Darnell JS. Free Clinics in the United StatesA Nationwide Survey. Arch Intern Med. 2010;170(11):946-953. doi:10.1001/archinternmed.2010.107
Copyright 2010 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2010
Since an increasing proportion of the US population is without health insurance, a network of free clinics has gradually developed to provide care for the uninsured. Despite widespread concern about the uninsured and the viability of the safety net, free clinics have been overlooked and poorly studied, leaving old assumptions and beliefs largely unchallenged. As a result, policy discussions have been forestalled and potentially fruitful collaborations between free clinics and other safety net providers have been hindered. The objective of this study is to describe the attributes of free clinics and measure their contribution to the safety net.
National mail survey of all known free clinics in the United States. The main outcome measures were organizational structures, operations, revenue sources, patient profiles, services, and staffing.
The study represents the first census of free clinics in 40 years and garnered a 75.9% response rate. Overall, 1007 free clinics operated in 49 states and the District of Columbia. Annually, these clinics provided care for 1.8 million individuals, accounting for 3.5 million medical and dental visits. The mean operating budget was $287 810. Overall, 58.7% received no government revenue. Clinics were open a mean of 18 hours per week and generally provided chronic disease management (73.2%), physical examinations (81.4%), urgent/acute care (62.3%), and medications (86.5%).
Free clinics operate largely outside of the safety net system. However, they have become an established and meaningful contributor to it. Policymakers should consider integrating the free clinic network with other safety net providers or providing direct financial support.
Our nation's 46 million uninsured1 often delay or forego needed health care because the cost is prohibitive.2- 6 Traditional sources of primary care include private physicians, federally qualified health centers (FQHCs), public clinics, hospital outpatient departments, and emergency departments. More important, most require cost-sharing. The mean cost to an uninsured patient for a physician visit—the usual source of care for one-third of the uninsured7—has been reported to be more than $50.8- 10 The FQHCs are required to use a fee scale based on a patient's annual income and family size: fees range from $5 to $24 for patients whose income is at the poverty level to $87 for patients whose income is twice that level.11,12 Moreover, FQHCs bill patients. Public clinics also collect fees,13 ranging from $22 for patients whose income is at the poverty level to $97 at twice the poverty threshold.11 Aside from cost considerations, care is frequently difficult to find, especially for those with the least resources.8
On the margins of the formal health care safety net for uninsured people, free clinics serve to partly offset these costs and access problems. Structured as private, nonprofit organizations, free clinics offer basic health care services to uninsured patients by licensed volunteer clinicians at little or no cost. Very little is known about free clinics despite their being one of the few viable options for uninsured people with limited funds.
Free clinics have evolved from ad hoc “outlaw force[s] in medicine,”14(p156) treating drug addicts and runaway youth,15- 18 and shunned by the American Medical Association,14 to an established component of the health system.19- 22 The American Medical Association began to support free clinics in 1994.23 Also in the 1990s, a $12 million Robert Wood Johnson Foundation initiative supported their development.24 Finally, through the Health Insurance Portability and Accountability Act of 1996, the US Congress has extended medical malpractice protection for volunteer free clinic health care professionals.
The free clinic literature is dominated by accounts of individual clinics rather than the sector as a whole. These accounts cover free clinics' birth and development,25- 32 programs and services,33- 35 patients,18,36- 43 staff or volunteers,44- 47 and care quality.19,35,48,49 A few studies describe the free clinic movement,50- 52 free clinics in general,53 volunteerism,24,54- 56 medical student–run clinics,29,57- 62 and free clinics in one state,27,63,64 region,22 or nation.65- 68 The last census (59 clinics) was conducted in 1967-1969.66
Three national studies54,65,67 and 1 regional study22 are the only attempts in recent decades to characterize the free clinic sector. Each study, however, has substantial limitations. None applies standard criteria to define free clinics, which means that these studies likely describe a mixture of free clinics, low-cost charitable clinics, and federally supported clinics. In addition, all the studies have limited reach; 2 studies54,65 use outdated sources and 1 study54 combines 2 distinct models: free clinics and “free clinics without walls” in which physicians (usually specialists) provide free care in their offices. To address these limitations, I conducted a national survey of all known free clinics in which I examined their structures and operations, funding sources, caseload, staffing, and range of services. I sought to evaluate the extent to which these providers are functioning as a meaningful component of the safety net system.
Data sources included member lists from all national, regional, and state free clinic associations; a mailing list from Volunteers in Health Care (a now-defunct organization that supported free clinics); publicly available directories; Guidestar (http://www2.guidestar.org/), a database of more than 1.5 million Internal Revenue Service–recognized nonprofits; the Medical Student–Run Clinics of America; state primary care associations, area health education centers, and medical schools; the Internet; and survey respondents, who were asked to list free clinics in their communities. This latter snowball sampling technique generated both previously identified and previously unidentified free clinics. Altogether, these disparate sources yielded a list of 2545 potential free clinics.
An organization was operationally defined as a “free clinic” if it met all the following criteria: being a private, nonprofit organization or program component of a nonprofit; providing medical, dental, or mental health services and/or medications directly to patients; serving mostly (>50%) uninsured patients; charging no fees or nominal fees of not more than $20; not billing patients, denying services, or rescheduling appointments if the patient could not pay the requested fee/donation; and not being recognized as a FQHC or Title X family planning clinic. Clinics that received reimbursement from any third party and clinics that used salaried staff were included if the other criteria were met. Clinics that provided only pregnancy testing and/or counseling services, sexually transmitted disease testing, or human immunodeficiency virus testing were excluded. By excluding clinics supported directly by federal programs, setting a specific dollar threshold on what amount is considered “nominal,” and excluding clinics that condition services on payment (regardless of the dollar amount), this study devised a definition of free clinics that draws a clear line between free clinics and other kinds of ambulatory care safety net providers. These criteria yielded a study population of 1188 free clinics.
A 70-item, 12-page questionnaire requested information on operations, patients, services, staff and volunteers, and future plans. More than 50 free clinic experts and practitioners, government officials, foundation staff, academics, and health policymakers commented on draft versions. A revised draft survey was pretested at 23 clinics. The final survey booklet contained mostly closed-ended items.
The survey was administered between October 7, 2005, and December 15, 2006. All clinics were contacted at least twice and some up to 6 times. Except in 6 cases, all correspondence was sent to a named individual (typically the clinic director or medical director). Contacts included: (1) FedEx envelope69 containing the cover letter, survey, and self-addressed stamped envelope; (2) postcard; (3) e-mail/fax; (4) letter with replacement survey and self-addressed stamped envelope; (5) letter with nonmonetary incentive (ie, pen with inscription, “If you’ve seen one free clinic, you’ve seen one free clinic”); and (6) telephone call. Survey methods were approved by the School of Social Service Administration/Chapin Hall Institutional Review Board at The University of Chicago.
I examined clinics' organizational structure and operations, the number and characteristics of their patients, the number and type of services available on-site, the cost of care, and the number and composition of staff and volunteers. Means were computed for continuous variables and frequencies for categorical variables.
Potential unit nonresponse bias was explored using univariate statistics and multivariate logistic regression analyses of clinic founding year, geography, and population size of areas surrounding the clinic. The extent of item nonresponse was investigated by dividing the frequency of item response by the number of eligible respondents. A “don't know” response was treated as missing. Nonresponse rates of 10% or higher were considered “high.”
Of the 1188 surveys mailed, 945 (79.5%) were returned; of these, 764 clinics (80.9%) were determined eligible for the study and 181 (19.2%) were determined ineligible because they did not meet the study criteria. Clinics were excluded for numerous reasons. The most frequent was that the clinic charged more than the $20 maximum fee determined to be nominal (n = 36). In addition, 28 clinics were excluded because they were duplicate entries on the mailing list; 21 because the respondents indicated that the clinic was not a free clinic; 16 because they bill patients; 15 because they serve mostly insured patients; and 13 because they were a FQHC. The remaining 52 exclusions were for various other reasons, such as the clinic was closed, but no single reason accounted for more than 5% of the total number of ineligible clinics. Attempts to reach 15 clinics were unsuccessful; 1 or more mailings were returned undeliverable. The study achieved a 75.9% response rate. Overall, 1007 free clinics are known to exist throughout 49 states and the District of Columbia, with Alaska as the lone exception. In a logistic regression model exploring clinic founding year, geographic region, and population size, no factors were found to be statistically significant predictors of unit nonresponse (results are available from J.S.D.).
Item nonresponse rates exceeding 10% were observed for a small number of items and attributable overwhelmingly to valid “don't know” responses rather than refusals.
Organizational characteristics provide insight into the operational capacity of free clinics to respond to the health care needs of uninsured residents. Free clinics operate under a range of organizational structures (Table 1). Most operate as medical clinics in buildings that are generally rented; a few clinics own their buildings. Most are independent entities. Among affiliated clinics, the most frequently cited affiliation is with a hospital. The mean origination year is 1995; most free clinics have formed since 1990. The mean number of weekly hours that free clinics reported being open to see patients is 18 (median, 11) although weekly hours open varied widely: 215 clinics (28.6%) reported being open 5 or fewer hours per week, and 188 clinics (25.0%) reported being open, on average, 41 hours per week. Clinics are generally open during daytime hours. Most clinics also reported having evening hours. Free clinics reported scheduling appointments (67.3%) and/or allowing walk-in appointments (71.0%). The mean wait time to obtain an appointment for new and established patients is 12 days and 11 days, respectively. These wait times suggest that demand for free care exceeds clinic capacity. Perhaps in response to high demand, more than half of the clinics reported conducting eligibility screening based on insurance status, income, and residency before new patients can qualify to receive services.
Free clinics reported widely ranging operating budgets, with a mean (SD) of $287 810 (624 884) (median, $125 000). Free clinics receive funding from diverse sources, including private charitable donations (90.6% of the clinics), civic groups (66.8%), churches (66.3%), foundations (65.1%), and corporations (55.1%), whereas federal, state, and/or local grants support some of the operating costs for a few free clinics. Overall, 58.7% received no government revenue, and even among the largest clinics (ie, those in the top 25% of annual visits) 43.2% did not report receiving government revenue.
Free clinics serve patients with attributes that impede their access to primary care: uninsured, inability to pay, racial/ethnic minority, limited English proficiency, noncitizenship, and lack of housing (Table 2). These characteristics also increase their risk of poor health outcomes. Free clinics reported serving a mean (SD) of 747.0 (1183.4) new patients per clinic per year and 1796.0 (2872.4) total unduplicated patients. Overall, the 1007 free clinics serve about 1.8 million mostly uninsured patients annually. Free clinics reported providing a mean of 3217.0 (6001.7) medical visits and 825.0 (1367.7) dental visits per clinic per year. Collectively, they are estimated to provide 3.1 million medical visits and nearly 300 000 dental visits annually.
The scope of services available on-site and by referral provides information about the extent to which free clinics are equipped to handle patients' health problems. Clinics were provided a list of 22 types of services and asked to specify whether each service was offered on-site, by referral, or not available. Overall, free clinics provide a fairly limited range of primary care services, reproductive health services, and services for other selected health conditions on-site (Table 3). The mean number of services is 8.4 (median, 8.0). Most free clinics provide medications (86.5%), physical examinations (81.4%), health education (77.4%), chronic disease management (73.2%), and urgent/acute care (62.3%). Clinics open full-time offer the broadest scope of services, with most supplementing the aforementioned services with gynecological care (73.0%), laboratory services (55.8%), case management (56.9%), vision screening (53.5%), and tuberculosis care (51.7%). Except for the 188 full-time clinics (25.0%) that offer comprehensive services, free clinics do not appear to be an appropriate substitute for other comprehensive primary care providers. This is especially true for women because most free clinics do not directly provide any reproductive health services (eg, only 46.2% offer gynecological care). Most free clinics reported offering medications from a dispensary (65.9%) rather than a licensed pharmacy (25.3%), including free samples obtained from pharmaceutical manufacturers (86.8%), pharmaceuticals purchased with the assistance of corporate patient assistance programs (77.3%), direct purchases from manufacturers (54.9%), or outside pharmacies (52.2%). Free clinics reported using individual volunteer health care providers (34.5%); community health care providers such as health centers, health departments, and public hospitals (53.8%); and health care providers from a single hospital or physician group (31.1%) to deliver free services unavailable on-site. Among all responding clinics, the mean annual number of referrals is 362 (median, 118).
The minimum amounts charged by private physicians, health centers, and public clinics are considerably more than the $9.30 mean fee/donation requested by 45.9% of free clinics; 54.1% of free clinics charge nothing (Table 4). The commitment to making free or low-cost health care available extends even to services many free clinics do not themselves offer. For example, most free clinics reported making arrangements for patients to receive free laboratory and radiographic services (80.7% and 63.4%, respectively), although few offered these services on-site (laboratory, 43.9%; radiography, 8.8%).
Free clinics' service capacity can be measured, in part, by who is providing care (Table 5). The status of staff and providers (paid or volunteer) provides insight into the clinic's permanency, potential responsiveness to as-yet-unmet needs, and ability to expand. Nearly all clinics reported that volunteer health care professionals provided some health care services (97.7%). The mean annual number of volunteer hours per clinic was 4237 (median, 2087). This mean equates to 2.4 volunteer hours per patient (including clinical services and administrative functions). Among volunteers, the health care provider type cited most frequently is physician (82.1%), 95.0% of whom are board certified. Free clinics also reported using other volunteer health professionals, including nurses (72.6%) and nurse practitioners/physician assistants (54.9%). There were fewer social workers (25.6%) and psychologists (12.0%) in volunteer positions. More than three-quarters of the clinics reported using paid staff (77.5%), either full-time (54.6%) or part-time (61.1%). Notably, about two-thirds employ a paid executive director (65.8%), and about half pay administrative staff (48.9%).
To my knowledge, this study is the first systematic (ie, definitionally rigorous and sectorally comprehensive) overview of free clinics in 40 years. Its results depart substantially from those of a 2005 national free clinic survey,65 with the most likely explanation being the different methods used in the present study. Unlike the previous survey, the present study used numerous disparate data sources to identify the population of free clinics, applied uniform criteria based on a standard definition to evaluate eligibility, and elicited comprehensive information from 764 clinics based on a census of all known free clinics.
Because they relied on a single source—the National Free Clinic Directory, a directory of self-identified free clinics—to identify their sample (n = 355), Nadkarni and Philbrick's 2005 free clinic survey65 is vulnerable to undercoverage and voluntary selection bias. Because they did not verify the status of the clinics listed in the directory, their results are biased because some clinics that are included among the respondents are not, in fact, free clinics. My review of the directory revealed that 54 of the clinics listed in the source do not meet the definitional criteria used in this study. Some clinics on the list are FQHCs (n = 19); charge more than $20, bill patients, or deny/reschedule care if a patient cannot pay (n = 28); serve mostly insured patients (n = 3); are “free clinics without walls” (n = 1); or are public clinics (n = 3). If all 54 clinics actually participated in the Nadkarni and Philbrick survey, nearly 20% of their sample (281 [19.2%]) would be contaminated with clinics that are not strictly free clinics.
The present description suggests that free clinics are a much more important component of the ambulatory care safety net than generally recognized. For instance, the Institute of Medicine's seminal study on the safety net70 did not mention free clinics. The present results suggest that this is a major oversight in a context where more than 1000 free clinics are estimated to serve 1.8 million mostly uninsured patients and provide more than 3 million medical visits annually. These numbers may be compared with the 6 million uninsured (of 15 million total) served in 2006 by the $1.8 billion federal health center program (http://bphc.hrsa.gov).
Free clinics tend to serve similar patients (mostly uninsured, nonelderly adults; women; and minorities with low incomes) but have diverse organizational structures, operations, scopes of services, and compositions of staff. This diversity suggests that there is a high degree of variability across the sector in terms of individual clinics' capacity to satisfy the basic health care needs of uninsured patients. All clinics rely extensively on volunteer licensed health care professionals to deliver services and on private donations for operating budgets, 2 conditions that impede expansion efforts.
The niche that free clinics occupy in the ambulatory health care safety net may be appreciated more fully by comparing free clinics with health centers, which have been the focus of our health care delivery solutions for the poor and uninsured. Operating with smaller (mostly privately financed) budgets, free clinics provide a more limited scope of services, use mostly volunteers, and charge patients little or nothing (Table 6).
Free clinics suggest an alternative model of primary care to the underserved, and the merits of the free clinic model ought to be discussed as viable options to serve the uninsured. The limited data about the quality of care provided by free clinics constrains the debate, but it is worth noting that a recent study of a volunteer-based free clinic that is open just 2 nights per week documented clinically meaningful improvements in chronic disease outcomes after adopting the chronic care model.49
Of interest, study findings challenge the belief that free clinics would be phased out if comprehensive health insurance reform were enacted.67 This belief rests on several assumptions: that there would be no coverage gaps; that free clinics are “temporary”; that patients, if given a choice, would not choose free clinics; and that free clinics would not be interested in participating in third-party programs. This study suggests otherwise. Throughout their history, free clinics have served as gap-fillers, targeting patients who are underserved by mainstream medicine. They also focus on providing services less readily available elsewhere, such as medications, eyeglasses, and health education. Hundreds of free clinics have existed for a decade or more, employ staff, own their premises, manage a budget of more than $750 000, and serve thousands of patients annually. In response to universal insurance, it would seem likely that these clinics would adapt rather than close. There is some evidence to suggest that newly insured patients would return to free clinics. Although most free clinics currently exclude the insured, a mean of 7.5% of free clinic patients actually do have insurance. Therefore, these patients probably have alternative sources of care and nevertheless select free clinics. In addition, the survey data suggest that free clinics may be interested in participating in third-party insurance programs once their insured caseload is large enough to warrant participation. For example, in the highest-volume clinics (top 10% of annual visits), 8 (13.1%) reported billing insurers vs only 1 clinic in the bottom 10%.
Some study limitations should be acknowledged. Sampling bias could have occurred if existing free clinics were not included in the cases examined, but this likelihood was reduced because the population was surveyed. In addition, clinics were identified using numerous disparate sources. Nevertheless, the smallest, youngest, and least formal free clinics were more likely to be excluded.
To reduce item nonresponse bias and measurement error, rigorous pretest procedures were adopted to extract poorly worded questions or response categories. Partially closed responses were used to lessen the possibility of systematic biases resulting from loaded questions.
The cross-sectional design can suggest associations between variables, but cannot establish causality. Also, only free clinics that were currently operating were described.
The findings should stimulate further academic inquiry. Research is needed to assess changes in the sector over time, care quality, and the reasons why patients choose (or end up in) free clinics. Longitudinal data collection and analysis should be a top research priority.
Free clinics provide a range of preventive and general medical care for an estimated 10% of the working-age adult uninsured population who seek care.72,73 In light of free clinics' population reach, service limits, and staffing and financial constraints coupled with their extensive practice of making referrals and collaborating with safety net providers for diagnostic services and specialty care, a prudent next step would be to establish federal or state demonstration programs to promote and evaluate collaborations between free clinics and other safety net providers. Any new demonstration program must be designed to avoid the pitfalls of the Healthy Community Access Program, a federal demonstration program that aimed to improve access to care through coordinated delivery systems but, ultimately, was judged “ineffective,” largely because of its unclear purpose, poor design, and lack of accountability.74 Free clinics have passed the point in history when they can exist below the radar. At the same time, policymakers and other safety net providers must acknowledge the important role that free clinics play. Formal integration of free clinics into the safety net has the potential to strengthen the overall health system, which is important regardless of the outcome of the national health reform debate.
Correspondence: Julie S. Darnell, PhD, MHSA, Division of Health Policy and Administration, School of Public Health, University of Illinois at Chicago, 1603 W Taylor St, Room 754, M/C 923, Chicago, IL 60612 (email@example.com).
Accepted for Publication: December 4, 2009.
Financial Disclosure: None reported.
Funding/Support: This research was supported by dissertation grant R36 HS15555-01 from the Agency for Healthcare Research and Quality. Supplemental funding for the study was provided by contracts 03-1062-919 and 06-1067-200 awarded by the Kaiser Commission on Medicaid and the Uninsured and contract 20051854 from the California Endowment and fellowships awarded by the Society for Social Work and Research, the National Association of Social Workers, and the Chicago Center of Excellence in Health Promotion Economics at The University of Chicago. Preliminary work on the study was supported by research training grant 5 T32 HS000084 from the Agency for Healthcare Research and Quality and a fellowship provided by the School of Social Service Administration at The University of Chicago.
Role of the Sponsors: The Kaiser Commission on Medicaid and the Uninsured and the California Endowment provided feedback on the draft survey. Otherwise, there was no involvement from the sponsors or funders in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
Additional Contributions: I thank the free clinic staff and volunteers who generously completed the surveys. I thank my dissertation committee: Michael R. Sosin, PhD (chair), and Willard G. Manning, PhD, The University of Chicago; Edward F. Lawlor, PhD, and Sarah Gehlert, PhD, of Washington University, St Louis, Missouri; and Laurence E. Lynn Jr, PhD, of The University of Texas at Austin. In addition, I thank Charles L. Bennett, MD, PhD, MPP, of Northwestern University, Evanston, Illinois, and Jack Zwanziger, PhD, of The University of Illinois at Chicago for providing comments on draft versions of the manuscript and David Jemielity, MPhil, for his assistance with editing.