Treatment of Obsessive-Compulsive Disorder in a Nationwide Survey of Office-Based Physician Practice

This study was conducted between July 2012 and August 2013, and institutional review board approval was obtained. The NAMCS is conducted annually by the National Center for Health Statistics (5). It samples a nationally representative group of visits to non–federally employed office-based physicians who are primarily engaged in direct patient care. The annual NAMCS physician target sample includes all physicians in databases maintained by the American Medical Association and the American Osteopathic Association. These physicians represent an array of medical specialties (that is, both primary and specialty care). Visits to other health and mental health care providers are not included in the survey.

The NAMCS uses a multistage probability sample design involving primary sampling units (a county, a group of adjacent counties, or a standard metropolitan statistical area), physician practices within primary sampling units, and patient visits within physician practices. During one week, attending physicians or office staff complete a one-page form that documents demographic, clinical, and treatment characteristics for selected patient visits. Each visit is weighted by sampling probability, adjustment for nonresponse, physician specialty, and geographic location to obtain nationally representative estimates. Following National Center for Health Statistics recommendations, we combined data from contiguous survey years 2003 to 2010 to establish a larger base from which to derive more stable estimates. Response rates across survey years varied from 58.3% to 66.9% (median=61.6%).

Diagnoses (recorded in three data fields as “primary” and the last two fields as “other”) were made by the treating physicians according to ICD-9-CM. We limited the sample to outpatient visits in which either a primary diagnosis of OCD (ICD-9-CM code 300.3) was assigned or OCD was assigned as a secondary or tertiary diagnosis and the other diagnoses were an anxiety disorder other than posttraumatic stress disorder, a mood disorder other than bipolar disorder, an adjustment disorder, or a general medical disorder. Visits by patients with a comorbid primary diagnosis of a mood disorder and other anxiety disorders were included in our sample because these commonly co-occur with OCD (6) and have first-line treatments similar to those for OCD (that is, antidepressants or CBT). Visits by patients with comorbid primary diagnoses of schizophrenia, bipolar disorder, and other psychotic disorders were excluded because first-line treatments for these disorders are different from those for OCD.

Data were collected on patient age, sex, race, and ethnicity. Visits were also classified according to whether the physician had seen the patient before. Data regarding sources of payment for the visit were collapsed into four non–mutually exclusive categories: public insurance, private insurance, self-pay, and a residual category “other” that included no charge, uncompensated care, workers' compensation, and unknown payment source. Specialty of physicians was classified as psychiatry, primary care (general, family, and internal medicine), or other specialty.

Visits that included psychotropic medications were classified into six medication groups: SRIs, other antidepressants, benzodiazepines, antipsychotics, mood stabilizers, and stimulants. Psychotherapy visits included psychotherapy, mental health counseling, or social problem counseling. Visit duration was recorded in minutes and included only time spent in face-to-face contact with the patient. A study comparing NAMCS ratings with direct observation found high specificity; however, NAMCS ratings moderately overestimated the time spent with patients (7). To account for this tendency in recording visit duration, we analyzed only those psychotherapy visits that were 45 minutes or longer (7).

Sample means, standard errors, and 95% confidence intervals (CIs) were calculated for characteristics of persons in treatment for OCD. We used STATA, version 12.0, to accommodate the complex sampling design and weights from NAMCS. Sample means and CIs are reported averaged over the study period (2003–2010). Logistic regression analyses were used to determine predictors of treatment with SRIs or any psychotherapy (lasting more than 45 minutes) of an unspecified type.

Between 2003 and 2010, a total of 316 patient visits with a diagnosis of OCD were sampled in the NAMCS, representing approximately 728,644 annual office visits (619,871 to a psychiatrist, 79,990 to a primary care physician, and 28,783 to another physician specialty). Table 1 shows visit characteristics. Most of the visits (unweighted N=208, 66%) were for a primary diagnosis of OCD; 30% (unweighted N=96) were for a primary diagnosis of a mood, adjustment, or anxiety disorder; and 4% (unweighted N=12) were for a general medical diagnosis. Most visits were by patients who had been previously seen by the same physician six or more times in the previous 12 months. Psychotropic medications were prescribed in most visits, primarily SRIs, followed by benzodiazepines, antipsychotics, stimulants, and mood stabilizers. Thirty-nine percent of the visits included psychotherapy lasting more than 45 minutes of an unspecified type.

In a sensitivity analysis, we examined whether receipt of psychotherapy or SRI medications varied by age, gender, race-ethnicity, payment type, and physician specialty (psychiatrist versus other specialty). Persons who self-paid for treatment had higher odds of receiving psychotherapy, compared with persons with private insurance (OR=2.84, CI=1.28–6.27, p=.01). Persons who saw a psychiatrist had higher odds of receiving psychotherapy, compared with persons who saw physicians in other specialties (OR=113.23, CI=13.22–970.00, p<.001). Persons age 36–50 had higher odds of receiving psychotherapy than persons age 18–35 (OR=2.19, CI=1.13–4.27, p=.02). No differences were found in receipt of psychotherapy by gender or race-ethnicity. No differences were found in receipt of SRI medications by age, gender, race-ethnicity, payment, or physician type.

This is the first national study examining OCD treatment in office-based medical practice. Even though major U.S. epidemiological studies have found similar prevalence rates of OCD among non-Hispanic blacks and whites (8,9), we found a large imbalance in the ethnic and racial distribution of OCD visits: 91% were by non-Hispanic white patients, 1% by non-Hispanic black patients, 3% by Hispanic patients, and 4% by other non-Hispanic patients. This imbalance has also been reported in OCD treatment and clinical trials (10). More research is needed to identify barriers to outpatient treatment for members of racial-ethnic minority groups who have OCD, such as a lack of knowledge about OCD, mistrust of providers, and lack of proximity to specialized treatment centers (11), and to develop strategies for combating these treatment disparities.

Consistent with previous studies in more specialized settings (4), individuals with OCD who received office-based medical treatment were more commonly treated with medications than with psychotherapy. We found that those who self-paid for treatment or saw a psychiatrist had higher odds of receiving psychotherapy, suggesting that financial barriers (11) and physician specialty play a role. Given the psychotherapy training that psychiatrists receive, we were not surprised to find that medical specialty was associated with the treatment received. We also found that patients with OCD were relatively high-intensity users of mental health services and that OCD was largely treated by psychiatrists in office-based medical practice, even though studies in primary care settings estimate a 12-month prevalence of OCD between 1.9% and 2.2% (12) and SRIs are a first-line treatment for OCD that are commonly prescribed in primary care (13).

This study had several limitations related to the source of the data. First, the sample was restricted to office-based visits and did not include visits to nonphysician mental health care professionals, who likely account for a substantial proportion of psychotherapy for OCD. Second, NAMCS data are cross-sectional, and data on previous treatments, treatment response, and specific type of evidence-based psychotherapy (such as CBT) were not collected, which is common in large administrative data sets (14). Third, the level of analysis in the NAMCS is the visit rather than the individual patient, which may affect clinical interpretation of visit distributions if the number of visits per treatment episode varies across patient characteristics. Fourth, incomplete response to the survey by eligible physicians opens the potential for selection effects that may have biased the reported estimates. Last, diagnoses in the NAMCS are based on the independent judgment of the clinician and are not subject to expert validation.

In broad terms, treatment of OCD in office-based medical practice mainly comprises SRI treatment by psychiatrists. Even though the prevalence of OCD in primary care is estimated to range up to 2.2% and primary care physicians commonly prescribe SRIs, OCD treatment by nonpsychiatrists was uncommon. Findings from this study represent an assessment of the treatment of OCD before health care reform under the Affordable Care Act (ACA). It is likely that practice patterns will change under the ACA, given estimates that by 2019 the ACA will extend health insurance to nearly four million previously uninsured individuals with severe mental disorders. Given the potential shift to the delivery of more mental health care by primary care physicians and nurse practitioners (15), efforts to better understand the recognition and treatment of OCD in primary care are warranted.

This study was funded by the New York State Office of Mental Health Policy Scholar Award program and by grant K24 MH091555 from the National Institute of Mental Health.

Dr. Simpson has received research funds for clinical trials from Janssen Pharmaceuticals and Transcept Pharmaceuticals, consultation fees from Quintiles, Inc., and book royalties from Cambridge University Press and Up To Date, Inc. The other authors report no competing interests.