ASSOCIATIONS BETWEEN CLINICAL VARIABLES AND PSYCHOLOGICAL SYMPTOMS IN RHEUMATOID ARTHRITIS: A NETWORK SCIENCE PERSPECTIVE

Background: Rheumatoid arthritis (RA) is associated with an increased prevalence of common mental disorders, including anxiety and depression (Matcham et al. 2013). Borsboom's (2017) network theory of mental illness is gaining traction as a model for depression which incorporates biological, social and psychological aspects of depression by looking at interactions between individual symptoms and other variables within a wider network. Currently applications have been restricted to psychiatric samples, however, this model may potentially help gain a clearer understanding of the causal links clinical variables and mental health symptoms in RA. Objectives: To study the prescriptions of NSAIDs in patients with rheumatoid arthritis (RA), psoriatic arthritis (PsA) and ankylosing spondylitis (AS) registered in ICEBIO and matched controls, and explore their relationship with disease activity measures. In addition, to explore the impact of initial TNFα-inhibitor therapy on NSAID prescription rates. Objectives: To test the feasibility of a network analysis approach to examine associations between clinical variables and mental health symptoms in RA. Methods: The data used are from patients attending rheumatology clinics at King’s College Hospital who completed patient reported outcomes (PROs) electronically via the Integrating Mental and Physical Healthcare (IMPARTS) system. Over 1,000 patients completed PROs via IMPARTS, with a subsample of 211 extracted for this analysis where psychological screening and inflammatory markers were recorded concurrently (<14days). The screening tools used were the two-item Patient Health Questionnaire (PHQ2) and the two-item Generalised Anxiety Disorder (GAD2), which assess the core symptoms of depression and anxiety: low pleasure/interest, low mood, high anxiety, uncontrollability of worry. Additional data recorded were joint counts and visual analogue scales for pain, fatigue and global disease activity. Missing data were imputed using multiple imputation. Network analysis was conducted using the ggraph package in R based on the regularised correlations between variables. With a graphical network model of variables created to calculate centrality values. Results: Figure 1 below illustrates that the symptoms with the most connections were PHQ1 (low pleasure/interest), GAD2 (uncontrollable worry), pain and global disease activity. As expected the strongest connections were between PHQ1 (low pleasure/interest) with PHQ2 (low mood), GAD1 (high anxiety) with GAD2 (uncontrollable worry), pain with patient global, tender joints with swollen joints, and ESR with CRP. The results highlight pain and PHQ2 (low mood) as having both the highest degree (3.9 & 3.8, respectively) and betweenness centrality (22 & 10, respectively). This indicates that these are the variables with both the highest number of connections and providing the shortest pathway between other symptoms and so may act as key variables linking inflammation and mental health. Pain and global disease activity had the highest closeness centrality (0.033 & 0.032, respectively), illustrating that they have the shortest path with all other symptoms and capture the influence of both inflammation and mental health. Tender and swollen joints have weak connections to the mental health variables, suggesting that that extra-articular aspects of pain may be important. Conclusion: Inflammation in RA does not appear to have a strong influence on mental health, with pain providing the main connection between these areas of the network. Concerning the symptoms of mental health considered, all were strongly connected but low mood provided the main connection between clinical and psychological variables. This indicates mood as potentially a key variable in RA, which is easy to monitor in routine care. Disclosure of Interests: Hsiu Yen Tung: None declared, Sam Norton: None declared, Faith Matcham: None declared, James Galloway Consultant for: Pfizer Inc, Matthew Hotopf: None declared REFERENCES: Borsboom, D. (2017). A network theory of mental disorders. World Psychiatry, 16(1), 5-13. doi:10.1002/wps.20375Matcham, F., Rayner, L., Steer, S., & Hotopf, M. (2013). The prevalence of depression in rheumatoid arthritis: a systematic review and meta-analysis. Rheumatology (Oxford), 52(12), 2136-2148. doi:10.1093/rheumatology/ket169 DOI: 10.1136/annrheumdis-2019-eular.2095Citation: Ann Rheum Dis, volume 78, supplement 2, year 2019, page A119Session: When rheumatoid arthritis (RA) does not walk alone: new data on comorbidities in RA (Scientific Abstracts)