Portability of an algorithm to identify rheumatoid arthritis in electronic health records

Abstract

Electronic phenotype algorithms allow rapid identification of case populations in multiple sites with little retraining.

A collaborative approach to developing an electronic health record phenotyping algorithm for drug-induced liver injury

“…This approach has led to the development of several validated algorithms to identify individuals with specific phenotypes (ie, EHR phenotyping algorithms). [22][23][24][25][26] Many of these studies have also demonstrated the ability to share EHR phenotyping algorithms among multiple institutions, 21 22 26 26a although they usually develop and validate an algorithm at one institution before implementation at other sites. In contrast, in this study, two institutions (Columbia University (CU) and Mayo Clinic (Mayo)) developed DILI EHR phenotyping algorithms separately from one another with project goals and disease case definitions informed by different organizations (eMERGE/iSAEC and DILIN, respectively).…”

Retrospective Derivation and Validation of an Automated Electronic Search Algorithm to Identify Post operative Cardiovascular and Thromboembolic Complications

“…At our institution, the search strategy performed with high levels of sensitivity and specificity; however, the point estimates cannot be expected to express the true limits for these values in other settings. In this sense, it is important to note that although the derivation and validation strategy could potentially be enhanced by specifically tailoring the algorithm to better suit other institutions, the portability of free-text search algorithms was previously demonstrated across different institutions and EMRs with minimal need for institution-specific algorithm optimization [8]. The efficiency with which clinical and translational research can now be performed is not only of tremendous value to researchers but can also indirectly enhance the quality of patient care by providing information about risk factors for adverse outcomes of interest.…”

“…While this study did not determine generalizability to other institutions, appropriate measures were taken to include broad free-text, natural language search criteria that are expected to accommodate for variations in medical terminology across institutions. Furthermore, there is prior evidence to the portability of similar electronic search tools in the literature [8].…”

“…These studies have all demonstrated that electronic searches can achieve sensitivities and specificities greater than 90% when compared to manual search efforts. Additionally, a previous study demonstrated portability of such electronic search tools, potentially allowing for application of search algorithms at external institutions [8]. However, there is limited available literature on the derivation and validation of an electronic search technique for identifying complications in the postoperative setting and the effectiveness of such methodology in this realm, specifically when compared to manual chart review of a prospectively-collected electronic database.…”

Improving the power of genetic association tests with imperfect phenotype derived from electronic medical records

“…For example, the Informatics for Integrating Biology and the Bedside (i2b2) Center, an NIH-funded National Center for Biomedical Computing based at Partners HealthCare System, has developed algorithms for several phenotypes including rheumatoid arthritis (RA), Crohn's disease, and ulcerative colitis (Liao et al 2010; Carroll et al 2012; Ananthakrishnan et al 2013). In existing EMR-based genetic studies, the probability is thresholded to classify individuals as cases and controls for subsequent analyses (Liao et al 2010; Kurreeman et al 2011).…”