Featured PROMIS Project
How Well Do Generic Patient PROMIS Instruments Capture Health Status in Rheumatoid Arthritis?
Bartlett SJ1,2, Orbai AM2, Jones M2, Bingham CO2
Susan J. Bartlett, PhD; Susan.firstname.lastname@example.org; 1McGill University, Montreal Canada; 2Johns Hopkins School of Medicine
Objective: PROMIS offers precise, reliable generic measurement of physical, mental and social health across chronic conditions. However, little is known about the validity and performance of PROMIS instruments in rheumatoid arthritis (RA), where high levels of pain and fatigue and low physical function and mood are common.
Methods: Data were from the baseline visit of the first 125 RA patients enrolled in an ongoing study to evaluate the impact of systematically integrating broader assessment of patient-valued symptoms and impacts into ongoing arthritis care. Patients completed PROMIS computerized adaptive tests (CATs) assessing pain, fatigue, physical function, mood, sleep and social roles/activities using a tablet computer linked to Assessment Center in the waiting room immediately prior to the visit. Legacy measures (100 mm VAS for pain and fatigue, MHAQ) were also obtained as well as traditional clinical indicators of disease activity. Relationships between PROMIS and legacy measures were assessed using correlation and regression, and across CDAI disease activity levels with ANOVA.
Results: Patients were mostly female (79%) and white (86%) with a mean (SD) age of 56 (13) and disease duration of 12 (9) yr; 10% were diagnosed <= 2 yr. PROMIS CATs included an average 63 (12) items requiring 12 (5-32) minutes to complete. PROMIS CATs for pain (intensity, impact), fatigue, physical function, anxiety, depression, anger, sleep (disturbance, impairment) and social roles/activities (participation, satisfaction) correlated highly with pain VAS (rho's=.82-.83), fatigue VAS (rho=.86) and mHAQ (rho=-.74) (p's<.001). A dose-response relationship was evident in PROMIS measures across remission, low and moderate disease activity levels (except anger). Floor effects were common in legacy measures (16%, 11% and 43% for pain VAS, fatigue VAS and MHAQ, respectively) and were not common in PROMIS instruments.
Conclusions: These data contribute preliminary evidence of convergent and known groups validity and demonstrate generic PROMIS instruments can reliably assess health status, symptoms and impacts at health visits in people with RA. PROMIS CATs can be completed relatively quickly with results available immediately, and appear to address some of the well-recognized limitations (non-linearity, floor effects) of existing legacy measures in RA.