The Centers for Disease Control and Prevention¹ estimates that a quarter of adults in the United States report having a mental illness at any given time and about half will experience mental illness during their lifetime. In the wake of the Connecticut school shooting and other recent mass shootings, policy makers and the public have called for increased access to mental health services.² For example, President Obama’s “Now Is the Time” proposal, released in January 2013, called for better mental health services, including programs to identify diagnosable mental health problems early so that patients can be referred for treatment, and increased training of mental health professionals.²

Psychiatrists play an important role in the diagnosis and treatment of patients with mental illnesses particularly because of their training and ability to prescribe medications.³ One issue that advocates for increased mental health access have neglected to explore is limited access owing to psychiatrists’ refusal to accept insurance.⁴ In previous studies, we and others have shown that overall physician acceptance rates for private noncapitated insurance was high but declining modestly in recent years.^5- 7 Little is known about specialty differences in insurance acceptance rates, but prior reports suggest that nonacceptance of insurance may be particularly high for psychiatrists.

To assess this potential problem, we used a national survey of office-based physicians to answer 2 research questions: (1) what are recent trends in acceptance of insurance by psychiatrists compared with physicians in other specialties? and (2) what are the characteristics of psychiatrists who do not accept insurance? Nonacceptance of insurance may have implications on access to care for patients with mental illness and may limit the ability of policy makers to implement proposals to improve access for patients. Understanding the characteristics and predictors of psychiatrists who do not accept insurance may also help the medical community understand where and why this might be a particular problem.

The National Ambulatory Medical Care Survey (NAMCS) is a nationally representative survey administered by the Centers for Disease Control and Prevention’s National Center for Health Statistics. It contains information about physicians practicing in non–federally funded, non–hospital-based offices throughout the United States.^8,9 Physicians in the fields of anesthesiology, radiology, and pathology are excluded. Approximately 90% of outpatient visits in the United States involve the office-based physicians represented in the NAMCS.

The NAMCS uses a multistage sampling design based on geographic primary sampling units and physicians practicing within those units. Physicians are identified from the American Medical Association’s and the American Osteopathic Association’s master files. Each physician responder is then weighted based on the sampling design so the data can be used for national estimates.

The NAMCS collects information on physician and practice characteristics in a telephone survey. For this study, we used data from the 2005 to 2010 physician surveys.

We restricted our sample to physicians who reported that they accepted new patients. Among all physicians, 95.3% accepted new patients across the study years. This percentage was lower among psychiatrists (87.7%). For our analysis of physician acceptance of Medicare, we excluded pediatricians because they rarely care for Medicare patients. We included pediatricians in all other analyses.

Our main outcome variables were physician acceptance of new patients with private noncapitated insurance, Medicare, or Medicaid. The NAMCS collects acceptance rates of private capitated insurance, but we did not include this in our analysis because only a very small proportion of psychiatrists are paid through capitated arrangements. Acceptance rates (by physicians of all specialties) for new patients with capitated insurance are substantially lower than for those with noncapitated insurance (55.9% vs 88.9%).

Our main independent variables were physician specialty and 2-year groupings (2005-2006, 2007-2008, and 2009-2010). In our comparison of psychiatrists who accept private noncapitated insurance, Medicare, and Medicaid vs those who do not, we looked at practice region (Northeast, Midwest, South, and West), and practice size (solo or group).

We report percentages of physicians who accepted new patients by insurance type and by 2-year grouping. All percentages are weighted at the physician level to reflect the sampling design of the NAMCS so that national estimates can be reliably calculated. We used ordinary least squares regression to estimate linear trends over time in physician acceptance of new patients by insurance type and by physician specialty (psychiatrists vs all other specialties) and to test for differences in the slope of these 2 trend lines. For 2009 and 2010, we report the percentages of physicians who accepted new patients with each of the 3 types of insurance (private noncapitated, Medicare, and Medicaid) for 15 specialties.

We used the Pearson χ² test to compare acceptance rates for psychiatrists with those for physicians of all other specialties in 2009 and 2010 and to compare characteristics between psychiatrists who accept each insurance type and those who do not. To test for independent associations between practice characteristics and psychiatrist acceptance of insurance, we used multivariable logistic regression with insurance acceptance as the dependent variable and practice region and size as the predictor variables. Analyses were performed using Stata statistical software, version 11.0 (StataCorp).

The Weill Cornell Medical College Institutional Review Board approved this study.

The mean number of physicians surveyed from each year between 2005 and 2010 was 1250 (range, 1058-1357 physicians); 5.5% of these physicians were psychiatrists.

The percentage of psychiatrists who accepted private noncapitated insurance was lower than that for other physicians in all years and decreased by 17.0%, from 72.3% (95% CI, 61.9%-80.7%) in 2005-2006 to 55.3% (95% CI, 46.6%-63.8%) in 2009-2010 (P = .01 for trend across years; Figure 1A). The percentage of physicians in other specialties who accepted private noncapitated insurance decreased by 4.4%, from 93.1% (95% CI, 91.5%-94.5%) in 2005-2006 to 88.7% (86.4%-90.7%) in 2009-2010 (P < .001 for trend across years). The slopes of the trends for psychiatrists and other physicians were significantly different (P = .002).

The percentage of psychiatrists who accepted Medicare was lower than that for other physicians in all years and decreased by 19.5%, from 74.3% (95% CI, 64.6%-82.0%) in 2005-2006 to 54.8% (46.6%-62.7%) in 2009-2010 (P < .001 for trend across years; Figure 1B). The percentage of other physicians (excluding pediatricians) who accepted Medicare was unchanged over the years (87.8% [95% CI, 86.3%-84.4%] in 2005-2006 and 86.1% [84.4%-87.7%] in 2009-2010; P = .14 for trend across years). The slopes of the trends for psychiatrists and other physicians were significantly different (P < .001).

Psychiatrists’ Medicaid acceptance rates were lower than those for other physicians across all years (Figure 1C), but these rates did not decline significantly from 2005-2006 to 2009-2010 for psychiatrists (49.3% [95% CI, 41.2%-57.5%] in 2005-2006 and 43.1% [34.9%-51.7%] in 2009-2010; P = .29) or other physicians (75.1% [75.1%-77.7%] in 2005 and 73.0% [70.3%-75.5%] in 2010; P = .21).

In 2009-2010, 55.3% (95% CI, 46.7%-63.7%) of psychiatrists accepted private noncapitated insurance vs a mean of 88.7% (86.4%-90.7%; P < .01) (Figure 2) for physicians of all other specialties. The specialty with the highest acceptance rate for private noncapitated insurance was cardiology (93.6%; 95% CI, 88.3%-96.6%). Among general internists, who like psychiatrists, primarily provide evaluation and management services, 91.3% (95% CI, 84.0%-95.5%) accepted private noncapitated insurance.

During that same period, 54.8% (95% CI, 46.6%-62.7%) of psychiatrists accepted Medicare, compared with a mean of 86.1% (84.4%-87.7%) for other physicians. Again, the specialty with the highest rate of Medicare acceptance was cardiology (99.4%; 95% CI, 95.6%-99.9%). Among general internists, 95.0% (89.7%-97.6%) accepted Medicare.

Psychiatrists had a lower rate of Medicaid acceptance (43.1%; 95% CI, 34.9%-51.7%) than general internists (60.9%; 51.0%-69.9%).

Psychiatrists in the Midwest were more likely to accept private noncapitated insurance (85.1%; 95% CI, 59.5%-95.7%) than those in the Northeast (48.5%; 33.8%-63.4%), South (43.0%; 29.1%-58.1%), or West (57.8%; 38.3%-75.1%; P = .02) (Table 1). This regional difference was not seen for Medicare or Medicaid acceptance.

More psychiatrists than other physicians practiced in solo practices (60.1% [95% CI, 51.7%-68.0%] vs 33.1% [30.3%-36.1%]). Psychiatrists in solo practice were less likely to accept all types of insurance (private fee-for-service plans, 43.0% [95% CI, 32.7%-54.0%] of solo vs 74.9% [58.5%-86.4%] of group practitioners [P = .002]; Medicare, 45.0% [34.3%-56.3%] vs 69.5% [55.6%-80.6%], respectively [P = .01]; and Medicaid, 26.8% [18.1%-37.9%] vs 67.3% [53.3%-78.8%], respectively [P < .001]).

In multivariable analysis, we found that a practice in the Midwest was independently associated with higher rates of acceptance of private noncapitated insurance (adjusted odds ratio, 4.91; 95% CI, 1.11-21.62) (Table 2). This association was not seen for other regions or for other types of insurance. We also found that practicing in a group practice was independently associated with higher rates of acceptance for private noncapitated insurance (adjusted odds ratio, 3.75; 95% CI, 1.48-9.51), Medicare (adjusted odds ratio, 3.03; 95% CI, 1.34-6.85), and Medicaid (adjusted odds ratio, 6.55; 95% CI, 2.98-14.44).

In this analysis of a national survey of office-based physicians, acceptance rates for all types of insurance were significantly lower for psychiatrists than for physicians of other specialties. For private noncapitated insurance and for Medicare, acceptance rates by psychiatrists have dropped since 2005. In 2009-2010, almost half of all psychiatrists did not accept private noncapitated insurance, and more than half did not accept Medicare or Medicaid. To our knowledge, no prior studies have documented such a striking difference in insurance acceptance rates between psychiatrists and physicians in other specialties. These low rates of acceptance may affect recent calls for increased access to mental health services, and if the trend of declining acceptance rates continues then the impact may be even more significant.

Low reimbursement has been cited as a reason why physicians do not accept insurance.^6,10,11 However, reimbursement rates for office-based psychiatric treatment are similar to those for office-based medical evaluation and management, such as primary care services.¹² What may be unique to psychiatry is the time it takes to provide counseling and therapy. Primary care physicians probably can see and provide management to patients in shorter visits (eg, 10 or 20 minutes) than psychiatrists, especially if psychiatrists want to provide psychotherapy along with medication management. As a result, psychiatrists may not be able to see as many patients in a day as physicians in other specialties.Psychiatrists who want to provide psychotherapy may opt not to accept insurance.

A shortage of psychiatrists may also be a potential reason why many do not accept insurance. Faulkner et al¹³ recently reported a 14% decline in the number of graduates from psychiatry training programs during an 8-year period from 2000 to 2008. These declines coupled with an aging workforce (55% of psychiatrists are aged 55 or older) may mean that the supply of psychiatrists cannot meet the demand for their care.¹⁴ As a result, many psychiatrists may have so much demand for their services that they do not need to accept insurance.

Our data show that most office-based psychiatrists practice in solo practices and that solo practice was independently associated with nonacceptance of insurance. Solo practices often can function with much less infrastructure than larger single-specialty or multispecialty group practices. As a result, they may have little incentive to hire staff to interact with insurance companies.

Unfortunately, the NAMCS does not ask physicians why they do not accept insurance, so all of these possible explanations for our findings are hypothetical. Another limitation of our analysis is that the NAMCS does not collect specialty information or insurance acceptance information on physicians practicing in hospital outpatient departments. Nevertheless, the physicians surveyed in the NAMCS represent the physicians that see approximately 90% of outpatient visits in the United States, and the NAMCS includes information on physicians practicing in mental health centers and federally qualified health centers. Finally, the NAMCS surveys specialists based on the proportion of physicians practicing in that specialty. Although the sample of physicians is weighted to reflect these proportions, the sample of psychiatrists in the NAMCS is relatively small compared with other specialties (eg, family practice and internal medicine). As a result, the confidence intervals for each point estimate are relatively wide.

It is unknown whether low acceptance rates of insurance by psychiatrists affect access to timely care for patients with mental health problems. Many mental health conditions, such as depression and anxiety, are often managed by patients’ primary care physicians; limits of insurance acceptance may affect patients who have more complicated psychiatric diagnoses than a primary care physician can manage. In addition, many psychiatrists who do not accept insurance still provide bills to their patients who can submit them for out-of-network reimbursement. However, out-of-network benefits often cover only a portion of the bill at the usual and customary rate; patients generally have higher out-of-pocket costs when they see out-of-network physicians.^15,16 More research is needed to understand whether low rates of insurance acceptance are a barrier that patients face when trying to seek psychiatric care and a barrier that physicians face when trying to refer their patients for psychiatric care.

Nonetheless, our findings suggest that policies to improve access to timely psychiatric care may be limited because many psychiatrists do not accept insurance. If, in fact, future work shows that psychiatrists do not take insurance because of low reimbursement, unbalanced supply and demand, and/or administrative hurdles, policy makers, payers, and the medical community should explore ways to overcome these obstacles.

One approach is to increase the supply of psychiatrists with policies that encourage students to pursue psychiatry (eg, loan forgiveness program) and incentives for nonphysician providers (eg, nurse practitioners) to pursue areas that focus on mental health, although these solutions will not immediately solve the supply shortage.

Another area that may offer lessons for improving access to psychiatry is the policy, payment, and medical community’s response to the primary care shortage. Facing this looming shortage, Medicare and Medicaid have increased payment for primary care and are incentivizing the implementation of new delivery models.^17- 19 These system redesigns, such as accountable care organizations and patient-centered medical homes, may provide the financial incentives for innovations (eg, paying psychiatrists to provide telephone or electronic consultations to primary care physicians) and more and better use of nonphysician staff (eg, social workers and case managers).^20- 22 Although these innovations may not necessarily increase the level of insurance acceptance, they may ensure access to care for patients with mental illness.