Physician Practice Variation in Electronic Health Record Documentation.

Abstract

Adoption of electronic health records (EHRs) was motivated by the expectation that they would improve quality and decrease costs of care. EHRs’ value, however, depends on how they are used, which likely explains the heterogeneous benefits observed in the literature. This dissertation uses mixed methods to explore a critical component of EHR use in primary care: variation in EHR documentation, defined as differences in how users record or remove information.

The first chapter delineates a conceptual framework of variation in EHR documentation that includes five different forms of variation and five levels where the forms may materialize. This chapter focuses on potentially harmful variation by detailing how non-patient factors foster variation that interferes with clinical decision support, care coordination, and population health management, jeopardizing the efficient delivery of high-quality healthcare.

The second chapter measures variation in one form of variation, completion of documentation, in a national sample of primary care practices. Using data from a major EHR vendor, this chapter finds differences in how variably providers complete fifteen different clinical documentation categories and identifies patient’s problems, the provider’s assessment and diagnosis, the social history, the review of systems, and communication about lab and test results as the most varied. The majority of variation exists across providers in the same practice, suggesting providers are making different decisions about documentation for comparable patients.

The final chapter explores the context of this variation with semi-structured interviews, finding that variation in EHR documentation is perceived as a commonplace phenomenon resulting from a flexible EHR design that allows users to develop different documentation styles. Variation reportedly introduced inefficiencies into care delivery and created patient safety and care quality risks from missed or misinterpreted information. Respondents identified additional training, ongoing meetings, and improvements in EHR design as effective strategies to prevent harm.

Widespread variation in EHR documentation can interfere with care delivery by obscuring the location and meaning of patient information. In order to realize gains from adopting EHRs, practices, vendors, and policymakers must collaboratively develop better interfaces and clearer guidelines to support their effective use.