In a proof-of-concept analysis, researchers show that machine learning methods paired with longitudinal patient-reported outcomes (PRO) data were able to classify subsequent rheumatoid arthritis (RA) disease activity after beginning treatment with a biologic
Combining machine learning (ML) with patient-reported outcomes (PROs) may improve characterization of disease activity among patients with rheumatoid arthritis (RA) beginning treatment, found researchers of a new study published in ACR Open Rheumatology.
In a proof-of-concept analysis, researchers were able to show that ML methods paired with longitudinal patient-reported data were able to classify subsequent disease activity after beginning treatment with a biologic.
“These results have important applications for real-world evidence generation in as much as they provide a framework whereby patients might initiate a new medication at an office visit with a clinician, at which time the Clinical Disease Activity Index [CDAI] could be measured and follow-up information could be collected only from patients (eg, on a smartphone app) to assess treatment response,” explained the researchers. “Given the dramatic growth in telehealth-delivered care in rheumatology beginning in 2020, our findings potentially also have implications for improving the efficiency of remote patient monitoring.”
According to the researchers, all variables used in the ML models are available to rheumatologists in their electronic health record systems or are short PROs that can easily be captured in a remote patient monitoring program.
Among the 500 patients, all initiating treatment with either golimumab or infliximab, 36% achieved low-disease activity (LDA)—indicated by a CDAI score of 10 or less. The CDAI has 4 components: patient global, physician global, tender joint count, and swollen joint count. The group found that the positive predictive value (PPV) to accurately classify LDA among the patients exceeded 80% at a sensitivity rate of 60% or greater for the best performing models.
Among 8 PROs from the Patient-Reported Outcomes Measurement Information System (PROMIS) and the Short Form 36 (SF-36), several were considered useful for classification, although not including information from SF-36 had a minimal effect on model performance.
Although disease activity measures continue to be the most accepted measures in RA, PROs are increasingly being included as considerations of care. Across all classification models, PROMIS, baseline CDAI, and patient global were the most important features. PROMIS measures included social participation, pain interference, pain intensity, and physical function. The researchers noted that a significant number of patients did not have follow-up PRO data within the prespecified 1-month timeframe compared with the CDAI.
“Our findings add to a growing body of literature showing that in cohorts of patients with RA, PROs, such as those contained in the PROMIS system, correlate moderately or strongly with traditional measures of RA disease activity (eg, CDAI) and other patient-reported assessment,” wrote the researchers. “There will understandably be some discordance between patients’ and physicians’ ratings of RA disease activity. For example, in a large trial of 793 patients initiating certolizumab pegol randomized to management with the Routine Assessment of Patient Index Data 3 (RAPID3) versus the CDAI, the RAPID3 was found to be a less sensitive method to detect improvement in RA disease activity as measured by the Disease Activity Score in 28 joints.”
The researchers explained that data from that same trial showed that among patients with a phenotype characterized by depression, anxiety, and other features that was linked with concomitant fibromyalgia, there was a prominent difference in outcomes comparing management using RAPID3 and CDAI.
Reference
Curtis JR, Su Y, Black S, et al. Machine learning applied to patient-reported outcomes to classify physician-derived measures of rheumatoid arthritis disease activity. ACR Open Rheumatol. Published online October 11, 2022. doi:10.1002/acr2.11499
Bringing Connectivity to the Specialty Pharmacy Workflow
May 2nd 2024In a session during the final full day of conference activity at AXS24, experts from CVS Health and Surescripts emphasized the need to simplify the prescribing workflow for specialty medication through proactive messaging, automation, and interoperability.
Read More
Emily Goldberg Shares Insights as a Genetic Counselor for Breast Cancer Risk Screening
October 30th 2023On this episode of Managed Care Cast, Emily Goldberg, MS, CGC, a genetic counselor at JScreen, breaks down how genetic screening for breast cancer works and why it is so important to increase awareness and education around these screening tools available to patients who may be at risk for cancer.
Listen
AUA to Focus on Inclusive Care, Robotic Surgeries, and Future of Urology at 2024 Annual Meeting
May 1st 2024The American Urological Association (AUA) 2024 Annual Meeting will highlight the latest innovations and future trends in urology, featuring dynamic plenary sessions, educational opportunities for providers, and discussions on cutting-edge treatments and global health initiatives.
Read More
Examining Telehealth Uptake to Increase Equitable Care Access
January 26th 2023To mark the publication of The American Journal of Managed Care®’s 12th annual health IT issue, on this episode of Managed Care Cast, we speak with Christopher M. Whaley, PhD, health care economist at the RAND Corporation, who focuses on health economics issues, including the influence of the COVID-19 pandemic on health care delivery.
Listen
Forging a Patient-Centric Path to Revolutionize and Redefine Value-Based Care
April 30th 2024Optum Life Sciences and Takeda Pharmaceuticals are partnering on an innovative virtual care pilot program for inflammatory bowel disease meant to both continue the mission of the current value-based health care landscape and raise the bar for personalized care delivery optimization.
Read More