Background: Context - the combined effect of factors such as physician type, clinical setting, and guideline characteristics - influences the ability of automated decision support (ADS) to improve physician compliance with clinical practice guidelines (CPGs). Our goal was to determine whether research about the utility of ADS for promoting CPG compliance is contextually applicable to United States physicians.

Methods: We extracted information about physicians, settings, and guidelines from all articles published in the last 10 years that describe original research about the use of ADS to promote CPG compliance. The extracted information was compared to the range of practice contexts seen in the United States.

Results: Nearly two-thirds (65.3%) of papers described studies conducted in an academic setting, but only 11% of physicians report academic affiliations (p<0.0001). Physician reimbursement structure is often not reported. Salaried physicians were explicitly included as subjects in 14% of articles, but make up 45% of US physicians (p<0.0001). There are little data about the generalizability of ADS research to emergency care settings (6% of articles), and nursing home or skilled nursing facilities (10% of articles). Finally, ADS has not been studied at all in several epidemiologically important disease categories.

Conclusion: The literature does not adequately address some physician, setting, and guideline contexts. Before making policy or spending decisions based on the effectiveness of ADS, additional research is needed to determine whether ADS research can be generalized to under-represented contexts.