People with disability and privacy in precision medicine research: what's at stake?

Re-identification from data used in precision medicine research is presumed to create minimal risk but may disproportionately impact health disparity populations. We consider plausible privacy risks and the negative ramifications thereof for people with disabilities, the largest health disparity population in the USA, and suggest measures to address these concerns.

Human heuristics for AI-generated language are flawed.

Human communication is increasingly intermixed with language generated by AI. Across chat, email, and social media, AI systems suggest words, complete sentences, or produce entire conversations. AI-generated language is often not identified as such but presented as language written by humans, raising concerns about novel forms of deception and manipulation. Here, we study how humans discern whether verbal self-presentations, one of the most personal and consequential forms of language, were generated by AI. In six experiments, participants (N = 4,600) were unable to detect self-presentations generated by state-of-the-art AI language models in professional, hospitality, and dating contexts. A computational analysis of language features shows that human judgments of AI-generated language are hindered by intuitive but flawed heuristics such as associating first-person pronouns, use of contractions, or family topics with human-written language. We experimentally demonstrate that these heuristics make human judgment of AI-generated language predictable and manipulable, allowing AI systems to produce text perceived as "more human than human." We discuss solutions, such as AI accents, to reduce the deceptive potential of language generated by AI, limiting the subversion of human intuition.

The role of genetic risk factors of Alzheimer's disease in synaptic dysfunction.

Alzheimer's disease (AD) is a neurodegenerative disease characterized by the progressive deterioration of cognitive functions. Due to the extended global life expectancy, the prevalence of AD is increasing among aging populations worldwide. While AD is a multifactorial disease, synaptic dysfunction is one of the major neuropathological changes that occur early in AD, before clinical symptoms appear, and is associated with the progression of cognitive deterioration. However, the underlying pathological mechanisms leading to this synaptic dysfunction remains unclear. Recent large-scale genomic analyses have identified more than 40 genetic risk factors that are associated with AD. In this review, we discuss the functional roles of these genes in synaptogenesis and synaptic functions under physiological conditions, and how their functions are dysregulated in AD. This will provide insights into the contributions of these encoded proteins to synaptic dysfunction during AD pathogenesis.

Making sense of breast cancer risk estimates.

Individual probabilistic assessments on the risk of cancer, primary or secondary, will not be understood by most patients. That is the essence of our arguments in this paper. Greater understanding can be achieved by extensive, intensive, and detailed counseling. But since probability itself is a concept that easily escapes our everyday intuition-consider the famous Monte Hall paradox-then it would also be wise to advise patients and potential patients, to not put undue weight on any probabilistic assessment. Such assessments can be of value to the epidemiologist in the investigation of different potential etiologies describing cancer evolution or to the clinical trialist as a way to maximize design efficiency. But to an ordinary individual we cannot anticipate that these assessments will be correctly interpreted.

Pooling controls from nested case-control studies with the proportional risks model.

The standard approach to regression modeling for cause-specific hazards with prospective competing risks data specifies separate models for each failure type. An alternative proposed by Lunn and McNeil (1995) assumes the cause-specific hazards are proportional across causes. This may be more efficient than the standard approach, and allows the comparison of covariate effects across causes. In this paper, we extend Lunn and McNeil (1995) to nested case-control studies, accommodating scenarios with additional matching and non-proportionality. We also consider the case where data for different causes are obtained from different studies conducted in the same cohort. It is demonstrated that while only modest gains in efficiency are possible in full cohort analyses, substantial gains may be attained in nested case-control analyses for failure types that are relatively rare. Extensive simulation studies are conducted and real data analyses are provided using the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) study.

Fast and flexible inference for joint models of multivariate longitudinal and survival data using integrated nested Laplace approximations.

Modeling longitudinal and survival data jointly offers many advantages such as addressing measurement error and missing data in the longitudinal processes, understanding and quantifying the association between the longitudinal markers and the survival events, and predicting the risk of events based on the longitudinal markers. A joint model involves multiple submodels (one for each longitudinal/survival outcome) usually linked together through correlated or shared random effects. Their estimation is computationally expensive (particularly due to a multidimensional integration of the likelihood over the random effects distribution) so that inference methods become rapidly intractable, and restricts applications of joint models to a small number of longitudinal markers and/or random effects. We introduce a Bayesian approximation based on the integrated nested Laplace approximation algorithm implemented in the R package R-INLA to alleviate the computational burden and allow the estimation of multivariate joint models with fewer restrictions. Our simulation studies show that R-INLA substantially reduces the computation time and the variability of the parameter estimates compared with alternative estimation strategies. We further apply the methodology to analyze five longitudinal markers (3 continuous, 1 count, 1 binary, and 16 random effects) and competing risks of death and transplantation in a clinical trial on primary biliary cholangitis. R-INLA provides a fast and reliable inference technique for applying joint models to the complex multivariate data encountered in health research.

Delineating genotype and parent-of-origin effect on the phenotype in MSH6-associated Lynch syndrome.

BACKGROUND: This study investigates the potential influence of genotype and parent-of-origin effects (POE) on the clinical manifestations of Lynch syndrome (LS) within families carrying (likely) disease-causing MSH6 germline variants. PATIENTS AND METHODS: A cohort of 1615 MSH6 variant carriers (310 LS families) was analyzed. Participants were categorized based on RNA expression and parental inheritance of the variant. Hazard ratios (HRs) were calculated using weighted Cox regression, considering external information to address ascertainment bias. The findings were cross-validated using the Prospective Lynch Syndrome Database (PLSD) for endometrial cancer (EC). RESULTS: No significant association was observed between genotype and colorectal cancer (CRC) risk (HR = 1.06, 95% confidence interval [CI]: 0.77-1.46). Patients lacking expected RNA expression exhibited a reduced risk of EC (Reference Cohort 1: HR = 0.68, 95% CI: 0.43-1.03; Reference Cohort 2: HR = 0.63, 95% CI: 0.46-0.87). However, these results could not be confirmed in the PLSD. Moreover, no association was found between POE and CRC risk (HR = 0.78, 95% CI: 0.52-1.17) or EC risk (Reference Cohort 1: HR = 0.93, 95% CI: 0.65-1.33; Reference Cohort 2: HR = 0.8, 95% CI: 0.64-1.19). DISCUSSION AND CONCLUSION: No evidence of POE was detected in MSH6 families. While RNA expression may be linked to varying risks of EC, further investigation is required to explore this observation.

A transformer model for cause-specific hazard prediction.

BACKGROUD: Modelling discrete-time cause-specific hazards in the presence of competing events and non-proportional hazards is a challenging task in many domains. Survival analysis in longitudinal cohorts often requires such models; notably when the data is gathered at discrete points in time and the predicted events display complex dynamics. Current models often rely on strong assumptions of proportional hazards, that is rarely verified in practice; or do not handle sequential data in a meaningful way. This study proposes a Transformer architecture for the prediction of cause-specific hazards in discrete-time competing risks. Contrary to Multilayer perceptrons that were already used for this task (DeepHit), the Transformer architecture is especially suited for handling complex relationships in sequential data, having displayed state-of-the-art performance in numerous tasks with few underlying assumptions on the task at hand. RESULTS: Using synthetic datasets of 2000-50,000 patients, we showed that our Transformer model surpassed the CoxPH, PyDTS, and DeepHit models for the prediction of cause-specific hazard, especially when the proportional assumption did not hold. The error along simulated time outlined the ability of our model to anticipate the evolution of cause-specific hazards at later time steps where few events are observed. It was also superior to current models for prediction of dementia and other psychiatric conditions in the English longitudinal study of ageing cohort using the integrated brier score and the time-dependent concordance index. We also displayed the explainability of our model's prediction using the integrated gradients method. CONCLUSIONS: Our model provided state-of-the-art prediction of cause-specific hazards, without adopting prior parametric assumptions on the hazard rates. It outperformed other models in non-proportional hazards settings for both the synthetic dataset and the longitudinal cohort study. We also observed that basic models such as CoxPH were more suited to extremely simple settings than deep learning models. Our model is therefore especially suited for survival analysis on longitudinal cohorts with complex dynamics of the covariate-to-outcome relationship, which are common in clinical practice. The integrated gradients provided the importance scores of input variables, which indicated variables guiding the model in its prediction. This model is ready to be utilized for time-to-event prediction in longitudinal cohorts.

Deep self-supervised machine learning algorithms with a novel feature elimination and selection approaches for blood test-based multi-dimensional health risks classification.

BACKGROUND: Blood test is extensively performed for screening, diagnoses and surveillance purposes. Although it is possible to automatically evaluate the raw blood test data with the advanced deep self-supervised machine learning approaches, it has not been profoundly investigated and implemented yet. RESULTS: This paper proposes deep machine learning algorithms with multi-dimensional adaptive feature elimination, self-feature weighting and novel feature selection approaches. To classify the health risks based on the processed data with the deep layers, four machine learning algorithms having various properties from being utterly model free to gradient driven are modified. CONCLUSIONS: The results show that the proposed deep machine learning algorithms can remove the unnecessary features, assign self-importance weights, selects their most informative ones and classify the health risks automatically from the worst-case low to worst-case high values.

Disentangling sex-dependent effects of APOE on diverse trajectories of cognitive decline in Alzheimer's disease.

Current diagnostic systems for Alzheimer's disease (AD) rely upon clinical signs and symptoms, despite the fact that the multiplicity of clinical symptoms renders various neuropsychological assessments inadequate to reflect the underlying pathophysiological mechanisms. Since putative neuroimaging biomarkers play a crucial role in understanding the etiology of AD, we sought to stratify the diverse relationships between AD biomarkers and cognitive decline in the aging population and uncover risk factors contributing to the diversities in AD. To do so, we capitalized on a large amount of neuroimaging data from the ADNI study to examine the inflection points along the dynamic relationship between cognitive decline trajectories and whole-brain neuroimaging biomarkers, using a state-of-the-art statistical model of change point detection. Our findings indicated that the temporal relationship between AD biomarkers and cognitive decline may differ depending on the synergistic effect of genetic risk and biological sex. Specifically, tauopathy-PET biomarkers exhibit a more dynamic and age-dependent association with Mini-Mental State Examination scores (p<0.05), with inflection points at 72, 78, and 83 years old, compared with amyloid-PET and neurodegeneration (cortical thickness from MRI) biomarkers. In the landscape of health disparities in AD, our analysis indicated that biological sex moderates the rate of cognitive decline associated with APOE4 genotype. Meanwhile, we found that higher education levels may moderate the effect of APOE4, acting as a marker of cognitive reserve.

Can metformin prevent cancer relative to sulfonylureas? A target trial emulation accounting for competing risks and poor overlap via double/debiased machine learning estimators.

There is mounting interest in the possibility that metformin, indicated for glycemic control in type 2 diabetes, has a range of additional beneficial effects. Randomized trials have shown that metformin prevents adverse cardiovascular events, and metformin use has also been associated with reduced cognitive decline and cancer incidence. In this paper, we dig more deeply into whether metformin prevents cancer by emulating target randomized trials comparing metformin to sulfonylureas as first line diabetes therapy using data from Clinical Practice Research Datalink, a U.K. primary care database (1987-2018). We included individuals with diabetes, no prior cancer diagnosis, no chronic kidney disease, and no prior diabetes therapy who initiated metformin (N=93353) or a sulfonylurea (N=13864). In our cohort, the estimated overlap weighted additive separable direct effect of metformin compared to sulfonylureas on cancer risk at 6 years was -1% (.95 CI=-2.2%, 0.1%), which is consistent with metformin providing no direct protection against cancer incidence or substantial protection. The analysis faced two methodological challenges-poor overlap, and pre-cancer death as a competing risk. To address these issues while minimizing nuisance model misspecification, we develop and apply double/debiased machine learning estimators of overlap weighted separable effects in addition to more traditional effect estimates.

Causal Effects of Competing Obstetrical Interventions: Mediators of Placental Abruption and Perinatal Mortality.

Placental abruption, the premature placental separation, confers increased perinatal mortality risk with preterm delivery as an important pathway through which the risk appears mediated. While pregnancies complicated by abruption are often delivered through an obstetrical intervention, many deliver spontaneously. We examined the contributions of clinician-initiated (PTDIND) and spontaneous (PTDSPT) preterm delivery at <37 weeks as competing causal mediators of the abruption-perinatal mortality association. Using the Consortium for Safe Labor (2002-2008) data (n = 203,990; 1.6% with abruption), we applied a potential outcomes-based mediation analysis to decompose the total effect into direct and mediator-specific indirect effects through PTDIND and PTDSPT. Each mediated effect describes the reduction in the counterfactual mortality risk if that preterm delivery subtype was shifted from its distribution under abruption to without abruption. The total effect risk ratio (RR) of abruption on perinatal mortality was 5.4 (95% confidence interval [CI] 4.6, 6.3). The indirect effect RRs for PTDIND and PTDSPT were 1.5 (95% CI: 1.4, 1.6) and 1.5 (95% CI: 1.5, 1.6), respectively; these corresponded to mediated proportions of 25% each. These findings underscore that spontaneous and clinician-initiated preterm deliveries each play essential roles in shaping perinatal mortality risks associated with placental abruption.

Death ORACL: An Algorithm to Predict Death Using Insurance Claims Data.

The inability to identify dates of death in insurance claims data is the United States is a major limitation to retrospective claims-based research. While deaths result in disenrollment, disenrollment can also occur due to changes in insurance providers. We created an algorithm to differentiate between disenrollment from health plans due to death and disenrollment for other reasons. We identified 5,259,735 adults who disenrolled from private insurance between 2007 and 2018. Using death dates ascertained from the Social Security Death Index, inpatient discharge status, and death indicators in the administrative data, 7.6% of all disenrollments were classified as resulting from death. We used elastic net regression to build an algorithm using claims data in the year prior to disenrollment; candidate predictors included medical conditions, individual demographic characteristics, treatment utilization, and structural factors related to health insurance eligibility and coding. Using a predicted probability threshold of 0.9 (selected to reflect the corresponding known prevalence of mortality), internal validation found that the algorithm classified death at disenrollment with a positive predictive value of 0.815, sensitivity of 0.721 and specificity of 0.986 (AUC=0.97). Independent data sources were used for external validation and for an applied example. Code for implementation is publicly available.

Does breast cancer modify the long-term relationship between lifestyle behaviors and mortality? A prospective analysis of breast cancer survivors and population-based controls.

BACKGROUND: Modifiable lifestyle factors are known to impact survival. It is less clear whether this differs between postmenopausal women ever diagnosed with breast cancer and unaffected women. METHODS: Women diagnosed with breast cancer and unaffected women of comparable age were recruited from 2002 to 2005 and followed up until 2020. Using baseline information, a lifestyle adherence score (range 0-8; categorized as low [0-3.74], moderate [3.75-4.74], and high [≥4.75]) was created based on the 2018 World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) cancer prevention recommendations. Cox regression and competing risks analysis were used to analyze the association of adherence to WCRF/AICR lifestyle recommendations with overall mortality and with death due to cardiovascular diseases and cancer, respectively. RESULTS: A total of 8584 women were included (2785 with breast cancer and 5799 without). With a median follow-up of 16.1 years there were 2006 total deaths. Among the deaths of known causes (98.6%), 445 were cardiovascular-related and 1004 were cancer-related. The average lifestyle score was 4.2. There was no differential effect of lifestyle score by case-control status on mortality. After adjusting for covariates, moderate (hazard ratio [HR], 0.66; 95% confidence interval [CI], 0.57-0.76) and high (HR, 0.54; 95% CI, 0.47-0.63) adherence to WCRF/AICR lifestyle recommendations were significantly associated with a decrease in overall mortality. Similarly, in competing risks analysis, moderate and high adherence were associated with decreased mortality from cardiovascular diseases and from cancer. CONCLUSIONS: A healthy lifestyle can substantially reduce mortality risk in women. With low adherence to all WCRF/AICR guidelines in about a third of study participants, health interventions are warranted.

Benefit-to-radiation-risk of low-dose computed tomography lung cancer screening.

BACKGROUND: The US National Lung Screening Trial (NLST) and Dutch-Belgian NELSON randomized controlled trials have shown significant mortality reductions from low-dose computed tomography (CT) lung cancer screening (LCS). NLST, ITALUNG, and COSMOS trials have provided detailed dosimetry data for LCS. METHODS: LCS trial mortality benefit results, organ dose and effective dose data, and Biological Effects of Ionizing Radiation, Report VII (BEIR VII) organ dose-to-cancer-mortality risk data are used to estimate benefit-to-radiation-risk ratios of the NLST, ITALUNG, and COSMOS trials. Data from those trials also are used to estimate benefit-to-radiation-risk ratios for longer-term LCS corresponding to scenarios recommended by United States Preventive Services Task Force and the American Cancer Society. RESULTS: Including only screening doses, NLST benefit-to-radiation-risk ratios are 12:1 for males, 19:1 for females, and 16:1 overall. Including both screening and estimated follow-up doses, benefit-to-radiation-risk ratios for NLST are 9:1 for males, 13:1 for females, and 12:1 overall. For the ITALUNG trial, the benefit-to-radiation-risk ratio is 58-63:1. For the COSMOS trial, assuming sex-specific mortality benefits like those of the NELSON trial, the benefit-to-radiation-risk ratio is 23:1. Assuming a conservative 20% mortality benefit, annual screening in people 50-79 years old with a 20+ pack-year history of smoking has benefit-to-radiation-risk ratios of 23:1 (with follow-up doses adding 40% to screening doses) to 29:1 (with follow-up adding 10%) based on COSMOS dose data. CONCLUSIONS: Based on linear, no threshold BEIR VII dose-risk estimates, benefit-to-radiation-risk ratios for LCS are highly favorable. Results emphasize the importance of using modern CT technologies, maintaining low diagnostic follow-up rates, and minimizing both screening and diagnostic follow-up doses. PLAIN LANGUAGE SUMMARY: The benefits of lung cancer screening significantly outweigh estimates of future harms associated with exposure to radiation during screening and diagnostic follow-up examinations. Our findings emphasize the importance of lung cancer screening practices using state-of-the-art computed tomography scanners and specialized low-dose lung screening and diagnostic follow-up techniques.

Age at diagnosis and social risks among Black cancer survivors: Results from the Detroit Research on Cancer Survivors cohort.

BACKGROUND: Social risks are common among cancer survivors who have the fewest financial resources; however, little is known about how prevalence differs by age at diagnosis, despite younger survivors' relatively low incomes and wealth. METHODS: The authors used data from 3703 participants in the Detroit Research on Cancer Survivors (ROCS) cohort of Black cancer survivors. Participants self-reported several forms of social risks, including food insecurity, housing instability, utility shut-offs, not getting care because of cost or lack of transportation, and feeling unsafe in their home neighborhood. Modified Poisson models were used to estimate prevalence ratios and 95% confidence intervals (CIs) of social risks by age at diagnosis, controlling for demographic, socioeconomic, and cancer-related factors. RESULTS: Overall, 35% of participants reported at least one social risk, and 17% reported two or more risks. Social risk prevalence was highest among young adults aged 20-39 years (47%) followed by those aged 40-54 years (43%), 55-64 years (38%), and 65 years and older (24%; p for trend < .001). Compared with survivors who were aged 65 years and older at diagnosis, adjusted prevalence ratios for any social risk were 1.75 (95% CI, 1.42-2.16) for survivors aged 20-39 years, 1.76 (95% CI, 1.52-2.03) for survivors aged 40-54 years, and 1.41 (95% CI, 1.23-1.60) for survivors aged 55-64 years at diagnosis. Similar associations were observed for individual social risks and experiencing two or more risks. CONCLUSIONS: In this population of Black cancer survivors, social risks were inversely associated with age at diagnosis. Diagnosis in young adulthood and middle age should be considered a risk factor for social risks and should be prioritized in work to reduce the financial effects of cancer on financially vulnerable cancer survivors.

A competing risks model to estimate the risk of graft failure and patient death after kidney transplantation using continuous donor-recipient age combinations.

Graft failure and recipient death with functioning graft are important competing outcomes after kidney transplantation. Risk prediction models typically censor for the competing outcome thereby overestimating the cumulative incidence. The magnitude of this overestimation is not well described in real-world transplant data. This retrospective cohort study analyzed data from the European Collaborative Transplant Study (n = 125 250) and from the American Scientific Registry of Transplant Recipients (n = 190 258). Separate cause-specific hazard models using donor and recipient age as continuous predictors were developed for graft failure and recipient death. The hazard of graft failure increased quadratically with increasing donor age and decreased decaying with increasing recipient age. The hazard of recipient death increased linearly with increasing donor and recipient age. The cumulative incidence overestimation due to competing risk-censoring was largest in high-risk populations for both outcomes (old donors/recipients), sometimes amounting to 8.4 and 18.8 percentage points for graft failure and recipient death, respectively. In our illustrative model for posttransplant risk prediction, the absolute risk of graft failure and death is overestimated when censoring for the competing event, mainly in older donors and recipients. Prediction models for absolute risks should treat graft failure and death as competing events.

The Role of Pancreatitis Risk Genes in Endocrine Insufficiency Development After Acute Pancreatitis in Children.

BACKGROUND & AIMS: Acute pancreatitis (AP) is increasingly recognized as a risk factor for diabetes mellitus (DM). We aimed to study the association of pancreatitis genes with pancreatic endocrine insufficiency (pre-DM and DM) development post-AP in children. METHODS: This was an observational cohort study that enrolled subjects ≤21 years with their first episode of AP and followed them for 12 months for the development of pancreatic endocrine insufficiency. Pancreatitis risk genes (CASR, CEL, CFTR, CLDN2, CPA1, CTRC, PRSS1, SBDS, SPINK1, and UBR1) were sequenced. A genetic risk score was derived from all genes with univariable P < .15. RESULTS: A total 120 subjects with AP were genotyped. Sixty-three subjects (52.5%) had at least 1 reportable variant identified. For modeling the development of pancreatic endocrine insufficiency at 1 year, 6 were excluded (2 with DM at baseline, 3 with total pancreatectomy, and 1 death). From this group of 114, 95 remained normoglycemic and 19 (17%) developed endocrine insufficiency (4 DM, 15 pre-DM). Severe AP (58% vs 20%; P = .001) and at least 1 gene affected (79% vs 47%; P = .01) were enriched among the endocrine-insufficient group. Those with versus without endocrine insufficiency were similar in age, sex, race, ethnicity, body mass index, and AP recurrence. A model for pre-DM/DM development included AP severity (odds ratio, 5.17 [1.66-16.15]; P = .005) and genetic risk score (odds ratio, 4.89 [1.83-13.08]; P = .002) and had an area under the curve of 0.74. CONCLUSIONS: In this cohort of children with AP, pancreatitis risk genes and AP disease severity were associated with pre-DM or DM development post-AP.

Longitudinal Outcomes Associated With Metabolic Dysfunction-Associated Steatotic Liver Disease: A Meta-analysis of 129 Studies.

BACKGROUND & AIMS: The progression of metabolic dysfunction-associated steatotic liver disease (MASLD) has been found to manifest in a series of hepatic and extrahepatic complications. A comprehensive meta-analysis of the longitudinal outcomes associated with MASLD has yet to be conducted. METHODS: To investigate the longitudinal outcomes associated with MASLD, Medline and Embase databases were searched to identify original studies that evaluated the longitudinal risks of incident clinical outcomes among MASLD patients compared with non-MASLD individuals. DerSimonian Laird random-effects meta-analysis was performed. Pooled effect estimates were calculated, and heterogeneity among studies was evaluated. RESULTS: One hundred twenty-nine studies were included in the meta-analysis. Meta-analysis revealed a significant increase in the risk of cardiovascular outcomes (hazard ratio [HR], 1.43; 95% confidence interval [CI], 1.27-1.60; P < .01), various metabolic outcomes such as incident hypertension (HR, 1.75; 95% CI, 1.46-2.08; P < .01), diabetes (HR, 2.56; 95% CI, 2.10-3.13; P < .01), pre-diabetes (HR, 1.69; 95% CI, 1.22-2.35; P < .01), metabolic syndrome (HR, 2.57; 95% CI, 1.13-5.85; P = .02), chronic kidney disease (HR, 1.38; 95% CI, 1.27-1.50; P < .01), as well as all cancers (HR, 1.54; 95% CI, 1.35-1.76; P < .01) among MASLD patients compared with non-MASLD individuals. By subgroup analysis, MASLD patients with advanced liver disease (HR, 3.60; 95% CI, 2.10-6.18; P < .01) were also found to be associated with a significantly greater risk (P = .02) of incident diabetes than those with less severe MASLD (HR, 1.63; 95% CI, 1.0-2.45; P = .02) when compared with non-MASLD. CONCLUSIONS: The present study emphasizes the association between MASLD and its clinical outcomes including cardiovascular, metabolic, oncologic, and other outcomes. The multisystemic nature of MASLD found in this analysis requires treatment targets to reduce systemic events and end organ complications.

Semiparametric marginal regression for clustered competing risks data with missing cause of failure.

Clustered competing risks data are commonly encountered in multicenter studies. The analysis of such data is often complicated due to informative cluster size (ICS), a situation where the outcomes under study are associated with the size of the cluster. In addition, the cause of failure is frequently incompletely observed in real-world settings. To the best of our knowledge, there is no methodology for population-averaged analysis with clustered competing risks data with an ICS and missing causes of failure. To address this problem, we consider the semiparametric marginal proportional cause-specific hazards model and propose a maximum partial pseudolikelihood estimator under a missing at random assumption. To make the latter assumption more plausible in practice, we allow for auxiliary variables that may be related to the probability of missingness. The proposed method does not impose assumptions regarding the within-cluster dependence and allows for ICS. The asymptotic properties of the proposed estimators for both regression coefficients and infinite-dimensional parameters, such as the marginal cumulative incidence functions, are rigorously established. Simulation studies show that the proposed method performs well and that methods that ignore the within-cluster dependence and the ICS lead to invalid inferences. The proposed method is applied to competing risks data from a large multicenter HIV study in sub-Saharan Africa where a significant portion of causes of failure is missing.