Premium
Using medical records to supplement a claims‐based comparative effectiveness analysis of antidepressants
Author(s) -
Croghan Thomas W.,
Esposito Dominick,
Daniel Gregory,
Wahl Peter,
Stoto Michael A.
Publication year - 2010
Publication title -
pharmacoepidemiology and drug safety
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.023
H-Index - 96
eISSN - 1099-1557
pISSN - 1053-8569
DOI - 10.1002/pds.1988
Subject(s) - medicine , medical record , depression (economics) , exploratory research , anxiety , psychiatry , health plan , family medicine , health care , sociology , anthropology , economics , macroeconomics , economic growth
Purpose Because health insurance claims lack clinical information, comparative effectiveness research studies that rely on these data may be challenging to interpret and may result in biased inference. We conducted an exploratory study to determine if medical information contained in patient charts could offer clinical details that would assist in interpreting the results of a claims‐based comparative effectiveness study of selective serotonin reuptake inhibitors (SSRIs). Methods Retrospective review of 457 charts of patients initiating SSRI treatment. Descriptive data elements included patient diagnosis, symptoms of depressive and anxiety disorders, provider's assessment, and medication treatment and side effects. Results Most subjects were excluded from the study because their charts were not accessible (58.7%), they did not have a follow‐up visit (55.6%), providers could not be contacted (58.0%), or providers refused participation in the study (36.5%). Among those included in the study, most patients were noted to have depression, but most charts lacked information on the majority of depression symptoms at baseline and follow‐up. Few concomitant symptoms, side effects, and other important clinical and treatment characteristics were recorded. Conclusions Inability to obtain charts due to plan or provider refusal, lack of available information in charts at key times in the course of illness, and missing data elements posed considerable challenges and prevented firm conclusions beyond those drawn from the parent, claims‐based study. Copyright © 2010 John Wiley & Sons, Ltd.