At 3-month follow-up, a computerized intervention designed to promote CHD prevention medication use and adherence increased self-reported adherence by 25 percentage points and reduced 10-year predicted global CHD risk by 1.1 percentage points, with a larger effect seen in a pre-specified high risk sub-group. The effect on predicted CHD risk was mediated, in part, by intent to reduce CHD risk and by self-reported adherence.
These findings are similar to those reported in previous studies of more resource-intensive interventions for CHD prevention. For instance, multiple systematic reviews have shown 10-25 absolute percentage point increases in adherence, with the largest effects in studies using multiple telephone or in-person counseling sessions . Furthermore, a recent systematic evidence review found that global risk information, in repeated doses or with repeated counseling, reduced predicted CHD risk by 0.2 to 2 absolute percentage points . What is noteworthy about this study is that it achieved its effects using a computerized decision aid and automated message library to perform the functions of education and counseling.
Also noteworthy about this study are the reported effects of the intervention on aspirin adherence. Given the risk based recommendations for aspirin therapy  and individuals' desire for inexpensive, familiar, trusted interventions , it is not surprising that the intervention had its greatest effect on initiation of and adherence to aspirin.
What is somewhat surprising about this study is the apparent lack of effect of our intervention on cholesterol-lowering medication adherence. Not only are cholesterol guidelines risk-based, but a recent systematic review  found consistent reductions in total and LDL cholesterol in patients who received predicted CHD risk information and counseling. We suspect the lack of effect of the intervention on cholesterol is due in large part to the small number of participants (n = 20) choosing, initiating and adhering to cholesterol therapy, which limits the ability to detect any effect.
To better understand the effects of our intervention, investigators performed mediation analyses. These analyses suggest that both intent to reduce CHD risk and self-reported adherence to medications partially mediate reductions in 10-year predicted CHD risk. The proportional effects, however, are small (~20%), raising questions about other potential mediators such as diet and physical activity.
Interpretation of these findings should proceed with acknowledgement of the following limitations. First, this was a relatively small study not designed to look at subgroup effects. Subgroup analyses are small and underpowered, and should be used only for hypothesis generation. Second, physicians saw patients in both the intervention and control groups, which may have resulted in contamination between study groups. This may have reduced the effects of our intervention. Third, the effect of the intervention was diluted by the presence of participants (10%) who had no options for CHD risk reduction. Fourth, investigators measured adherence primarily by self-report (although with correlation with biomarkers). Based on studies comparing objective and subjective measures of adherence , it is likely we slightly overestimated adherence. Fifth, although randomization is expected to result in equally distributed characteristics between the intervention and control groups, the difference in education between groups raises questions about unmeasured confounders. Sixth, the study team conducted this study in one academic medicine clinic, and results might not generalize to other clinics, physicians, or patient populations. Finally, this was a short-term study with only 3 months of follow-up; longer follow-up likely would affect results.