only one third of patients in primary care settings³¹ schedule such a visit. A simple calculation demonstrates, however, that if primary physicians in our audit spent 30 additional minutes each year performing a complete physical examination for the approximately 54,000 patients currently seen for acute care only, there would be little or no time remaining to care for sick patients. Since the audited group represents only a part of the whole, the systemwide impact would be even greater. Additionally, when one author (D.L.H.) systematically invited all adults aged older than 50 years to schedule a complete physical examination, most did not do so (unpublished observations).

We agree with Frame²⁰ that it is neither feasible nor necessary to insist that all patients schedule preventive visits to receive preventive care. The delivery of preventive services during acute care visits has been advocated as a necessary strategy to deliver adequate services to entire patient populations.^21,22 People want physicians to provide more preventive services,³² and many patients who have not scheduled preventive visits will accept them if the physician offers.^4,33 Preventive services delivery during illness visits is common in community practice,³³ though it is more prevalent in high-risk than average-risk patients.³⁴ Other primary care–based studies show that the best results are accomplished when these services are systematically offered to all patients during illness visits, regardless of risk status.^4,23,35,36 Offering preventive care during acute care visits can be integrated into office practice^4,33,34 and can be effective.^4,36

Conclusions

Coordinating efforts to offer evidence-based preventive services to all patients seen for acute care visits is a potential strategy to increase the delivery of this care. Implementation of this strategy could result in increased screening for patients enrolled in HMOs and for FFS patients with coverage. In 1996, Dean Health System began implementing a systemwide adult health maintenance guideline that emphasizes assessment and delivery during acute care visits and includes provider reports of guideline adherence. After 3 years of implementation, providers reported that more than 60,000 patients (mean = 1775 per month) were treated in accordance with the guideline. Follow-up audits are planned to monitor the outcome of this strategy to assess and offer preventive services during acute care visits.

Acknowledgments

We would like to thank Kurt C. Stange, MD, PhD, for constructive suggestions during the preparation of this manuscript, and Marcie Berger, MD; Jonathon Berkhoff, MD; Gene Bettinger; and Suzanne Hallett for technical assistance.

Preventive care to the well patient has become an accepted activity in primary care.¹ Clinical preventive services require a high standard of proof of effectiveness.² Although evidence-based guidelines¹ and goals³ have been published, optimal delivery has been achieved only infrequently.⁴ The formidable list of its potential barriers⁴ includes physician factors (lack of consensus, motivation, compensation, time), patient factors (age, race, sex, patient acceptance, insurance status, visit frequency, not scheduling preventive visits), and system factors (absence of paramedical assistance, disorganized medical records, fragmentation of care, lack of a systematic delivery program). Strategies to improve the delivery of preventive services must recognize the complexity and uniqueness of each office practice^5-7 and advocate the use of quality improvement methodologies.⁸

Previous studies have shown that insurance coverage significantly affects preventive service delivery,⁹ but the fear that health maintenance organization (HMO) patients would receive fewer preventive services than fee-for-service (FFS) patients¹⁰ is unfounded.¹¹ The favorable effect of HMO enrollment on preventive service delivery is probably because of insurance coverage rather than self-selection.¹¹ Many studies of FFS and HMO care delivery have compared one physician group seeing FFS patients exclusively with another group seeing HMO patients.^12-18 In that study design, differences may be related to confounding physician and system factors rather than the type of insurance coverage.¹⁹

Physicians have been trained to deliver preventive care during the annual complete physical examination. The effectiveness of the traditional complete physical as a vehicle for preventive service delivery has been questioned recently,²⁰ however, and delivery of preventive services during acute care visits has been advocated as a more effective alternative.^21,22 Indeed, systematic delivery strategies that include the offer of these services during acute care visits have achieved the highest levels of preventive service delivery in primary care settings.^4,23 But these studies were performed in a small number of private practice settings, and it is unclear whether the same results apply in the larger organized systems of care.²⁴

As part of a program to increase delivery of adult clinical preventive services, we determined the major factors associated with the performance of specified preventive services in a large physician-owned health delivery system that included both HMO and FFS patients. We particularly sought to measure the quantity of preventive services associated with insurance status (HMO vs FFS) and visit type (preventive visits vs acute care only).

Methods

Practice Setting

We studied a multispecialty group practice with 15 offices located in Dane County, Wisconsin. The service area is representative of the county and contains one midsize city (190,000 population), several smaller cities and towns, and a surrounding rural area. During the audit period there was a county population of 351,362, of whom 149,225 (42%) were enrolled in HMOs, 7027 (2%) were covered by Medicaid, and 42,193 (12%) were uninsured. The group practice owned and operated its own prepaid health plan (DeanCare HMO), which was clinically managed by the physician-owners who were compensated for their services on the basis of a discounted FFS formula that did not provide financial incentives to manage HMO and FFS patients differently. The HMO paid for all preventive services deemed appropriate by the physician, while many FFS insurance carriers did not.

Audit of Computerized Billing Records

The group practice computerized billing system contained a record for every encounter. The records include provider and patient codes, the site of service, diagnosis codes, and any billable tests obtained within the group’s clinical laboratory and radiology facilities. Only group practice sites that used these facilities exclusively were audited. Patient demographic information was obtained by cross-referencing the patient code with a master enrollment file. Computerized billing data for 1985 to 1995 were audited. For each patient, yearly data were abstracted that included a flag for the presence/absence (1/0) of a billable preventive procedure. From this data file, a set of positive criteria was constructed to determine the prevalence of preventive service delivery on the basis of age, sex, and frequency criteria Table 1. Positive criteria were modeled after those developed by the National Committee on Quality Assurance.²⁵

The audited population included all adult outpatients (aged 18 years and older) who visited a defined panel of primary care providers at 10 sites at least once in 1995. In addition to age and sex, the following data were abstracted: billable preventive procedures (Papanicolaou test, screening mammogram, cholesterol test, tetanus immunization, fecal occult blood test, and sigmoidoscopy), total visits to primary care providers, insurance type (HMO, FFS, Medicare, Medicaid, or workers’ compensation), and visit type (acute care–only or scheduled preventive care visit). Cholesterol testing as part of a chemistry panel was counted separately from single cholesterol determinations or lipid profiles, which were classified as screening.

Computer Audit Validation Using Medical Record Reviews

At one study site, the computer audit results were validated by manual review of 200 randomly selected medical records. The validation audit found good agreement (±2.5%) for documentation of preventive procedures; the proportion of positive criteria for the computerized audit was always within 2.5% of that for the manual chart review of the same patient population. Computerized visit frequency counts were highly correlated with medical record review counts (R = 0.93; P <.0001). Medical record review also showed that, except for sigmoidoscopy, approximately 98% of audited preventive procedures had been performed for screening, not for symptoms. (Sixty-seven percent of sigmoidoscopies had been done for screening.)

A separate review of a stratified sample of 245 medical records⁴ found evidence that screening was performed outside the system of care (and not recorded in the transaction files) in 0.5% or fewer of preventive procedures; the only exception was Papanicolaou tests, which had been obtained outside the system in 13.8% of women aged 20 to 29 years and 4.5% of those aged 30 to 39 years.

Statistical Testing

We explored univariate associations between age, sex, visit frequency, insurance type, and visit type using analysis of variance for continuous dependent variables and the Fisher exact test for binary categorical variables. We used logistic regression to test whether preventive services were independently associated with those same variables. For the statistical testing, visit frequency was defined as all visits to primary care providers in 1995, and presence of a visit scheduled specifically for preventive care was coded as positive if one occurred during 1994 or 1995 (P <.05 was reported as significant).

Results

Thirty-five family practice physicians, 33 general internists, 15 obstetrician/gynecologists, and 12 physician assistants supervised by the primary care physicians made up the primary care provider panel. This panel encountered 75,621 outpatients at least once in 1995. Characteristics of this patient group are presented in Table 2.

Age was positively associated with men (mean = 47 years vs 45 years for women), visit frequency (R = .073), HMO insurance (mean = 45 years vs 41 years for FFS) and acute care–visits (mean = 48 years vs 43 years for preventive care visit) (P <.001 for all comparisons). Women were positively associated with visit frequency (mean = 3.4 visits per year vs 2.6 visits for men) and preventive care visits (52% of all women scheduled such visits vs 22% of men ).

HMO membership was equal for both sexes. Visit frequency was positively associated with HMO membership (mean = 3.2 visits vs 2.6 for FFS) and preventive care visits (mean = 3.1 visits vs 3.0 for acute care only) (P <.001 for both). HMO members scheduled more preventive visits than FFS patients in 1995 (36% and 31%, respectively; P <.001).

Table 3 presents the results for the Papanicolaou test, mammography, cholesterol, tetanus immunization, fecal occult blood testing, and sigmoidoscopy. In general, older patients received more preventive services than younger patients; differences by sex were inconsistent. Patients with 3 or more visits had consistently more positive criteria (7% to 12% more) than patients encountered only once or twice in 1995. HMO patients had 6% to 13% more screening than FFS patients. Associations with visit type were greatest: Patients who scheduled at least one preventive visit had 9% to 45% more positive criteria than patients seen only for acute care. Because of the large sample size, all these differences were statistically significant. Even a trivial positive association of age and mammography use achieved statistical significance (P = .01) in the multivariate model. Logistic regression analyses showed that, for every preventive service audited, the associations of visit frequency and type were statistically independent (P <.001 in all cases).

Table 4 illustrates the profound interaction between insurance type and visit type. HMO patients who scheduled a preventive visit had the highest rates of screening, while FFS patients seen only for acute care had rates that were 14% to 50% lower. FFS patients who scheduled a preventive visit had rates approaching the high rates of screening of the HMO/preventive visit group. Interestingly, HMO patients seen only for acute care had low rates of screening, closer to those of FFS/acute care–only than to the HMO patients who scheduled a preventive visit.

Discussion

The audit methodology we used in this study measured the delivery of adult clinical preventive services in a multispecialty, multisite group practice health system that treats both FFS and HMO patients. The audited group included all patients having at least one face-to-face encounter with a primary care provider in 1995. This methodology excluded patients who were not seen during the audited year. Peripheral sites where billable services were not recorded were also excluded. Positive audit criteria were determined according to nationally recognized norms.²⁵

Limitations

Two random medical record reviews validated the computerized billing file audit showing that most recorded preventive procedures were performed for screening, and few patients had received any preventive services elsewhere. (The exceptions were the approximately 33% of sigmoidoscopies that were done for symptoms and a small percentage of Papanicolaou tests that were performed outside the system.) The magnitude of these discrepancies was insufficient to alter the conclusions of our study. It is possible that physicians did not code for some preventive visits scheduled by FFS patients whose insurance did not cover prevention. If so, our audit could have underestimated the association of scheduling preventive visits with the delivery of such services for FFS patients. There was no incentive for physicians to undercode preventive visits for HMO patients. The system of care did not record assignment of patients to individual physicians. Therefore, it was not possible to audit individual providers. The system’s database did not record demographic information, such as education, income, ethnic origin, or race. Thus, we could not analyze health system performance in regard to socioeconomic status, which is an important additional predictor of preventive service use.²⁶ Nonbillable services, such as smoking cessation counseling and blood pressure testing, were not measurable using our methodology.

FFS Versus HMO Insurance

With one exception,²⁷ studies of HMO and FFS care have compared one group treating HMO patients with another group providing care for FFS patients. That type of study design raises the concern that system factors (different provider group attitudes, training, system access, protocols, and so forth) were responsible for the differences reported between HMO and FFS care. It is unlikely that this happened in our study, because both HMO and FFS patients were seen within the same system of care by the same physicians whose compensation formula did not discriminate between insurance types. In this system of care, HMO members had insurance to pay for any preventive service offered by the physician, while coverage for prevention was not uniformly available to FFS policy holders.

Our results that show HMO patients received more preventive services than FFS patients agree with the results of the National Health Interview Survey (NHIS),¹¹ which included a representative sample of the US population. It is unlikely that HMO self-selection accounted for the higher preventive services delivery rates for HMO patients in these studies. After controlling for factors correlated with selection into HMOs, health status, and use of medical services, results of the NHIS were not altered. In our study, only half of the HMO patients scheduled a preventive visit; the other half had lower screening rates Table 4.

The most important determinant of access to health care is having health insurance.²⁸ We believe the most simple explanation for the superiority of preventive service delivery to HMO patients in our system of care is that FFS patients without preventive coverage are reluctant to pay out-of-pocket expenses for these services. This belief is supported by the results of a recent study in managed care settings showing that physician compensation method was not significantly related to use, while plan benefit level was positively related to increased service delivery.⁹ Thus, providing insurance coverage to pay for preventive care is one potential strategy for increasing delivery.

Preventive Visits Versus Acute Care

Visit frequency has previously been shown to have a positive association with the delivery of preventive services.^4,29 Our study confirms the association with visit frequency, but found a statistically independent and greater association with visit type. Similar to results in other primary care studies,³⁰ we found that a scheduled preventive visit was strongly predictive of the delivery of preventive services.

One strategy, then, for rectifying this discrepancy is to insist that all patients schedule a yearly preventive visit. Currently only one third of patients in primary care settings³¹ schedule such a visit. A simple calculation demonstrates, however, that if primary physicians in our audit spent 30 additional minutes each year performing a complete physical examination for the approximately 54,000 patients currently seen for acute care only, there would be little or no time remaining to care for sick patients. Since the audited group represents only a part of the whole, the systemwide impact would be even greater. Additionally, when one author (D.L.H.) systematically invited all adults aged older than 50 years to schedule a complete physical examination, most did not do so (unpublished observations).

We agree with Frame²⁰ that it is neither feasible nor necessary to insist that all patients schedule preventive visits to receive preventive care. The delivery of preventive services during acute care visits has been advocated as a necessary strategy to deliver adequate services to entire patient populations.^21,22 People want physicians to provide more preventive services,³² and many patients who have not scheduled preventive visits will accept them if the physician offers.^4,33 Preventive services delivery during illness visits is common in community practice,³³ though it is more prevalent in high-risk than average-risk patients.³⁴ Other primary care–based studies show that the best results are accomplished when these services are systematically offered to all patients during illness visits, regardless of risk status.^4,23,35,36 Offering preventive care during acute care visits can be integrated into office practice^4,33,34 and can be effective.^4,36

Conclusions

Coordinating efforts to offer evidence-based preventive services to all patients seen for acute care visits is a potential strategy to increase the delivery of this care. Implementation of this strategy could result in increased screening for patients enrolled in HMOs and for FFS patients with coverage. In 1996, Dean Health System began implementing a systemwide adult health maintenance guideline that emphasizes assessment and delivery during acute care visits and includes provider reports of guideline adherence. After 3 years of implementation, providers reported that more than 60,000 patients (mean = 1775 per month) were treated in accordance with the guideline. Follow-up audits are planned to monitor the outcome of this strategy to assess and offer preventive services during acute care visits.

Acknowledgments

We would like to thank Kurt C. Stange, MD, PhD, for constructive suggestions during the preparation of this manuscript, and Marcie Berger, MD; Jonathon Berkhoff, MD; Gene Bettinger; and Suzanne Hallett for technical assistance.

Preventive care to the well patient has become an accepted activity in primary care.¹ Clinical preventive services require a high standard of proof of effectiveness.² Although evidence-based guidelines¹ and goals³ have been published, optimal delivery has been achieved only infrequently.⁴ The formidable list of its potential barriers⁴ includes physician factors (lack of consensus, motivation, compensation, time), patient factors (age, race, sex, patient acceptance, insurance status, visit frequency, not scheduling preventive visits), and system factors (absence of paramedical assistance, disorganized medical records, fragmentation of care, lack of a systematic delivery program). Strategies to improve the delivery of preventive services must recognize the complexity and uniqueness of each office practice^5-7 and advocate the use of quality improvement methodologies.⁸

Previous studies have shown that insurance coverage significantly affects preventive service delivery,⁹ but the fear that health maintenance organization (HMO) patients would receive fewer preventive services than fee-for-service (FFS) patients¹⁰ is unfounded.¹¹ The favorable effect of HMO enrollment on preventive service delivery is probably because of insurance coverage rather than self-selection.¹¹ Many studies of FFS and HMO care delivery have compared one physician group seeing FFS patients exclusively with another group seeing HMO patients.^12-18 In that study design, differences may be related to confounding physician and system factors rather than the type of insurance coverage.¹⁹

Physicians have been trained to deliver preventive care during the annual complete physical examination. The effectiveness of the traditional complete physical as a vehicle for preventive service delivery has been questioned recently,²⁰ however, and delivery of preventive services during acute care visits has been advocated as a more effective alternative.^21,22 Indeed, systematic delivery strategies that include the offer of these services during acute care visits have achieved the highest levels of preventive service delivery in primary care settings.^4,23 But these studies were performed in a small number of private practice settings, and it is unclear whether the same results apply in the larger organized systems of care.²⁴

As part of a program to increase delivery of adult clinical preventive services, we determined the major factors associated with the performance of specified preventive services in a large physician-owned health delivery system that included both HMO and FFS patients. We particularly sought to measure the quantity of preventive services associated with insurance status (HMO vs FFS) and visit type (preventive visits vs acute care only).

Methods

Practice Setting

We studied a multispecialty group practice with 15 offices located in Dane County, Wisconsin. The service area is representative of the county and contains one midsize city (190,000 population), several smaller cities and towns, and a surrounding rural area. During the audit period there was a county population of 351,362, of whom 149,225 (42%) were enrolled in HMOs, 7027 (2%) were covered by Medicaid, and 42,193 (12%) were uninsured. The group practice owned and operated its own prepaid health plan (DeanCare HMO), which was clinically managed by the physician-owners who were compensated for their services on the basis of a discounted FFS formula that did not provide financial incentives to manage HMO and FFS patients differently. The HMO paid for all preventive services deemed appropriate by the physician, while many FFS insurance carriers did not.

Audit of Computerized Billing Records

The group practice computerized billing system contained a record for every encounter. The records include provider and patient codes, the site of service, diagnosis codes, and any billable tests obtained within the group’s clinical laboratory and radiology facilities. Only group practice sites that used these facilities exclusively were audited. Patient demographic information was obtained by cross-referencing the patient code with a master enrollment file. Computerized billing data for 1985 to 1995 were audited. For each patient, yearly data were abstracted that included a flag for the presence/absence (1/0) of a billable preventive procedure. From this data file, a set of positive criteria was constructed to determine the prevalence of preventive service delivery on the basis of age, sex, and frequency criteria Table 1. Positive criteria were modeled after those developed by the National Committee on Quality Assurance.²⁵

The audited population included all adult outpatients (aged 18 years and older) who visited a defined panel of primary care providers at 10 sites at least once in 1995. In addition to age and sex, the following data were abstracted: billable preventive procedures (Papanicolaou test, screening mammogram, cholesterol test, tetanus immunization, fecal occult blood test, and sigmoidoscopy), total visits to primary care providers, insurance type (HMO, FFS, Medicare, Medicaid, or workers’ compensation), and visit type (acute care–only or scheduled preventive care visit). Cholesterol testing as part of a chemistry panel was counted separately from single cholesterol determinations or lipid profiles, which were classified as screening.

Computer Audit Validation Using Medical Record Reviews

At one study site, the computer audit results were validated by manual review of 200 randomly selected medical records. The validation audit found good agreement (±2.5%) for documentation of preventive procedures; the proportion of positive criteria for the computerized audit was always within 2.5% of that for the manual chart review of the same patient population. Computerized visit frequency counts were highly correlated with medical record review counts (R = 0.93; P <.0001). Medical record review also showed that, except for sigmoidoscopy, approximately 98% of audited preventive procedures had been performed for screening, not for symptoms. (Sixty-seven percent of sigmoidoscopies had been done for screening.)

A separate review of a stratified sample of 245 medical records⁴ found evidence that screening was performed outside the system of care (and not recorded in the transaction files) in 0.5% or fewer of preventive procedures; the only exception was Papanicolaou tests, which had been obtained outside the system in 13.8% of women aged 20 to 29 years and 4.5% of those aged 30 to 39 years.

Statistical Testing

We explored univariate associations between age, sex, visit frequency, insurance type, and visit type using analysis of variance for continuous dependent variables and the Fisher exact test for binary categorical variables. We used logistic regression to test whether preventive services were independently associated with those same variables. For the statistical testing, visit frequency was defined as all visits to primary care providers in 1995, and presence of a visit scheduled specifically for preventive care was coded as positive if one occurred during 1994 or 1995 (P <.05 was reported as significant).

Results

Thirty-five family practice physicians, 33 general internists, 15 obstetrician/gynecologists, and 12 physician assistants supervised by the primary care physicians made up the primary care provider panel. This panel encountered 75,621 outpatients at least once in 1995. Characteristics of this patient group are presented in Table 2.

Age was positively associated with men (mean = 47 years vs 45 years for women), visit frequency (R = .073), HMO insurance (mean = 45 years vs 41 years for FFS) and acute care–visits (mean = 48 years vs 43 years for preventive care visit) (P <.001 for all comparisons). Women were positively associated with visit frequency (mean = 3.4 visits per year vs 2.6 visits for men) and preventive care visits (52% of all women scheduled such visits vs 22% of men ).

HMO membership was equal for both sexes. Visit frequency was positively associated with HMO membership (mean = 3.2 visits vs 2.6 for FFS) and preventive care visits (mean = 3.1 visits vs 3.0 for acute care only) (P <.001 for both). HMO members scheduled more preventive visits than FFS patients in 1995 (36% and 31%, respectively; P <.001).

Table 3 presents the results for the Papanicolaou test, mammography, cholesterol, tetanus immunization, fecal occult blood testing, and sigmoidoscopy. In general, older patients received more preventive services than younger patients; differences by sex were inconsistent. Patients with 3 or more visits had consistently more positive criteria (7% to 12% more) than patients encountered only once or twice in 1995. HMO patients had 6% to 13% more screening than FFS patients. Associations with visit type were greatest: Patients who scheduled at least one preventive visit had 9% to 45% more positive criteria than patients seen only for acute care. Because of the large sample size, all these differences were statistically significant. Even a trivial positive association of age and mammography use achieved statistical significance (P = .01) in the multivariate model. Logistic regression analyses showed that, for every preventive service audited, the associations of visit frequency and type were statistically independent (P <.001 in all cases).

Table 4 illustrates the profound interaction between insurance type and visit type. HMO patients who scheduled a preventive visit had the highest rates of screening, while FFS patients seen only for acute care had rates that were 14% to 50% lower. FFS patients who scheduled a preventive visit had rates approaching the high rates of screening of the HMO/preventive visit group. Interestingly, HMO patients seen only for acute care had low rates of screening, closer to those of FFS/acute care–only than to the HMO patients who scheduled a preventive visit.

Discussion

The audit methodology we used in this study measured the delivery of adult clinical preventive services in a multispecialty, multisite group practice health system that treats both FFS and HMO patients. The audited group included all patients having at least one face-to-face encounter with a primary care provider in 1995. This methodology excluded patients who were not seen during the audited year. Peripheral sites where billable services were not recorded were also excluded. Positive audit criteria were determined according to nationally recognized norms.²⁵

Limitations

Two random medical record reviews validated the computerized billing file audit showing that most recorded preventive procedures were performed for screening, and few patients had received any preventive services elsewhere. (The exceptions were the approximately 33% of sigmoidoscopies that were done for symptoms and a small percentage of Papanicolaou tests that were performed outside the system.) The magnitude of these discrepancies was insufficient to alter the conclusions of our study. It is possible that physicians did not code for some preventive visits scheduled by FFS patients whose insurance did not cover prevention. If so, our audit could have underestimated the association of scheduling preventive visits with the delivery of such services for FFS patients. There was no incentive for physicians to undercode preventive visits for HMO patients. The system of care did not record assignment of patients to individual physicians. Therefore, it was not possible to audit individual providers. The system’s database did not record demographic information, such as education, income, ethnic origin, or race. Thus, we could not analyze health system performance in regard to socioeconomic status, which is an important additional predictor of preventive service use.²⁶ Nonbillable services, such as smoking cessation counseling and blood pressure testing, were not measurable using our methodology.

FFS Versus HMO Insurance

With one exception,²⁷ studies of HMO and FFS care have compared one group treating HMO patients with another group providing care for FFS patients. That type of study design raises the concern that system factors (different provider group attitudes, training, system access, protocols, and so forth) were responsible for the differences reported between HMO and FFS care. It is unlikely that this happened in our study, because both HMO and FFS patients were seen within the same system of care by the same physicians whose compensation formula did not discriminate between insurance types. In this system of care, HMO members had insurance to pay for any preventive service offered by the physician, while coverage for prevention was not uniformly available to FFS policy holders.

Our results that show HMO patients received more preventive services than FFS patients agree with the results of the National Health Interview Survey (NHIS),¹¹ which included a representative sample of the US population. It is unlikely that HMO self-selection accounted for the higher preventive services delivery rates for HMO patients in these studies. After controlling for factors correlated with selection into HMOs, health status, and use of medical services, results of the NHIS were not altered. In our study, only half of the HMO patients scheduled a preventive visit; the other half had lower screening rates Table 4.

The most important determinant of access to health care is having health insurance.²⁸ We believe the most simple explanation for the superiority of preventive service delivery to HMO patients in our system of care is that FFS patients without preventive coverage are reluctant to pay out-of-pocket expenses for these services. This belief is supported by the results of a recent study in managed care settings showing that physician compensation method was not significantly related to use, while plan benefit level was positively related to increased service delivery.⁹ Thus, providing insurance coverage to pay for preventive care is one potential strategy for increasing delivery.

Preventive Visits Versus Acute Care

Visit frequency has previously been shown to have a positive association with the delivery of preventive services.^4,29 Our study confirms the association with visit frequency, but found a statistically independent and greater association with visit type. Similar to results in other primary care studies,³⁰ we found that a scheduled preventive visit was strongly predictive of the delivery of preventive services.

One strategy, then, for rectifying this discrepancy is to insist that all patients schedule a yearly preventive visit. Currently only one third of patients in primary care settings³¹ schedule such a visit. A simple calculation demonstrates, however, that if primary physicians in our audit spent 30 additional minutes each year performing a complete physical examination for the approximately 54,000 patients currently seen for acute care only, there would be little or no time remaining to care for sick patients. Since the audited group represents only a part of the whole, the systemwide impact would be even greater. Additionally, when one author (D.L.H.) systematically invited all adults aged older than 50 years to schedule a complete physical examination, most did not do so (unpublished observations).

We agree with Frame²⁰ that it is neither feasible nor necessary to insist that all patients schedule preventive visits to receive preventive care. The delivery of preventive services during acute care visits has been advocated as a necessary strategy to deliver adequate services to entire patient populations.^21,22 People want physicians to provide more preventive services,³² and many patients who have not scheduled preventive visits will accept them if the physician offers.^4,33 Preventive services delivery during illness visits is common in community practice,³³ though it is more prevalent in high-risk than average-risk patients.³⁴ Other primary care–based studies show that the best results are accomplished when these services are systematically offered to all patients during illness visits, regardless of risk status.^4,23,35,36 Offering preventive care during acute care visits can be integrated into office practice^4,33,34 and can be effective.^4,36

Conclusions

Coordinating efforts to offer evidence-based preventive services to all patients seen for acute care visits is a potential strategy to increase the delivery of this care. Implementation of this strategy could result in increased screening for patients enrolled in HMOs and for FFS patients with coverage. In 1996, Dean Health System began implementing a systemwide adult health maintenance guideline that emphasizes assessment and delivery during acute care visits and includes provider reports of guideline adherence. After 3 years of implementation, providers reported that more than 60,000 patients (mean = 1775 per month) were treated in accordance with the guideline. Follow-up audits are planned to monitor the outcome of this strategy to assess and offer preventive services during acute care visits.

Acknowledgments

We would like to thank Kurt C. Stange, MD, PhD, for constructive suggestions during the preparation of this manuscript, and Marcie Berger, MD; Jonathon Berkhoff, MD; Gene Bettinger; and Suzanne Hallett for technical assistance.