Objective: This paper presents the results of a systematic literature review and a formal analysis of the scientific quality of empirical research on computerized provider order-entry (CPOE) applications.
Design: Formal, systematic review techniques were used to search the literature, determine study relevance, and evaluate study quality.
Measurement: A search of multiple databases from 1976 through mid-2007 yielded a final set of 46 articles. Relevance criteria included: (1) a direct comparison of a CPOE system with a non-CPOE system; (2) implementation in a clinical setting; and (3) clinically relevant outcomes.
Results: Study quality varied widely. Three major areas were identified for improvement in future studies: (1) internal validity, especially in terms of study designs, blinding, and instrumentation bias; (2) construct validity of the phenomenon of CPOE itself; and (3) measurement strategies, including reliability and validity assessments.
Conclusions: The evidence for the impact of CPOE needs to be improved to support scientific generalizability. Several common confounds are found in this literature. Future researchers will want to address them to improve the strength of the inference between CPOE and clinical outcomes. Discussion focuses on methods to improve future CPOE research.