Probabilistic pathways from chronic pain to suicidal ideation in Qatar: a Bayesian-Network study.

backgroundChronic pain frequently co-occurs with depressive and anxiety symptoms and doubles or triples the risk of suicidal ideation. Yet the joint probabilistic associations linking these conditions remain under-explored in Middle Eastern settings. We therefore sought to map the probabilistic interdependencies between chronic pain, mental-health symptoms, and demographic factors in Qatari adults, with a focus on associations converging on suicidal thoughts.

methodsWe analyzed questionnaire data from 2363 Qatari adults enrolled in the Qatar Biobank. A data-driven Bayesian Network (BN) was learned with hill-climbing and Bayesian Information Criterion (BIC) scoring. Domain-knowledge blacklists constrained demographic nodes as exogenous and suicidal ideation as a terminal node. Arc stability was confirmed by 200-replicate bootstrap resampling (arc-strength threshold ≥ 0.50). Conditional probability tables yielded odds ratios (ORs) with 95% confidence intervals (CIs) for each directed edge. Because the study is cross-sectional, all directed edges reflect conditional independence relationships and probabilistic co-occurrence rather than established causal mechanisms.

resultsThe final BN contained 46 directed edges; bootstrap resampling confirmed 38 as stable (arc-strength ≥ 0.50). Self-regret (OR = 9.45; 95% CI: 5.78-15.45; arc-strength = 0.71) and psychomotor change (OR = 6.68; 95% CI: 4.03-11.07; arc-strength = 0.44) were the strongest direct probabilistic associates of suicidal thoughts. A co-occurring symptom cluster-sleep problems → anhedonia → depression-was probabilistically linked to elevated suicide risk (depression → self-regret OR = 8.32). Fatigue was probabilistically associated with all-body pain (OR = 2.45) and with self-regret (OR = 3.69), illustrating a potential physical-psychological co-occurrence pattern. Age and gender moderated several associations, with younger women showing the highest conditional probability of the full symptom cluster.

conclusionsBN modelling revealed complex, non-linear probabilistic associations between pain, mental health, and suicidality in a Middle-Eastern population. Because the design is cross-sectional, causal directionality cannot be established; longitudinal validation is needed. The high-risk symptom clusters identified-especially those centred on self-regret and psychomotor disturbance-are hypothesis-generating targets for integrated pain-mental-health screening. Tailoring screening to demographic sub-groups could inform future intervention development in Qatar.