|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Visits to the emergency department by Medicaid recipients for nonemergency problems are common and contribute to rising health care costs. However, such patients may have few alternatives. We conducted a telephone survey of 953 ambulatory care sites in 10 cities to determine the availability of appointments for Medicaid recipients with common problems.
Methods Research assistants telephoned all ambulatory care clinics and a stratified sample of private primary care practices in the catchment area served by the hospital emergency department in each city. The assistants identified themselves as Medicaid recipients seeking care for one of three problems (low back pain, dysuria, or sore throat) and asked a standardized series of questions. Data were collected on appointments or walk-in visits authorized at any time, within two days after the call, or after 5 p.m.; copayment requirements; and reasons appointments could not be made. If an appointment was made, it was canceled at the end of each call or shortly thereafter. Several weeks later, private-practice sites in six of the cities were recontacted; the research assistants identified themselves as patients with private insurance and the same problem.
Results An appointment or an authorization for a walk-in visit was obtained from 418 of the 953 practice sites (44 percent); 47 of the sites (5 percent) could not be contacted. Appointment rates for the different types of sites ranged from 72 percent for free-standing urgent care centers to 34 percent for private practices. "Not accepting Medicaid" was the most common reason given for not granting an appointment or walk-in visit. Only 72 of the sites (8 percent) offered after-hours care within two working days after the call without a cash copayment. Sixty percent of the 330 private practices that were recontacted agreed to see a patient with private insurance within two working days, but only 26 percent agreed to see a patient with Medicaid within two days (P<0.001).
Conclusions Medicaid recipients in urban areas have limited access to outpatient care apart from that offered by hospital emergency departments.
If Medicaid recipients cannot gain ready access to ambulatory care, proposals to discourage the use of emergency departments for such care may further compromise the health of this underserved population. We conducted a telephone survey in 10 cities to address three questions: Can Medicaid recipients with common symptoms readily obtain appointments for outpatient care? If difficulty is encountered, are the reasons financial or nonfinancial? And when Medicaid recipients are denied outpatient appointments, where are they told to seek treatment?
Methods
Study Sites
Research teams in 10 academic emergency departments that provide a substantial amount of ambulatory care to the poor participated in the study. These departments are located in the following hospitals: Boston City Hospital, Boston; Johns Hopkins University Hospital, Baltimore; John K. Doyne Hospital, Milwaukee; Lincoln Medical and Mental Health Center, Bronx, N.Y.; Oregon Health Sciences University, Portland; Regional Medical Center, Memphis, Tenn.; Tampa General Hospital, Tampa, Fla.; Texas Tech Health Sciences Center, El Paso; University of California, Davis, Medical Center, Sacramento; and the hospitals affiliated with Wright State University, Dayton, Ohio. In the 10 states where the participating hospitals are located, Medicaid reimbursement rates for a limited office visit range from $17 to $33, according to Medicaid fee schedules. Reimbursement rates for the same care rendered in a hospital emergency department range from $14 to $27. Four of the 10 states (Tennessee, Wisconsin, California, and Florida) permit physicians to charge a small copayment; the other 6 do not.
Identification of Practices
Each research team used the local telephone directory and the emergency department's referral list to identify all ambulatory care clinics and private primary care practices in the geographic area served by the department. Each site of care was categorized as a university- or hospital-affiliated primary care clinic, a publicly funded or subsidized clinic (e.g., a county health department or federally funded neighborhood health center), a clinic funded by a private charity, a free-standing urgent care center, or a private physician's office. If the total number of practices in an area exceeded 120, we stratified private physicians' offices according to the physician's degree (M.D. or D.O.) and specialty (family practice, general practice, or internal medicine) and called a proportionate, random sample of each subgroup. Group practices with a common telephone number were considered a single site of care.
Training of Interviewers
Each research assistant received a training manual that outlined the rationale and design of the study and included guidelines for making telephone calls, an interview flow sheet, and a standardized form for reporting data. Additional training was provided by site supervisors. Mock interviews were conducted. Once the instruction had been completed, each assistant viewed a videotape depicting six simulated calls of increasing complexity and scored the calls with an interview flow sheet. Analysis of these ratings showed a high degree of interrater agreement (kappa statistics for key outcomes ranged from 0.91 to 1.0)11.
Interview Procedure
The telephone calls simulated requests for appointments from Medicaid recipients. One of three physically uncomfortable symptoms (low back pain, dysuria, or sore throat) was randomly assigned to each call. All calls were made between 9 a.m. and 3 p.m. Monday through Wednesday. After reaching a person authorized to accept the call, the research assistant reported the symptom and requested an appointment or permission to come in without an appointment "as soon as possible." Information about insurance (i.e., Medicaid, Medicare, or private insurance) was not volunteered, but it was always requested. If an appointment could be obtained, the caller asked to be seen within two working days. If this request was granted, the caller asked to be seen after 5 p.m. If copayment was required, the caller asked whether it could be waived.
Whenever a request was denied, the caller asked for the reason and also asked where else he or she could go to be seen. If no alternative was recommended, this was noted. If more than one option was offered, the first was recorded. If an appointment was made, it was canceled at the end of the call or shortly thereafter. If a busy signal was obtained, the caller waited five minutes before trying again. If five attempts were unsuccessful, the caller tried again the following day. If five additional attempts were unsuccessful, the effort to reach the site was abandoned.
Approximately three weeks after this initial series of calls, research assistants in six cities (Baltimore, the Bronx, Dayton, Memphis, Portland, and Sacramento) recontacted all the private practice sites on their lists and identified themselves as patients with private insurance. This second round of calls was otherwise identical to the calls described above.
Statistical Analysis
Data were analyzed in aggregate according to city and practice type. A chi-square analysis was used to compare rates of access according to subgroup, and two-tailed P values less than 0.05 were considered to indicate statistical significance. Ninety-five percent confidence intervals were calculated for the proportion of ambulatory care sites that granted appointments in response to the three requests (an appointment, an appointment within two days, and an appointment after 5 p.m.)11.
Informed Consent and Other Considerations
The study did not involve subjects enrolled in experimental protocols and posed no risk to clinic personnel. Staff of the institutional review board at the University of Tennessee determined that our protocol did not require a review (Cox CE: personal communication). No institutional review board at the other participating institutions required that informed consent be obtained from the staff of the clinics that were contacted. Nonetheless, strict measures were taken to protect the identity of every private practitioner and ambulatory care clinic. All analyses were conducted in aggregate. Identifying information was stored separately and has since been destroyed. All appointments were canceled at or shortly after the end of the telephone calls. To the best of our knowledge, no patient was denied access to care, and no practice lost an opportunity to treat a patient as a result of this study.
Results
A total of 953 ambulatory care sites were called (Table 1). Fewer sites were called in smaller cities, such as Dayton, and in catchment areas with relatively few private practices, such as those of Lincoln Medical Center in the Bronx and Boston City Hospital. Our sample included 119 hospital-affiliated clinics, 85 publicly funded or subsidized health care centers, 60 nonprofit charity clinics, 88 free-standing urgent care centers, and 601 private primary care practices.
|
|
|
Urgent Appointments
Fewer requests for an appointment were granted when the research assistants asked to be seen within two working days. Eighty of the 418 providers who could accommodate our callers at a later date could not schedule an earlier visit. Only 128 sites, including 45 free-standing urgent care centers, could accommodate our callers after 5 p.m. At other sites, either the practice was closed after 5 p.m. or the evening hours were fully booked.
Copayments
Mandatory copayments were another potential barrier to care. A copayment was requested by 168 of the 418 sites (40 percent) that initially agreed to give our callers an appointment or walk-in visit. The median amount requested was $45 (range, $1 to $200). Staff at 134 of these 168 practices (80 percent) stated that this fee would not be waived if the patient was unable to pay. Sixty-one of the 128 sites that agreed to see our callers after 5 p.m. (48 percent) required a copayment. Staff at 56 of these 61 practices (92 percent) stated that the fee would not be waived if the patient was unable to pay. Only 72 of the 953 primary care sites (8 percent) offered after-hours care within two working days without a cash copayment (Figure 1).
|
One third of the practices that did not schedule an appointment or agree to a walk-in visit also did not recommend an alternative site for care. Sixteen percent advised the callers to go to an emergency department. Other recommendations included contacting another doctor's office (12 percent), an urgent care center (12 percent), or a doctor's referral service (6 percent).
Private Insurance
Research assistants in six cities had substantially greater success obtaining an appointment when they called the private practices again and identified themselves as patients with private insurance. Sixty percent of the 330 private practices that were recontacted agreed to see a patient with private insurance within two working days, whereas only 26 percent of these same practices had granted a request for an appointment when the caller was thought to have Medicaid (P<0.001). Practices could not be contacted by telephone in less than 5 percent of instances.
Discussion
The Medicaid program was created in 1965 to make health care more accessible to the poor12. It is widely believed that many Medicaid recipients still encounter substantial barriers to primary care, although this view has not been verified1,2,3,4,5,6,7,8,9,10. Because of the considerable interest in restricting access to nonemergency care in hospital emergency departments, it is particularly important to determine the availability of other types of outpatient care for Medicaid recipients.
The study technique we used -- training research assistants to pose as patients and asking them to call for an appointment -- was used by Dallek to assess the effect of Medicare payment reform in Los Angeles County13. We asked research assistants to pose as patients because it was the only practical way to assess directly the experience of Medicaid recipients when they try to schedule outpatient appointments. We found that access to appointments for such patients was limited, especially in private physicians' offices.
Our study has several limitations. First, our findings are based on telephone calls made within the catchment areas of urban emergency departments that provide a large volume of ambulatory care to the poor. Access to outpatient care for Medicaid patients in these settings may differ from that in suburban communities, small towns, or rural areas. It is noteworthy that the use of emergency departments by the poor and uninsured has increased more in rural communities than in urban areas, suggesting that access to care in rural communities may be even more limited1.
Second, our callers identified themselves as new patients. Medicaid recipients who have an established relationship with a primary care provider may have less difficulty in obtaining appointments. Nonetheless, many practices did not accept Medicaid recipients, and some that did could not provide appointments because they were overbooked or closed after 5 p.m. These factors would presumably also affect Medicaid recipients who already have a primary care provider.
Third, we did not categorize ambulatory care practices according to the volume of care provided to Medicaid recipients, the annual number of outpatient visits, or other utilization criteria. Such information is not generally available to patients seeking care. Instead, we considered each clinic or private practitioner as a potential point of contact with the local health care system. We did not call more than one physician at each group practice because we assumed that if the group accepted Medicaid recipients and a particular doctor was not available, the receptionist would arrange an appointment with another member of the group. When this occurred, we recorded the call as a successful contact.
Fourth, a clinic receptionist who suspected that the caller was a research assistant posing as a new patient may have altered his or her responses. Given the difficulty our assistants had in obtaining appointments, however, it is unlikely that this was a confounding problem. No research assistant was challenged, and none of the investigators received a complaint.
Finally, we included only one city (Dayton, Ohio) where large numbers of Medicaid recipients were being treated under a system of managed care. At the time of our study, all but 1 of the 10 states were experimenting with this approach, but most were doing so in small pilot projects12. The effect of managed care on Medicaid recipients' access to outpatient treatment is unclear, but the early results are not encouraging14. In Minneapolis, a city where a large number of patients receiving Medicaid and general assistance have been assigned to health maintenance organizations, overall use of emergency departments has not declined substantially (Ling L: personal communication).
It is not appropriate to draw conclusions about the availability of care from any single physician or clinic on the basis of one telephone call. At sites that treat Medicaid recipients, the availability of appointments may vary from day to day or even hour to hour. We believe that an aggregate analysis overcomes these limitations.
Although financial considerations limit the access of Medicaid recipients to outpatient care, nonfinancial factors also play a part. Forty percent of the 330 private primary care practices we called in six cities could not accommodate a request to be seen within two working days, even when the research assistant posed as a private patient. Heavy patient bookings, limited office hours, and the small number of primary care providers in some neighborhoods are also barriers to care15. Many primary care clinics are not open in the evening. Limited hours may present a substantial barrier to patients with children, jobs, or other responsibilities that make it difficult or impossible to schedule an appointment during the day4.
About half the ambulatory care sites that could accommodate a caller after 5 p.m. requested a substantial copayment. According to federal law, Medicaid beneficiaries can be required to make only small copayments; however, the maximum is unspecified16.
Seeking care in a hospital emergency department is rarely a matter of convenience. Many emergency departments are crowded with acutely ill and injured patients, as well as those who have been admitted to the hospital and are waiting for a bed1,17,18. Overcrowding results in a prolonged waiting period for patients with relatively minor illnesses or injuries1,4,18. According to reports from California, long waits cause many walk-in patients to leave the emergency department without being seen6,7.
Efforts to limit emergency department visits when alternative sources of care are lacking may be fraught with problems19,20,21. The medical conditions used in our study were selected from a list of 51 that are not considered serious enough to warrant treatment in the emergency department of the University of California, Davis, Medical Center in Sacramento22. Only 22 percent of the 120 ambulatory care sites we studied in Sacramento were able or willing to treat a Medicaid recipient within two days after the call.
Our results suggest that Medicaid recipients in urban areas have limited access to ambulatory care outside hospital emergency departments. These findings should be verified by a larger and more systematic survey before steps are taken to restrict access to emergency department care. Mandatory copayments, triage of patients who are not considered sick enough for emergency department care, limits on insurance reimbursement, and the use of gatekeepers to restrict emergency department visits may temporarily control costs but could adversely affect the health of Medicaid recipients2,3,12,14,19,21,22. If financial and administrative barriers to outpatient care in the emergency department are imposed before workable alternatives are in place, many of the poor will have nowhere to go.
Supported by a grant from the Robert Wood Johnson Foundation.
We are indebted to Kris Arheart, Ph.D., and Knox Todd, M.D., for their help with the statistical analysis; to William B. Applegate, M.D., for his review of an early draft of this paper; and to Annie Lea Schuster, who supervised this grant for the Robert Wood Johnson Foundation.
Source Information
Dr. Kellermann, as the chairman of the study, assumes full responsibility for the overall content and integrity of the manuscript.The members of the Medicaid Access Study Group are as follows: University of Tennessee, Memphis -- A.L. Kellermann, C. Conway, R. Wolcott, B.B. Hackman, and S. Bogan; Boston University, Boston -- E. Bernstein, S. Fish, and T. Kerl; University of South Florida, Tampa -- T. Mitchell and C. Ramsey; Johns Hopkins University, Baltimore -- G.D. Kelen and J.B. Shahan; New York Medical College, Bronx, New York -- H. Osborn and L. Bazan; Medical College of Wisconsin, Milwaukee -- S. Hargarten and S. Laurence; Oregon Health Sciences University, Portland -- J.R. Hedges, G. Henkel, S. Winter, and A. Yekrang; Texas Tech University, El Paso -- L. Binder, B. Kempton, and D. Harriman; University of California, Davis, Sacramento -- G. Wintemute and H. Smith; and Wright State University, Dayton, Ohio -- G. Hamilton and L. Morris.
Address reprint requests to Dr. Arthur L. Kellermann at the Division of Emergency Medicine, Department of Surgery, Emory University School of Medicine, 1462 Clifton Rd., N.E., Atlanta, GA 30322.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Access of Medicaid Recipients to Outpatient Care
Derlet R. W., Kinser D., Bessinger C.D., Pollock S. G., Pipas J.E., Kellermann A. L., Wintemute G., Binder L.
Extract |
Full Text
N Engl J Med 1994;
331:877-878, Sep 29, 1994.
Correspondence
Costs of Visits to Emergency Departments
Baier C. L., Auerbach S. B., Becker K. A., Hearst N., Lucey C., Williams R. M.
Extract |
Full Text
N Engl J Med 1996;
335:209-211, Jul 18, 1996.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
