Enrollment factors and preexisting medical conditions in a prepaid capitated plan: a comparison of university and nonuniversity clinic enrollees.
There are, however, several potential problems for academic centers affiliating with capitated plans. Previous studies have suggested a more costly case-mix of enrollees in teaching hospitals when compared with enrollees in nonteaching hospitals,(5,11,12) though the evidence varies." For example, Kosecoff et al(l4) found significant chronic morbidity among patients in 15 general internal medicine clinics in teaching hospitals, though Lion and Altman" found only a 10% to 15% increase in complexity among patients in hospital outpatient departments compared with patients in private practice settings, based on physician assessment of patient illness.
There are many reasons insured patients with significant health problems might seek care at an academic medical center, among which are the center's state-of-the-art technology and the availability of subspecialists and special research programs.(16-17) The importance of a preexisting physician-patient relationship in prepaid plan enrollment decisions has been emphasized in previous studies. 18-10 This relationship is less likely in university clinics, where because of the nature of medical training, physician turnover is higher than in private practice settings. No study has examined the factors influencing capitated plan members to enroll in university clinics as compared with those factors influencing members to enroll in nonuniversity clinics.
It was hypothesized that a university practice might enroll a different population than nonuniversity practices because the patients were sicker and needed more resources. This enrollment selection would adversely affect the economic viability of medical practices in academic centers that care for a large patient population belonging to prepaid plans. To explore this hypothesis, university and nonuniversity enrollees were compared at the time of enrollment into a prepaid plan, focusing on the following factors: demographic characteristics, self-reported health status and preexisting medical conditions, and previous health care utilization during the year prior to the survey. The same enrollees were also compared with regard to factors influencing selection of health care plan, clinic, and physician. METHODS Description of Prepaid Plan
The survey was distributed to a random sample of new enrollees of a 5-year-old prepaid capitated network-model health maintenance organization (HMO) in western Washington State in 1986. The plan was affiliated with clinics employing five or more physicians and was open to enrollment only through employers.
Under the plan's guidelines each enrollee was required to choose a group practice affiliated with the plan and to choose a primary care physician (family physician, internist, pediatrician, or obstetrician-gynecologist) at or before the time of first utilization of care under the plan. The term clinic in this paper refers to the individual practice groups studied. University clinic enrollees were enrollees who were assigned to a primary care physician at the university outpatient facilities and, with few exceptions, obtained all medical care at university facilities. Six nonuniversity clinics affiliated with the capitated plan were chosen as candidates for the study because they represented a spectrum of practice types ranging from primary-care-based clinics to multispecialty groups in urban, suburban, and rural settings. Five of the six nonuniversity clinics agreed to participate. The number of enrollees at each participating clinic ranged from 1953 to 5080. The clinic that chose not to participate was a suburban family practice clinic that had affiliated with the plan relatively recently. Three of the participating clinics (including the university) were large multispecialty urban clinics, one was a small suburban multispecialty clinic, and two were suburban and rural practices where family practitioners were the predominant primary care providers.
All clinics were located within 60 miles of Seattle, Washington. There were 2871 enrollees at the University of Washington clinics and 15,203 enrollees at the five nonuniversity clinics studied (average 3005 enrollees per clinic). Population Studied
The central administration of the prepaid plan provided a randomly generated list of new enrollees who had joined the plan after January 1, 1986, from which a subset of enrollees was randomly selected. Because the primary goal of the study was to compare new enrollees in university clinics with new enrollees in nonuniversity clinics, the sample included 400 enrollees in the university clinics (14% of total university enrollees) and 759 enrollees in the nonuniversity clinics (5% of the nonuniversity enrollees). A power calculation predicted that a sample size of 200 per group would have a 90% chance of detecting a 15% difference (0.6 points on a 4-point scale) between groups (P < .05).(21)
A five-page 53-item questionnaire was developed based on previous literature(l5,18,19) and on the investigators' experience with capitated health care plans. Respondents were asked to provide basic demographic data and information about prior medical conditions and health care utilization during the previous year. Self-reported utilization and medical conditions were chosen as indicators of morbidity. Respondents rated the factors influential in choosing an insurance plan, clinic, and physician, using a 4-point Likert-type scale (1 = not very important to 4 = extremely important). Respondents could also list other factors not named on the questionnaire. A pilot study of the questionnaire was conducted on a sample of ten enrollees making clinic visits at the university and was modified based on their comments. Most respondents were able to complete the questionnaire in 10 to 15 minutes. A questionnaire, a cover letter signed by the network medical director of the capitated plan, and a prepaid return envelope were mailed to the home address of each selected enrollee, directed to the person in the household who made most of the health insurance decisions. At the suggestion of plan administrators, who had achieved adequate return rates by addressing market surveys to female members of enrolled households, the survey was mailed to the female member of the household if two adult enrollees lived at the same address. A duplicate questionnaire was sent to households from which no response had been received within 2 months. Analysis Results were coded and analyzed to calculate the frequency of responses to individual items. Analysis of variance, chi-square, and Z tests for significant differences between proportions were used to determine significant differences between university and nonuniversity enrollees. 22 Parametric statistical methods were employed, with each question treated as a separate variable. Pregnancy was included in utilization statistics, since participating medical groups were responsible for the costs of pregnancy just as for medical illness. RESULTS Population Studied
A total of 676 of the 1159 mailed questionnaires were returned; 618 were usable (214 of 400 from the university clinics, 404 of 759 from nonuniversity clinics) and 58 were either incomplete or returned unanswered because of change of address or disenrollment from the plan (12 from university clinics and 46 from nonuniversity clinics). The overall corrected response rate was 56.1% with no significant difference in response rates among university clinic and nonuniversity clinic enrollees. The demographic characteristics of respondents differed somewhat in university and nonuniversity settings (Table 1). Compared with nonuniversity clinic enrollees, university clinic enrollees were slightly younger (34.1 vs 37.0 years, analysis of variance, P = .0 1) and reported a higher level of formal education. More than two thirds of the respondents in both settings were women, probably because of the manner in which surveys were addressed. Demographic data from the plan's records revealed the proportion of female enrollees in university and nonuniversity clinic settings to be 53.3% and 54.5%, respectively.
University clinic respondents did not differ significantly from nonuniversity clinic respondents with regard to the specialty of their primary care provider ( X.sup.2 = 4.53, NS). A majority of the respondents were assigned to a family physician as their primary care provider, though the proportion of respondents who reported they were assigned to a family physician ranged from 6% to 88% among the six clinics studied. University clinic respondents were significantly more likely than nonuniversity clinic respondents to see a female physician X2 = 18.82, P < .01) as their primary care provider.
Reasons for Choice of Plan
Respondents'reasons for choosing the plan are displayed in Table 2. Based on responses to the questionnaire, university enrollees reported fewer preferences regarding plan, clinic, and physician. For example, university clinic enrollees were significantly less likely to have chosen the plan because of such benefits as lack of bills for services, coverage of prescription drugs, or the plan's choice of physicians. The two groups did not differ with regard to importance placed on other benefits and aspects of the plan or with regard to satisfaction with the plan. Enrollees were equally satisfied with the plan overall (3.8 and 4.0 on a scale of 1 to 5 for university and nonuniversity, respectively, where 4 = very satisfied and 5 = extremely satisfied), and were not more likely to disenroll in either of the two settings.
Reasons for Clinic Selection
Respondents' reasons for choosing a clinic are summarized in Table 3. Factors considered significantly less important by university clinic enrollees in choosing a clinic were a previous affiliation with a physician in the clinic and a recommendation by the respondents' employer or friends. Reasons for Physician Selection Respondents' reasons for choosing their physician are summarized in Table 4. University clinic enrollees were significantly less likely to report choosing a physician on the basis of a family member's familiarity with the physician, location of the physician's office, physician availability, or others' advice.
Other enrollee preferences regarding clinic and physician are summarized in Table 5. The majority of both university and nonuniversity clinic enrollees planned to retain their current clinic affiliation (83.3% and 88.0%, respectively), a statistically insignificant difference. University clinic enrollees were more likely than nonuniversity clinic enrollees to prefer a female physician (34.7% and 19.8%, respectively).
Preexisting Conditions and Health Care Utilization University clinic enrollees reported the same number of preexisting conditions and the same degree of recent morbidity within the previous year, with two exceptions: university clinic enrollees were significantly less likely than nonuniversity clinic enrollees to report a hospitalization or two or more visits for pregnancy during the previous year (Table 6) and reported significantly fewer hospitalizations for other diagnoses during the previous year. University clinic enrollees also reported slightly-more back or orthopedic problems and hypertension, though these rates for these conditions did not differ significantly from nonuniversity clinic enrollees.
There were no significant differences with regard to nonpregnancy-related morbidity, number of health problems, number of physician visits, number of days spent in bed, or number of days during which usual activities were curtailed (Table 7). Very few respondents provided any responses to the "other" option on each question.
New university clinic enrollees joining this prepaid capitated plan differed very little from new nonuniversity clinic enrollees with regard to three preexisting characteristics (demographic profile, prior health status, and recent health care utilization) and with regard to factors influencing the three enrollment decisions (choice of plan, clinic, and physician). Demographic Characteristics The younger mean age and higher level of education among university clinic enrollees was not surprising, given the young age of physicians at the university and the close proximity of university clinics to university employees. Data from plan administrators indicated that 43% of university enrollees were university employees or their family members. Demographic data provided by the capitated plan about all the enrollees in the six clinics studied parallel findings in this study, suggesting that the university clinics served a younger population than the nonuniversity clinics. It has been noted in previous studies that younger physicians tend to see younger patients.13 This difference does not confer any competitive advantage or disadvantage, since capitation rates are age adjusted. Preexisting Conditions and Utilization Based on the self-reported medical conditions assessed in this survey, university clinic enrollees in the study did not represent a more complex case mix than nonuniversity clinic enrollees. This finding may be the result of the younger mean age and the high proportion of university employees among university clinic enrollees. University clinics offer easy access to complex and costly medical services such as infertility programs and psoriasis care. The results of this enrollee survey, however, do not support the original hypothesis that university clinics attract sicker patients whose care increases the economic risks to clinics caring for them under a prepaid capitated plan. The relatively low response rate (56%) indicates the need for caution in interpreting these results. Nonetheless, morbidity statistics provided by the plan for each of the six clinics studied confirm this trend, showing that nonuniversity clinics averaged 294 hospital days per 1000 enrollees (range 238 to 345 for the five nonuniversity clinics), and the university clinic enrollees average 263. Nonuniversity clinics averaged 24 obstetric cases per 1000 enrollees (range 17 to 25 for the five nonuniversity clinics), whereas university clinic enrollees averaged 13. Although this survey failed to identify a significantly higher level of self-reported utilization among the university clinic enrollees studied, the number of enrollees reporting high utilization was small. Given the low incidence of high utilization in all six clinics studied, these findings could reflect the problem of finding a "needle in the haystack," since the 618 respondents comprised only 3.4% of the 18,074 enrollees in the six clinics studied. There is evidence in previous studies that a small number of persons with severe illnesses may account for a large portion of hospital costs, that these high cost patients are more likely to receive care at referral hospitals than at community hospitals, and that teaching hospitals are more likely than nonteaching hospitals to manage patients with repeat hospitalizations and complex life-maintenance requirements in a prepaid network.(24,25)
Factors influencing Enrollment Decisions Based on these data, university clinic enrollees were less attracted to the plan than nonuniversity clinic enrollees by the lack of bills for services and by prescription drug coverage. It is unknown whether planning to disenroll from the plan represented dissatisfaction or other factors such as employer change, moves, or job changes. The relative satisfaction of enrollees parallels findings of consumer satisfaction studies in other HMOs.(26)
University clinics also attracted enrollees who had fewer medical preferences about their clinic and primary care physician, with one exception: university respondents were more likely to prefer a female physician. This stated preference may have arisen because the university clinics had more female providers on their staff than the nonuniversity clinics. University respondents were more likely to state that their decisions about plan, clinic, and physician were not based on the recommendations of others, perhaps because university clinic enrollees were younger, newer to the community, and thus had fewer acquaintances who could recommend a source of medical care. One study27 found that a previous relationship with a particular physician was an important factor in enrollment decisions by potential enrollees in a prepaid plan. Several enrollees commented that they perceived little freedom of choice in their selection of health plan or physician. Physician availability also varied by site and by the urgency of the medical problem for which the enrollee sought care. CONCLUSIONS Based on responses to a mailed survey to a random sample of new enrollees in a 5-year-old capitated health plan in western Washington State, when compared with nonuniversity clinic enrollees, the university clinic enrollees (1) were younger, with more formal education, (2) reported less recent utilization of health services, and (3) reported fewer strong preferences with regard to plan, clinic, and physician characteristics. In this study, contrary to expectations, self-reported utilization and preexisting morbidity among university clinic enrollees was lower than anticipated. These findings notwithstanding, common sense dictates that clinics affiliating with capitated plans, especially clinics providing many specialty services, should still ascertain the target enrollee population and insure themselves with stop-loss provisions. Potential economic risk can arise if care is delivered to even a small number of prepaid patients with complex medical conditions. Such enrollees would be likely to find the combination of a comprehensive array of services at a referral facility and extensive prepaid health care coverage particularly attractive. In today's rapidly changing health insurance marketplace, the assumption that university clinics always draw the sickest patients needs to be supported by empirical data. Future studies could focus on a homogeneous group of enrollees, for example, enrollees with the same employer who could choose from among university-affiliated and nonuniversity-affiliated clinics for their health care, and sample a larger population. References 1 .Moore GT: HMOs and medical education: Fashioning a marriage.
Health Aff 1986; 5(l):147-153 2. Shine KI: Innovations in ambulatory care education. N Engl J Med
1986; 314:52-53 3. Rieselbach RE, Jackson TC: in support of a linkage between the
funding of graduate medical education and care of the indigent. N
Engl J Med 1986; 314:32-34 4. Heyssel R, Seidel H: The Johns Hopkins experience in Columbia,
Maryland. N Engl J Med 1976; 295:1225-1231 5. Hoft RH, Glaser RJ: The problems and benefits of associating academic
medical centers with health maintenance organizations. N
Engl J Med 1982; 307:1681-1689 6. Capitated systems seen as main source for sales. Am Med News
1986, April 4, p 33, col 1-4 7. Traska MR: Capitated contracts, competition hit HMO markets.
Hospitals 1987; 61(7):52 8. Wiegert HT: The academic challenge to address health care costs.
Fam Med 1985; 17:123-124 9. Smith GA: A family practice center experience with an HMO. J Fam
Pract 1981; 13:541-552 10. Eisenberg JM: The internist as gatekeeper. Ann Intern Med 1985;
102:537-543 11, Schwartz MD, Newhouse JP, Williams AP: is the teaching hospital
an endangered species? N Engl J Med 1985; 313:157-162 12. Ament RP, Kobrinski EJ, Wood WR: Case mix complexity differences
between teaching and non-teaching hospitals. J Med Educ
1981; 56:894-903 13. Sloan FA, Valvona J: Uncovering the high costs of teaching hospitals.
Health Aff 1986; 5(3):68-85 14. Kosecoff J, Fink A, Brook RH, et al: General medical care and the
education of internists in university hospitals: An evaluation of the
teaching hospital's great medicine group position, Ann Intern Med
1985; 102:250-257 15. Lion J, Altman S: Case-mix differences between hospital outpatient
departments and private practice. Health Care Finance Rev 1982;
4:89-98 16. Kirz HL, Larsen C: Cost and benefits of medical student training to a
health maintenance organization. JAMA 1986; 256:734-739 17. Herrmann TJ, Matthews CW, Segadelli LJ: Weighing the views of a
university hospital and medical school regarding an HMO. J Med
Educ 1983; 58:686-694 18. Berki SE, Ashcraft MLF: HMO enrollment: Who joins what and why.
A review of the literature. Milbank Mem Fund 1980; 58:588-632 19. Wersinger RP, Sorenson AA: Demographic characteristics and prior
utilization experience of HMO enrollees compared with total membership.
Med Care 1982; 20:1188-1196 20, Kleinman JH: A comment on demographic characteristics and prior
utilization experience of HMO enrollees compared with total membership.
Med Care 1983; 21:1034-1035 21. Nye NH, Hull CH, Jenkins JG, et al (eds): Statistical Package for the
Social Sciences. New York, McGraw Hill, 1975 22. Feigle P. A graphical aid for determining sample size when comparing
two independent proportions. Biometrics 1978; 34:111-122 23. Rosenblatt RA, Cherkin DC, Schneeweiss R, et al: The structure
and content of family practice: Current status and future trends. J
Fam Pract 1982; 15:681-722 24. Schroeder SA, Showstack JA, Roberts HE: Frequency and clinical
description of high-cost patients in 17 acute care hospitals. New
Engl J Med 1979; 300:1306-1309 25. Zook CJ, Moore FD: High-cost users of medical care. N Engl J Med
1980; 302:996-1002 26. Taylor H, Kagay M: The HMO report card: A closer look, Health Aff
1986; 5(l):81-89 27. Grazier KL, Richardson WC, Martin DP, Diehr P: Factors affecting
choice of health care plans. Health Serv Res 1986; 20:659-682
|Printer friendly Cite/link Email Feedback|
|Author:||Ellsbury, Kathleen E.; Larson, Eric B.; Raskind, Wendy; Montano, Daniel E.; Kuykendall, David|
|Publication:||Journal of Family Practice|
|Date:||Nov 1, 1989|
|Previous Article:||Graded exercise stress test training in family practice and internal medicine residencies.|
|Next Article:||Building an Ambulatory Clinical Information System in a family practice residency.|