A substantial proportion of treatment for mental and substance use disorders is in the private sector, with private insurance accounting for about 24% of mental disorder expenditures and 10% of substance use disorder expenditures in 2003 (1). With the recent federal parity legislation and health care reform, the role of private insurance may increase in the future. Under the parity law, plans that cover behavioral services must provide coverage equivalent to that of general medical services. Under health reform, the essential benefits package required for plans in the insurance exchange includes behavioral health and thus must also meet parity requirements.

Treatment for mental and substance use disorders under private insurance is carved out by 72% of private health plans to specialty managed behavioral health care organizations (MBHOs) (2). MBHOs maintain networks of approved or credentialed treatment practitioners who provide office-based treatment ("in-network practitioners"). Health plans select networks to ensure competence and high-quality care and to make available a variety of practitioners in terms of demographic characteristics, geographic location, and skills (3,4). In addition, networks allow plans to negotiate lower fees with practitioners in exchange for patient volume. Using their insurance benefit, patients may seek treatment from in-network practitioners. In some plans patients may see an out-of-network practitioner at a higher cost compared with seeing one who participates in the network, and of course all patients may obtain treatment without using their insurance benefit, paying wholly out of pocket.

Since managed behavioral health care first became common in the 1990s, there has been concern about its impact on practitioners (3,5,6,7,8). Some have suggested that managed care would favor practitioners who charge lower fees or who have less training or different skills—choosing social workers and counselors over psychiatrists and psychologists, for example (3,7,9). With the rise in the use of prescription medications for treating mental disorders, studies predicted that psychiatrists in particular would be limited to a role of medication management (9,10). Despite these long-standing assumptions about managed behavioral health care, little is known, other than types of practitioner (9,11,12), about the practitioners in MBHO networks who treat mental and substance use disorders for privately insured patients. Several studies have specifically examined the changing role of psychiatrists (10,13,14), yet we know less about other practitioners.

The question of who is providing care is relevant to understanding behavioral health treatment in the private sector. Staffing mix (types of practitioners, as well as specialization) is important to consider when aiming to improve efficiency without sacrificing quality of care (15), and it is increasingly important as pay-for-performance initiatives are making their way into behavioral health (16). In this study we aimed to address several gaps in the literature by considering the types of outpatient care that are provided by different types of in-network practitioners in a national MBHO and further, how practitioner types and expertise are related to patient diagnoses of mental disorder, substance use disorder, or both.

We analyzed 2004 data from a large national MBHO. Data included in-network practitioner credentialing data, which were based on information practitioners provided when applying for MBHO network inclusion; enrollment files for patients with behavioral health insurance through this MBHO; and claims files. Analyses included all in-network practitioners who submitted at least one claim in 2004 (N=28,897) and all patients treated by them who had at least one outpatient behavioral health visit in 2004 (N=132,466), representing 844,071 outpatient claims. Statistical significance was determined by chi square tests and t tests. The study received institutional review board approval.

Practitioners were coded into mutually exclusive types: psychiatrist, psychologist, licensed social worker, advanced practice nurse, or master's-level and licensed counselor. A few practitioners with multiple credentials were coded by their highest degree. Specialized training in treatment of major psychiatric disorders or specialized alcohol-drug training was ascertained by the item, "Please check the areas in which you have at least 1,500 hours of training and experience and wish to provide services. Your experience should be evidenced in your work history." On the basis of primary and secondary diagnoses on their 2004 claims, each patient was categorized as having a mental disorder (N=128,173), a substance use disorder (N=1,565), or co-occurring mental and substance use disorders (N=2,728).

The MBHO had 28,897 in-network practitioners who saw patients with mental or substance use disorders in 2004. The network was well represented by all practitioner types except nurses (1.9%, N=559). Psychologists (N=9,191) and social workers (N=8,248) each accounted for about 30% of the network, and psychiatrists (N=5,703) and counselors (N=5,184) each accounted for nearly 20%. About half (N=14,513) reported specialized training specifically for treatment of major psychiatric disorders, and just over one-third (N=10,064) had training for treatment of alcohol and drug problems. Nurses were least likely to have specialized alcohol-drug training (25.9%, N=145), whereas social workers (37.1%, N=3,054) and counselors (36.8%, N=1,905) were most likely (p<.001). Psychiatrists were most likely to report having specialized training in major psychiatric disorders (70.5%, N=4,019), and counselors were least likely to do so (30.8%, N=1,596, p<.001).

Network practitioners saw a mean±SD of 5.6±10.1 patients from this MBHO on average during 2004, with an average of 7.3±6.6 visits per patient. These caseload characteristics ranged widely, with a maximum of 334 patients per practitioner in 2004 and a maximum of 128 visits per patient. Psychiatrists had more patients (9.8±18.3) and fewer visits (4.9±4.8) than other practitioner types (p<.001 for visits to all practitioners except nurses, for which the comparison was not significant).

Eighty-four percent (N=709,079) of claims were for individual therapy, of which 7.5% (N=63,357) included medication management and 76.5% (N=645,722) did not. Only 16.1% (N=135,886) of total claims included medication management. As expected, medication management claims were almost exclusively submitted by psychiatrists and nurses, and these visits made up 62.5% (N=128,946) of claims by psychiatrists and 43.9% (N=5,972) of claims by nurses. However, a majority of visits for both psychiatrists and nurses involved some form of psychotherapy (67.2%, N=138,550, and 69.6%, N=9,465, respectively), with or without medication management. Individual sessions constituted the bulk of the work performed by psychologists, social workers, and counselors. Group sessions accounted for very few practitioner visits. Family sessions were rarely provided by psychiatrists or nurses. [A table showing types of outpatient visits by practitioner type is available as an online supplement to this column at ps.psychiatryonline.org.]

These privately insured patients were most likely to have a visit with a psychiatrist (38.3%, N=50,783) or a psychologist (34.9%, N=46,382) compared with other types of practitioner. When stratified by patient type, patients with co-occurring mental and substance use disorders (65.2%, N=1,779) were more likely than those with a mental disorder (37.7%, N=48,321) or those with a substance use disorder (40.1%, N=628) to have a psychiatrist claim (p<.001). Patients with a substance use disorder were less likely to see a psychologist (19.2%, N=300) than were patients with co-occurring disorders (30.0%, N=818) or those with a mental disorder only (35.2%, N=45,117) (p<.001). Furthermore, patients with co-occurring disorders were more likely to see psychiatrists than all other practitioner types, and patients with a mental disorder were most likely to see psychiatrists and psychologists.

As expected, patients with a substance use disorder or co-occurring disorders were more likely to see practitioners who had specialized training for alcohol-drug treatment (67.8%, N=1,061, and 64.7%, N=1,765, respectively) than were patients with a mental disorder (43.4%, N=55,627, p<.001). Patients with a mental disorder or co-occurring disorders were more likely to see practitioners indicating specialized training in major psychiatric disorders (mental disorder, 61.7%, N=79,083; co-occurring disorders, 74.0%, N=2,019; versus substance use disorder only, 56.5%, N=884) (p<.001).

In this population, the overwhelming majority of claims were for individual counseling, with or without medication management, and only 16.1% of all claims involved medication management. Although two-thirds of claims for psychiatrists involved medication management, we also found that two-thirds of their claims involved some form of psychotherapy (an overlap of medication management and individual psychotherapy of about 30%). Despite potential financial disincentives for psychiatrists to conduct psychotherapy (17), our findings show that billed claims for psychotherapy by psychiatrists were common. The proportion of medication management visits in this study were lower than found in a 2000 study, where 26% of patients of sampled psychiatrists received no psychotherapy, indicating that visits were for medication (14).

Practitioners saw an average of 5.6 unique patients in this MBHO in 2004, but practitioners contract with many MBHOs; thus this figure is unlikely to represent their full caseload (18). On average, these practitioners had seven claims per patient, a treatment length comparable with national estimates of authorizations and visits (12,19). The range of visits per patient indicates some flexibility in authorized and covered patient visits. The number of visits per patient varied by practitioner type, and we found that psychiatrists had the fewest visits per patient.

Whether a patient had only mental disorder claims, only substance use disorder claims, or claims for both was related to practitioner type and expertise. One would expect patients with mental disorders to be most likely to receive medications, which was suggested by the finding of greater use of psychiatrists by patients with mental disorders. Patients with both mental and substance use disorder claims would be expected to have more complicated cases, and our findings suggest that. Patients with co-occurring disorders were vastly more likely to see a psychiatrist—similar to high-severity cases in the National Comorbidity Survey Replication household survey (20)—and saw practitioners with specialized training in both areas. At the same time, one-third of patients with substance use disorders saw practitioners who did not report specialized training in alcohol or drug use disorders. Further distinguishing diagnoses of serious mental illness might identify additional patterns.

Although this study contributes to our understanding of practitioners and treatment in private MBHOs, we could not examine the specific content or quality of treatment offered or out-of-network care, and little was known about the patients themselves, such as severity of illness. Specialized training was self-reported by the practitioner. Broader access-to-care issues, such as geographic access or treatment authorization, are also key to consider in future analyses. These data stemmed from only one MBHO, but it has national distribution; most providers are in multiple networks, so these characteristics may be mirrored elsewhere.

On the whole, the MBHO network we studied appears to represent a range of practitioner types, with most outpatient utilization for psychotherapy, instead of only medication management. The patterns shown for more complex patients indicated use of appropriately qualified practitioners with substantial experience on average. Finally, it appears that the broader skill set of psychiatrists is still being tapped, with provision of both medication management and psychotherapy.

This study was funded by grant P50-DA010233 from the National Institute on Drug Abuse. The authors thank Galina Zolotusky, M.S., for statistical programming; they also thank the managed behavioral health care organization that provided access to the data.

The authors report no competing interests.