Antidepressant drug use: differences between psychiatrists and general practitioners. Results from a drug utilization observation study with fluoxetine.

Between 1990 and 1993, a series of drug utilization observation studies with fluoxetine (Flx) were conducted in Germany in several waves. 3,158 patients treated by psychiatrists/neurologists (PN) and 15,601 patients treated by general practitioners/internists (GPI) were included; data collection at start and end of treatment focussed on diagnoses, symptoms, prescription, comedication, efficacy (CGI, Zung scale), and adverse events. Differences between PN and GPI patients were of major interest. For more than 90% of both the PN and the GPI cases. Fix was used for the indication of "depression", with a dosis of 20 mg/day. More PN patients (47%) than GPI patients (28%) were diagnosed as "endogenous"; GPI patients more often presented with first episodes (36 vs. 24%). "suicidal ideation" was less prominent compared to PN subjects (17 vs. 28%). Psychotropic comedication was regarded as necessary in 39% (PN) and 10% (GPI) of the cases. Early treatment termination because of "remission/major improvement" was observed in 13% (PN) vs. 21% (GPI) and because of "adverse events" in 11% (PN) vs. 3% (GPI) of the patients. At observation end, 53% (PN) vs. 74% (GPI) were rated as "symptom-free/markedly improved" (CGI); self-ratings reflected comparable results, marked improvements over time, but still PN/GPI differences at the end. "Suicidality" related to depression was more pronounced in the PN group at both points in time. 24% (PN) vs. 6% (GPI) of the cases reported "routine" adverse events, while in 2% (PN) and 1% (GPI) "serious" adverse events were observed. (For all the above comparisons p < 0.001 to < 0.0001.) These findings reveal that-under routine conditions handled by PNs and GPIs-Fix shows an efficacy and safety consistent with clinical trial data. The body of data suggests that PN patients present with more severe depression and more suicidality, require more comedication, and end up with a poorer outcome. Differences in the physicians' perception of psychiatric and somatic symptomatology and their treatment routines may also have something to do with the PN/GPI group differences observed.