148 - The Identification of Rheumatoid Arthritis Patients for Comparative Effectiveness Research Using An Electronic Health Record and Mathematical Modeling

Sunday, November 6, 2011: 9:00 AM-6:00 PM
Hall F2 - Poster Hall (McCormick Place West)
Aarat M. Patel1, Larry W. Moreland2, Melissa Saul3, Stephen R. Wisniewski3 and Marc C. Levesque4, 1Univ of Pittsburgh Med Ctr / Children's Hospital of Pittsburgh of UPMC, Pittsburgh, PA, 2University of Pittsburgh, Pittsburgh, PA, 3Univ of Pittsburgh, Pittsburgh, PA, 4Univ of Pittsburgh Med Ctr, Pittsburgh, PA
Presentation Number: 148

Background/Purpose: To develop an electronic search algorithm with a high specificity and sensitivity to accurately identify rheumatoid arthritis (RA) patients in a large health care system linked by an electronic health record (EHR) system.

Method:  Records from the Medical Archival Retrieval System (MARS) at the University of Pittsburgh Medical Center (UPMC) were used to identify potential RA patients.  We searched the UPMC MARS system for subjects with a 714.0 International Classification of Diseases, 9th revision (ICD-9) code and used a recursive partitioning method to develop a search algorithm with a high specificity and sensitivity for the identification of RA patients; the recursive partitioning method used the 714.0 ICD-9 code and tested 35 additional variables (serology, inflammatory markers, medications, specific words in physician notes, etc).  During the development and validation of the algorithm, patients were classified into those likely or unlikely to have RA and representative sets of these patient records were reviewed to determine if subjects met the 1987 and/or 2010 RA diagnostic criteria.

Result:  We initially analyzed records from 2009 to study the effect of clinical setting (inpatient vs. outpatient rheumatology clinic) on the identification of RA patients.  For inpatients, there was a low PPV of a 714.0 (39.0%) whereas for outpatient-rheumatology subjects there was a high PPV of a 714.0 for the identification of RA patients (87.3%), (n=95, p<0.0001; Fisher’s exact test).  When the records of subjects with and without a 714.0 ICD-9 code were analyzed (n=400), the sensitivity, specificity, PPV and NPV of a 714.0 among outpatient-rheumatology patients was 98%, 88%, 87% and 98%, respectively.  We next used recursive partitioning to test whether other variables besides a 714.0 could be used to create a more specific algorithm for identifying RA patients among outpatient rheumatology patients seen in 2009.  We found that a search algorithm with 3 variables 1.) 714.0 ICD-9 code, 2.) ratio of “rheumatoid arthritis” in a physician note per rheumatology visit and 3.) ratio of “RA” in a physician note per rheumatology visit improved the specificity for identifying RA patients.  The sensitivity, specificity, PPV and NPV of the algorithm was 93%, 95%, 94%, 95% (n=400).  Validation of this algorithm with analysis of an additional 400 subjects produced similar results (95%, 96%, 96%, 95%, respectively).  We used this algorithm to analyze records from outpatient-rheumatology patients evaluated in 2010 (n=17,571) to identify 2,610 patients with RA.

Conclusion: The ICD-9 code for RA (714.0) alone was not reliable for identifying RA patients in the inpatient setting and had suboptimal specificity in the outpatient rheumatology setting. We developed and validated a simple algorithm using recursive partitioning that used 3 variables to identify RA patients with a high specificity, sensitivity, PPV and NPV.  This simple algorithm represents a substantial improvement in terms of sensitivity and specificity over existing published algorithms.  Using an EHR and this electronic search algorithm will enable large-scale comparative effectiveness studies on the treatment and management of RA in “real-world” clinical settings.

Keywords: diagnosis and rheumatoid arthritis (RA)

Disclosure: A. M. Patel, None; L. W. Moreland, None; M. Saul, None; S. R. Wisniewski, None; M. C. Levesque, None.