Cohort Definition Name : Earliest event of Rheumatoid Arthritis
Contributor name : Joel Swerdel
Contributor OrcId : 0000-0002-6976-2594’
Logic Description : Earliest occurrence of Rheumatoid Arthritis indexed on diagnosis (condition or observation) date, for the first time in history cohort exit is the end of continuous observation
Recommended study application : target
Assertion statement : This cohort definition was executed on at least one real person-level observational health data source and resulted in a cohort with at least 1 person.
Target Clinical Description : Rheumatoid arthritis (RA) is a chronic inflammatory disease that primarily involves the joints. The global RA prevalence estimate was 0.46% (95% confidence interval [CI] 0.39–0.54; I2 = 99.9%).[i] The pathophysiology of RA involves chronic inflammation of the synovial membrane, which can destroy articular cartilage and juxta-articular bone. Extra-articular manifestations of RA include rheumatoid nodules, pulmonary involvement or rheumatoid vasculitis, and systemic comorbidities. Cardiovascular disease is a common consequence of chronic inflammation and the primary cause of death in people with RA. Causes of RA are unknown, however genetic, epigenetic, and environmental risk factors have been identified. The disease affects women 2-3 times more often than men and occurs at any age. While there is no cure for RA, the management of RA over the last 15 years has changed dramatically due to the advent of novel therapeutics, introduction of early therapy, the development of new classification criteria, and application of new effective treatment strategies[ii]. Availability of these effective treatments and approaches has led to remission of symptoms in many patients.
Evaluation conclusion : We developed a prevalent cohort definition for rheumatoid arthritis (RA) using a concept set of 15 concepts from the condition and observation domains which incorporated all those found from the literature review, the analysis of PHOEBE, and orphan concepts in cohort diagnostics. We evaluated the cohorts across 12 data sources.The algorithm retrieves subjects from all 12 databases tested. There were no significant conditions prior to index to justify using index date reclassification. Drug use prior to the index date did indicate that the condition may have been prevalent in a subset of subjects prior to index but due to the long lead time to diagnosis in RA we did not think this would significantly correct the index date. We developed a more specific cohort requiring a second diagnosis or observation code for RA in the time period 31-365 days after index. This cohort improves the specificity of the algorithm albeit at the expense of sensitivity as determined by PheValuator. Performance characteristics were determined for 10 of the 12 databases. The remaining databases did not contain enough subjects to produce an accurate diagnostic model. Using A, sensitivity ranged from about 43% in Germany to about 93% in Pharmetrics while positive predictive value ranged from about 54% in CPRD to about 98% in Germany
Thanks @Azza_Shoaibi and @jswerdel for sharing this phenotype for RA. I plan to re-use it as a nesting indication in our HowOften effort for OHDSI2023.
A couple comments when making use of this:
- The RA conceptset contains both conditions and observations, so I needed to remember to look in both domains. For the observation concepts, there is some risk that some data partner may have a VALUE as a qualifier to the concept that I’m not making use of (e.g. VALUE_AS_CONCEPT = ‘No’ would invalidate the observation using it as a positive occurrence).
- In investigating the conceptset, I noted that if you ‘optimized’ it in ATLAS with the latest vocabulary, a few concepts were redundant (specifically, ‘Seropositive rheumatoid arthritis’ and ‘seronegative rheumatoid arthritis’ are children of ‘Rheumatoid arthritis’). However, I very much distinctively remember our Barcelona study-a-thon when we had identified these two concepts and how they did NOT map up into ‘Rheumatoid arthritis’ within SNOMED. So this is a nice illustration that we need to recognize that the vocabularies are evolving …in this case, SNOMED expanded their relationships over time. In this case, no harm in a ‘sub-optimal’ representation, they are logically equivalent. But a good general note that as vocabulary versions come out, it’s useful to check in your phenotypes to see how they are impacted. I know @Dymshyts is working on some tools for the community that could help with this specific problem.
Imported to the OHDSI Phenotype Library. It may be expected to be found with id = 858 in the next release. Thank you