Session Information
Session Type: ACR Poster Session C
Session Time: 9:00AM-11:00AM
Background/Purpose: The patient global assessment (PGA) is widely used in disease activity scores for RA, including the CDAI, SDAI, and DAS28. While pain has been shown to be a major component of this score, other specific contributors of the patient experience to this score remain less clear. We examined the contribution of multiple domains of health-related quality of life measured by Patient-Reported Outcomes Measurement Information System (PROMIS) computer adapted testing. These measures by nature have heavily correlated structures which may lead to erroneous results from regression. Here, a factor analysis was used to identify latent variables. Using these latent variables along with clinical characteristics, we developed a multivariable model to explain the relative contributions of these measures to the PGA.
Methods: We conducted a cross sectional analysis of a cohort of 196 patients with RA receiving guideline-based care in an academic rheumatology clinic. PROMIS computer-adapted tests, including domains of physical function, pain interference, social participation, fatigue, sleep disturbance, depression, and anxiety were used. Factor analysis was used to identify factor structures. A multivariable linear regression model was then developed to identify an optimal model including these factors and other clinical characteristics.
Results: Among 196 patients (mean age 54.8 +/- 13.4 years, 81% female, mean RA duration of 11 +/- 9.6 years, mean PGA 29). Most (69%) were well-controlled, in remission or low disease activity. Factor analysis revealed three major factors. The first, referred to as “function”, had highest loading of physical function, followed by pain interference, fatigue, and social participation. The second, referred to as “emotional distress”, had equal loading of depression and anxiety. A third had loading only on sleep disturbance. “Function” alone, in univariable analysis, explained 52% of the variance in PGA. A multiple linear regression including “function”, sleep disturbance, swollen joint count, sex, BMI, and an interaction term of each of the three factors with one another explained 59% of the variance in PGA.
Conclusion: A single factor including fatigue, pain interference, physical function, and social participation explained most of the variance in the PGA in univariable regression. Emotional distress was not associated with the PGA when other factors were adjusted for. These data suggest that the PGA was primarily related to one latent variable related to “function”, and was influenced by “emotional distress” only insofar as it modified other symptoms.
|
Univariable |
Multivariable – Adj R2 = 0.59 |
||||
|
β |
p-value |
R-squared |
β |
p-value |
|
|
“Function” |
20.8 |
<0.0001 |
0.52 |
18.0 |
<0.0001 |
|
“Emotional distress” |
11.7 |
<0.0001 |
0.17 |
——- |
——– |
|
“Sleep disturbance” |
23.1 |
<0.0001 |
0.39 |
6.5 |
0.01 |
|
“Function” x “Emotional” |
0.8 |
0.68 |
0.001 |
-4.9 |
0.008 |
|
“Function” x “Sleep” |
1.9 |
0.44 |
0.003 |
4.9 |
0.02 |
|
“Sleep” x “Psych” |
2.5 |
0.32 |
0.005 |
7.8 |
0.001 |
|
Swollen Joints |
2.69 |
<0.0001 |
0.11 |
0.9 |
0.04 |
|
BMI |
0.55 |
0.04 |
0.02 |
-0.3 |
0.1 |
|
Sex |
0.44 |
0.93 |
0.00 |
5.5 |
0.1 |
To cite this abstract in AMA style:
Craig E, Bartlett SJ, Perin J, Zeger S, Bingham III CO. Function, Pain, Fatigue, and Participation Are the Primary Contributors to the Patient Global Assessment in Rheumatoid Arthritis [abstract]. Arthritis Rheumatol. 2017; 69 (suppl 10). https://acrabstracts.org/abstract/function-pain-fatigue-and-participation-are-the-primary-contributors-to-the-patient-global-assessment-in-rheumatoid-arthritis/. Accessed .« Back to 2017 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/function-pain-fatigue-and-participation-are-the-primary-contributors-to-the-patient-global-assessment-in-rheumatoid-arthritis/