Disparities in severe infection risk associated with SLE: evidence from a causal forest analysis

Severe infections are a primary cause of morbidity and premature mortality in patients with SLE. Although SLE patients are known to have an elevated average risk of severe infection, little is known about how this excess risk varies by patient characteristics. Population-average estimates may mask important clinical heterogeneity. Identifying factors that place certain patients at disproportionately higher risk is essential for personalized risk assessment. The aim of this study was to quantify the heterogeneity in the excess severe infection rate due to SLE and to identify the key patient-level clinical and demographic risk modifiers that drive this heterogeneity.

We conducted a population-based matched-cohort study using health administrative data from British Columbia, Canada (1990-2024). We identified 10 517 incident SLE patients using a validated algorithm and matched them 1:5-52 585 non-SLE controls based on birth year, sex and index year. To measure the heterogeneity of infection risk, we employed causal forests to estimate personalized excess infection risk for each SLE patient. Generalized additive models (GAMs) were used to characterize the potentially non-linear relationships between patient profiles and the estimated excess risk of severe infection.

Using causal forests, the estimated individual-level excess risk of severe infection among SLE patients averaged 21.12 per 1000 person-years (SE = 3.14), with substantial variation across individuals (SD = 19.73 per 1000 person-years). Variable importance analyses highlight baseline number of outpatient visits, number of hospitalizations, Charlson comorbidity index, age, rural residence, hypertension and income as key effect modifiers. GAMs analyses reveal higher excess risk is associated with prior infection, rural residence, hypertension, cardiovascular medication, congestive heart failure, glucocorticoid use and depression. Higher income and oral contraceptive pill/hormone replacement therapy are associated with lower excess infection risk.

Significant disparities exist in the excess infection risk associated with SLE. Focusing only on population-average risk can obscure vulnerable subgroups with disproportionately high infection rates. Modern causal machine-learning methods can support personalized risk stratification, helping clinicians move beyond average risk estimates towards targeted surveillance and prevention for high-risk SLE patients.