Patients With Rheumatoid Arthritis Report Inconsistent Medicare Claims History

March 10, 2021
Lana Dykes

Older adults who self-report rheumatoid arthritis (RA) do not have a Medicare claims history consistent with their diagnosis, according to study.

Older adults who self-report rheumatoid arthritis (RA) do not have a Medicare claims history consistent with their diagnosis, according to a study published in ACR Open Rheumatology.1 Additionally, revisions to self-reported questions may be able to better identify RA in national health surveys.

Investigators determined physician diagnosis of RA using multiple gold-standard measures based on nationally representative Medicare claims for adults aged 65 years and older. The study also verified whether additional questions about medication usage and recent physician visits can improve the positive predictive value (PPV) accuracy in self-reported RA. As the American College of Rheumatology (ACR) guidelines recommend continuing follow-up care to assess treatment for those with active RA, questions about physician visits were used to decide if correspondence improved self-reported RA and Medicare claims. The accuracy of self-reported RA was based on 3 different claims-based RA indicators.

According to the study, self-reported health conditions and well-defined chronic conditions show varying agreement with medical records and administrative data. Interestingly, patients reporting non-chronic diseases, such as cancer and emphysema, have been shown to present accurate and useful information for identifying a disease-free population.

Investigators curated the study sample from the 2004, 2008, and 2012 waves, the most recent surveys that included self-reported RA and RA treatment, of the US Health and Retirement Study (HRS), a longitudinal study of US residents aged 50 and older. Participants were followed from entry until death, with new subjects included every 6 years and then subsequently surveyed every 2 years. To validate conditions, investigators measured self-reported answers against a gold standard, using a combination of sensitivity, specificity, PPV, negative predictive value (NPV), and Cohen’s κ. Previous studies on RA have indicated that most self-reported, physician-diagnosed cases do not qualify as having RA according to medical records. According to the PPV, on the higher end of the spectrum, 41% of people with RA had confirmation of their disease based on the gold standard. For studies reporting sensitivity of self-report ranged from 54% to 100% and κ results ranged from 0.06 to 0.46. It can be assumed that the generalizable PPV studies include a sample prevalence that can be representative of the actual population prevalence.

Participants 65 years or older were excluded, as well as those who had linked Medicare data and had received Parts A and B Medicare coverage for 11+ months per year for 4 years from 2005 to 2008 or 2009 to 2012. Investigators excluded all prevalent cases of RA in 2004, which left an eligible sample size of 4228. Further exclusions, including nonresponse, death, or withdrawal reduced the sample to 3768 participants. Of the remaining subjects, the mean age was 73.7 years in 2004, 58.8% were women, and 88.5% were Caucasian. In total, 345 participants had self-reported RA in 2008 or 2012. Of these incidents, 57 had a single RA claim, 41 had 2 or more claims and 16 met the strict RA criterion. Investigators identified 245 respondents with a single RA claim and 29 respondents who met the strict RA algorithm. Additionally, 161 participants were taking medication for arthritis. As far as doctor visits were concerned, 152 patients reported seeing a physician for arthritis in the past 2 years and 111 respondents reported RA, medication usage, and seeing a doctor in the past 2 years.

As low PPV and κ values of self-reported RA create problems for investigators studying RA, investigators exemplified previous RA studies and added additional survey questions related to medication usage in order to improve PPV. Studies indicated that disease-modifying antirheumatic drugs (DMARDs) increase PPV and κ statistics of self-reported RA. The studies included the Center for Disease Control and Prevention (CDC) ICD‐9‐CM case definition of arthritis and other rheumatic conditions, the Chronic Conditions Warehouse (CCW) algorithm for establishing osteoarthritis (OA) and RA, and an algorithm used validate methods for identifying patients with RA using administrative data assessed using the PPV of ICD code–based algorithms. All 3 concluded that self-reported RA combined with DMARD use were able to identify cases of RA.

A limitation of the study is that investigators excluded HRS patients over 65 years of age who did not have linked Medicare claims as well as those who did not have linked Medicare claims and those who did not have full non-HMO Parts A and B coverage within the 4 years of lookback periods. The sample was also greatly reduced from the original HRS sample size, which reduces generalizability. In addition, participants who were excluded due to dropout, death, and nonparticipation may alter results.

“Our analysis expands on findings from prior studies by testing the validity of self‐reported RA in a US‐based nationally representative sample. We found that self‐reported RA has low validity for identifying survey respondents with RA on the basis of Medicare claims data, as indicated by three different claims‐based RA diagnostic algorithms,” investigators concluded.


Booth MJ, Clauw D, Janevic MR, Kobayashi LC, Piette JD. Validation of Self-Reported Rheumatoid Arthritis Using Medicare Claims: A Nationally Representative Longitudinal Study of Older Adults [published online ahead of print, 2021 Feb 23]. ACR Open Rheumatol. 2021;10.1002/acr2.11229. doi:10.1002/acr2.11229