Two interactive computer-based systems have been evaluated: a teaching program with text and animated graphics and a multiple-choice knowledge-assessment program (KAP) with optional prescriptive feedback. One hundred seventy-four routine-attending insulin-dependent diabetes mellitus (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM) patients were allocated to active and control groups to determine the effect of these programs on knowledge and control after a 4- to 6-mo follow-up period. Interactive computer teaching (ICT) resulted in a significant knowledge increment in both IDDM and NIDDM patients (P less than .05), together with a mean fall of 0.8 and 0.7%, respectively, in HbA1c (P less than .05 and P greater than .1), but no changes were observed in respective control groups. The KAP with feedback also produced a significant knowledge increment in both IDDM and NIDDM patients (P less than .05), of similar magnitude to the ICT program, and a mean fall in HbA1c of 1.2 and 1.3%, respectively (P less than .05), with no changes in the corresponding control groups. Even when KAP was used without prescriptive feedback, smaller but significant mean falls in HbA1c of 0.7 and 0.8% (P less than .05) were seen in IDDM and NIDDM patients, respectively, suggesting a motivational effect resulting from program participation.(ABSTRACT TRUNCATED AT 250 WORDS)