Physician Practice Variation in Electronic Health Record Use

Substantial public and private funding has been invested to increase physician adoption of electronic health records (EHRs), with expectations of EHR use leading to higher-quality, lower-cost health care. While there is a growing body of evidence supporting these benefits, there is also recognition of the unintended ways in which EHR use impedes delivery of this care. One such domain is provider-to-provider variation in EHR documentation -- that is, differences in the content, structure, or location of the same patient information in the EHR. Such variation in how information is documented in the EHR may challenge users’ abilities to find and act on relevant details of the patient’s history, as well as impede tools designed to automatically extract information from the patient chart.

The objective of this AHRQ funded dissertation project was as follows:

• Examine the prevalence of variation in EHR documentation in physician practice, its causes, effects, and strategies to mitigate its potential for harm.

The project focused on the following three research questions:

• For core categories of clinical documentation, are there any categories that reflect high variation across primary care physicians (PCPs) in the same practice?
• What are the perceived causes of such variation in EHR documentation and how, if at all, do PCPs and staff perceive that variation affects their ability to use their EHR to deliver high-quality care?
• What strategies could be implemented in primary care practices to prevent or mitigate the negative consequences of variation in EHR documentation?

The research team used data from a national EHR vendor to quantify the extent of physician-to-physician variation in 15 categories of clinical documentation for 809 PCPs in 237 practices. Once they identified documentation categories with high variation, they conducted semi-structured interviews with physicians and staff to explore the causes and consequences of such variation, as well as to identify strategies to prevent or mitigate negative consequences.

The study found substantial variation in documentation of patients’ problems, providers’ assessments and diagnoses, patients’ social histories, the review of systems, and communication about lab and test results. Multilevel modeling revealed that most of the observed variation could be explained by differences in documentation across providers in the same practice, suggesting providers make different decisions about documentation for comparable patients. Interviewees perceived variation as common, while also attributing inefficiencies in care delivery and risks to safety and quality from missed or misinterpreted information to variation in documentation within their practices. Respondents identified additional training, ongoing meetings, and improvements in EHR design as potential strategies to prevent harm.