Quality of Care and Outcomes of Adults With Asthma Treated by Specialists and Generalists in Managed Care FREE

WITH MANAGED health care, an established feature in the United States, worries about quality of care have increased. Because managed care organizations (MCOs) realize some of their efficiency by controlling access to specialty care, there is concern that these limitations will affect health outcomes, particularly for patients with serious or chronic illness. However, few published studies have shown differences in outcomes of care provided by generalist vs specialist physicians.

Studies suggest that specialist physicians are more knowledgeable about conditions that fall within their area of expertise^1- 3 and are more reliable than generalists at diagnosing them.^4- 6 Some reports suggest that specialists treating acute illnesses obtain better patient outcomes,^7- 8 but for chronic illness the impact has been more elusive. One report from the Medical Outcomes Study⁹ showed similar outcomes for patients with uncomplicated hypertension or type 2 diabetes mellitus treated by specialists (cardiologists and endocrinologists, respectively) vs generalists. Another report found better outcomes when depression was treated by mental health care providers.⁵

Asthma provides an excellent example for comparison of generalist and specialist care. Asthma is a common chronic illness that affects approximately 9 million to 12 million people in the United States and that is associated annually with approximately 1.8 million emergency department (ED) visits, 500 000 hospitalizations, and 3 million workdays lost.¹⁰ Asthma care is provided by generalists and specialists in a variety of treatment models. Studies of asthma care have shown that intensive treatment and education are related to better adherence to clinical guidelines and to improved functional outcomes.^11- 12 However, the strategies tested have combined specialty care with intensive education and drug therapy, making the effects of the different components difficult to separate. A few studies that have looked specifically at the effect of the physician on treatment and outcomes suggest that patients treated by specialists^13- 15 or switching to specialist care¹⁶ had fewer symptoms and ED visits and improved quality of life, albeit at a potentially higher cost.¹¹ Inpatient studies of generalist and specialist care have also shown improved outcomes for patients treated by specialists.^17- 18 However, previous studies have not been adjusted for measures of disease severity or have lacked the patient's perspective. In addition, none have examined differences in outcomes between allergists and pulmonologists or between generalists with different levels of asthma management expertise.

In conjunction with a consortium of managed care companies and purchaser organizations, we conducted a large study of quality of care and outcomes of patients with asthma enrolled in managed health care.^19- 20 In this study, we investigated whether there were differences in process of care and outcomes attributable to the specialty of the health care provider.

This study was part of a project undertaken by the Managed Health Care Association Outcomes Management System Consortium. The asthma project involved several large employers and their managed care partners in a prospective cohort study to test the feasibility and usefulness of information on process and patient outcomes to improve the quality of health care.^19- 20

Patients were selected from enrollees in each MCO by using claims data or administrative information to apply the following 3 inclusion criteria: (1) aged 18 years or older as of September 1, 1993; (2) enrolled in the MCO at the time of sampling; and (3) at least 2 medical care encounters (ED visits or hospitalizations) with a diagnosis of asthma (International Classification of Diseases, Ninth Revision, Clinical Modification code 493.x) in the previous 24 months. This analysis used a subset of patients from the overall study consisting of 1954 patients who completed the baseline and 2 follow-up surveys and their 1078 matched physicians who responded to a separate physician survey (patients were asked to name the physician principally responsible for managing their asthma care). Compared with patients with matching physician information, patients without matching physician information (n = 1533) were similar in sex, asthma symptoms, and number of workdays lost, but were more likely to be older and white. Since the most salient physician data were collected on the year 1 survey, the primary analysis examined baseline data, with outcomes assessed at year 2.

Information was obtained by means of mailed, self-administered patient and physician surveys. The patientquestionnaires provided information on demographic char questionnaires provided information on demographic characteristics, general health status, asthma-specific health status, use of medical care resources, disability, and ratings of the quality of care. The physician survey provided information on specialty training and self-rated asthma expertise.

Nine outcome indicators assessed at year 2 included (1) canceled or rearranged activities due to asthma during the past 4 weeks (dichotomized as occurring at all vs not at all); (2) hospitalization or ED visits for asthma in the past 12 months (≥1 vs none); (3) frequency of asthma attacks (increased difficulty breathing accompanied by cough, wheezing, chest tightness, or other symptoms) in the past 4 weeks (≥3 attacks per week vs fewer); (4) workdays missed due to health in the past 4 weeks (≥1 vs none); (5) asthma symptom scores, based on 7 common asthma symptoms (cough, sputum production, chest tightness, wheezing, shortness of breath, nocturnal symptoms, and the chronicity of symptoms between attacks) in the past 4 weeks, with the symptom score calculated as a summary rating scale ranging from 0 to 5 and higher scores indicating more severe asthma symptoms; (6) changes in general health status assessed using the physical component score from the 36-item Short-Form Health Survey of the Medical Outcomes Study^21- 22 (higher score indicates better health); (7) changes in general health status using the mental component score from the same instrument; (8) rating of the quality of communication with physicians and nurses about asthma (good, fair, or poor vs excellent or very good); and (9) rating of satisfaction with the overall quality of care received for asthma during the past 12 months (good, fair, or poor vs excellent or very good).

Patient characteristics included age, sex, race, educational level (college graduate or postgraduate vs others), history of cigarette smoking (yes vs no), passive exposure to cigarette smoke (yes vs no), presence of chronic obstructive pulmonary disease (COPD) (yes vs no), and asthma symptom score (range, 1-5, with a higher score indicating more severe symptoms).

Treatment indicators were based on National Heart Lung and Blood Institute National Asthma Education and Prevention Program²³ guidelines for patients with asthma and included (1) having a metered dose inhaler for corticosteroids (ICS) and using it regularly (≥5 d/wk), intermittently (1-3 d/wk), or not at all; (2) having a peak flow meter; (3) reporting knowing everything they should about recognizing triggers, managing flare-ups, or adjusting medications; (4) having had allergy treatment and/or evaluation; and (5) having ever discussed asthma triggers with health care providers. Previous studies have shown improved health outcomes related to use of inhaled corticosteroids,^24- 27 and patient education.^28- 32

Physician specialty was categorized as generalist, asthma-experienced generalist, or specialist (pulmonologist or allergist). The variable for the asthma-experienced generalist was constructed from 2 questions completed by physicians. The first question was "What is your specialty?" with response options of internal medicine; family practice; ear, nose, and throat; pulmonology; allergy; and other. The second question was "Are you an asthma specialist?" with response options of yes or no. If a physician responded that he or she is not an asthma specialist and that his or her specialty is internal medicine or family practice, then that respondent was grouped as generalist, the reference group in this study. If a physician responded that he or she is an asthma specialist with a specialty in internal medicine or family practice, then he or she was grouped as an asthma-experienced generalist. Specialists included pulmonologists or allergists and were examined together and separately, compared with generalists.

The primary analysis examined the relationship of physician specialty to treatment indicators (assessed at year 1) and outcomes (reported at year 2). We hypothesized that patients treated by specialists would be more likely to report adherence to treatment indicators and would have better outcomes than patients cared for by generalists. We anticipated that in exploratory analyses, we would find similar levels of performance for pulmonary and allergy specialists.

The analysis included simple descriptive statistics and multivariable analyses. Frequency distributions and summary statistics were generated on the independent variables and dependent variables of interest. Univariate and bivariate analyses were performed to describe the data, including χ² and t tests, correlations, and the McNemar test. Mean values were computed for the effect of physician specialty on continuous outcome indicators.

Multivariable logistic regression analyses were performed on the relationships between physician specialty and the 9 outcome measures. In our first model, we controlled for demographics, asthma symptoms, presence of COPD, smoking and passive smoke exposure, and comorbid conditions that increase asthma symptoms.²³ To explore possible mechanisms for specialty-related differences, in a second model we further adjusted for quality of care indicators including possession of an ICS and peak flow meter, adequacy of information about asthma management, discussion of triggers, and allergy testing. A general linear model was used to analyze the relationships between selected outcomes and process of health care measures and physician specialty. All statistical computations were performed using commercially available software (SAS version 6.12; SAS Institute, Inc, Cary, NC).

Overall, the mean age of patients was 47.4 years (range, 18-94 years) (Table 1). Of the 1954 patients, 69.3% were female, 84.9% were white, and 35.0% had a college degree or postgraduate education. At baseline, 35.8% reported also having chronic bronchitis or emphysema (COPD), and 48.9% had smoked cigarettes.The mean asthma symptom score was 2.6. The association between patient baseline characteristics and physician specialty are presented in Table 1. Patients treated by asthma-experienced generalists or specialists differed from the patients of generalists in several respects. Compared with patients of generalists, patients treated by asthma-experienced generalists were more likely to be white (89.5% vs 80.8%; P<.003). Patients treated by pulmonologists were more likely to be white (88.7% vs 80.8%; P<.01) and older (mean age, 52.2 vs 47.4 years; P<.01) and to have COPD (44.9% vs 37.8%; P = .03) than patients of generalists. Patients of allergists were also more likely to be white (91.6% vs 80.8%; P<.001), but were younger (mean age, 44.7 vs 47.4 years; P<.001), had more education, and were less likely to have smoked (40.5% vs 50.8%; P<.001). They also had less severe asthma symptoms and were less likely to report having COPD (25.7% vs 37.8%; P<.001).

Associations between physician specialty and patient-reported variables of process of care and knowledge about asthma are presented in Table 2. For treatment indicators, overall 80.8% had been told by medical personnel how to avoid triggers; 76.3% reported knowing how to manage asthma flare-ups, medications, or triggers; 75.7% had had an allergy evaluation; 72.1% had an ICS; and 48.4% had a peak flow meter at home. Treatment indicators varied by physician specialty. Patients of asthma-experienced generalists had several indications of greater consistency of care with National Asthma Education Program guidelines compared with patients of generalists, including more knowledge about flare-ups, medication use, and triggers and having an ICS and peak flow meter. Patients of pulmonologists or allergists had more knowledge about asthma management and were more likely to have a peak flow meter. In addition, patients of allergists were more likely to have discussed triggers and to have had an allergy evaluation, whereas patients of pulmonologists were more likely to have an ICS.

In multivariable logistic regression controlling for baseline patient characteristics, including demographics, asthma symptoms, and comorbid illness (model 1 in Table 3), there were significant differences by physician specialty in patient outcomes assessed 1 year later. Compared with patients of generalists, patients of asthma-experienced generalists were significantly less likely to cancel activities (odds ratio [OR], 0.55; 95% confidence interval [CI], 0.35-0.88) or miss work (OR, 0.58; 95% CI, 0.35-0.94) in the month before follow-up, or to have been hospitalized or to have visited an ED for asthma (OR, 0.53; 95% CI, 0.29-0.96). Patients of pulmonologists rated communication with their physicians and overall quality of care significantly higher than patients of generalists, but they were also more likely to be hospitalized or have ED visits (OR, 1.59; 95% CI, 1.08-2.35). Patients of allergists were significantly less likely than patients of generalists to cancel activities (OR, 0.57; 95% CI, 0.40-0.82) and to be hospitalized or have ED visits (OR, 0.63; 95% CI, 0.40-0.97). They rated physician communication and overall quality of care significantly higher, and showed significant improvements in asthma symptoms and physical health status.

To explore why outcomes might vary by physician specialty, we further adjusted for quality-of-care indicators, including use of ICS and peak flow meters, adequacy of information about self-management, discussion of triggers, and allergy testing. (model 2 in Table 3). We found similar associations to those seen in the previous 2 models, although some relationships were no longer statistically significant. Patients treated by pulmonologists, on average, had significantly improved symptom scores and rated overall quality of care as better (OR, 0.54; 95% CI, 0.33-0.87). Patients of allergists were less likely to cancel activities (OR, 0.51; 95% CI, 0.34-0.77), or have hospitalization or ED use (OR, 0.55; 95% CI, 0.33-0.90). Patients of allergists also had greater improvement in asthma symptom scores (P = .07) and physical health scores (P<.001) and rated overall quality of care higher (OR, 0.57; 95% CI, 0.36-0.88). Thus, differences in care, as reflected by the quality indicators, explained some but not all of the benefits associated with specialist care.

With the growth of managed health care in the United States, concern has grown regarding measures to control costs. In particular, controls on access to inpatient care and specialty services have raised questions regarding the potential for adverse effects on patient health. The results of this study suggest that asthma care provided by specialists and experienced generalists was associated with better patient outcomes than was care provided by generalists. These results were consistent across a range of outcome measures that have relevance for patients, MCOs, employers, and other purchasers of health care.

In our study, patients treated by specialists were more likely to report having ICSs, peak flow meters, and adequate knowledge to manage their own condition. Patients of generalists who had substantial expertise in treating asthma also had higher rates for all of these process indicators. These patterns are consistent with treatment guidelines promulgated by the National Asthma Education Program.²³ With regard to outcomes, patients treated by a specialist or an asthma-experienced generalist had better outcomes during the ensuing year relative to patients treated by generalists, even when controlling for a variety of risk factors. The outcomes included a lower likelihood of canceled activities and workdays lost, fewer asthma symptoms, and better physical health as measured using the 36-item Short-Form Health Survey. The difference in workdays lost was equivalent to 0.6 days per month, or 7.2 days per year. Patients of specialists were most satisfied with the care they received.

This study is unusual in the opportunity it presented to explore differences in patient outcomes for 2 different asthma subspecialty groups, pulmonologists and allergists. As there may be differences in the training, experience, and approach to asthma care, we compared outcomes of patients treated by pulmonologists with those treated by allergists, and outcomes of both sets of patients with those of patients treated by generalists. Our results suggested that there are modest differences between the specialists that favor allergists. However, given the differences in patients seen by both specialties, our results should be considered preliminary and hypothesis generating rather than conclusive.

The performance of experienced generalists in this study is more noteworthy. Our results support the conclusion that many patients with asthma cared for by internists and family physicians would benefit from treatment by providers with more experience in treating asthma. However, it is premature to conclude that the sickest patients with asthma would do as well if treated by an experienced generalist as by a pulmonary or allergy specialist. Selection of a physician specialist is likely to depend on the severity of illness, the difficulty in managing the condition, patient preferences, and managed health care controls on access. Even after adjustment for symptom severity, there appears to be significant unmeasured risk for adverse outcomes. In addition, it is likely that patients' health status at baseline already reflected the quality of care being received. As a result, it seems likely that our results underestimate the differences in outcome attributable to specialty care.

Our study had several limitations. Although a disproportionate burden of asthma falls on disadvantaged minority urban populations,^33- 35 the sample in this study was composed of mostly white, well-educated adults insured through plans affiliated with prominent US companies. Thus, our findings are most generalizable to adults treated in managed health care settings. Also, as we oversampled patients who had been hospitalized or had had an ED visit in the 2 years before the study, asthma symptom scores were more severe in our study sample than for all adults with asthma. Measures of asthma symptoms and comorbidity were not sufficiently precise to adjust for all confounding effects of intrinsic disease severity, which limited our ability to demonstrate the benefit of treatment consistent with guidelines. In particular, we recognize the risks of confounding by indication when adjusting for treatments in studies of asthma. Adjustment for treatments likely accounts for their beneficial effects and the propensity to use treatments, the latter of which is driven in part by underlying disease severity or lack of asthma control. Thus, medications serve as indicators of asthma symptom severity. Without ideal measures of intrinsic asthma symptoms (in this and any other observational study), we do not know whether accounting for use of treatments leads to underadjustment or overadjustment for symptoms. Improvements are needed in predictive modeling of the natural course of this chronic condition to allow investigators to assess the contributions of different treatments to patient outcomes over time.

For the practicing primary care physician, our study raises more questions than it answers. If there is a volume-outcome relationship for asthma,³⁶ the threshold for a minimum number of patients to provide ongoing care to achieve or maintain expertise may lie above the 20 or more patients reported by more than two thirds of generalists. At the time of referral, when is a single consultation, shared care, or assumption of care the best course of action? The important issue of comanagement was not addressed in this study. More research is needed to address these questions.

What are the implications for MCO managers? Evidence from this report and recently updated asthma guidelines³⁷ suggest that it may be more effective to assign the sickest patients to specialty care. Determining who is most likely to benefit from specialty care may require new strategies of patient matching. On the other hand, it is likely that some of the benefits of subspecialty care reside in structural aspects of the practices rather than in physicians themselves, such as staff who are knowledgeable about asthma. There may be other ways to realize the same level of results when using nonspecialist resources.

This study did not answer a pivotal question of what is different about the practice of subspecialists and asthma-experienced generalists that contributed to improved outcomes for their patients. The differential outcomes that we observed may have arisen from differences in training, experience, services provided, or unmeasured patient selection effects. We have examined selected aspects of asthma treatment, and these did not fully explain the differences in outcomes. To pursue goals of accountability and information that can support quality improvement, a range of next steps will be important. These steps include attaining a better understanding of the differences in care provided by subspecialists, experienced generalists, and generalists with limited asthma experience. We need a better understanding of the referral and care-seeking practices that lead patients with asthma of similar severity to be treated by physicians with different levels of training and experience. Translating the information learned into new guidelines for training and practice should contribute toward a system that manages health care to ensure the best outcomes possible.

Accepted for publication April 9, 2001.

The study was supported by the Managed Health Care Association Outcomes Management System Project Consortium, Washington, DC.

We would like to acknowledge the guidance of the following present and past members of the Managed Health Care Association Outcomes Management System Consortium Asthma Project Subcommittee: William Glasheen, PhD, Trigon Blue Cross Blue Shield, Richmond, Va; Joseph M. Healy, Jr, PhD, Harvard Pilgrim Health Care, Wellesley, Mass; Allan Khoury, MD, PhD, Kaiser-Permanente/Ohio Region, Cleveland; John Hayes Mason, PhD, Blue Cross Blue Shield of Massachusetts, Boston; Robert McDonald, MD, Anthem Blue Cross Blue Shield, Indianapolis, Ind; Richard Neimeyer, Procter & Gamble, Cincinnati, Ohio; and Ann Saladino, formerly of Alliance Blue Cross Blue Shield, St Louis, Mo. This study was coordinated by the Health Outcomes Institute, Bloomington, Minn. Employer members of the Consortium participating in the Asthma Project were Ameritech, Chicago, Ill; Becton Dickinson and Company, Franklin Lakes, NJ; Commonwealth of Virginia, Digital Equipment Corporation, Maynard, Mass; GTE Service Corporation, Irving, Tex; HealthTrust, Inc, Nashville, Tenn; James River Corporation, Richmond; Marriott International Corporation, Bethesda, Md; Procter & Gamble, Cincinnati; Promus Companies, Memphis, Tenn; and Xerox Corporation, Stamford, Conn. Managed care organizations participating in the consortium as partners of the employers were Aetna Life Insurance Company, Hartford, Conn; Alliance Blue Cross Blue Shield (Mo), St Louis; Anthem Blue Cross Blue Shield (Ohio), Cleveland; Blue Cross Blue Shield of Illinois, Chicago; Blue Cross Blue Shield of Massachusetts, Boston; Blue Cross Blue Shield of Rochester, Rochester, NY; Fallon Community Health Plan, Worcester, Mass; Harvard Pilgrim Health Care, Wellesley, Mass; Intermountain Health Care, Salt Lake City, Utah; Kaiser Permanente/Ohio Region, Cleveland; Matthew Thornton Health Plan, Manchester, NH; The Prudential Health Care System, Newark, NJ; Trigon Blue Cross Blue Shield (Va), Richmond; United Health Care, Minnetonka, Minn; and USQA (US Healthcare), Philadelphia, Pa.

Corresponding author and reprints: Albert W. Wu, MD, MPH, Health Services Research and Development Center, The Johns Hopkins University, 624 N Broadway, Baltimore, MD 21205-1901 (e-mail: awu@jhsph.edu).