Enabling the diagnosis of occult psychiatric illness in the emergency department: A randomized, controlled trial of the computerized, self-administered PRIME-MD diagnostic system

Study objective: Undiagnosed mental illness is highly prevalent and produce s needless morbidity. Emergency department patients with vague or longstand ing complaints are at risk far occult mental illness, but are seldom diagno sed. We conducted this study to determine whether a previously validated, s elf-administered, computerized psychiatric interview (Primary Care Evaluati on of Mental Disorders [PRIME-MD]) could increase the detection of occult p sychiatric illness in the ED. Methods: This was a randomized, controlled trial of consecutive patients en rolled during convenient times at a university teaching hospital ED with an annual census of 38,000. ED house staff and attending physicians participa ted. Patients were those with nonspecific complaints potentially associated with occult psychiatric illness (eg, long-standing headache, abdominal or back pain). Exclusion criteria were known psychiatric illness, complaint, o r medication; and straightforward reason for the ED visit. Consenting subje cts completed the PRIME-MD questionnaire in the waiting room, and were rand omly assigned to either the "report" (report results given to physician) or "no-report" groups. PRIME-MD results were clipped to the front of the char t of report group patients. There was no other intervention. The main outco me measures were the percentage of all patients and percentage of patients with a PRIME-MD diagnosis who received a psychiatric diagnosis, consultatio n, or referral from the emergency physician. Results: A total of 339 (5.1%) of all patients were approached; 230 consent ed to participate in the study, and 218 completed the PRIME-MD session and were randomly assigned to study groups. Ninety-two patients in the report g roup and 98 cases in the no-report group were analyzed. Patients were omitt ed for the following reasons: left without being seen (8), mistakenly enrol led (10), or unretrievable medical records (10). Seventy nine (42%) patient s received a psychiatric diagnosis from PRIME-MO and 9 (5%) from the physic ian. The frequencies of physician psychiatric diagnosis and referral did no t differ between groups (diagnosis: report 9%, no report 9%, Delta 0%, 95% confidence interval [CI] -13, 14; referral: report 9%, no report 7%, Delta 2%, 95% CI -11, 16). Conclusion: ED patients willingly completed the PRIME-MD questionnaire (in a median time of 7 minutes), which frequently diagnosed psychiatric conditi ons. Despite this, physicians rarely diagnosed or treated these conditions regardless of whether they were provided with the PRIME-MO diagnoses.