Background/Purpose

Comorbidity in RA can delay diagnosis and influence treatment decisions. It is known to affect RA outcomes, and can confound data analysis. Most RA observational cohort studies have used either the generic and weighted Charlson Comorbidity Index (CCI), one of few validated tools but complicated and designed for medical inpatient settings, or non-standardised but simple comorbidity counts. We have previously reported on the feasibility of the latter and its value in assessing impact of comorbidity on survival and function [ref]. There is currently no standardised, uncomplicated and validated instrument for recording and collecting comorbidity data which is relevant to contemporary and routine rheumatology practice. Is it feasible to measure comorbidity in daily rheumatology practice? The purpose of the study is to evaluate the feasibility of a simple method based on ICD10 systems (chapters) to measure comorbidity in RA.

Methods

Clinicians involved in a UK inception observational cohort study of RA (the Early RA Study, ERAS, n=1465, median follow up 10yrs) completed a simple and specific outcome form at regular intervals which included indicating the presence or not of the main ICD10 systems (n=15), with space to add details as free text (completed in 91%).

Results

More than 90% of all comorbidities reported covered 10 systems in order of frequency – Non Cardiac Vascular (to comply with WHO classification), Cardiovascular, Endocrine, Gastro Intestinal & Hepatic, Respiratory, Psychiatric, Malignancies, Renal, Dermatology, Opthalmology. As musculoskeletal and extra articular RA conditions would be managed within the specialty, these were identified and remained as a separate group and not part of this analysis. 75% of all individual comorbidities recorded  in the ICD10 systems included only 20 specific medical conditions, and the most common 2-3 specific conditions in each ICD10 system made up most (70-80%) of each system. For the other less common systems, there was a wider range of individual conditions of roughly equal frequencies. At least one comorbidity at baseline, 3, 5 & 10yr follow up was present in 21%, 40%, 50% and 78% respectively based on a simple numeric score, compared to weighted CCI of 11%, 27%, 36%, 52%. Mean scores were 0.21, 0.53, 0.72, 1.08 and 0.13, 0.4, 0.55, 0.86 respectively, with modest correlations since they measure different aspects: Spearman’s rho of around 0.72. Kappa statistics were 19.3, 25.1, 24.6, 22.9 respectively. Multiple morbidity can be measured from ICD10 (>1 major condition) and was present in <1%, 11%, 17% & 29% at baseline, 3, 5, 10yrs.

Conclusion

This study has shown the feasibility of collecting comorbidity data in a relatively simple way using standard definitions, and its value in identifying multi-morbidity.  It is this latter group with complex disease that requires prompt identification, multiple speciality input and coordination of patient care routine clinical settings.

Ref: Norton S et al. A study of baseline prevalence and cumulative incidence of comorbidity and extra-articular manifestations in RA and their impact on outcome. Rheumatology (Oxford) 2013;52(1): 99-110

ACR Meeting Abstracts - https://acrabstracts.org/abstract/comorbidity-in-rheumatoid-arthritis-it-is-feasible-to-record-concomitant-medical-conditions-and-multi-morbidity-in-observational-research-studies-can-this-be-extended-to-routine-clinical-setti/