Medication utilization patterns in patients with post-COVID syndrome (PCS): Implications for polypharmacy and drug-drug interactions.

BACKGROUND: Post-COVID syndrome (PCS) causes lasting symptoms like fatigue and cognitive issues. PCS treatment is nonspecific, focusing on symptom management, potentially increasing the risk of polypharmacy. OBJECTIVES: To describe medication use patterns among patients with Post-COVID Syndrome (PCS) and estimate the prevalence of polypharmacy, potential drug-drug interactions, and anticholinergic/sedative burden. METHODS: A cross-sectional analysis of baseline data from the Quebec Action for Post-COVID cohort, consisting of individuals self-identifying with persistent COVID-19 symptoms beyond 12 weeks. Medications were categorized using Anatomical Therapeutic Classification (ATC) codes. Polypharmacy was defined as using 5 or more concurrent medications. The Anticholinergic and Sedative Burden Catalog assessed anticholinergic and sedative loads. The Lexi-Interact checker identified potential drug-drug interactions, which were categorized into 3 severity tiers. RESULTS: Out of 414 respondents, 154 (average age 47.7 years) were prescribed medications related to persistent COVID-19 symptoms. Drugs targeting the nervous system were predominant at 54.5%. The median number of medications was 2, while 11.7% reported polypharmacy. Over half of the participants prescribed medications used at least 1 anticholinergic or sedative medication, and 25% had the potential risk for clinically significant drug-drug interactions, primarily needing therapy monitoring. CONCLUSIONS: Our study reveals prescription patterns for PCS, underscoring the targeted management of nervous system symptoms. The risks associated with polypharmacy, potential drug-drug interactions, and anticholinergic/sedative burden stress the importance of judicious prescribing. While limitations like recall bias and a regional cohort are present, the findings underscore the imperative need for vigilant PCS symptom management.

An umbrella review of the literature on the effectiveness of goal setting interventions in improving health outcomes in chronic conditions.

PURPOSE: To identify the contexts in which goal setting has been used in chronic disease management interventions and to estimate the magnitude of its effect on improvement of health outcomes. METHODS: The strength of evidence and extent of potential bias in the published systematic reviews of goal setting interventions in chronic conditions were summarized using AMSTAR2 quality appraisal tool, number of participants, 95% prediction intervals, and between-study heterogeneity. Components of goal setting interventions were also extracted. RESULTS: Nine publications and 35 meta-analysis models were identified, investigating 25 health outcomes. Of the 35 meta-analyses, none found strong evidence and three provided some suggestive evidence on symptom reduction and perceived well-being. There was weak evidence for effects on eight health outcomes (HbA1c, self-efficacy, depression, anxiety, distress, medication adherence, health-related quality of life and physical activity), with the rest classified as non-significant. Half of the meta-analyses had high level of heterogeneity. CONCLUSION: Goal setting by itself affects outcomes of chronic diseases only to a small degree. This is not unexpected finding as changing outcomes in chronic diseases requires a complex and individualized approach. Implementing goal setting in a standardized way in the management of chronic conditions would seem to be a way forward.IMPLICATIONS FOR REHABILITATIONThe link between goal setting and health outcomes seems to be weak.Some levels of positive behavioural change could be of benefits to patients as seen by improved self-efficacy, patients' satisfaction and overall quality of life.Systematic and consistent application of personalized goal-oriented interventions considering patient's readiness to change could better predict improved outcomes.Incorporation of various goal setting components while actively engaging patient and/or their care givers in the process could appraise how goal setting could help with challenges in faced by people living with chronic conditions in different areas.

The association between anticholinergic/sedative burden and physical frailty in people aging with HIV.

OBJECTIVE: This study aimed to estimate the strength of the association between anticholinergic/sedative burden and concurrent physical frailty in people aging with HIV. DESIGN: This cross-sectional analysis examined baseline data from 824 adults with a mean age of 53 enrolled in the Positive Brain Health Now study. METHODS: Anticholinergic medications were identified using four methods: Anticholinergic Cognitive Burden (ACB) Scale, Anticholinergic Risk Scale (ARS), Anticholinergic Drug Scale (ADS), and the anticholinergic list of the Anticholinergic and Sedative Burden Catalog (ACSBC). Sedatives were identified using the Sedative Load Model (SLM) and the sedative list of the ACSBC. Physical frailty was assessed using a modified Fried Frailty Phenotype (FFP) based on self-report items. Multivariable logistic regression models, adjusted for sociodemographic factors, lifestyle considerations, HIV-related variables, comorbidities, and co-medication use, were used to estimate odds ratios (ORs). RESULTS: Anticholinergic burden demonstrated associations with frailty across various methods: total anticholinergic burden (OR range: 1.22-1.32; 95% confidence interval (CI) range: 1.03-1.66), sedative burden (OR range: 1.18-1.24; 95% CI range: 1.02-1.45), high anticholinergic burden (OR range: 2.12-2.74; 95% CI range: 1.03-6.19), and high sedative burden (OR range: 1.94-2.18; 95% CI: 1.01-4.34). CONCLUSION: The anticholinergic and sedative burdens may represent modifiable risk factors for frailty in people aging with HIV. Future studies should evaluate the effects of reducing anticholinergic and sedative burdens on frailty outcomes and explore the prognostic value of diverse scoring methods.

The Patient Generated Index (PGI) as an early-warning system for predicting brain health challenges: a prospective cohort study for people living with Human Immunodeficiency Virus (HIV).

PURPOSE: In research people are often asked to fill out questionnaires about their health and functioning and some of the questions refer to serious health concerns. Typically, these concerns are not identified until the statistician analyses the data. An alternative is to use an individualized measure, the Patient Generated Index (PGI) where people are asked to self-nominate areas of concern which can then be dealt with in real-time. This study estimates the extent to which self-nominated areas of concern related to mood, anxiety and cognition predict the presence or occurrence of brain health outcomes such as depression, anxiety, psychological distress, or cognitive impairment among people aging with HIV at study entry and for successive assessments over 27 months. METHODS: The data comes from participants enrolled in the Positive Brain Health Now (+ BHN) cohort (n = 856). We analyzed the self-nominated areas that participants wrote on the PGI and classified them into seven sentiment groups according to the type of sentiment expressed: emotional, interpersonal, anxiety, depressogenic, somatic, cognitive and positive sentiments. Tokenization was used to convert qualitative data into quantifiable tokens. A longitudinal design was used to link these sentiment groups to the presence or emergence of brain health outcomes as assessed using standardized measures of these constructs: the Hospital Anxiety and Depression Scale (HADS), the Mental Health Index (MHI) of the RAND-36, the Communicating Cognitive Concerns Questionnaire (C3Q) and the Brief Cognitive Ability Measure (B-CAM). Logistic regressions were used to estimate the goodness of fit of each model using the c-statistic. RESULTS: Emotional sentiments predicted all of the brain health outcomes at all visits with adjusted odds ratios (OR) ranging from 1.61 to 2.00 and c-statistics > 0.73 (good to excellent prediction). Nominating an anxiety sentiment was specific to predicting anxiety and psychological distress (OR 1.65 & 1.52); nominating a cognitive concern was specific to predicting self-reported cognitive ability (OR 4.78). Positive sentiments were predictive of good cognitive function (OR 0.36) and protective of depressive symptoms (OR 0.55). CONCLUSIONS: This study indicates the value of using this semi-qualitative approach as an early-warning system in predicting brain health outcomes.

The Prognostic Utility of Anticholinergic Burden Scales: An Integrative Review and Gap Analysis.

BACKGROUND: Anticholinergic drugs are commonly prescribed, especially to older adults. Anticholinergic burden scales (ABS) have been used to evaluate the cumulative effects of multiple anticholinergics. However, studies have shown inconsistent results regarding the association between anticholinergic burden assessed with ABS and adverse clinical outcomes such as cognitive impairment, functional decline, and frailty. This review aims to identify gaps in research on the development, validation, and evaluation of ABS, and provide recommendations for future studies. METHOD: A comprehensive search of five databases (MEDLINE, Embase, PsychInfo, CINAHL, CENTRAL) was conducted for relevant studies published from inception until 25 May 2023. Two reviewers screened for eligibility and assessed the quality of studies using different tools based on the study design and stage of the review framework. Research evidence was evaluated, and gaps were identified and grouped into evidence, knowledge, and methodological gaps, using evidence tables to summarize data. RESULTS: Several evidence, knowledge, and methodological gaps in existing development, validation, and evaluation studies of ABS were identified. There is no universally accepted scale, and there is a need to define a clinically relevant threshold for measuring total anticholinergic burden. The current evidence has limitations, underrepresenting low- and middle-income countries, younger individuals, and populations with cognitive disabilities. The impact of anticholinergic burden on frailty is also understudied. Existing evaluation studies provide limited evidence on the benefit of reducing anticholinergic burden on clinical outcomes or the safety of anticholinergic deprescribing. There is also uncertainty regarding optimal reduction, clinically significant anticholinergic burden thresholds, and cost effectiveness. CONCLUSIONS: Future research recommendations to bridge knowledge gaps include developing a risk assessment framework, refining ABS scales, establishing a standardized consensus scale, and creating a longitudinal measure of cumulative anticholinergic risk. Strategies to minimize bias, consider frailty, and promote multidisciplinary and multinational collaborations are also necessary to improve patient outcomes.

Translocation of bacterial LPS is associated with self-reported cognitive abilities in men living with HIV receiving antiretroviral therapy.

BACKGROUND: Gut damage allows translocation of bacterial lipopolysaccharide (LPS) and fungal ß-D-glucan (BDG) into the blood. This microbial translocation contributes to systemic inflammation and risk of non-AIDS comorbidities in people living with HIV, including those receiving antiretroviral therapy (ART). We assessed whether markers of gut damage and microbial translocation were associated with cognition in ART-treated PLWH. METHODS: Eighty ART-treated men living with HIV from the Positive Brain Health Now Canadian cohort were included. Brief cognitive ability measure (B-CAM) and 20-item patient deficit questionnaire (PDQ) were administered to all participants. Three groups were selected based on their B-CAM levels. We excluded participants who received proton pump inhibitors or antiacids in the past 3 months. Cannabis users were also excluded. Plasma levels of intestinal fatty acid binding protein (I-FABP), regenerating islet-derived protein 3 α (REG3α), and lipopolysaccharides (LPS = were quantified by ELISA, while 1-3-ß-D-glucan BDG) levels were assessed using the Fungitell assay. Univariable, multivariable, and splines analyses were performed. RESULTS: Plasma levels of I-FABP, REG3α, LPS and BDG were not different between groups of low, intermediate and high B-CAM levels. However, LPS and REG3α levels were higher in participants with PDQ higher than the median. Multivariable analyses showed that LPS association with PDQ, but not B-CAM, was independent of age and level of education. I-FABP, REG3α, and BDG levels were not associated with B-CAM nor PDQ levels in multivariable analyses. CONCLUSION: In this well characterized cohort of ART-treated men living with HIV, bacterial but not fungal translocation was associated with presence of cognitive difficulties. These results need replication in larger samples.

Improving Palliative Care and Medical Assistance in Dying Practice in Canada: How Patients-Partners Could Contribute to Continuing Medical Education.

Medical Assistance in Dying (MAiD) is still considered an evolving practice in Canada. Practitioners are facing the challenge of staying up to date and hence need efficient continuing medical education (CME). A patient-partner has been recently invited as a keynote speaker to CME activities in Canada to share her perspectives and views about patient engagement in palliative care and MAiD practice, calling for compassion. To our knowledge, few data exist on patient-partners' contribution to CME on these topics. Based on that experience, we discuss different issues on patient engagement's contribution in such CME events and call for further research.

"Alone, there is nobody": A qualitative study of the lived experience of loneliness in older men living with HIV.

Loneliness has been shown to be a predictor of poor health and early mortality in the general population. Older men living with human immunodeficiency virus (HIV) are at heightened risk of experiencing loneliness. Here, we aim to describe the lived experience of loneliness in older men living with HIV and identify targets for intervention. We used grounded theory with a theoretical framework of narrative phenomenology to focus data collection and analysis on significant experiences related to loneliness. Based on individual narrative interviews with 10 older men living with HIV, experiences of loneliness related to "multiple losses," "being invisible" and "hiding out" as emergent themes. Participants also described living with loneliness by "finding meaning," "creating social experiences," "pursuing interests and things to 'live for'" and attending events in which "everyone is welcome." The discussion situates experiences of loneliness within the accumulation of losses and stigmas over time and how the participants strategies for living with loneliness could inform interventions to reduce loneliness in older men living with HIV at individual and societal levels.

Machine Learning Approaches to Understand Cognitive Phenotypes in People With HIV.

Cognitive disorders are prevalent in people with HIV (PWH) despite antiretroviral therapy. Given the heterogeneity of cognitive disorders in PWH in the current era and evidence that these disorders have different etiologies and risk factors, scientific rationale is growing for using data-driven models to identify biologically defined subtypes (biotypes) of these disorders. Here, we discuss the state of science using machine learning to understand cognitive phenotypes in PWH and their associated comorbidities, biological mechanisms, and risk factors. We also discuss methods, example applications, challenges, and what will be required from the field to successfully incorporate machine learning in research on cognitive disorders in PWH. These topics were discussed at the National Institute of Mental Health meeting on "Biotypes of CNS Complications in People Living with HIV" held in October 2021. These ongoing research initiatives seek to explain the heterogeneity of cognitive phenotypes in PWH and their associated biological mechanisms to facilitate clinical management and tailored interventions.

Effect of autogenic training on quality of life and symptoms in people living with HIV: A mixed method randomized controlled trial.

OBJECTIVE: Quality of life of people living with HIV is strongly affected by sleep problems, fatigue, pain, anxiety, and depression symptoms. This study set out to evaluate the effects of autogenic training (AT) on quality of life and symptoms within this population. METHODS: A mixed method randomized controlled trial was conducted. Participants were randomly assigned to either the AT group (n = 32) or the control group (CG) (n = 31). Quality of life and symptoms were measured in both groups three times: prior to intervention (T0), immediately after intervention (T3), and three months after intervention (T6). Fourteen individual interviews were conducted. RESULTS: Results show a significant improvement in social and mental dimensions of quality of life for the two groups at T6. They also show a significant improvement in sleep for AT participants at T3. Qualitative results are consistent with quantitative ones. CONCLUSION: AT seems to improve sleep quality and could improve some dimensions of quality of life and other symptoms among people living with HIV. Further studies are needed to confirm these results. TRIAL REGISTRATION NUMBER: NCT01901016.

Development of a self-report measure of cognitive change: assessment of interpretability in two samples, people with HIV and people without HIV.

AIM: The overall aim of this study was to develop a method of measuring change in cognitive ability from the person's perspective. METHODS: Cognitive change items came from an item pool that was used to develop the Communicating Cognitive Concerns Questionnaire (C3Q). The change items were administered to a test sample of 211 people with HIV + and a sample of 484 people drawn from the general population (HIV- sample). Rasch analysis was used to identify items that formed a linear continuum and correlations with measures of related constructs were used to support the interpretability of the new measure. RESULTS: Eleven of the original 12 change items fit the unidimensional Rasch model in both samples with a near similar ordering of the items. The average value for cognitive change of the HIV + sample was greater than that of the HIV- sample. Values on C3Q-Change correlated most highly (> 0.7) with current self-reported cognitive status and measures of depression and anxiety (> 0.6). The lowest correlation was with performance-based cognitive ability (r = 0.2). CONCLUSION: The items of C3Q-Change fit a strong measurement model and related to converging constructs in an expected way. Further work needs to be done to assess the meaning of self-reported cognitive change in relationship to measured change and to examine sources of differential item functioning.

Impact of Hepatitis C Virus Cure on Depressive Symptoms in the Human Immunodeficiency Virus-Hepatitis C Virus Coinfected Population in Canada.

BACKGROUND: Depression is common in people with human immunodeficiency virus (HIV) and hepatitis C virus (HCV), with biological and psychosocial mechanisms at play. Direct acting antivirals (DAA) result in high rates of sustained virologic response (SVR), with minimal side-effects. We assessed the impact of SVR on presence of depressive symptoms in the HIV-HCV coinfected population in Canada during the second-generation DAA era (2013-2020). METHODS: We used data from the Canadian CoInfection Cohort (CCC), a multicenter prospective cohort of people with a HIV and HCV coinfection, and its associated sub-study on food security. Because depression screening was performed only in the sub-study, we predicted Center for Epidemiologic Studies Depression Scale-10 classes in the CCC using a random forest classifier and corrected for misclassification. We included participants who achieved SVR and fit a segmented modified Poisson model using an interrupted time series design, adjusting for time-varying confounders. RESULTS: We included 470 participants; 58% had predicted depressive symptoms at baseline. The median follow-up was 2.4 years (interquartile range [IQR]: 1.0-4.5.) pre-SVR and 1.4 years (IQR: 0.6-2.5) post-SVR. The pre-SVR trend suggested depressive symptoms changed little over time, with no immediate level change at SVR. However, post-SVR trends showed a reduction of 5% per year (risk ratio: 0.95 (95% confidence interval [CI]: .94-.96)) in the prevalence of depressive symptoms. CONCLUSIONS: In the DAA era, predicted depressive symptoms declined over time following SVR. These improvements reflect possible changes in biological pathways and/or better general health. If such improvements in depression symptoms are durable, this provides an additional reason for treatment and early cure of HCV.

Effectiveness of a personalized health profile on specificity of self-management goals among people living with HIV in Canada: findings from a blinded pragmatic randomized controlled trial.

PURPOSE: To estimate among people living with chronic HIV, to what extent providing feedback on their health outcomes will affect the number and specificity of patient-formulated self-management goals. METHODS: A personalized feedback profile was produced for individuals enrolled in a Canadian HIV Brain Health Now study. Goal specificity was measured by total number of specific words (matched to a domain-specific developed lexicon) per person-words using text mining techniques. RESULTS: Of 176 participants enrolled and randomly assigned to feedback and control groups, 110 responses were received. The average number of goals was similar for both groups (3.7 vs 3.9). The number of specific words used in the goals formulated by the feedback and control group were 642 and 739, respectively. Specific nouns and actionable verbs were present to some extent and "measurable" and "time-bound" words were mainly missing. Negative binomial regression showed no difference in goal specificity among groups (RR = 0.93, 95% CI 0.78-1.10). Goals set by both groups overlapped in 8 areas and had little difference in rank. CONCLUSION: Personalized feedback profile did not help with formulation of high-quality goals. Text mining has the potential to help with difficulties of goal evaluation outside of the face-to-face setting. With more data and use of learning models automated answers could be generated to provide a more dynamic platform.

Safety and Tolerability of Oral Cannabinoids in People Living with HIV on Long-Term ART: A Randomized, Open-Label, Interventional Pilot Clinical Trial (CTNPT 028).

BACKGROUND: With anti-inflammatory properties, cannabinoids may be a potential strategy to reduce immune activation in people living with HIV (PLWH) but more information on their safety and tolerability is needed. METHODS: We conducted an open-label interventional pilot study at the McGill University Health Centre in Montreal, Canada. PLWH were randomized to oral Δ9-tetrahydrocannabinol (THC): cannabidiol (CBD) combination (THC 2.5 mg/CBD 2.5 mg) or CBD-only capsules (CBD 200 mg). Individuals titrated doses as tolerated to a maximum daily dose THC 15 mg/CBD 15 mg or 800 mg CBD, respectively, for 12 weeks. The primary outcome was the percentage of participants without any significant toxicity based on the WHO toxicity scale (Grades 0-2 scores). RESULTS: Out of ten individuals, eight completed the study. Two from the CBD-only arm were withdrawn for safety concerns: phlebotomy aggravating pre-existing anemia and severe hepatitis on 800 mg CBD with newly discovered pancreatic adenocarcinoma, respectively. Seven did not have any significant toxicity. Cannabinoids did not alter hematology/biochemistry profiles. CD4 count, CD4/CD8 ratio, and HIV suppression remained stable. Most adverse effects were mild-moderate. CONCLUSIONS: In PLWH, cannabinoids seem generally safe and well-tolerated, though larger studies are needed. Screening for occult liver pathology should be performed and hepatic enzymes monitored, especially with high CBD doses.

Development of a Prototype for a Bilingual Patient-Reported Outcome Measure of the Important Health Aspects of Quality of Life in People Living with HIV: The Preference Based HIV Index (PB-HIV).

(1) Background: The aim of this project was to develop a short, HIV-specific, health-related quality of life measure with a scoring system based on patient preferences for the different dimensions of the Preference-Based HIV Index (PB-HIV). (2) Methods: This study is a cross-sectional analysis of data from the Canadian Positive Brain Health Now cohort (n = 854; mean age 53 years). Items from the standardized measures were mapped to the areas from the Patient-Generated Index and formed the domains. A Rasch analysis was used to identify the best performing item to represent each dimension. Each item was then regressed on self-rated health (scored 0 to 100) and the regression parameters were used as scaling weights to form an index score for the prototype measure. (3) Results: Seven independent dimensions with three declarative statements ordered as response options formed the PB-HIV Index (pain, fatigue, memory/concentration, sleep, physical appearance/body image, depression, motivation). Regression parameters from a multivariable model yielded a measure with a scoring range from 0 (worst health) to 100 (perfect health). (4) Conclusions: Preference-based measures are optimal, as the total score reflects gains in some dimensions balanced against losses in others. The PB-HIV Index is the first HIV-specific preference-based measure.

Efficacy potential of Goal Management Training to improve cognitive function in older people living with HIV.

Goal Management Training® (GMT) teaches strategies to reduce cognitive load and improve focus in everyday tasks. The aim of this study was to ascertain feasibility, acceptability, and efficacy potential of GMT for people (≥50 years) with stable HIV infection scoring low on tests of cognitive ability. A two-sample, parallel, controlled trial was carried out. Feasibility was demonstrated, as 21/30 participants in the GMT group attended ≥8 of the 9 sessions and completed at least half of the homework. There was no change on the primary performance-based cognitive outcomes in the GMT group or in the control group (n = 23). There was a meaningful improvement in self-reported cognition in those adherent to the intervention. GMT is a promising intervention for people aging with HIV who are dealing with cognitive difficulties affecting their everyday life and should be further investigated.

Relationship between reward-related evoked potentials and real-world motivation in older people living with human immunodeficiency virus.

Apathy, a clinical disorder characterized by low motivation, is prevalent in people living with Human Immunodeficiency Virus (HIV). It affects mental and physical health-related quality-of-life, medication adherence, and is associated with cognitive decline. However, the causes of apathy and the underlying brain mechanisms in HIV are unknown. Brain responses to reward may be relevant to understanding apathy and might serve as biomarkers for diagnosis or treatment response. Electroencephalogram (EEG) responses to gain and loss feedback in simple guessing tasks have been related to apathy in neurodegenerative conditions and healthy individuals. The primary aim of this study is to contribute evidence regarding the relationship between two EEG correlates of reward processing, the Reward Positivity, and the Feedback-P300, and real-world motivated behavior indicated by self-reported hours engaged in goal-directed leisure activities per week, in older individuals with well-controlled HIV infection. High-density EEG was collected from 75 participants while they performed a guessing task with gain or loss feedback. We found that a later component of reward processing, the Feedback-P300, was related to real-world engagement, while the earlier Reward Positivity was not. The Feedback-P300 measured with EEG holds promise as a biomarker for motivated behavior in older people living with HIV. These findings lay the groundwork for a better understanding of the neurobiology of apathy in this condition.

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.

Predicting the presence of depressive symptoms in the HIV-HCV co-infected population in Canada using supervised machine learning.

BACKGROUND: Depression is common in the human immunodeficiency virus (HIV)-hepatitis C virus (HCV) co-infected population. Demographic, behavioural, and clinical data collected in research settings may be of help in identifying those at risk for clinical depression. We aimed to predict the presence of depressive symptoms indicative of a risk of depression and identify important classification predictors using supervised machine learning. METHODS: We used data from the Canadian Co-infection Cohort, a multicentre prospective cohort, and its associated sub-study on Food Security (FS). The Center for Epidemiologic Studies Depression Scale-10 (CES-D-10) was administered in the FS sub-study; participants were classified as being at risk for clinical depression if scores ≥ 10. We developed two random forest algorithms using the training data (80%) and tenfold cross validation to predict the CES-D-10 classes-1. Full algorithm with all candidate predictors (137 predictors) and 2. Reduced algorithm using a subset of predictors based on expert opinion (46 predictors). We evaluated the algorithm performances in the testing data using area under the receiver operating characteristic curves (AUC) and generated predictor importance plots. RESULTS: We included 1,934 FS sub-study visits from 717 participants who were predominantly male (73%), white (76%), unemployed (73%), and high school educated (52%). At the first visit, median age was 49 years (IQR:43-54) and 53% reported presence of depressive symptoms with CES-D-10 scores ≥ 10. The full algorithm had an AUC of 0.82 (95% CI:0.78-0.86) and the reduced algorithm of 0.76 (95% CI:0.71-0.81). Employment, HIV clinical stage, revenue source, body mass index, and education were the five most important predictors. CONCLUSION: We developed a prediction algorithm that could be instrumental in identifying individuals at risk for depression in the HIV-HCV co-infected population in research settings. Development of such machine learning algorithms using research data with rich predictor information can be useful for retrospective analyses of unanswered questions regarding impact of depressive symptoms on clinical and patient-centred outcomes among vulnerable populations.