Brooks RG, Menachemi N, Clawson A, Beitsch L. Availability of Physician Services in Florida, RevisitedThe Effect of the Professional Liability Insurance Market on Access to Health Care. Arch Intern Med. 2005;165(18):2136-2141. doi:10.1001/archinte.165.18.2136
Access to care remains a key part of improving health care outcomes in the United States. Recent reports have suggested that the number of physicians able to meet the demands for access to care may be decreasing.
We surveyed physicians practicing in rural and urban/suburban areas of Florida in 2004 to determine whether changes were occurring in health care service delivery. Secondary outcomes assessed included changes in professional liability insurance and their possible effects on changes in service delivery.
Overall, 727 (54.4%) of responding physicians stated that the delivery of services had been decreased or eliminated in the previous year. The most commonly eliminated services were nursing home coverage (42.1%), vaginal deliveries (29.1%), cesarean deliveries (26.0%), emergency department coverage (22.8%), and mental health services (21.2%). Surgical specialists (70.2%) and general surgeons (68.5%) were the groups with the highest number of decreased or eliminated services, but this trend was broad, with 63.6% of obstetrician/gynecologists and 60.2% of family medicine physicians also decreasing or eliminating services. Decreases in services seem to be related to changes in professional liability insurance premiums when assessed by both percentage of change and total premium increases for physicians. Rural and urban/suburban physicians did not differ significantly in these assessments.
The findings suggest that physicians across Florida have continued to decrease or eliminate important health services and that these decreases seem to be related to the difficulty of finding or paying for professional liability insurance.
Recent articles in the scientific literature1- 5 and the lay press6 have highlighted the growing acknowledgment that the demand for physicians is likely to outstrip the available supply in the near future. Although much of the national dialogue on the topic of the physician workforce has revolved around the actual number of available physicians in the United States, the problem with physician service availability is more deeply embedded, with changing practice patterns such as numbers of patients seen per week,7 types of services provided for patients,8 and time spent dealing with administrative issues9,10 all playing a role in restricting patients’ access to needed medical care.
Simultaneous with this growing consensus of the need for more physicians in coming years, there has been a nationwide battle over medical malpractice and its effects on the professional liability insurance (PLI) market.11,12 Growing concern about the effects of the PLI crisis on patient access to health care has been noted, but little has been reported in the way of formal, systematic assessment of how physicians may be changing their services to adapt to the unstable PLI market.13 Are physicians decreasing or eliminating services so that they can afford PLI and remain in practice? If so, are there patterns that predict changes in services to patients? The answers to these and other questions about the impact of PLI market instability and its effect on health service delivery remain unclear. These issues are important to clinicians and policymakers alike because both are concerned with the delivery of high-quality care to patients throughout the country.
In 2003, at the height of the instability in the Florida PLI market, we surveyed, and later reported on,8,14 adverse changes in service provision by physicians in rural areas. Since then, policymakers in Florida have introduced and passed legislation designed to reform and stabilize the PLI market.15 In light of these recent legislative and market changes, the present study was initiated to revisit the trends in availability of services in rural areas and simultaneously to compare these trends with service delivery in urban/suburban locations of the state.
The survey method included 2 groups of physicians. Briefly, using a method identical to the one previously reported8,14 we identified all rural physicians in Florida. Using the same database, we obtained a comparison group of nonrural physicians by randomly selecting a specialty-stratified sample of urban/suburban physicians.
We obtained a list of all licensed physicians from the Florida State Department of Health. For purposes of this study, we used information on all allopathic or osteopathic physicians who had a clear, active license and a practice address in Florida. From this master list, we first identified rural physicians by their practice addresses who met any of the following 3 criteria by practicing in (1) the 33 statutorily designated rural counties in Florida, (2) rural areas of otherwise nonrural counties as designated by the rural urban commuting area codes,16 or (3) an area defined as rural according to the current Health Resources and Services Administration list of defined Florida rural ZIP codes.17 These 3 sources were used to capture an accurate and broad sampling of rural physicians in the state.
Next, we selected our urban/suburban comparison group based on matched specialty status. The specialty status of both the rural and urban/suburban groups was ascertained through self-reported status to the Florida State Department of Health or in some cases by assessment of residency training and board-certification status using the Web-based state practitioner profiling system. Ultimately, this selection process resulted in a database of 3588 physicians.
We mailed initial surveys in July 2004 and follow-up surveys to nonrespondents in September 2004. The survey included questions on physician and patient demographics, the physicians’ training and practice scope, recent changes in services offered, PLI premium changes, satisfaction with practice, and future practice plans.
After the first mailing, a number of surveys were returned as undeliverable, primarily because of unknown or changed addresses or incorrect practice location. We made numerous efforts to obtain updated mailing information, and surveys were remailed to those physicians with available updated practice location. Surveys were tracked by unique serial numbers, and participants who did not return a survey after 4 weeks were mailed a second copy of the survey (in September 2004).
All analyses were conducted using the statistical software program SPSS version 13.0 (SPSS Inc, Chicago, Ill). To be consistent with standard survey analytical strategies, we evaluated each question based on the responses received for that question. Therefore, the total number of individual respondents varied from question to question. Differences between rural and urban/suburban responses were analyzed using independent-sample t tests from continuous variables and the Fisher exact test or χ2 test for dichotomous and categorical variables, respectively. In all cases, we used 2-tailed tests, and P<.05 was the level of statistical significance.
The protocol was approved by the Florida State University, Tallahassee, institutional review board, and the Florida State University Survey Research Laboratory conducted the mailing and data entry.
Ultimately, 3413 surveys are believed to have reached their intended participants; of those surveys, 1346 were returned (response rate, 39.4%). The response rate for rural physicians was 685 (40.6%) of 1686 and for urban/suburban physicians, 661 (38.3%) of 1727. Among the specialties, only rural family physicians responded significantly more frequently than their urban/suburban counterparts (176/359 vs 132/372; P<.001). Demographic and practice characteristics of the respondents are shown in Table 1. In addition, we noted that insurance coverage for patients seen at these practices was private pay, 32.7%; Medicare, 37.1%; Medicaid, 12.9%; self-pay, 9.6%; and other, 6.7%. Ninety-five percent of respondents stated that they accept new private-pay patients (rural physicians, 94.7%; urban/suburban physicians, 95.9%; P = .31); 85.9% accept new patients covered by Medicare (rural, 87.7%; urban/suburban, 84.1%; P = .07); and only 53.7% accept new patients covered by Medicaid (rural, 56.8%; urban/suburban, 50.4%; P = .02).
Overall, 727 (54.4%) of 1336 responding physicians stated that they had decreased or eliminated the delivery of patient services in the previous year. This group included 380 (56%) rural and 347 (52.4%) urban/suburban physicians (P = .22). The decrease in or elimination of services based on procedure type for all physicians combined is outlined in Table 2. The most common services eliminated were nursing home coverage (42.1%), vaginal deliveries (29.1%), cesarean deliveries (26.0%), emergency department coverage (22.8%), and mental health services (21.2%). In addition to outright elimination, a number of physicians responded that they had decreased services in these areas as well (Table 2). Physicians who provided less “invasive,” or diagnostic, procedures (such as mammography, cytology [Papanicolaou smears], electrocardiography, radiology, and vaccine administration) all reported an increase in services performed.
The distribution of decreased or eliminated services by physician specialty types is presented in Table 3. Surgical specialists (70.2%) and general surgeons (68.5%), respectively, had the highest number of decreased or eliminated services. Obstetrician/gynecologists (63.6%) and family medicine physicians (60.2%) were also commonly represented in this group. Pediatricians were the only group in which less than half (41.5%) of the respondents indicated that they had not decreased or eliminated services in the last year. No significant differences were noted between rural and urban/suburban physicians in decreasing/eliminating services. The decrease/elimination of services by specific procedure is also shown in Table 3 for these 7 clinical disciplines. Changes in service delivery for these responding clinicians were broad in scope, with decrease in or elimination of services ranging from 33.4% for office-based surgical procedures to 59.5% for nursing home coverage.
A total of 597 (87.4%) rural and 568 (86.2%) urban/suburban physicians stated that they are currently covered by PLI. Of the approximately 13% of physicians without this insurance, 44 physicians (25.1%) stated that their PLI stopped within the last year, and 65 (37.1%) stated that their insurance stopped 1 to 2 years before the survey. No differences were seen between rural and urban/suburban physicians in the timing of their discontinuation of PLI.
Of the physicians who carry PLI, 938 responded to the request to estimate the percentage of change in their premiums in the previous year. This resulted in a mean increase in premiums of 64.8% for rural physicians (SEM, 10.6) and a 45.3% increase for urban/suburban physicians (SEM, 3.3) (P = .08). When asked, 911 physicians reported their actual current annual PLI premium. Four hundred sixty-nine rural physicians reported a mean premium of $31 699, and 442 urban/suburban physicians had a mean premium of $34 954.
Changes in health care services seemed to be related to changes in PLI premiums. Overall, physicians who had premium changes in the highest quartile (increase >50%) (61.1%) were more likely to indicate that they had decreased or eliminated services compared with those in the lowest quartile (increase <15%) (51.4%; P = .01). Similarly, when analyzed by actual premium paid, those physicians in the top quartile (increase >$40 000) (63.8%) were more likely to cut services compared with those who were in the lowest quartile (increase <$16 000) (51.1%; P = .003). Similarly, we noted statistically significant relationships between increases in PLI premiums and decrease in or elimination of services for rural physicians (66.2% vs 48.1%; P = .002) and for actual premiums for urban/suburban physicians (64.7% vs 43.0%; P = .001).
Next, we asked the 727 physicians who had stated that they had decreased or eliminated medical services within the last year, “to what extent has the difficulty paying for medical liability insurance played a role in decreasing or eliminating these services?” Of the 713 responding physicians, 416 (58.3%) replied that it had played “a large role” and 216 (30.3%) said “some role.” In addition, this same group suggested that “personal or family reasons” played “a large role” (11.1%) or “some role” (22.6%) in their changing service delivery. Inability to maintain skills played a role according to 15.7% of respondents, and decreased demand for services was a factor for 15.4% of this group. No significant differences were seen between rural and urban/suburban physicians in these trends.
Overall, 1320 physicians (670 rural and 650 urban/suburban) responded when asked about satisfaction with their current practice. Almost 20% (19.6%) of physicians said they were “very satisfied,” and 35.5% said they were “somewhat satisfied.” Slightly fewer than 22% (21.8%) said they were “somewhat dissatisfied,” and 9.5% were “very dissatisfied.” No significant differences existed between rural and urban/suburban physicians in satisfaction with practice.
When asked how long they intended to stay in their current practice, 146 (11.2%) of 1307 responding physicians stated “less than 2 years” and 251 (19.2%) said less than 5 years. Almost 27% stated that they intended to leave their practices in 5 to 10 years; 30.8%, in 10 to 20 years; and 12.1%, in more than 20 years. Almost 23% of rural physicians stated that they were intent on leaving their practices in 2 to 5 years compared with 15.4% of their urban/suburban counterparts (P = .002).
Among those who intended to leave their practices within 2 years, reasons for doing so included practice issues (44.1%), early retirement (30.1%), planned retirement (16.1%), family or personal reasons (9.8%), community issues (4.9%), and other reasons (22.4%). When this group (those leaving their practices within 2 years) was also asked, “to what extent has the inability to find medical liability insurance played a role” in their leaving, 39.9% said it had played “a large role,” and 21.7%, “some role.” Similarly, when asked “to what extent has the inability to pay for medical liability insurance played a role” in leaving within 2 years, 60.4% said it had played “a large role” and 17.4% “some role” in this decision. Neither of these reasons was significantly different between rural and urban/suburban physicians.
Additional analyses revealed that physicians who stated they were “somewhat” or “very” dissatisfied with their current practice were also more likely to have decreased or eliminated services within the last year. Almost 74% (73.5%) of physicians who were dissatisfied, vs 43.6% of those who were satisfied, decreased or eliminated services (P<.001). Those who were dissatisfied had seen a mean increase in PLI premiums of 61%, which was significantly higher than those who were satisfied (44.7%; P = .007). Last, those who were dissatisfied were more likely to state that they intended to leave their current practice within the next 2 years (21.4% vs 4.6% of those “satisfied”; P<.001). All 3 of these relationships between dissatisfaction and decreased services, PLI premiums, and intent to leave held true when examining only rural or only urban/suburban physicians.
Much of the recent debate on the availability of health services has focused on the actual number of physicians in the workforce and whether the current medical “pipeline” can produce enough physicians to meet the demands of an aging and technologically sophisticated American population.1- 5 Equally important, however, are questions that relate to what services those who practice medicine actually perform for patients and for whom they will be provided. In an effort to better understand the patterns of health care service delivery for patients in Florida, we surveyed physicians in both 2003 and 2004. Our initial research, conducted in 2003 at the height of the instability in the PLI markets in Florida, showed that physician services were being significantly curtailed in rural areas.8,14 The findings of the current study, conducted 1 year later, bring new concerns that the adverse changes in health care service delivery in this state are geographically broader and more persistent than previously identified.
These trends have continued despite attempts by policymakers to influence and stabilize the PLI market through new medical malpractice legislation.15 Although it is likely that any major change in PLI premiums resulting from the 2003 statutory revisions might take several years to take effect, one would have hoped that a year after the enacted changes physicians would have been optimistic enough to have stabilized or increased their patterns of health care service delivery. Based on the current study, such was not the case in Florida. These patterns of decreased or eliminated services and procedures suggest that additional work addressing barriers to health care, including those caused by PLI market instability, is needed if the goal of providing health care to all is to be realized.
Several findings of the current study regarding access to health care require specific comment. First, it is of great concern that the total number of physicians practicing in rural areas of Florida seems to be smaller than the previous year. Using the same criteria and data source8 to identify practicing rural physicians, 10% fewer physicians were available in 2004 than in 2003. Whether this finding represents a trend of physician movement out of rural areas of Florida or is an isolated event, the numbers of rural physicians needs to be carefully monitored.
Second, in another finding that causes concern, about 1 in 7 responding physicians stated that they do not accept new patients covered by Medicare, and almost 50% do not accept new patients with Medicaid coverage. The latter finding bodes poorly for patients in areas already burdened by a lack of health care providers.
Third, a broad decrease and elimination of services was noted for physicians in both rural and urban/suburban areas. For the rural physician population, who had been surveyed 1 year earlier, the current finding represents a continuing decline in services. For example, when surveyed in 2003, of 76 rural family medicine physicians who had been performing vaginal deliveries, 53 (69.7%) decreased or eliminated them.8 When resurveyed in 2004, only 31(14.1%) of 220 rural family medicine physician respondents were performing deliveries, and 23 (74.2%) of those had decreased or eliminated them. As in the previous year, the numbers of high-risk surgical and invasive procedures were most likely to decline. Important services such as vaginal and cesarean deliveries, hospital-based surgical procedures, and emergency department and mental health coverage were all noted to have been decreased or eliminated by a broad array of practicing physicians. In addition, nursing home coverage was newly assessed in the current study and was found to be frequently decreased or eliminated by many physicians as well. Moreover, similar to last year’s assessment, the decline in provision of services seems to span most of the clinical specialty areas of physicians. Although surgeons (both specialty and general) had the highest rates of declining service provision, primary care specialties and obstetrics/gynecology were affected as well.
Also similar to the previous report,8 the decline in services seems to be related to PLI issues. Physicians were more likely to decrease or eliminate services when their PLI rates increased the most. Similarly, a correlation exists between those who paid the highest premiums and those who decreased or eliminated services. The fact that physicians who pay the highest PLI premiums practice in higher-risk specialties (ie, surgery) likely contributes to this relationship. In addition, the present study identified a trend that numerous physicians are in practice without PLI coverage. The percentage of Florida physicians “going bare,” 13%, is slightly higher than the previous finding, 10%, in this key state. We suspect, however, that this figure underrepresents the percentage of Florida physicians without PLI coverage, because the sample was weighted more to primary care specialties and less to higher-risk, higher-premium–paying subspecialties (eg, neurosurgery, high-risk obstetrics).
The survey also evaluated physician satisfaction and future plans for practice. In the case of rural physicians, who had been surveyed the year before, the general feeling of satisfaction (or dissatisfaction) was unchanged. Given the policy changes in the state and the general diminution of publicity on the medical malpractice crisis, this lack of change in general feeling in itself may be a harbinger of lingering doubts about the overall practice climate in Florida. This feeling may also be reflected in the fact that, overall, 11% of physicians in the present study said they intend to leave their practices within 2 years and another 19% within 5 years. The trend was stronger for rural practice, suggesting continued pressures on this group of physicians and the rural communities they serve.
Although we believe that the trends described herein show important evidence of changes in access to health care in Florida, we do recognize a number of limitations to this study. Even though the participation rate of almost 40% is similar to that shown in many major surveys of physicians,18,19 the possibility exists that physicians who were more adversely affected by the changing PLI situation were more likely to respond to the survey than those who were not. In addition, responses were based on self-reported data and were not independently verified. Last, the findings of the present study represent a select group of practicing physicians in Florida and may not be representative of physicians in other communities or states.
Despite these limitations, we believe the findings presented herein suggest strongly that physicians across Florida have continued to decrease or eliminate important health care services. This trend seems to be affecting a broad array of services and types of physicians, both generalists and specialists. Given the importance of access to health care for vulnerable populations, these statewide trends suggest the need for additional attention by physician leaders and policymakers to the ongoing effects of the PLI market.
Correspondence: Robert G. Brooks, MD, Florida State University, College of Medicine, 1115 W Call St, Tallahassee, FL 32306-4300 (Robert.Brooks@med.fsu.edu).
Accepted for Publication: June 2, 2005.
Financial Disclosure: None.
Funding/Support: This study was supported in part by funding from the Center for Rural Health Research and Policy of the Florida State University College of Medicine, Tallahassee.