A model to measure self-assessed proficiency in electronic medical records: Validation using maturity survey data...

Adoption of electronic medical records (EMRs) does not necessarily translate to proficiency -referred to here as EMR maturity. To realize the full benefit of wide scale EMR adoption, the focus must shift from adoption to advancing mature use. This calls for validated assessment models so that researchers, health system planners and digital health developers can better understand what contributes to maturity among physicians. This research aims to validate a measurement model for self-assessed EMR maturity among community-based physicians.

As part of an Ontario government-funded EMR adoption program, the EMR Maturity Model for community-based practices was adapted from a hospital-based EMR maturity model. A survey instrument was developed on the foundation of the new model and revised by experts and stakeholders. Content validity, face validity and user acceptance were established before survey administration. Internal consistency and construct validity of the model were tested after survey data were collected. Finally, physicians' comments collected via the survey were qualitatively analyzed to provide additional insights that can be applied to refinement of the model and survey.

As of August 1, 2019, 1588 physicians completed the survey. Ordinal alpha tests for reliability and content validity yielded an alpha value of 0.86 across all key measures specifically associated with maturity. Among most of these, there was a pattern of weak to moderate significant (p < .0001) positive Spearman inter-correlations. One factor was extracted for items measuring dimensions of maturity and all factor loadings of the key measures were greater than 0.40. The fit of the one-factor model was moderately adequate. This indicates the model is valid and reliable, with consistency across key measures for measuring one factor: maturity.

This is the first known validated model published in English that measures EMR maturity among community-based physicians. While the model is shown to be valid and reliable statistically and qualitative analysis supports this, there is room for improvement. Both the statistical analysis and portions of the qualitative analysis suggest areas of exploration to strengthen the model and survey. Future efforts will include refining the survey to improve user interface and accrue further data, as the sample to date is insufficient for generalizability.