preview

The Use Of Medical Records For The Purposes Of Scientific Research

Better Essays

Introduction The use of medical records for the purposes of scientific research is not a new methodological concept. Physician Alvan Feinstein and co-authors wrote a series of seminal articles articulating the problems associated with medical record reviews in cancer research in 1969 (Feinstein, Pritchett, & Schimpff, 1969a, 1969b). What has changed, however, is the advent of new technology associated with medical records, most notably the rise and proliferation of the electronic health record (EHR). The capabilities of EHRs to integrate patient, clinical, and system-level data into computer-based systems has led to the use of clinical EHRs for numerous research applications including observational, comparative effectiveness, and …show more content…

For example, Brennan & Watson (2012) found over 1600 adults over the age of 30 miscoded as having attended outpatient child and adolescent psychiatry services in the National Health System EHR database, a likely error in data imputation (Brennan & Watson, 2012). Data abstraction methods that use automated abstraction techniques for EHRs may also have issues related to data accuracy. For example, when compared to manual chart abstraction of EHR data, automated data extraction techniques have variable positive predictive values, some as low as 20% (Kahn et al., 2012; Mullooly, Donahue, DeStefano, Baggs, & Eriksen, 2008).
Complete Data and the Abscense of Evidence Data completeness in EHR data quality represents the overall completeness of the data, also defined as the degree of missing values within the database or dataset (Schafer & Graham, 2002). Data completeness is context driven, meaning not all missing data is missing for the same reason. For example, cosmologist Martin Rees in pondering the abscense of data indicating extraterrestrial life succinctly described the problem as “absence of evidence is not evidence of absence” suggest that evaluation of missing data requires key consideration of missing data context (Oliver, Billingham, Breu, Guggenbichler, &

Get Access