Current issues of ACP Journal Club are published in Annals of Internal Medicine


PRIME-MD identified mental and eating disorders in primary care

ACP J Club. 1995 May-June;122:73. doi:10.7326/ACPJC-1995-122-3-073

Source Citation

Spitzer RL, Williams JB, Kroenke K, et al. Utility of a new procedure for diagnosing mental disorders in primary care. The PRIME-MD 1000 Study. JAMA. 1994 Dec 14;272: 1749-56.



To evaluate PRIME-MD (Primary Care Evaluation of Mental Disorders) for diagnosing mental (mood, anxiety, somatoform, and alcohol-related) and eating disorders by primary care physicians (PCPs).


Blinded comparison of diagnoses by PCPs using the PRIME-MD questionnaires and by mental health professionals using semi-structured telephone interviews.


4 primary care clinics (31 PCPs).


Of 1360 patients who came to the clinical centers for medical care, 1000 (74%) were eligible (mean age, 55 y; 60% women; 58% white). Exclusion criteria included refusal; inability to comply because of dementia, language, or severity of illness; or age < 16 years.

Description of Test and Diagnostic Standard

PRIME-MD includes a 1-page patient questionnaire of 26 yes-no questions and an assessment of over-all health. The questions assess symptoms and signs in the previous month that have been divided into 5 categories (mood, anxiety, somatoform, alcohol-related, and eating disorders). The PCP interviewed participants on the basis of positive questionnaire responses using a 12-page Clinician Evaluation Guide. Criteria for diagnosis were based on the Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition (DSM-III-R),and the CAGE questionnaire for alcohol dependence. Mental health professionals used semi-structured telephone interviews to assess psychopathology and, after blinding was removed, participant evaluation of the process.

Main Outcome Measures

Likelihood ratios for mental disorders.

Main Results

Participants were considered to be symptom-screen negative (19%) or symptom-screen positive with no psychiatric diagnosis (42%) or to have subthreshold psychiatric diagnosis (13%) or threshold psychiatric diagnosis (26%). Of those who were symptom-screen positive, mean interview time was 11.4 minutes for participants who received a PRIME-MD diagnosis (n = 377) and 5.6 minutes for participants without a PRIME- MD diagnosis (n = 413). {Likelihood ratios for positive diagnosis and chance- corrected agreement (kappa) were 6.9 and 0.71 for any psychiatric diagnosis, 8.4 and 0.61 for any mood disorder, 6.9 and 0.55 for any anxiety disorder, 40.5 and 0.71 for probable alcohol abuse or dependence, and 73.0 and 0.73 for any eating disorder} (numbers calculated from data in article).


Positive results on the PRIME-MD questionnaire were associated with substantially increased probability of mood, anxiety, somatoform, and, particularly, alcohol-related and eating disorders in primary care.

Source of funding: Roerig and Pratt Pharmaceuticals.

For article reprint and PRIME-MD materials: Dr. R.L. Spitzer, Biometrics Research Department, Unit 74, New York State Psychiatric Institute, 722 West 168 Street, New York, NY 10032. FAX 212-543-5525.


What should PCPs do about screening for psychiatric illness? Screening is worthwhile if, first, the disorder causes a substantial burden of illness. Mental health disorders are common in primary care, cause major suffering and disability, and often are unrecognized and untreated.

Second, an accurate test must be available. Spitzer and colleagues describe a new instrument that pairs a highly sensitive initial screen with a more specific follow-up evaluation. Their assessment is comprehensive, and their findings appear valid and generalizable. The sensitivity of the screening instrument, however, is limited in several areas. For instance, the sensitivity for major depressive or panic disorder is 57% (> 40% of patients with major depressive disorder or panic disorder were not identified).

Third, effective treatment must be available. Treatment for some mental health disorders is effective; for others, it is not, notably the subthreshold diagnoses (those not meeting DSM-III-R criteria), which comprise one third of the patients with diagnoses. Although subthreshold diagnoses represent substantial disability and health service use, interobserver agreement about these diagnoses is low, the burden of illness is smaller, and the value of treatment has not been established (1).

Successful screening programs must also be affordable and practical. The instrument described by Spitzer and colleagues required < 9 minutes of clinician time, but in the typical primary care office visit, 9 minutes cannot be spared; therefore, in most cases, a separate office visit would be required.

Improving the recognition and treatment of mental health disorders in primary care is complex and requires a multifaceted intervention (2); a screening tool such as PRIME-MD is an important element. Equally important are enabling strategies that provide PCPs with the time, logistic support, and resources necessary to care for these patients.

Paul Gorman, MD
Portland Medical Center Portland, Oregon