Search

BACKGROUND: As Medicaid clients have come to be enrolled in managed care, concerns have arisen about the ability of private sector systems to meet the needs of enrollees with substance abuse problems. OBJECTIVES: This project describes treatment initiation and duration for Medicaid and commercial substance abuse treatment clients in a large health maintenance organization (HMO). RESEARCH DESIGN: This study was a prospective secondary analysis of information from HMO databases. Subjects were 641 adult Medicaid clients who contacted the HMO's addiction medicine department in 1996 or 1997 and commercial HMO addiction medicine patients (n=447). First contact with addiction medicine during the study period was the index event. MEASURES: Chief dependent variables were initiation and duration of substance abuse treatment after the index event. RESULTS: Logistic regression showed that longer HMO enrollment predicted treatment initiation after substance abuse assessment, but Medicaid status was not a significant predictor. A competing risks analysis using Cox proportional hazards models indicated that once subjects had initiated, Medicaid was not significantly related to exit from substance abuse treatment. Analysis of health plan disenrollment by Medicaid clients indicated that the most common reason was loss of Medicaid eligibility. CONCLUSIONS: These results raise the possibility that state Medicaid policies may make it difficult for clients to obtain suitable chemical dependency treatment services. [ABSTRACT FROM AUTHOR]

Clinical uncertainty is a source of variation in medical decision-making as well as a source of work-related stress. Increasing enrollment in organized health care systems has intensified interest in understanding referral utilization as well as issues such as physician dissatisfaction and burnout. We examined whether primary care physicians' affective reactions to uncertainty and their job characteristics were associated with use of referrals and burnout. Data came from mail surveys of primary care physicians practicing in two large group model health maintenance organizations (HMOs) in the USA. Consistent with past research, we found that younger physicians had higher referral rates than older physicians, and that general internists had higher rates than either family practitioners or pediatricians. Greater stress from uncertainty increased referrals and referrals were negatively correlated with heavier work demands (patient visits per hour). Greater stress from uncertainty, perceived workload (too high) and a sense of loss of control over the practice environment were associated with higher levels of burnout. [ABSTRACT FROM AUTHOR]

Research suggests that female physicians have a more patient-centered communication style, characterized by more collaboration with the patient and discussion of treatment options. These same communication characteristics have been associated with patient satisfaction and malpractice prevention. Few empirical studies, however, have examined whether gender affects malpractice risk. The primary objective of this study was to determine whether gender was associated with malpractice among health maintenance organization (HMO) physicians. A secondary objective was to examine how malpractice affected physicians' psychological well-being. Data came from mail surveys of physicians practicing in two group model HMOs. The average response rate was 80% (N = 608). Female physicians were significantly less likely than male physicians to have been sued. Only 26% of the female physicians had been sued compared with 49% of the male physicians. These differences remained after adjusting for physician age, time in practice, specialty, workload, and characteristics of physicians' patient panels. Sued physicians were significantly more at risk for burnout than never-sued physicians and were less satisfied with the HMO. These findings did not differ by gender. [ABSTRACT FROM AUTHOR]

The purpose of this study was to determine whether specialty alcohol and other drug (AOD) treatment is associated with reduced subsequent medical care costs. AOD treatment costs and medical costs in a group model health maintenance organization (HMO) were collected for up to 6 years on 1,472 HMO members who were recommended for specialty AOD treatment, and on 738 members without AOD diagnoses or treatment. Addiction Severity Index measures were also obtained from a sample of 293 of those recommended for treatment. Changes in medical costs did not differ between treatment and comparison groups. Nor did individuals with improved treatment outcomes have greater reductions in medical costs. AOD treatment costs were not inversely related to subsequent medical costs, except for a subgroup with recent AOD treatment. In the interviewed sample, better treatment outcomes did not predict lower subsequent medical costs. Multiple treatment episodes may hold promise for producing cost-offsets.

We compared perceptions about the practice environment and the job satisfaction of physician assistants (PAs), nurse practitioners (NPs), and primary care physicians in a large group-model HMO. The data source was a self-administered mail survey (average response rate = 79%). PA/NPs and primary care physicians reported that professional autonomy was not a problem and were satisfied with most aspects of practice in this setting. Common areas of dissatisfaction included patient load and amount of time with patients. PA/NPs were more likely than the physicians to experience stress on a daily basis, however, and were less likely to report that they would choose the practice setting again. They also were significantly less satisfied than the physicians with their incomes and fringe benefits. Our findings suggest that more attention should be given to practice conditions and compensation of PAs and NPs in managed care.

A survey of 8,034 primary care patients in a health maintenance organization examined the relationship between alcohol consumption and health care costs and service use. Costs were estimated from service use data for 1 year before and 2 years after study enrollment. No strong, consistent relationships were identified between multiple indicators of drinking patterns and either health care costs or service use. Compared with total costs among very light drinkers, former drinkers were higher, lifetime abstainers were similar, and persons in the higher drinking levels tended to have lower but not significantly different costs. Drinking patterns did not appear to be an important predictor of short-term health care costs or service use in this setting. Further study of former drinkers is warranted to examine the role of alcohol-related illnesses in the decision to quit drinking.

Men and women differ in their use of alcohol, in their rates of chronic illnesses and psychological symptoms, and in the social support they receive. In this paper, we assess how the latter three factors are associated with alcohol use, and how these associations differ by gender. Respondents were 3,074 male and 3,947 female randomly selected Health Maintenance Organization members who responded to a mail survey in 1990. Hierarchical multiple regression analyses indicate that social support is associated with alcohol consumption in similar ways for both genders, yet the associations between some demographic, physical health/functioning, and psychological well-being measures are different for men and women. Men with fewer role limits due to physical health drank more, while women with better psychological well-being drank less. Poor psychological well-being may be a modifiable risk factor for increased alcohol use among women; practitioners should be alert for greater consumption among men with few functional limitations and good health.

Objective: To identify the factors that predict professional satisfaction, organizational commitment, and burnout among physicians working for health maintenance organizations (HMOs)., Methods and Participants: Data came from mail surveys of Kaiser Permanente physicians in the Northwest and Ohio regions. The average response rate was 80% (n = 608)., Results: The single most important predictor for all 3 outcomes was a sense of control over the practice environment. Other significant predictors included perceived work demands, social support from colleagues, and satisfaction with resources. The relative importance of these predictors varied, depending on the outcome under consideration. All 3 outcomes were also related to physician age and specialty. Older physicians had higher levels of satisfaction and commitment and lower levels of burnout. Pediatricians were more satisfied and committed to the HMO and were less likely to burn out., Conclusions: Physicians who perceive greater control over the practice environment, who perceive that their work demands are reasonable, and who have more support from colleagues have higher levels of satisfaction, commitment to the HMO, and psychological well-being. Interventions and administrative changes that give physicians more control over how they do their professional work and that enhance social supports are likely to improve both physician morale and performance.

This study examined whether a brief intervention to reduce hazardous alcohol consumption among primary care patients reduced use of medical care. In a parent, randomized controlled trial, at-risk drinkers identified in HMO outpatient waiting rooms were randomly assigned to receive usual care or brief clinician advice plus a 15-minute motivational counseling session. The current study (n = 514) examined the groups' use of outpatient and inpatient medical services during two years after intervention. Although the intervention reduced alcohol consumption at six-month follow-up, intervention and control groups made similar numbers of outpatient visits (M = 17.7 vs. 18.3, respectively; p = .47), were equally likely to be hospitalized (21.2% vs. 22.0%; p = .81), and among those hospitalized, had similar lengths of stay (4.7 vs. 6.6 days; p = .37). Although brief interventions to reduce hazardous drinking may potentially reduce medical care utilization, more evidence is needed to substantiate their practicality and cost-effectiveness.

Introduction: The study was designed to test a brief intervention for reducing alcohol consumption among moderate to heavy (hazardous) drinkers in a busy HMO primary care setting., Methods: In a randomized controlled trial, hazardous drinkers (n = 516) were identified by the AUDIT screening questionnaire. Intervention included brief clinician advice (30 seconds), a 15-minute motivational session by counselors, and printed materials., Results: At six-month follow-up, intervention subjects reported fewer total standard drinks in the past three months (176 versus 216, P = .04, one-tailed) and fewer drinking days per week (2.8 versus 3.3, P = .02) than controls, but similar drinks per drinking day (3.3 versus 3.5; P = .13). At 12 months, intervention subjects again reported fewer drinking days per week (2.7 versus 3.1; P = .04) than controls, but similar numbers of standard drinks (157 versus 179; P = .13) and drinks per drinking day (3.6 versus 3.3; P = .20). Intervention subjects were somewhat more likely than controls to report drinking within daily recommended limits (< or = 3 for men, < or = 2 for women) at both six months (79% versus 71%; P = .06) and 12 months (80% versus 73%; P = .07), but did not differ significantly from controls on other drinking outcomes (percent abstinent, frequency of drinking > or = 6 drinks per drinking occasion, estimated peak blood alcohol concentration), or use of medical care in the year following intervention., Conclusions: A one-time, brief motivational intervention using minimal clinician time supplemented by trained counselors resulted in a modest reduction in frequency of alcohol consumption in a busy primary care population. Future research should focus on strengthening and maintaining intervention effects.

Objective: To reduce variability in primary care physicians' use of procedures for imaging the lumbar spine., Design: Controlled intervention using clinical practice guideline and practice pattern feedback., Study Sample: Sixty-seven internists and 28 family practitioners in a large, group-model HMO., Measurements and Main Results: Intervention group physicians received the clinical practice guideline for low back pain, followed after 4 months by three bimonthly feedback reports on their current use rates for lumber spine x-rays and computed tomography and magnetic resonance imaging scans of the lumbar spine. Control group physicians received neither the guideline nor the feedback reports. Automated radiology utilization data were used to compare intervention and control group physicians' changes in use rates and variability in use rates over the course of the study period. Neither the guideline alone nor the guideline plus feedback was associated with a significant decrease in use rates or in the variability in use rates for the lumbar spine imaging procedures under study., Conclusions: Clinical practice guidelines and practice pattern feedback fail to achieve their goals when features of the practice setting and patient expectations and behavior are not identified and addressed.

Health maintenance organizations have employed physician assistants, nurse practitioners, and other nonphysician providers for decades, yet there is little information on how satisfied these providers are with this form of practice. This paper examines how physician assistants evaluate their experience practicing in a large group model health maintenance organization and compares their attitudes and satisfaction levels with those of other nonphysician providers-nurse practitioners, optometrists, mental health therapists, and chemical dependency counselors. The data source is a 1992 survey of 5,000 nonphysician employees of a health maintenance organization. The survey instrument was a self-administrated questionnaire that included both structured and open-ended questions. The response rate averaged 88 percent for physician assistants and the other non-physician providers. Physician assistants expressed the most satisfaction with the amount of responsibility, support from coworkers, job security, working hours, supervision, and task variety. They were less satisfied with workload, control over the pace of work, and opportunities for advancement. Most physician assistants were also satisfied with pay and fringe benefits. Compared with other nonphysician providers, chemical dependency counselors expressed the highest levels of satisfaction across the various dimensions of work and optometrists the lowest. Nurse practitioners, chemical dependency counselors, and mental health professionals also tended to be satisfied with most aspects of practice in this setting. In a number of instances, they were more satisfied than the physician assistants. The findings are consistent with other studies that found health maintenance organizations to be favorable practice settings for physician assistants. The limits of physician assistant involvement and their role satisfaction and efficient use in HMOs are more likely to relate to physician attitudes and acceptance than to lack of support by coworkers and other attributes of the work environment.

Despite well-recognized gender differences in patterns of social network support, few studies have explored whether the pathways by which social support affects mortality risk differ for men and women. In a 15-year follow-up study of elderly HMO members, we found that network size affected men's mortality risk indirectly, through their health status, while no such indirect effect was found for women. The data also suggested that network size had a direct protective effect on mortality risk for both men and women, with men gaining protection at a lower level of network size than women. These findings confirm the need for a gender-specific approach to further research on this subject, and suggest the need to measure variables that capture the different meaning and value of social network participation for men and women.

Research on adults has documented that use of medical services decreases after initiation of treatment for alcohol problems, but little is known about this relationship among adolescents. We studied utilization and costs of care following participation in the Adolescent Chemical Health Program (ACHP) of Kaiser Permanente, Northwest Region, in 1986-88. Three groups of adolescents (and their parents) were identified: adolescents who were assessed and initiated treatment in ACHP (n = 561), adolescents who were assessed and recommended for treatment but did not return for treatment (n = 278), and adolescents with no known substance use problems (n = 381). Medical records were reviewed for 1 year pre- and 1.5 years postassessment. After adjusting for preassessment medical visits, severity of alcohol and drug use, gender, and age, analyses suggested that substance user treatment was not associated with reduced use of medical services or costs by either adolescents or parents.

If HMOs are to remain viable, they will have to control and eliminate conditions that contribute to the burnout of their physicians. Based on the belief that interventions should be based on a thorough understanding of how an HMO affects burnout, physicians in a large, prepaid group practice were asked to respond to objective and open-ended questions about their background and career, work conditions, professional autonomy, patient care, and stress and well-being. Findings about burnout prevalence, factors related to burnout, and burnout outcomes are presented and used as a starting point for a discussion of possible interventions.

HMOs are under increasing pressure to expand benefits and services for treatment of adolescents who abuse alcohol and drugs. Little information exists, however, on these programs. This article describes a comprehensive adolescent chemical dependency treatment program within an HMO and presents data on use, costs, and results. Characteristics of clients were similar to adolescents seen in community treatment programs. Less than 1% of the HMO adolescent population sought treatment, and the mean number of outpatient treatment visits was 9.7. The median was between two and three visits, and the mode was one visit. Thirty-four percent of the adolescents required residential treatment, and 65% of the adolescents completed the recommended 28-day stay. The additional premium cost per member per month for adolescent chemical dependency treatment was approximately $0.28. A telephone follow-up survey of a random sample of treated adolescents found that most adolescents had reduced their use of alcohol and drugs and had made improvements in other areas of their lives. Few adolescents, however, met the program's goal of total abstinence.

Kaiser Permanente Northwest Region is a prepaid group practice health maintenance organization. Among the employed are 65 physician assistants (PAs) who work in primary care and in certain subspecialties. Kaiser Permanente was one of the first managed health care systems to use PAs and has contributed to the historical documentation of their effectiveness. An interest in experimenting with new forms of health care delivery as well as enabling State legislation has contributed to an expanded role for nonphysician providers. Together with nurse practitioners, PAs comprise 20 percent of the primary care staff and write 25 percent of the prescriptions for the membership. The use of PAs in managed health care settings will likely increase to meet growing primary care demands. Analysts have found the cost of a PA ranges from 25 percent to 53 percent of the cost of a physician. PAs are capable of providing care for 86 percent of the diagnoses seen in outpatient primary care setting, and patient acceptance is high.

This study identified consistently high and low users of medical care services in a group of older HMO members continuously enrolled for six years. Consistently high users made up 26% of the sample, but accounted for more than 50% of total outpatient contacts and hospital admissions. Average ambulatory care costs were more than four times greater for the high users compared with the low users. Consistently high users were older than consistently low users, but did not differ significantly in other sociodemographic characteristics. Compared with the low users, the high users reported more total medical conditions and were more likely to indicate they had arthritis, high blood pressure, heart conditions, and other chronic problems. They perceived themselves to be in poorer health and reported higher levels of psychologic distress. The low users tended to be less satisfied, but the two user groups were not significantly different regarding use of services outside the HMO.

Smoking behavior of consistently high and low users of medical care services were compared in two groups of older health maintenance organization (HMO) members continuously enrolled for five years and a subgroup who were continuously enrolled for 10 years. Smokers and former smokers, combined, were more likely than never-smokers to be consistently high users of ambulatory services (52 percent vs 34 percent in the five-year group, and 45 percent vs 30 percent in the 10-year group).

An original goal of the Kaiser-Permanente Neighborhood Health Center Project was to organize the project so that a medically indigent population would be able to utilize fully and appropriately the services of a complex medical care program. A special program of outreach services was designed as the principal means to achieve this goal. This study was made to determine the effects of these outreach services on (1) the use of or nonuse of ambulatory care services; (2) the volume and type of services used; (3) the patterns of use; and (4) the appointment-keeping behavior of the project population for a 12-month period. Outreach and medical care services were provided to an average of 7,000 persons in 1,500 low-income families who were enrolled as health plan members in the Kaiser-Permanente Medical Care Program. Project participants were randomly divided into two groups, one with and one without services, and utilization data were collected from their medical and administrative records. The findings suggest that outreach intervention had a positive effect on access to care. Persons who received outreach services were more likely to contact the medical care system; these persons also showed a substantial difference in the volume of services they used, when compared to those without outreach services. Outreach workers were less successful in changing utilization patterns, although slight differences were found in the direction of more appropriate use. Persons with outreach services were more likely to have made contacts with their regular physician, to have made a smaller proportion of walk-in contacts, to have had a higher proportion of regularly scheduled contacts, and to have made a higher proportion of continuing visits. Outreach workers also had little or no effect on appointment-keeping behavior.

The purpose of this study was to determine HMO physicians' receptivity to special organized programs dealing with sociomedical and behavioral problems. The study population consisted of full-time physicians in a large prepaid group practice HMO, and the data were obtained in 1977 by means of a self-administered structured questionnaire. Most physicians favored special organized services for alcoholism, drug abuse, obesity, disturbances in sexual relations, and the like. Except for alcohol and drug abuse, favoring organized services for one problem did not correlate highly across problem areas. Specialty, AMA membership, and political orientation were the main characteristics that differentiated physicians on their receptivity to organized programs for sociomedical problems. Social background and professional training and experience may be more important than the practice setting in influencing physicians' receptivity to these types of services.

When evaluating the effectiveness of medical care programs, one concern is whether receipt of care is based upon health care needs or upon socioeconomic status. This study describes the relation between health status and socioeconomic status and attempts to determine which has the greater effect on ambulatory care utilization. The study setting was an operating HMO serving a cross-sectional membership of nearly 200,000 persons. Outpatient utilization data were derived from the medical records of a five per cent sample of health plan members for 1969 and 1970. Social, economic, situational, and attitudinal data were provided by 2,603 respondents in a household interview survey. Since a population's perceived health status may reflect health need, information from the survey provided measures of health status that ranged from specific symptoms and complaints to a general measure of perceived health status. Although the findings varied somewhat according to which variables were considered, they generally showed health status to correlate more highly than socioeconomic factors with the utilization of services in this medical care system. An exception was the use of preventive services, which was not significantly related to health status measures but rather, for women, to education and, to a lesser extent, income.

The study examines the attitudes of physicians in a large HMO toward expanded activities for pharmacists and attempts to determine the physicians and their practice characteristics that account for these attitudes. The setting was the Oregon Region of the Kaiser-Permanente Medical Care Program. The data were obtained from a survey of the physicians serving the membership. The findings indicated that about two of every three physicians favored expanded activities for pharmacists. The specific activities most favored were those least threatening to physicians' roles. Their attitudes were most strongly related to whether they favored prescribing privileges for physician assistants and nurse practitioners. The physicians and their practice characteristics accounted for a small amount of the variation in physicians' attitudes. Large HMOs may be favorable settings for expanding the activities of pharmacists.

A special program of outreach services was implemented to assist a poverty population to appropriately use health services in the Kaiser-Permanente Medical Care Program. A study was conducted to determine the effect of outreach workers' intervention on the use of preventive services by this population. Intially, families were divided into two groups, one with and one without outreach workers. Outreach workers (neighborhood health coordinators) were trained in prevention and health education. They were then assigned to specific subgroups of the poverty population to teach the importance of preventive services and to motivate persons to use these services. This paper focuses on the effect of outreach workers' services on the use of selected preventive care services (immunizations and tine test) by preschool children from poverty families. Preschool children in families with coordinator services had higher use rates for preventive care. The sub-group for which outreach workers were specially trained to focus on preventive procedures for the pre-school group had markedly higher use rates for preventive care. The findings suggest that special intervention programs, using indigenous and nonprofessional outreach workers, can increase preventive service utilization by poverty groups.

The relationships between duration of membership and the use of outpatient medical care services for disadvantaged and general membership groups in a prepaid group practice were analyzed. The effects of duration of membership upon utilization rates during a six-year study period were estimated by retrospective cohort analyses. Rates were computed for successive three-month intervals. There were no significant start-up effects on use for new enrollees in either study population. This finding is in contrast to what has been reported in other studies. New HMOs or existing HMOs with many new members may not require the additional resources previously thought necessary. Average cohort utilization rates were higher for the disadvantaged cohorts, with the exception of young males 0 to 14 years old. We also examined the persistence of utilization patterns over time. Consistently high users and nonusers were found in both populations. In general, these findings suggest that length of time does not affect ambulatory care use. Age, sex, socioeconomic background and health status are more important than duration of membership in accounting for differences in utilization.

Data from a 6-year follow-up period were used to analyze the mortality of adults who had been continuously enrolled in a prepaid group practice health maintenance organization (HMO) for the previous 7 years. Thirteen percent of the HMO members were classified as consistently high users of outpatient care. Even after adjustment for age, sex, and cigarette smoking, this high user group's mortality rates during the follow-up period were significantly greater than those of the other HMO enrollees. The HMO members as a whole had mortality rates marginally lower than the mortality of the general Oregon population. It can be concluded that HMO enrollees who consistently use high quantities of health care resources generally have significant medical problems in addition to psychosocial difficulties. Assuming that mortality reflects health, long-term HMO members as a group are not much healthier than ordinary Oregonians.

This article explores the potential for collaboration between investigators in institutions outside of the VA and those engaged in research within the VA. The focus is on the potential for collaborative work in health services research; our perspective is that of researchers in a freestanding HMO research center affiliated with the Veterans Administration's Northwest Health Services Research and Development Field Program. The paper begins with a review of the reasons that make collaboration between VA researchers and other health services researchers so appropriate at this time. An example of collaboration is presented, drawing on the experience of the Northwest Field Program and the Kaiser Permanente Center for Health Research. Finally, some difficulties inherent in collaboration between VA and other health services researchers are discussed.

Use by physicians of medical care resources to treat similar patients and morbidities has implications for the efficiency and effectiveness of medical care. This study examined the variation in primary care physicians' use of outpatient laboratory, radiology, drugs, and hospital admissions; the relationships among the uses of these resources by physicians; and uses of these resources by physicians over time. The setting was a large HMO and the physicians were internists (n = 30), pediatricians (n = 20), and obstetrician/gynecologists (n = 17). Resource use was a dollar rate for laboratory, radiology, and drug, and an admission rate per contact for the hospital. The physicians' use rates were derived from the computerized utilization record of a sample of HMO members for 1976-1978. Substantial variation occurred in the rates of use of each resource by physicians in each specialty. The relationships among the resource rates were similar for internists and pediatricians, and were somewhat stronger than expected clinically. There were also substantial consistencies over time in the levels of use of laboratory and drugs by internists and pediatricians. Findings suggest a potential for more judicious use of ambulatory care resources by primary care physicians in this setting.

Criteria used in this study established that 13% of long-term adult members of a prepaid group practice health maintenance organization (HMO) were consistently high users of outpatient medical care services. This population accounted for 31% of the total doctor office visits (DOVs), 35% of the hospital admissions, and 30% of the outpatient surgical services for long-term members. The most frequent reason for DOVs in this high user group was treatment and/or follow-up of chronic conditions. Patterns of utilization were unrelated to marital status, income, occupation, and perceived social class. Smoking and alcohol use also were not associated with utilization patterns. However, the consistently high users were more likely to perceive their health status as fair or poor and to report a higher number of physical symptoms. They were also more likely to be characterized by a higher degree of psychological distress, especially depression. Contacts with the HMO's mental health department constituted less than 1% of their total medical care contacts, and only 13% made at least one mental health contact over the study period. The findings are discussed in terms of their health and medical care implications.

This study seeks to compare clients' reports of satisfaction/dissatisfaction with an HMO and its medical care program to providers' perceptions of client satisfaction. The main conclusion is that providers are fairly aware of medical care system attributes with which clients are highly satisfied and dissatisfied. Those with direct patient contact, especially physicians and nurses, are most accurate in their perceptions. Personnel without direct patient contact are about equally accurate in their perceptions of client satisfaction, though personnel without patient contact perceive relatively more dissatisfaction than clients report. Most approaches to medical care evaluation examine costs and efficiency and, more recently, the technical quality of care. Less attention has been given to the assessment of client satisfaction, and almost no attention has been given to what providers perceive as the assessment of clients. Clearly, HMOs' viability is determined largely by their ability to identify and correct major problems affecting their enrollments.

This study examined the attitudes of physicians working in health maintenance organizations toward the use of nurse practitioners and physician assistants. It also explored some of the underlying reasons for these attitudes: effect upon quality of care, risk of malpractice, role threat and gender bias. The setting was a health maintenance organization serving 270,000 members. The data were derived from a survey of physicians' attitudes and behavior. Physicians from internal medicine, pediatrics and obstetrics-gynecology were the study population. Internists and pediatricians had favorable attitudes toward both nurse practitioners and physician assistants. Obstetrician-gynecologists had somewhat less favorable attitudes. Physicians in all three specialties favored nurse practitioners more than physician assistants. Physicians felt that nurse practitioners were more likely to increase the quality of care and less likely to increase the risk of malpractice. Nurse practitioners were not seen as a greater role threat. Some gender bias appeared to be present, but it did not appear to constrain the use of nurse practitioners. Large, multi-specialty, prepaid group practice health maintenance organizations may be favorable settings for nurse practitioners and physician assistants to practice primary care.