Benchmark dose profiles for bivariate exposures.

While benchmark dose (BMD) methodology is well-established for settings with a single exposure, these methods cannot easily handle multidimensional exposures with nonlinear effects. We propose a framework for BMD analysis to characterize the joint effect of a two-dimensional exposure on a continuous outcome using a generalized additive model while adjusting for potential confounders via propensity scores. This leads to a dose-response surface which can be summarized in two dimensions by a contour plot in which combinations of exposures leading to the same expected effect are identified. In our motivating study of prenatal alcohol exposure, cognitive deficits in children are found to be associated with both the frequency of drinking as well as the amount of alcohol consumed on each drinking day during pregnancy. The general methodological framework is useful for a broad range of settings, including combinations of environmental stressors, such as chemical mixtures, and in explorations of the impact of dose rate rather than simply cumulative exposure on adverse outcomes.

A dose-response analysis of the effects of prenatal alcohol exposure on cognitive development.

BACKGROUND: Most studies of the effects of prenatal alcohol exposure (PAE) on cognitive function have assumed that the dose-response curve is linear. However, data from a few animal and human studies suggest that there may be an inflection point in the dose-response curve above which PAE effects are markedly stronger and that there may be differences associated with pattern of exposure, assessed in terms of alcohol dose per drinking occasion and drinking frequency. METHODS: We performed second-order confirmatory factor analysis on data obtained at school age, adolescence, and early adulthood from 2227 participants in six US longitudinal cohorts to derive a composite measure of cognitive function. Regression models were constructed to examine effects of PAE on cognitive function, adjusted for propensity scores. Analyses based on a single predictor (absolute alcohol (AA)/day) were compared with analyses based on two predictors (dose/occasion and drinking frequency), using (1) linear models and (2) nonparametric general additive models (GAM) that allow for both linear and nonlinear effects. RESULTS: The single-predictor GAM model showed virtually no nonlinearity in the effect of AA/day on cognitive function. However, the two-predictor GAM model revealed differential effects of maternal drinking pattern. Among offspring of infrequent drinkers, PAE effects on cognitive function were markedly stronger in those whose mothers drank more than ~3 drinks/occasion, and the effect of dose/occasion was strongest among the very frequent drinkers. Frequency of drinking did not appear to alter the PAE effect on cognitive function among participants born to mothers who limited their drinking to ~1 drink/occasion or less. CONCLUSIONS: These findings suggest that linear models based on total AA/day are appropriate for assessing whether PAE affects a given cognitive outcome. However, examination of alcohol dose/occasion and drinking frequency is needed to fully characterize the impact of different levels of alcohol intake on cognitive impairment.

A retrospective and prospective study of biostatistics in Canada.

Biostatistics is foundational to public health research and Canada has a history of high impact contributions both in seminal methodological advances and in the rigorous application of methods for the design or analysis of public health studies. In this article, we provide a brief and personal review of selected contributions from Canadian biostatisticians to fields such as survival and life history analysis, sampling, clinical trial methodology, environmental risk assessment, infectious disease epidemiology, and early work on prediction. We also provide a brief look forward at the upcoming needs and future directions of biostatistical research.

RéSUMé: La biostatistique est fondamentale pour la recherche en santé publique et le Canada a un historique de contributions à fort impact, tant dans les avancées méthodologiques majeures que dans l'application rigoureuse de méthodes pour la conception ou l'analyse d'études de santé publique. Dans cet article, nous présentons un examen bref et personnel des contributions des biostatisticiens canadiens dans des domaines tels que l'analyse de la survie et de l'histoire de vie, l'échantillonnage, la méthodologie des essais cliniques, le risque environnemental, l'épidémiologie des maladies infectieuses et les premiers travaux sur la prédiction et la classification. Nous fournissons également un bref aperçu des besoins à venir et des orientations futures de la recherche biostatistique.

Two-phase designs with failure time processes subject to nonsusceptibility.

Epidemiological studies based on 2-phase designs help ensure efficient use of limited resources in situations where certain covariates are prohibitively expensive to measure for a full cohort. Typically, these designs involve 2 steps: In phase I, data on an outcome and inexpensive covariates are acquired, and in phase II, a subsample is chosen in which the costly variable of interest is measured. For right-censored data, 2-phase designs have been primarily based on the Cox model. We develop efficient 2-phase design strategies for settings involving a fraction of long-term survivors due to nonsusceptibility. Using mixture models accommodating a nonsusceptible fraction, we consider 3 regression frameworks, including (a) a logistic "cure" model, (b) a proportional hazards model for those who are susceptible, and (c) regression models for susceptibility and failure time in those susceptible. Importantly, we introduce a novel class of bivariate residual-dependent designs to address the unique challenges presented in scenario (c), which involves 2 parameters of interest. Extensive simulation studies demonstrate the superiority of our approach over various phase II subsampling schemes. We illustrate the method through applications to the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial.

Statistical and Scientific Considerations Concerning the Interpretation, Replicability, and Transportability of Research Findings.

To advance scientific understanding of disease processes and related intervention effects, study results should be free from bias and replicable. More broadly, investigators seek results that are transportable, that is, applicable to a perceived study population as well as in other environments and populations. We review fundamental statistical issues that arise in the analysis of observational data from disease cohorts and other sources and discuss how these issues affect the transportability and replicability of research results. Much of the literature focuses on estimating average exposure or intervention effects at the population level, but we argue for more nuanced analyses of conditional effects that reflect the complexity of disease processes.

Derivation and Internal Validation of a Disease-Specific Cardiovascular Risk Prediction Model for Patients With Psoriatic Arthritis and Psoriasis.

OBJECTIVE: To address suboptimal cardiovascular risk prediction in patients with psoriatic disease (PsD), we developed and internally validated a five-year disease-specific cardiovascular risk prediction model. METHODS: We analyzed data from a prospective cohort of participants with PsD without a history of cardiovascular events. Traditional cardiovascular risk factors and PsD-related measures of disease activity were considered as potential predictors. The study outcome included nonfatal and fatal cardiovascular events. A base prediction model included 10 traditional cardiovascular risk factors. Eight PsD-related factors were assessed by adding them to the base model to create expanded models, which were controlled for PsD therapies. Variable selection was performed using Least Absolute Shrinkage and Selection Operator (LASSO) penalized regression with 10-fold cross-validation. Model performance was assessed using measures of discrimination and calibration and measures of sensitivity and specificity. RESULTS: Between 1992 and 2020, 85 of 1,336 participants developed cardiovascular events. Discrimination of the base model (with traditional cardiovascular risk factors alone) was excellent, with an area under the receiver operator characteristic curve (AUC) of 85.5 (95% confidence interval [CI] 81.9-89.1). Optimal models did not select any of the tested disease-specific factors. In a sensitivity analysis, which excluded lipid lowering and antihypertensive treatments, the number of damaged joints was selected in the expanded model. However, this model did not improve risk discrimination compared to the base model (AUC 85.5, 95% CI 82.0-89.1). CONCLUSION: Traditional cardiovascular risk factors alone are effective in predicting cardiovascular risk in patients with PsD. A risk score based on these factors performed well, indicating excellent discrimination and calibration.

Bayesian outcome selection modeling.

In psychiatric and social epidemiology studies, it is common to measure multiple different outcomes using a comprehensive battery of tests thought to be related to an underlying construct of interest. In the research that motivates our work, researchers wanted to assess the impact of in utero alcohol exposure on child cognition and neuropsychological development, which are evaluated using a range of different psychometric tests. Statistical analysis of the resulting multiple outcomes data can be challenging, because the outcomes measured on the same individual are not independent. Moreover, it is unclear, a priori, which outcomes are impacted by the exposure under study. While researchers will typically have some hypotheses about which outcomes are important, a framework is needed to help identify outcomes that are sensitive to the exposure and to quantify the associated treatment or exposure effects of interest. We propose such a framework using a modification of stochastic search variable selection, a popular Bayesian variable selection model and use it to quantify an overall effect of the exposure on the affected outcomes. The performance of the method is investigated empirically and an illustration is given through application using data from our motivating study.

Derivation of a Multivariable Psoriatic Arthritis Risk Estimation Tool (PRESTO): A Step Towards Prevention.

OBJECTIVE: A simple, scalable tool that identifies psoriasis patients at high risk for developing psoriatic arthritis (PsA) could improve early diagnosis. We aimed to develop a risk prediction model for the development of PsA and to assess its performance among patients with psoriasis. METHODS: We analyzed data from a prospective cohort of psoriasis patients without PsA at enrollment. Participants were assessed annually by a rheumatologist for the development of PsA. Information about their demographics, psoriasis characteristics, comorbidities, medications, and musculoskeletal symptoms was used to develop prediction models for PsA. Penalized binary regression models were used for variable selection while adjusting for psoriasis duration. Risks of developing PsA over 1- and 5-year time periods were estimated. Model performance was assessed by the area under the curve (AUC) and calibration plots. RESULTS: Among 635 psoriasis patients, 51 and 71 developed PsA during the 1-year and 5-year follow-up periods, respectively. The risk of developing PsA within 1 year was associated with younger age, male sex, family history of psoriasis, back stiffness, nail pitting, joint stiffness, use of biologic medications, patient global health, and pain severity (AUC 72.3). The risk of developing PsA within 5 years was associated with morning stiffness, psoriatic nail lesion, psoriasis severity, fatigue, pain, and use of systemic nonbiologic medication or phototherapy (AUC 74.9). Calibration plots showed reasonable agreement between predicted and observed probabilities. CONCLUSIONS: The development of PsA within clinically meaningful time frames can be predicted with reasonable accuracy for psoriasis patients using readily available clinical variables.

Analysis of secondary failure time responses in studies with response-dependent sampling schemes.

Response-dependent sampling is routinely used as an enrichment strategy in the design of family studies investigating the heritable nature of disease. In addition to the response of primary interest, investigators often wish to investigate the association between biomarkers and secondary responses related to possible comorbidities. Statistical analysis regarding genetic biomarkers and their association with the secondary outcome must address the biased sampling scheme involving the primary response. In this article, we develop composite likelihoods and two-stage estimation procedures for such secondary analyses in which the within-family dependence structure for the primary and secondary outcomes is modeled via a Gaussian copula. The dependence among responses within family members is modeled based on kinship coefficients. Auxiliary data from independent individuals are exploited by augmenting the composite likelihoods to increase precision of marginal parameter estimates and enhance the efficiency of estimators of the dependence parameters. Simulation studies are carried out to evaluate the finite sample performance of the proposed method, and an application to a motivating family study in psoriatic arthritis is given for illustration.

New late-emphasis and combination tests based on infimum and supremum logrank statistics with application in oncology trials.

Immunotherapy cancer clinical trials routinely feature an initial period during which the treatment is given without evident therapeutic benefit, which may be followed by a period during which an effective therapy reduces the hazard for event occurrence. The nature of this treatment effect is incompatible with the proportional hazards assumption, which has prompted much work on the development of alternative effect measures of frameworks for testing. We consider tests based on individual and combination of early- and late-emphasis infimum and supremum logrank statistics, describe how they can be implemented, and evaluate their performance in simulation studies. Through this work and illustrative applications we conclude that this class of test statistics offers a new and powerful framework for assessing treatment effects in cancer clinical trials involving immunotherapies.

Persistence of Biologics in the Treatment of Psoriatic Arthritis: Data From a Large Hospital-Based Longitudinal Cohort.

OBJECTIVE: To analyze the trends in biologics use at a specialized center over a period of 20 years. METHODS: We performed a retrospective analysis of 571 patients diagnosed with psoriatic arthritis enrolled in the Toronto cohort who initiated biologic therapy between January 1, 2000, and July 7, 2020. The probability of drug persistence over time was estimated nonparametrically. The time to discontinuation of first and second treatment was analyzed using Cox regression models, whereas a semiparametric failure time model with a gamma frailty was used to analyze the discontinuation of treatment over successive administrations of biologic therapy. RESULTS: The highest 3-year persistence probability was observed with certolizumab when used as first biologic treatment, while interleukin-17 inhibitors had the lowest probability. However, when used as second medication, certolizumab had the lowest drug survival even when accounting for selection bias. Depression and/or anxiety were associated with a higher rate of drug discontinuation due to all causes (relative risk [RR] 1.68, P = 0.01), while having higher education was associated with lower rates (RR 0.65, P = 0.03). In the analysis accommodating multiple courses of biologics, a higher tender joint count was associated with a higher rate of discontinuation due to all causes (RR 1.02, P = 0.01). Older age at the start of first treatment was associated with a higher rate of discontinuation due to side effects (RR 1.03, P = 0.01), while obesity had a protective role (RR 0.56, P = 0.05). CONCLUSION: Persistence in taking biologics depends on whether the biologic was used as first or second treatment. Depression and anxiety, higher tender joint count, and older age lead to drug discontinuation.

Multistate models as a framework for estimand specification in clinical trials of complex processes.

Intensity-based multistate models provide a useful framework for characterizing disease processes, the introduction of interventions, loss to followup, and other complications arising in the conduct of randomized trials studying complex life history processes. Within this framework we discuss the issues involved in the specification of estimands and show the limiting values of common estimators of marginal process features based on cumulative incidence function regression models. When intercurrent events arise we stress the need to carefully define the target estimand and the importance of avoiding targets of inference that are not interpretable in the real world. This has implications for analyses, but also the design of clinical trials where protocols may help in the interpretation of estimands based on marginal features.

Two-phase designs with current status data.

We consider the design and analysis of two-phase studies aiming to assess the relation between a fixed (eg, genetic) marker and an event time under current status observation. We consider a common setting in which a phase I sample is comprised of a large cohort of individuals with outcome (ie, current status) data and a vector of inexpensive covariates. Stored biospecimens for individuals in the phase I sample can be assayed to record the marker of interest for individuals selected in a phase II sub-sample. The design challenge is then to select the phase II sub-sample in order to maximize the precision of the marker effect on the time of interest under a proportional hazards model. This problem has not been examined before for current status data and the role of the assessment time is highlighted. Inference based on likelihood and inverse probability weighted estimating functions are considered, with designs centered on score-based residuals, extreme current status observations, or stratified sampling schemes. Data from a registry of patients with psoriatic arthritis is used in an illustration where we study the risk of diabetes as a comorbidity.

Spatial dependence modeling of latent susceptibility and time to joint damage in psoriatic arthritis.

Important scientific insights into chronic diseases affecting several organ systems can be gained from modeling spatial dependence of sites experiencing damage progression. We describe models and methods for studying spatial dependence of joint damage in psoriatic arthritis (PsA). Since a large number of joints may remain unaffected even among individuals with a long disease history, spatial dependence is first modeled in latent joint-specific indicators of susceptibility. Among susceptible joints, a Gaussian copula is adopted for dependence modeling of times to damage. Likelihood and composite likelihoods are developed for settings, where individuals are under intermittent observation and progression times are subject to type K interval censoring. Two-stage estimation procedures help mitigate the computational burden arising when a large number of processes (i.e., joints) are under consideration. Simulation studies confirm that the proposed methods provide valid inference, and an application to the motivating data from the University of Toronto Psoriatic Arthritis Clinic yields important insights which can help physicians distinguish PsA from arthritic conditions with different dependence patterns.

Musculoskeletal Surgery in Psoriatic Arthritis: Prevalence and Risk Factors.

OBJECTIVE: Despite medical therapy, damage occurs in patients with psoriatic arthritis (PsA) requiring musculoskeletal (MSK) surgery. We aimed to describe MSK surgery in patients with PsA and identify risk factors for undergoing first MSK surgery attributable to PsA. METHODS: A single-center cohort identified patients with PsA fulfilling Classification Criteria for Psoriatic Arthritis who had MSK surgery between January 1978 and December 2019 inclusive. Charts were reviewed to confirm surgeries were MSK-related and attributable to PsA. Descriptive statistics determined MSK surgery prevalence and types. Cox proportional hazards models evaluated clinical variables for undergoing first MSK surgery using time-dependent covariates. Using a dataset with 1-to-1 matching on markers of PsA disease severity, a Cox proportional hazards model evaluated the effect of targeted therapies, namely biologics on time to first MSK surgery. RESULTS: Of 1574 patients, 185 patients had 379 MSK surgeries related to PsA. The total number of damaged joints (hazard ratio [HR] 1.03, P < 0.001), tender/swollen joints (HR 1.04, P = 0.01), presence of nail lesions (HR 2.08, P < 0.01), higher Health Assessment Questionnaire scores (HR 2.01, P < 0.001), elevated erythrocyte sedimentation rate (HR 2.37, P = 0.02), and HLA-B27 positivity (HR 2.22, P = 0.048) were associated with increased risk of surgery, whereas higher Psoriasis Area Severity Index (HR 0.88, P < 0.002) conferred a protective effect in a multivariate model. The effect of biologics did not reach statistical significance. CONCLUSION: MSK surgery attributable to PsA is not rare, affecting 11.8% of patients. Markers of cumulative disease activity and damage are associated with a greater risk of requiring surgery.

Patterns of enteric infections in a population-wide cohort study of sequelae, British Columbia, Canada.

We assessed patterns of enteric infections caused by 14 pathogens, in a longitudinal cohort study of sequelae in British Columbia (BC) Canada, 2005-2014. Our population cohort of 5.8 million individuals was followed for an average of 7.5 years/person; during this time, 40 523 individuals experienced 42 308 incident laboratory-confirmed, provincially reported enteric infections (96.4 incident infections per 100 000 person-years). Most individuals (38 882/40 523; 96%) had only one, but 4% had multiple concurrent infections or more than one infection across the study. Among individuals with more than one infection, the pathogens and combinations occurring most frequently per individual matched the pathogens occurring most frequently in the BC population. An additional 298 557 new fee-for-service physician visits and hospitalisations for enteric infections, that did not coincide with a reported enteric infection, also occurred, and some may be potentially unreported enteric infections. Our findings demonstrate that sequelae risk analyses should explore the possible impacts of multiple infections, and that estimating risk for individuals who may have had a potentially unreported enteric infection is warranted.

Glass-ionomer and calcium silicate-based cements interactions with human dentine in health and disease: Two-photon fluorescence microscopy and Raman spectroscopy analysis.

OBJECTIVES: To investigate the potential mineralising effects of calcium silicate-based dentine replacement material (Biodentine™) in comparison with glass-ionomer cement (GIC) (Fuji IX™) on different human dentine substrates using a multimodal non-invasive optical assessment. METHODS: Cements were applied on artificially demineralised or naturally carious dentine and stored for 4 weeks in phosphate-rich media +/- tetracycline used for mineralisation labelling. Interfacial dentine was examined from the same sample and location before and after aging using two-photon fluorescence microscopy, fluorescence lifetime imaging (FLIM) and second harmonic generation (SHG) imaging. Additionally, Raman spectroscopy was used to detect changes in the mineral content of dentine. RESULTS: Significant changes in the fluorescence intensity and lifetime were detected in partially demineralised dentine and caries-affected dentine underneath both tested cements, after storage (p < 0.001). This was associated with a significant increase in the mineral content as indicated by the increased intensity of the phosphate Raman peak located at 959 cm-1 (p < 0.0001). Caries-infected dentine showed significant fluorescence changes under Biodentine™ after storage (p < 0.001), but not under GIC (p = 0.44). Tetracycline binding induced a reduction in the fluorescence lifetime with comparable increase in the fluorescence intensity in both cements' groups within the affected dentine (p < 0.001). Significance Two-photon fluorescence microscopy can be used efficiently for non-destructive in-vitro dentine caries characterisation providing a technique for studying the same dentine-cement interface over time and detect changes. Biodentine™ demonstrated comparable remineralising potential to GIC, in addition to inducing remineralisation of caries-infected dentine. This may suggest using Biodentine™ as part of minimally invasive operative dentistry (MID) in caries management.

Isolated axial disease in psoriatic arthritis and ankylosing spondylitis with psoriasis.

OBJECTIVES: To compare isolated axial psoriatic arthritis (PsA), axial PsA with peripheral involvement and isolated axial ankylosing spondylitis (AS) with psoriasis. To evaluate predictors for developing peripheral disease from isolated axial PsA over time. METHODS: Two PsA and AS cohorts identified patients with PsA with axial disease and isolated axial patients with AS with psoriasis. Logistic regression compared isolated axial PsA to axial PsA with peripheral involvement and isolated axial AS with psoriasis. Cox proportional hazards model evaluated predictors for developing peripheral disease from isolated axial PsA. RESULTS: Of 1576 patients with PsA, 2.03% had isolated axial disease and 29.38% had axial and peripheral disease. human leucocyte antigen HLA-B*27 positivity (OR 25.00, 95% CI 3.03 to 206.11) and lower Health Assessment Questionnaire scores (OR 0.004, 95% CI 0.00 to 0.28) were associated with isolated axial disease. HLA-B*27 also predicted peripheral disease development over time (HR 7.54, 95% CI 1.79 to 31.77). Of 1688 patients with AS, 4.86% had isolated axial disease with psoriasis. Isolated axial patients with PsA were older at diagnosis (OR 1.06, 95% CI 1.01 to 1.13), more likely to have nail lesions (OR 12.37, 95% CI 2.22 to 69.07) and less likely to have inflammatory back pain (OR 0.12, 95% CI 0.02 to 0.61) compared with patients with isolated axial AS with psoriasis. CONCLUSIONS: Isolated axial PsA and AS with psoriasis are uncommon. HLA-B*27 positivity is associated with isolated axial PsA and may identify those who develop peripheral disease over time. Isolated axial PsA is associated with better functional status. Isolated axial PsA appears clinically distinct from isolated axial AS with psoriasis.

Multistate models for the natural history of cancer progression.

BACKGROUND: Multistate models can be effectively used to characterise the natural history of cancer. Inference from such models has previously been useful for setting screening policies. METHODS: We introduce the basic elements of multistate models and the challenges of applying these models to cancer data. Through simulation studies, we examine (1) the impact of assuming time-homogeneous Markov transition intensities when the intensities depend on the time since entry to the current state (i.e., the process is time-inhomogenous semi-Markov) and (2) the effect on precancer risk estimation when observation times depend on an unmodelled intermediate disease state. RESULTS: In the settings we examined, we found that misspecifying a time-inhomogenous semi-Markov process as a time-homogeneous Markov process resulted in biased estimates of the mean sojourn times. When screen-detection of the intermediate disease leads to more frequent future screening assessments, there was minimal bias induced compared to when screen-detection of the intermediate disease leads to less frequent screening. CONCLUSIONS: Multistate models are useful for estimating parameters governing the process dynamics in cancer such as transition rates, sojourn time distributions, and absolute and relative risks. As with most statistical models, to avoid incorrect inference, care should be given to use the appropriate specifications and assumptions.

Adaptive response-dependent two-phase designs: Some results on robustness and efficiency.

Large cohort studies now routinely involve biobanks in which biospecimens are stored for use in future biomarker studies. In such settings, two-phase response-dependent sampling designs involve subsampling individuals in the cohort, assaying their biospecimen to measure an expensive biomarker, and using this data to estimate key parameters of interest under budgetary constraints. When analyses are based on inverse probability weighted estimating functions, recent work has described adaptive two-phase designs in which a preliminary phase of subsampling based on a standard design facilitates approximation of an optimal selection model for a second subsampling phase. In this article, we refine the definition of an optimal subsampling scheme within the framework of adaptive two-phase designs, describe how adaptive two-phase designs can be used when analyses are based on likelihood or conditional likelihood, and consider the setting of a continuous biomarker where the nuisance covariate distribution is estimated nonparametrically at the design stage and analysis stage as required; efficiency and robustness issues are investigated. We also explore these methods for the surrogate variable problem and describe a generalization to accommodate multiple stages of phase II subsampling. A study involving individuals with psoriatic arthritis is considered for illustration, where the aim is to assess the association between the biomarker MMP-3 and the development of joint damage.