Observational Health Data Sciences and Informatics

Version: 1.1

Date Submitted: October 1, 2018

Date Approved: October 23, 2018

Author(s): Jeff Smith, MD

Organization: Wincester Institute, Scotland, UK

Modality: (Rule-Based or Computational)

Current Implementation: http://www.ohdsi.org/web/atlas/#/cohortdefinition/100145}Atlas Phenotype

This definition is intended to capture patients with a first-observed diagnosis of chronic rheumatoid arthritis (RA), taking care to rule out patients with short-term joint pain or fibromyalgia.

This phenotype was developed by a group of 4 expert rheumatologists, who gathered and identified key condition codes that manifest in individuals with RA, along with drug combinations that almost exclusively identify RA patients.

The phenotype has been validated at the following sites, with the following metrics:

TP = True Positive FP = False Positive TN = True Negative FN = False Negative PPV = Positive Predictive Value NPV = Negative Predictive Value

Site A represents a large Midwestern academic medical center. Site B is a hospital in a mostly rural area in the Southeastern US. Site C represents proprietary EU patient data licensed by a large research organization

At each site, chart review was independently performed by 2 rheumatologists on a randomly selected subset (5%) of patients satisfying the phenotype definition.

At Site B, the small sample size was likely due to a lack of an on-site rheumatology clinic.

• Alternative blood tests besides C-reactive protein may be used to diagnose RA in some cases.
• Rheumatoid arthritis is known to disproportionately affect women more than men; hence, the resulting sample will not be expected to be balanced across gender.

i.e. Flowcharts or other diagrams that may be helpful to an end user

This phenotype will be re-reviewed in October 2022.