Fifty-one patients diagnosed as schizophrenics or schizoaffective according to DSM-III criteria were treated with conventional haloperidol regimens for 4 or 6 weeks. The clinical picture and extrapyramidal side effects were assessed by means of the Brief Psychiatric Rating Scale (BPRS), and the Simpson and Angus Scale (EPSE). Evaluations were made at admission and after 4 or 6 weeks of treatment. The clinical response to treatment was reported as the percent change in BPRS scores at the end of treatment from the BPRS scores at baseline. Haloperidol (HAL) and hydroxyhaloperidol (REDHAL) were determined by high-performance liquid chromatography (HPLC) with electrochemical detection in plasma. The mean total BPRS item score at the end of the study was significantly lower than at the beginning of the study. HAL and REDHAL levels were significantly related to the dose, and REDHAL levels were also related to HAL levels. There was no correlation between plasma HAL levels and the percent change in BPRS. The percent change in BPRS at the end of the study was negatively correlated with plasma REDHAL levels and REDHAL/HAL ratios and was positively correlated with the baseline BPRS total score. There was no significant correlation between the duration of illness and improvement, but patients with good improvement had significantly shorter duration of illness. Patients who improved also had higher baseline BPRS scores, lower REDHAL levels and REDHAL to HAL ratios, but not significantly different HAL levels. Therefore, the shorter the duration of the disease, the higher the baseline BPRS and the lower the reduced levels of its ratio to haloperidol levels, the higher the percent improvement.(ABSTRACT TRUNCATED AT 250 WORDS)