Photo Credit: SIphotography

The following is a summary of “Validation of Different Dementia Code-Based Definitions in a Population-Based Study of Rheumatoid Arthritis,” published in the October 2024 issue of Rheumatology by Vassilaki, et al. 

Researchers conducted a prospective study to examine the performance of various diagnostic dementia definitions in individuals with rheumatoid arthritis compared to those without rheumatoid arthritis.  

They involved 2,050 individuals (1,025 with rheumatoid arthritis) from a population-based cohort in southern Minnesota and compared the performance of 3 code-based dementia diagnostic algorithms against a medical record review diagnosis of dementia. For the overall comparison, each patient’s complete medical history was analyzed without specific time frames. Sensitivity analyses were also conducted using 1-, 2-, and 5-year windows around the date dementia was identified in the medical record, serving as the reference standard. 

The results showed that the algorithms performed similarly in individuals (+/−) rheumatoid arthritis. Generally, the algorithms exhibited high specificity, negative predictive values, and accuracy (greater than 88%) across all time windows studied. However, sensitivity and positive predictive values varied based on the specific algorithm and time frame used. Sensitivity ranged from 56.5% to 95.9%, while positive predictive values ranged from 55.2% to 83.1%. Notably, performance measures decreased as the time windows became more restrictive. 

Investigations concluded that code-based dementia diagnostic algorithms utilizing routinely collected electronic health record (EHR) data demonstrated good performance compared to medical record reviews, aiding future studies on identifying dementia in individuals with rheumatoid arthritis. 

Source: jrheum.org/content/51/10/978