Semi-supervised mixture multi-source exchangeability model for leveraging real-world data in clinical trials.

The traditional trial paradigm is often criticized as being slow, inefficient, and costly. Statistical approaches that leverage external trial data have emerged to make trials more efficient by augmenting the sample size. However, these approaches assume that external data are from previously conducted trials, leaving a rich source of untapped real-world data (RWD) that cannot yet be effectively leveraged. We propose a semi-supervised mixture (SS-MIX) multisource exchangeability model (MEM); a flexible, two-step Bayesian approach for incorporating RWD into randomized controlled trial analyses. The first step is a SS-MIX model on a modified propensity score and the second step is a MEM. The first step targets a representative subgroup of individuals from the trial population and the second step avoids borrowing when there are substantial differences in outcomes among the trial sample and the representative observational sample. When comparing the proposed approach to competing borrowing approaches in a simulation study, we find that our approach borrows efficiently when the trial and RWD are consistent, while mitigating bias when the trial and external data differ on either measured or unmeasured covariates. We illustrate the proposed approach with an application to a randomized controlled trial investigating intravenous hyperimmune immunoglobulin in hospitalized patients with influenza, while leveraging data from an external observational study to supplement a subgroup analysis by influenza subtype.

The Effect of HIV Treatment Interruption on Subsequent Immunological Response.

Recovery of CD4-positive T lymphocyte count after initiation of antiretroviral therapy (ART) has been thoroughly examined among people with human immunodeficiency virus infection. However, immunological response after restart of ART following care interruption is less well studied. We compared CD4 cell-count trends before disengagement from care and after ART reinitiation. Data were obtained from the East Africa International Epidemiology Databases to Evaluate AIDS (IeDEA) Collaboration (2001-2011; n = 62,534). CD4 cell-count trends before disengagement, during disengagement, and after ART reinitiation were simultaneously estimated through a linear mixed model with 2 subject-specific knots placed at the times of disengagement and treatment reinitiation. We also estimated CD4 trends conditional on the baseline CD4 value. A total of 10,961 patients returned to care after disengagement from care, with the median gap in care being 2.7 (interquartile range, 2.1-5.4) months. Our model showed that CD4 cell-count increases after ART reinitiation were much slower than those before disengagement. Assuming that disengagement from care occurred 12 months after ART initiation and a 3-month treatment gap, CD4 counts measured at 3 years since ART initiation would be lower by 36.5 cells/µL than those obtained under no disengagement. Given that poorer CD4 restoration is associated with increased mortality/morbidity, specific interventions targeted at better retention in care are urgently required.

Remdesivir for the Treatment of Covid-19 - Final Report.

BACKGROUND: Although several therapeutic agents have been evaluated for the treatment of coronavirus disease 2019 (Covid-19), no antiviral agents have yet been shown to be efficacious. METHODS: We conducted a double-blind, randomized, placebo-controlled trial of intravenous remdesivir in adults who were hospitalized with Covid-19 and had evidence of lower respiratory tract infection. Patients were randomly assigned to receive either remdesivir (200 mg loading dose on day 1, followed by 100 mg daily for up to 9 additional days) or placebo for up to 10 days. The primary outcome was the time to recovery, defined by either discharge from the hospital or hospitalization for infection-control purposes only. RESULTS: A total of 1062 patients underwent randomization (with 541 assigned to remdesivir and 521 to placebo). Those who received remdesivir had a median recovery time of 10 days (95% confidence interval [CI], 9 to 11), as compared with 15 days (95% CI, 13 to 18) among those who received placebo (rate ratio for recovery, 1.29; 95% CI, 1.12 to 1.49; P<0.001, by a log-rank test). In an analysis that used a proportional-odds model with an eight-category ordinal scale, the patients who received remdesivir were found to be more likely than those who received placebo to have clinical improvement at day 15 (odds ratio, 1.5; 95% CI, 1.2 to 1.9, after adjustment for actual disease severity). The Kaplan-Meier estimates of mortality were 6.7% with remdesivir and 11.9% with placebo by day 15 and 11.4% with remdesivir and 15.2% with placebo by day 29 (hazard ratio, 0.73; 95% CI, 0.52 to 1.03). Serious adverse events were reported in 131 of the 532 patients who received remdesivir (24.6%) and in 163 of the 516 patients who received placebo (31.6%). CONCLUSIONS: Our data show that remdesivir was superior to placebo in shortening the time to recovery in adults who were hospitalized with Covid-19 and had evidence of lower respiratory tract infection. (Funded by the National Institute of Allergy and Infectious Diseases and others; ACTT-1 ClinicalTrials.gov number, NCT04280705.).

Joint modeling of longitudinal and competing-risk data using cumulative incidence functions for the failure submodels accounting for potential failure cause misclassification through double sampling.

Most of the literature on joint modeling of longitudinal and competing-risk data is based on cause-specific hazards, although modeling of the cumulative incidence function (CIF) is an easier and more direct approach to evaluate the prognosis of an event. We propose a flexible class of shared parameter models to jointly model a normally distributed marker over time and multiple causes of failure using CIFs for the survival submodels, with CIFs depending on the "true" marker value over time (i.e., removing the measurement error). The generalized odds rate transformation is applied, thus a proportional subdistribution hazards model is a special case. The requirement that the all-cause CIF should be bounded by 1 is formally considered. The proposed models are extended to account for potential failure cause misclassification, where the true failure causes are available in a small random sample of individuals. We also provide a multistate representation of the whole population by defining mutually exclusive states based on the marker values and the competing risks. Based solely on the assumed joint model, we derive fully Bayesian posterior samples for state occupation and transition probabilities. The proposed approach is evaluated in a simulation study and, as an illustration, it is fitted to real data from people with HIV.

Contribution of alcohol use in HIV/hepatitis C virus co-infection to all-cause and cause-specific mortality: A collaboration of cohort studies.

Among persons with HIV (PWH), higher alcohol use and having hepatitis C virus (HCV) are separately associated with increased morbidity and mortality. We investigated whether the association between alcohol use and mortality among PWH is modified by HCV. Data were combined from European and North American cohorts of adult PWH who started antiretroviral therapy (ART). Self-reported alcohol use data, collected in diverse ways between cohorts, were converted to grams/day. Eligible PWH started ART during 2001-2017 and were followed from ART initiation for mortality. Interactions between the associations of baseline alcohol use (0, 0.1-20.0, >20.0 g/day) and HCV status were assessed using multivariable Cox models. Of 58,769 PWH, 29,711 (51%), 23,974 (41%) and 5084 (9%) self-reported alcohol use of 0 g/day, 0.1-20.0 g/day, and > 20.0 g/day, respectively, and 4799 (8%) had HCV at baseline. There were 844 deaths in 37,729 person-years and 2755 deaths in 443,121 person-years among those with and without HCV, respectively. Among PWH without HCV, adjusted hazard ratios (aHRs) for mortality were 1.18 (95% CI: 1.08-1.29) for 0.0 g/day and 1.84 (1.62-2.09) for >20.0 g/day compared with 0.1-20.0 g/day. This J-shaped pattern was absent among those with HCV: aHRs were 1.00 (0.86-1.17) for 0.0 g/day and 1.64 (1.33-2.02) for >20.0 g/day compared with 0.1-20.0 g/day (interaction p < .001). Among PWH without HCV, mortality was higher in both non-drinkers and heavy drinkers compared with moderate alcohol drinkers. Among those with HCV, mortality was higher in heavy drinkers but not non-drinkers, potentially due to differing reasons for not drinking (e.g. illness) between those with and without HCV.

Issues with the expected information matrix of linear mixed models provided by popular statistical packages under missingness at random dropout.

Likelihood-based methods ignoring missingness at random (MAR) produce consistent estimates provided that the whole likelihood model is correct. However, the expected information matrix (EIM) depends on the missingness mechanism. It has been shown that calculating the EIM by considering the missing data pattern as fixed (naive EIM) is incorrect under MAR, but the observed information matrix (OIM) is valid under any MAR missingness mechanism. In longitudinal studies, linear mixed models (LMMs) are routinely applied, often without any reference to missingness. However, most popular statistical packages currently provide precision measures for the fixed effects by inverting only the corresponding submatrix of the OIM (naive OIM), which is effectively equivalent to the naive EIM. In this paper, we analytically derive the correct form of the EIM of LMMs under MAR dropout to compare its differences with the naive EIM, which clarifies why the naive EIM fails under MAR. The asymptotic coverage rate of the naive EIM is numerically calculated for two parameters (population slope and slope difference between two groups) under various dropout mechanisms. The naive EIM can severely underestimate the true variance, especially when the degree of MAR dropout is high. Similar trends emerge under misspecified covariance structure, where, even the full OIM may lead to incorrect inferences and sandwich/bootstrap estimators are generally required. Results from simulation studies and application to real data led to similar conclusions. In LMMs, the full OIM should be preferred to the naive EIM/OIM, though if misspecified covariance structure is suspected, robust estimators should be used.

Prevalence of rheumatoid arthritis in Greece: results from the national health examination survey EMENO.

Rheumatoid arthritis (RA) is considered the most common form of autoimmune arthritis. The disease's prevalence is around 0.5-1% worldwide, but it seems to vary among different populations. The aim of this study was to estimate the prevalence of self-reported diagnosed RA in the general adult population in Greece. The data were derived from the Greek Health Examination Survey EMENO, a population-based survey performed between 2013 and 2016. Of the 6006 participants (response rate 72%), 5884 were eligible for this study. Prevalence estimates were calculated according to the study design. Prevalence of self-reported RA was estimated to be overall 0.5% (95% CI 0.4-0.7) being approximately three times higher in women than in men (0.7% vs 0.2%, p value = 0.004). A decrease in the prevalence of RA was observed in urban areas of the country. In contrast, higher disease rates were reported in individuals with lower socioeconomic status. Multivariable regression analysis showed that gender, age, and income were related to the occurrence of the disease. Osteoporosis and thyroid disease were the two comorbidities observed at statistically significant higher rates in individuals with self-reported RA. The prevalence of self-reported RA in Greece is similar to that reported in other European countries. Gender, age, and income are the main factors related to the disease's prevalence in Greece.

Individual risk prediction: Comparing random forests with Cox proportional-hazards model by a simulation study.

With big data becoming widely available in healthcare, machine learning algorithms such as random forest (RF) that ignores time-to-event information and random survival forest (RSF) that handles right-censored data are used for individual risk prediction alternatively to the Cox proportional hazards (Cox-PH) model. We aimed to systematically compare RF and RSF with Cox-PH. RSF with three split criteria [log-rank (RSF-LR), log-rank score (RSF-LRS), maximally selected rank statistics (RSF-MSR)]; RF, Cox-PH, and Cox-PH with splines (Cox-S) were evaluated through a simulation study based on real data. One hundred eighty scenarios were investigated assuming different associations between the predictors and the outcome (linear/linear and interactions/nonlinear/nonlinear and interactions), training sample sizes (500/1000/5000), censoring rates (50%/75%/93%), hazard functions (increasing/decreasing/constant), and number of predictors (seven, 15 including noise variables). Methods' performance was evaluated with time-dependent area under curve and integrated Brier score. In all scenarios, RF had the worst performance. In scenarios with a low number of events (⩽70), Cox-PH was at least noninferior to RSF, whereas under linearity assumption it outperformed RSF. Under the presence of interactions, RSF performed better than Cox-PH as the number of events increased whereas Cox-S reached at least similar performance with RSF under nonlinear effects. RSF-LRS performed slightly worse than RSF-LR and RSF-MSR when including noise variables and interaction effects. When applied to real data, models incorporating survival time performed better. Although RSF algorithms are a promising alternative to conventional Cox-PH as data complexity increases, they require a higher number of events for training. In time-to-event analysis, algorithms that consider survival time should be used.

Hepatitis B, C and human immunodeficiency virus knowledge among the general greek population: results from the Hprolipsis nationwide survey.

BACKGROUND: Although several studies on hepatitis B (HBV), C (HCV) and human immunodeficiency virus (HIV) infection have been conducted in Greece, little is known on the knowledge level of the Greek population towards these three infections. Our aim was to assess the knowledge level of the adult Greek general population about the HBV, HCV and HIV. METHODS: Data were derived from the first general population health survey, Hprolipsis. The sample was selected by multistage stratified random sampling. A standardized questionnaire was administered by trained interviewers during home visits. A knowledge score was constructed based on responses to 17 per infection selected items and categorized in three levels; high (12-17 correct replies) medium (6-11) and low (0-5). Among 8,341 eligible individuals, 6,006 were recruited (response rate: 72%) and 5,878 adults (≥ 18 years) were included in the analysis. The statistical analysis accounted for the study design. RESULTS: Only 30.4%, 21.6%, and 29.6% of the participants had a high overall knowledge level of HBV, HCV and HIV, respectively. These low percentages were mainly attributed to the high levels of misconception about transmission modes (65.9%, 67.2%, and 67.9%, respectively). Results showed that increasing age and living out of the big metropolitan cities were associated with decreased odds of having higher knowledge. Female gender, higher education level, higher monthly family income, higher medical risk score, history of testing and being born in Greece or Cyprus, were associated with increased odds of having higher knowledge. CONCLUSIONS: There are significant knowledge gaps in the Greek general population regarding modes of transmission, preventive measures and treatment availability for HBV, HCV and HIV. There is an urgent need for large scale but also localized awareness activities targeted to less privileged populations, to fill the gaps in knowledge and increase population engagement in preventive measures.

Human Immunotypes Impose Selection on Viral Genotypes Through Viral Epitope Specificity.

BACKGROUND: Understanding the genetic interplay between human hosts and infectious pathogens is crucial for how we interpret virulence factors. Here, we tested for associations between HIV and host genetics, and interactive genetic effects on viral load (VL) in HIV-positive antiretroviral treatment-naive clinical trial participants. METHODS: HIV genomes were sequenced and the encoded amino acid (AA) variants were associated with VL, human single nucleotide polymorphisms (SNPs), and imputed HLA alleles using generalized linear models with Bonferroni correction. RESULTS: Human (388 501 SNPs) and HIV (3010 variants) genetic data were available for 2122 persons. Four HIV variants were associated with VL (P < 1.66 × 10-5). Twelve HIV variants were associated with a range of 1-512 human SNPs (P < 4.28 × 10-11). We found 46 associations between HLA alleles and HIV variants (P < 1.29 × 10-7). HIV variants and immunotypes when analyzed separately were associated with lower VL, whereas the opposite was true when analyzed in concert. Epitope binding predictions supported our observations. CONCLUSIONS: Our results show the importance of immunotype specificity on viral antigenic determinants, and the identified genetic interplay emphasizes that viral and human genetics should be studied in the context of each other.Clinical Trials Registration: NCT00867048.

Withholding Primary Pneumocystis Pneumonia Prophylaxis in Virologically Suppressed Patients With Human Immunodeficiency Virus: An Emulation of a Pragmatic Trial in COHERE.

BACKGROUND: Using data from the COHERE collaboration, we investigated whether primary prophylaxis for pneumocystis pneumonia (PcP) might be withheld in all patients on antiretroviral therapy (ART) with suppressed plasma human immunodeficiency virus (HIV) RNA (≤400 copies/mL), irrespective of CD4 count. METHODS: We implemented an established causal inference approach whereby observational data are used to emulate a randomized trial. Patients taking PcP prophylaxis were eligible for the emulated trial if their CD4 count was ≤200 cells/µL in line with existing recommendations. We compared the following 2 strategies for stopping prophylaxis: (1) when CD4 count was >200 cells/µL for >3 months or (2) when the patient was virologically suppressed (2 consecutive HIV RNA ≤400 copies/mL). Patients were artificially censored if they did not comply with these stopping rules. We estimated the risk of primary PcP in patients on ART, using the hazard ratio (HR) to compare the stopping strategies by fitting a pooled logistic model, including inverse probability weights to adjust for the selection bias introduced by the artificial censoring. RESULTS: A total of 4813 patients (10 324 person-years) complied with eligibility conditions for the emulated trial. With primary PcP diagnosis as an endpoint, the adjusted HR (aHR) indicated a slightly lower, but not statistically significant, different risk for the strategy based on viral suppression alone compared with the existing guidelines (aHR, .8; 95% confidence interval, .6-1.1; P = .2). CONCLUSIONS: This study suggests that primary PcP prophylaxis might be safely withheld in confirmed virologically suppressed patients on ART, regardless of their CD4 count.

HIV continuum of care: expanding scope beyond a cross-sectional view to include time analysis: a systematic review.

INTRODUCTION: The continuum of care (CoC) model has been used to describe the main pillars of HIV care. This study aims to systematically review methods and elucidate gaps in the CoC analyses, especially in terms of the timing of the progression through steps, recognized nowadays as a critical parameter for an effective response to the epidemic. METHODS: A PubMed and EMBASE databases search up to December 2019 resulted in 1918 articles, of which 209 were included in this review; 84 studies presented in major HIV conferences were also included. Studies that did not provide explicit definitions, modelling studies and those reporting only on metrics for subpopulations or factors affecting a CoC stage were excluded. Included articles reported results on 1 to 6 CoC stages. RESULTS: Percentage treated and virally suppressed was reported in 78%, percentage diagnosed and retained in care in 58%, percentage linked to care in 54% and PLHIV in 36% of the articles. Information for all stages was provided in 23 studies. Only 6 articles use novel CoC estimates: One presents a dynamic CoC based on multistate analysis techniques, two base their time-to-next-stage estimates on a risk estimation method based on the cumulative incidence function, weighted for confounding and censoring and three studies estimated the HIV infection time based on mathematical modelling. CONCLUSION: A limited number of studies provide elaborated time analyses of the CoC. Although time analyses lack the straightforward interpretation of the cross-sectional CoC, they provide valuable insights for the timely response to the HIV epidemic. A future goal would be to develop a model that retains the simplicity of the cross-sectional CoC but also incorporates timing between stages.

Human Immunodeficiency Virus Continuum of Care in 11 European Union Countries at the End of 2016 Overall and by Key Population: Have We Made Progress?

BACKGROUND: High uptake of antiretroviral treatment (ART) is essential to reduce human immunodeficiency virus (HIV) transmission and related mortality; however, gaps in care exist. We aimed to construct the continuum of HIV care (CoC) in 2016 in 11 European Union (EU) countries, overall and by key population and sex. To estimate progress toward the Joint United Nations Programme on HIV/AIDS (UNAIDS) 90-90-90 target, we compared 2016 to 2013 estimates for the same countries, representing 73% of the population in the region. METHODS: A CoC with the following 4 stages was constructed: number of people living with HIV (PLHIV); proportion of PLHIV diagnosed; proportion of those diagnosed who ever initiated ART; and proportion of those ever treated who achieved viral suppression at their last visit. RESULTS: We estimated that 87% of PLHIV were diagnosed; 92% of those diagnosed had ever initiated ART; and 91% of those ever on ART, or 73% of all PLHIV, were virally suppressed. Corresponding figures for men having sex with men were: 86%, 93%, 93%, 74%; for people who inject drugs: 94%, 88%, 85%, 70%; and for heterosexuals: 86%, 92%, 91%, 72%. The proportion suppressed of all PLHIV ranged from 59% to 86% across countries. CONCLUSIONS: The EU is close to the 90-90-90 target and achieved the UNAIDS target of 73% of all PLHIV virally suppressed, significant progress since 2013 when 60% of all PLHIV were virally suppressed. Strengthening of testing programs and treatment support, along with prevention interventions, are needed to achieve HIV epidemic control.

Misspecifying the covariance structure in a linear mixed model under MAR drop-out.

Misspecification of the covariance structure in a linear mixed model (LMM) can lead to biased population parameters' estimates under MAR drop-out. In our motivating example of modeling CD4 cell counts during untreated HIV infection, random intercept and slope LMMs are frequently used. In this article, we evaluate the performance of LMMs with specific covariance structures, in terms of bias in the fixed effects estimates, under specific MAR drop-out mechanisms, and adopt a Bayesian model comparison criterion to discriminate between the examined approaches in real-data applications. We analytically show that using a random intercept and slope structure when the true one is more complex can lead to seriously biased estimates, with the degree of bias depending on the magnitude of the MAR drop-out. Under misspecified covariance structure, we compare in terms of induced bias the approach of adding a fractional Brownian motion (BM) process on top of random intercepts and slopes with the approach of using splines for the random effects. In general, the performance of both approaches was satisfactory, with the BM model leading to smaller bias in most cases. A simulation study is carried out to evaluate the performance of the proposed Bayesian criterion in identifying the model with the correct covariance structure. Overall, the proposed method performs better than the AIC and BIC criteria under our specific simulation setting. The models under consideration are applied to real data from the CASCADE study; the most plausible model is identified by all examined criteria.

High prevalence of cardiovascular risk factors in adults living in Greece: the EMENO National Health Examination Survey.

BACKGROUND: Nationwide data on cardiovascular risk factors prevalence is lacking in Greece. This work presents the findings of the national health examination survey EMENO (2013-2016) regarding the prevalence of hypertension, hypercholesterolemia, diabetes, obesity and smoking. METHODS: A random sample of adults (≥18 years) was drawn by multistage stratified random sampling based on 2011 Census. All EMENO participants with ≥1 measurement of interest [blood pressure (BP), fasting glucose, HbA1c, total cholesterol (TC), Body Mass Index (BMI)] were included. Hypertension was defined as BP ≥ 140/90 mmHg and/or antihypertensive treatment; diabetes as fasting glucose≥126 mg/dL and/or HbA1c ≥ 6.5% or self-reported diabetes; hypercholesterolemia as TC ≥ 190 mg/dL. Sampling weights were applied to adjust for study design and post-stratification weights to match sample age and sex distribution to population one. Non-response was adjusted by inverse probability weighting. RESULTS: Of 6006 EMENO participants, 4822 were included (51.5% females, median age:47.9 years). The prevalence of hypertension was 39.2%, higher in men (42.4%) than in women (36.1%); of hypercholesterolemia 60.2%, similar in men (59.5%) and women (60.9%); of diabetes 11.6%, similar men (12.4%) and women (10.9%); of obesity 32.1%, higher in women (33.5% vs 30.2%), although in subjects aged 18-40 year it was higher in men; of current smoking 38.2%, higher in men (44.0%) than in women (32.7%). The prevalence of all risk factors increased substantially with age, except smoking, which followed an inverse U shape. CONCLUSIONS: The burden of cardiovascular risk factors among Greek adults is alarming. There is considerable preventive potential and actions at health care and societal level are urgently needed.

Effect Estimates in Randomized Trials and Observational Studies: Comparing Apples With Apples.

Effect estimates from randomized trials and observational studies might not be directly comparable because of differences in study design, other than randomization, and in data analysis. We propose a 3-step procedure to facilitate meaningful comparisons of effect estimates from randomized trials and observational studies: 1) harmonization of the study protocols (eligibility criteria, treatment strategies, outcome, start and end of follow-up, causal contrast) so that the studies target the same causal effect, 2) harmonization of the data analysis to estimate the causal effect, and 3) sensitivity analyses to investigate the impact of discrepancies that could not be accounted for in the harmonization process. To illustrate our approach, we compared estimates of the effect of immediate with deferred initiation of antiretroviral therapy in individuals positive for the human immunodeficiency virus from the Strategic Timing of Antiretroviral Therapy (START) randomized trial and the observational HIV-CAUSAL Collaboration.

HIV-1 molecular transmission clusters in nine European countries and Canada: association with demographic and clinical factors.

BACKGROUND: Knowledge of HIV-1 molecular transmission clusters (MTCs) is important, especially in large-scale datasets, for designing prevention programmes and public health intervention strategies. We used a large-scale HIV-1 sequence dataset from nine European HIV cohorts and one Canadian, to identify MTCs and investigate factors associated with the probability of belonging to MTCs. METHODS: To identify MTCs, we applied maximum likelihood inferences on partial pol sequences from 8955 HIV-positive individuals linked to demographic and clinical data. MTCs were defined using two different criteria: clusters with bootstrap support >75% (phylogenetic confidence criterion) and clusters consisting of sequences from a specific region at a proportion of >75% (geographic criterion) compared to the total number of sequences within the network. Multivariable logistic regression analysis was used to assess factors associated with MTC clustering. RESULTS: Although 3700 (41%) sequences belonged to MTCs, proportions differed substantially by country and subtype, ranging from 7% among UK subtype C sequences to 63% among German subtype B sequences. The probability of belonging to an MTC was independently less likely for women than men (OR = 0.66; P < 0.001), older individuals (OR = 0.79 per 10-year increase in age; P < 0.001) and people of non-white ethnicity (OR = 0.44; P < 0.001 and OR = 0.70; P = 0.002 for black and 'other' versus white, respectively). It was also more likely among men who have sex with men (MSM) than other risk groups (OR = 0.62; P < 0.001 and OR = 0.69; P = 0.002 for people who inject drugs, and sex between men and women, respectively), subtype B (ORs 0.36-0.70 for A, C, CRF01 and CRF02 versus B; all P < 0.05), having a well-estimated date of seroconversion (OR = 1.44; P < 0.001), a later calendar year of sampling (ORs 2.01-2.61 for all post-2002 periods versus pre-2002; all P < 0.01), and being naïve to antiretroviral therapy at sampling (OR = 1.19; P = 0.010). CONCLUSIONS: A high proportion (>40%) of individuals belonged to MTCs. Notably, the HIV epidemic dispersal appears to be driven by subtype B viruses spread within MSM networks. Expansion of regional epidemics seems mainly associated with recent MTCs, rather than the growth of older, established ones. This information is important for designing prevention and public health intervention strategies.

Longitudinal and time-to-drop-out joint models can lead to seriously biased estimates when the drop-out mechanism is at random.

Missing data are common in longitudinal studies. Likelihood-based methods ignoring the missingness mechanism are unbiased provided missingness is at random (MAR); under not-at-random missingness (MNAR), joint modeling is commonly used, often as part of sensitivity analyses. In our motivating example of modeling CD4 count trajectories during untreated HIV infection, CD4 counts are mainly censored due to treatment initiation, with the nature of this mechanism remaining debatable. Here, we evaluate the bias in the disease progression marker's change over time (slope) of a specific class of joint models, termed shared-random-effects-models (SREMs), under MAR drop-out and propose an alternative SREM model. Our proposed model relates drop-out to both the observed marker's data and the corresponding random effects, in contrast to most SREMs, which assume that the marker and the drop-out processes are independent given the random effects. We analytically calculate the asymptotic bias in two SREMs under specific MAR drop-out mechanisms, showing that the bias in marker's slope increases as the drop-out probability increases. The performance of the proposed model, and other commonly used SREMs, is evaluated under specific MAR and MNAR scenarios through simulation studies. Under MAR, the proposed model yields nearly unbiased slope estimates, whereas the other SREMs yield seriously biased estimates. Under MNAR, the proposed model estimates are approximately unbiased, whereas those from the other SREMs are moderately to heavily biased, depending on the parameterization used. The examined models are also fitted to real data and results are compared/discussed in the light of our analytical and simulation-based findings.

Emulating a trial of joint dynamic strategies: An application to monitoring and treatment of HIV-positive individuals.

Decisions about when to start or switch a therapy often depend on the frequency with which individuals are monitored or tested. For example, the optimal time to switch antiretroviral therapy depends on the frequency with which HIV-positive individuals have HIV RNA measured. This paper describes an approach to use observational data for the comparison of joint monitoring and treatment strategies and applies the method to a clinically relevant question in HIV research: when can monitoring frequency be decreased and when should individuals switch from a first-line treatment regimen to a new regimen? We outline the target trial that would compare the dynamic strategies of interest and then describe how to emulate it using data from HIV-positive individuals included in the HIV-CAUSAL Collaboration and the Centers for AIDS Research Network of Integrated Clinical Systems. When, as in our example, few individuals follow the dynamic strategies of interest over long periods of follow-up, we describe how to leverage an additional assumption: no direct effect of monitoring on the outcome of interest. We compare our results with and without the "no direct effect" assumption. We found little differences on survival and AIDS-free survival between strategies where monitoring frequency was decreased at a CD4 threshold of 350 cells/µl compared with 500 cells/µl and where treatment was switched at an HIV-RNA threshold of 1000 copies/ml compared with 200 copies/ml. The "no direct effect" assumption resulted in efficiency improvements for the risk difference estimates ranging from an 7- to 53-fold increase in the effective sample size.

Psychotic (delusional) depression and completed suicide: a systematic review and meta-analysis.

BACKGROUND: It remains unclear whether psychotic features increase the risk of completed suicides in unipolar depression. The present systematic review coupled with a meta-analysis attempts to elucidate whether unipolar psychotic major depression (PMD) compared to non-PMD presents higher rates of suicides. METHODS: A systematic search was conducted in Scopus, PubMed, and "gray literature" for all studies providing data on completed suicides in PMD compared to non-PMD, and the findings were then subjected to meta-analysis. All articles were independently extracted by two authors using predefined data fields. RESULTS: Nine studies with 33,873 patients, among them 828 suicides, met our inclusion criteria. PMD compared to non-PMD presented a higher lifetime risk of completed suicides with fixed-effect pooled OR 1.21 (95% CI 1.04-1.40). In a sub-analysis excluding a very large study (weight = 86.62%), and comparing 681 PMD to 2106 non-PMD patients, an even higher pooled OR was found [fixed-effect OR 1.69 (95% CI 1.16-2.45)]. Our meta-analysis may provide evidence that the presence of psychosis increases the risk of suicide in patients suffering from severe depression. The data are inconclusive on the contribution of age, mood congruence, comorbidity, and suicide method on PMD's suicide risk. The lack of accurate diagnosis at the time of suicide, PMD's diagnostic instability, and the use of ICD-10 criteria constitute the main study limitations. CONCLUSIONS: The presence of psychosis in major depression should alert clinicians for the increased risk of completed suicide. Thus, the implementation of an effective treatment both for psychotic depression and patients' suicidality constitutes a supreme priority.