Tree-Based Machine Learning to Identify Predictors of Psoriasis Incidence at the Neighborhood Level: A Populational Study from Quebec, Canada.

BACKGROUND: Psoriasis is a major global health burden affecting ~ 60 million people worldwide. Existing studies on psoriasis focused on individual-level health behaviors (e.g. diet, alcohol consumption, smoking, exercise) and characteristics as drivers of psoriasis risk. However, it is increasingly recognized that health behavior arises in the context of larger social, cultural, economic and environmental determinants of health. We aimed to identify the top risk factors that significantly impact the incidence of psoriasis at the neighborhood level using populational data from the province of Quebec (Canada) and advanced tree-based machine learning (ML) techniques. METHODS: Adult psoriasis patients were identified using International Classification of Disease (ICD)-9/10 codes from Quebec (Canada) populational databases for years 1997-2015. Data on environmental and socioeconomic factors 1 year prior to psoriasis onset were obtained from the Canadian Urban Environment Health Consortium (CANUE) and Statistics Canada (StatCan) and were input as predictors into the gradient boosting ML. Model performance was evaluated using the area under the curve (AUC). Parsimonious models and partial dependence plots were determined to assess directionality of the relationship. RESULTS: The incidence of psoriasis varied geographically from 1.6 to 325.6/100,000 person-years in Quebec. The parsimonious model (top 9 predictors) had an AUC of 0.77 to predict high psoriasis incidence. Amongst top predictors, ultraviolet (UV) radiation, maximum daily temperature, proportion of females, soil moisture, urbanization, and distance to expressways had a negative association with psoriasis incidence. Nighttime light brightness had a positive association, whereas social and material deprivation indices suggested a higher psoriasis incidence in the middle socioeconomic class neighborhoods. CONCLUSION: This is the first study to highlight highly variable psoriasis incidence rates on a jurisdictional level and suggests that living environment, notably climate, vegetation, urbanization and neighborhood socioeconomic characteristics may have an association with psoriasis incidence.

Exposure to silica and systemic sclerosis: A retrospective cohort study based on the Canadian Scleroderma Research Group.

Introduction: Systemic sclerosis (SSc) is thought to be induced by an environmental trigger in genetically predisposed individuals. This study assessed the demographic and clinical characteristics and disease severity of silica exposed SSc patients. Methods: Data was obtained from the Canadian Scleroderma Research Group (CSRG) cohort, containing 1,439 patients (2004-2019). Univariate and multivariate logistic regression analyses were performed, to determine the phenotype and severity of silica-exposed SSc patients. Mortality was assessed using Cox Survival Regression and Kaplan-Meier analyses. Results: Among 1,439 patients (86.7% females), 95 patients reported exposure to silica. Those exposed were younger, of male sex and with more severe disease. Sex differences were observed where male patients exposed to silica were more likely to be Caucasian and smokers whereas female patients were younger at SSc diagnosis compared to unexposed. Multivariate regression, controlled for multiple confounders, showed that silica exposure was associated with a younger age at diagnosis and worse disease severity and mortality. Conclusion: Exposure to silica was reported in ∼7% of CSRG cohort and ∼20% of male patients and was associated with a worse prognosis in terms of age of diagnosis, organ involvement and mortality. Hence, screening for silica exposure among higher risk individuals may be beneficial and these patients may require closer monitoring for systemic disease.

Trajectories of systemic agent use and associated depression- and anxiety-related health care costs among patients with psoriasis.

Background: Systemic treatment patterns and related mental health disorders and economic burden among patients with psoriasis are largely unknown. Objective: To assess systemic treatment patterns and associated depression and anxiety-related health care costs among patients with psoriasis initiating a conventional systemic treatment (CST). Methods: Using a retrospective cohort design with sequence and cluster analyses, we assessed systemic treatment trajectories (CST and tumor necrosis factor inhibitors or ustekinumab, [TNFi/UST]) over a 2-year period following CST initiation. We compared health care costs between trajectories using 2-part models. Results: We included 781 patients and identified 8 trajectories: persistent methotrexate users, persistent acitretin users, early CST discontinuation, late methotrexate discontinuation, switch to TNFi/UST, adding TNFi/UST, discontinuation then restart on methotrexate, and discontinuation then restart on acitretin or multiple CST switches. Overall, 165 (21%) patients incurred depression- and anxiety-related health care costs (median annual cost, CAN$56; quartiles, $14-$127). Compared with persistent methotrexate users, adding a TNFi/UST (cost ratio, 3.63; 95% CI, 1.47-5.97) and discontinuation then restart on acitretin or multiple switches between systemic agents (cost ratio, 13.3; 95% CI 5.76-22.47) had higher costs. Limitations: Trajectory misclassification may have occured. These date represent an association, and causality cannot be inferred, particularly given the risk of confounding. Conclusion: Depression- and anxiety-related health care costs were high among patients adding TNFi/UST and those discontinuing then restarting on acitretin or experiencing multiple switches between systemic agents.

Sex Differences in the Patterns of Systemic Agent use Among Patients With Psoriasis: A Retrospective Cohort Study in Quebec, Canada.

Background: Sex differences exist in psoriasis manifestation and expectations from treatment with systemic agents, including, conventional systemic agents (CSA) and tumor necrosis factor inhibitors or ustekinumab (TNFi/UST). However, sex differences in patterns of systemic agent use, such as CSA discontinuation and switch from CSA to TNFi/UST have not been examined. Objectives: To assess sex differences in patterns of CSA use and identify factors associated with switch to (or add) a TNFi/UST and those associated with CSA discontinuation. Methods: We conducted a retrospective cohort study using the Quebec health administrative databases. We included patients with psoriasis initiating a CSA in 2002-2015. We excluded patients with a psoriasis diagnosis in the 3 years prior to the first diagnosis date between 2002 and 2015, and those with a systemic agent dispensation in the year prior to that date. We used Cox regression models with the Least Absolute Shrinkage and Selection Operator method to identify factors associated with Switch/add TNFi/UST, and those associated with CSA discontinuation. Separate analyses were performed for male and female patients. Results: We included 1,644 patients (55.7% females, mean age 60.3 years), among whom 60.4% discontinued their CSA and 7.4%, switched/added TNFi/UST (3.4% switched and 4.0% added) within a median of 0.78 years of follow-up. Among male and female patients, rates of Switch/add TNFi/UST per 1,000 person-year were 49.1 and 41.0 and rates of CSA discontinuation were 381.2 and 352.8. Clinical obesity in male patients (HR 3.53, 95% CI 1.20-10.35), and adjustment/somatoform/dissociative disorders (HR 3.17, 95% CI 1.28-7.85) and use of nonsteroidal anti-inflammatory drugs (HR 2.70, 95% CI 1.56-4.70) in female patients were associated with Switch/add TNFi/UST. Male patients followed by a rheumatologist (HR 0.66, 95% CI 0.46-0.94) and those with a prior hospitalization (HR 0.70, 95% CI 0.57-0.87) were at lower risk of CSA discontinuation, while those initiated on acitretin (vs methotrexate) were at higher risk to discontinue their CSA (HR 1.61, 95% CI 1.30-2.01). Female patients with rheumatoid arthritis comorbidity (HR 0.69, 95% CI 0.51-0.93), those with a dispensed lipid-lowering agent (HR 0.72, 95% CI 0.59-0.88) and hypoglycemic agent (HR 0.75, 95% CI 0.57-0.98) and those initiated on methotrexate (vs all other CSAs) were less likely to discontinue their CSA. Male and female patients entering the cohort between 2011 and 2015 were at reduced risk of CSA discontinuation compared to those entering the cohort before 2011. Conclusion: Most male and female patients discontinued their CSA within 1 year of follow-up. Our study highlighted sex differences in patients' characteristics associated with switch/add a TNFi/UST and CSA discontinuation; treatment switch and discontinuation may be indications of treatment failure in most patients.

The association between side effects and adherence to antidepressants among primary care community-dwelling older adults.

Aims: To evaluate the association between adherence to antidepressants and self-reported side effects while considering their tolerability among primary care community-dwelling older adults.Methods: This is a secondary analysis of data available for 137 individuals participating in the third wave of the Étude sur la Santé des Aînés - Services study (2015-2017) conducted among older adults aged 65 years and older. Adherence to antidepressants was assessed with the 4-item Medication Assessment Questionnaire. Side effects were also self-reported from a list of 20 potential side effects related to antidepressant use, while considering their tolerability (tolerable and non-tolerable) and were then grouped into seven categories specific to organ or function systems. Multilevel logistic regression analyses were carried out to assess the association between adherence and the presence of side effects accounting for participants nested within primary health clinics.Results: In this study, 69.3% of participants were adherent to their antidepressants and 30.7% were non-adherent. Participants reporting sleep disturbance (OR = 0.58, 95% CI = 0.47-0.72), gastrointestinal system (OR = 0.64, 95% CI = 0.45-0.92), and nervous system (OR = 0.60, 95% CI = 0.48-0.78) related side effects were less likely to be adherent to their antidepressants. Participants reporting palpitations were more likely to adhere to antidepressants (OR = 2.20, 95% CI = 1.03-4.67). With regards to severity, participants reporting non-tolerable nervous system related side effects were less likely to be adherent (OR = 0.37, 95% CI = 0.26-0.53) and those reporting non-tolerable gastrointestinal system related side effects reported higher adherence to antidepressants (OR = 1.82, 95% CI = 1.08-3.08).Conclusion: Adherence to antidepressants was associated with side effects, and more precisely with sleep disturbance, gastrointestinal and nervous systems side effects.

Patient and health system factors associated with hospital readmission in older adults without cognitive impairment.

OBJECTIVE: To study the factors associated with hospital readmission. METHODS: Data used in this study came from a population-based survey of older adults without cognitive impairment. Cox regression was used to assess the factors associated with readmission within a 2-year follow-up period. According to Andersen's model of healthcare seeking behavior, study variables considered included predisposing, enabling and need factors at the individual and health system levels. RESULTS: Of the 433 participants with an index hospitalization, 97% were discharged with a physical and 3% with a psychiatric disorder. During follow-up, 29% (128/433) were readmitted with a median time to readmission reaching 83 days. The risk of readmission was associated with the following: age, marital status, attraction index of the region of residence for psychiatric services, the presence of an anxio-depressive and other mental disorder, as well as a disorder of the musculoskeletal system. The presence of a physical and psychiatric comorbidity was also associated with readmission. CONCLUSIONS: Post-discharge follow-up of vulnerable populations with a history of mental disorders and improved availability of psychiatric services in the community are associated with a reduced risk of readmission.

Out-of-pocket costs and adherence to antihypertensive agents among older adults covered by the public drug insurance plan in Quebec.

OBJECTIVE: To evaluate the effect of patient out-of-pocket costs on adherence to antihypertensive agents (AHA) in community-dwelling older adults covered by the public drug insurance plan in Quebec. METHODS: This is a secondary analysis of data from the "Étude sur la santé des aînés" study (2005-2008) on community-dwelling older adults in Quebec aged 65 years and older (N=2,811). The final sample included 881 participants diagnosed with arterial hypertension and treated with AHA. Medication adherence was measured with the proportion of days covered over a 2-year follow-up period (<80% and ≥80%). Out-of-pocket costs for AHA, in Canadian dollars (CAD), at cohort entry were categorized as follows: $0, $0.01-$5.00, $5.01-$10.00, $10.01-$15.00 and $15.01-$36.00. Multivariable logistic regression models were constructed to study adherence to AHA as a function of out-of-pocket costs while controlling for several confounders. Models were also stratified by annual household income (<$15,000 CAD and ≥$15,000 CAD). RESULTS: In this study, 80.8% of participants were adherent to their AHA. Among participants reporting an annual household income <$15,000 CAD, those with an out-of-pocket cost of $10.01-$15.00 CAD were significantly less adherent to their AHA than those with no contribution (OR =0.175, 95% CI: 0.042-0.740). Among participants reporting an income of ≥$15,000 CAD, those with out-of-pocket costs of $0.01-$5.00 CAD (OR =0.194; 95% CI: 0.048-0.787), $5.01-$10.00 CAD (OR =0.146; 95% CI: 0.036-0.589), $10.01-$15.00 CAD (OR =0.192; 95% CI: 0.047-0.777) and $15.01-$36.00 CAD (OR =0.160, 95% CI: 0.039-0.655) were significantly less adherent to their AHA than participants with no contribution. CONCLUSION: Increased out-of-pocket costs are associated with non-adherence to AHA in older adults covered by a public drug insurance plan, more importantly in those reporting an annual household income ≥$15,000 CAD. A reduction in the amount of out-of-pocket costs and yearly maximum contribution for drugs may improve adherence to treatment.