Estimate Treatment Effects Research Articles

Learning to Be Fair: A Consequentialist Approach to Equitable Decision Making

In an attempt to make algorithms fair, the machine learning literature has largely focused on equalizing decisions, outcomes, or error rates across race or gender groups. To illustrate, consider a hypothetical government rideshare program that provides transportation assistance to low-income people with upcoming court dates. Following this literature, one might allocate rides to those with the highest estimated treatment effect per dollar while constraining spending to be equal across race groups. That approach, however, ignores the downstream consequences of such constraints and, as a result, can induce unexpected harm. For instance, if one demographic group lives farther from court, enforcing equal spending would necessarily mean fewer total rides provided and potentially more people penalized for missing court. Here we present an alternative framework for designing equitable algorithms that foregrounds the consequences of decisions. In our approach, one first elicits stakeholder preferences over the space of possible decisions and the resulting outcomes—such as preferences for balancing spending parity against court appearance rates. We then optimize over the space of decision policies, making trade-offs in a way that maximizes the elicited utility. To do so, we develop an algorithm for efficiently learning these optimal policies from data for a large family of expressive utility functions. In particular, we use a contextual bandit algorithm to explore the space of policies while solving a convex optimization problem at each step to estimate the best policy based on the available information. This consequentialist paradigm facilitates a more holistic approach to equitable decision making. This paper has been This paper was accepted by Catherine Tucker, Special Issue on The Human-Algorithm Connection. Funding: This work was supported by Stanford Impact Labs [Start-Up Impact Labs/GEJBU], the Stanford Institute for Human-Centered Artificial Intelligence, and Harvard Data Science Initiative. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.00345 .

Effect of a multicomponent, person-centred care intervention on client experience and HIV treatment outcomes in Zambia: a stepped-wedge, cluster-randomised trial.

Recipients of health services value not only convenience but also respectful, kind, and helpful providers. To date, research to improve person-centred HIV treatment has focused on making services easier to access (eg, differentiated service delivery) rather than the interpersonal experience of care. We developed and evaluated a person-centred care (PCC) intervention targeting practices of health-care workers. Using a stepped-wedge, cluster-randomised design, we randomly allocated 24 HIV clinics stratified by size in Zambia into four groups and introduced a PCC intervention that targeted caring aspects of the behaviour of health-care workers in one group every 6 months. The intervention entailed training and coaching for health-care workers on PCC practices (to capacitate), client experience assessment with feedback to facilities (to create opportunities), and small performance-based incentives (to motivate). In a probability sample of clients who were pre-trained on a client experience exit survey and masked to facility intervention status, we evaluated effects on client experience by use of mean score change and also proportion with poor encounters (ie, score of ≤8 on a 12-point survey instrument). We examined effects on missed visits (ie, >30 days late for next scheduled encounter) in all groups and retention in care at 15 months in group 1 and group 4 by use of electronic health records. We assessed effects on treatment success at 15 months (ie, HIV RNA concentration <400 copies per mL or adjudicated care status) in a prospectively enrolled subset of clients from group 1 and group 4. We estimated treatment effects with mixed-effects logistic regression, adjusting for sex, age, and baseline care status. This trial is registered at the Pan-African Clinical Trials Registry (202101847907585), and is completed. Between Aug 12, 2019, and Nov 30, 2021, 177 543 unique clients living with HIV made at least one visit to one of the 24 study clinics. The PCC intervention reduced the proportion of poor visits based on exit surveys from 147 (23·3%) of 632 during control periods to 33 (13·3%) of 249 during the first 6 months of intervention, and then to eight (3·5%) of 230 at 6 months or later (adjusted risk difference [aRD] for control vs ≥6 months intervention -16·9 percentage points, 95% CI -24·8 to -8·9). Among all adult scheduled appointments, the PCC intervention reduced the proportion of missed visits from 90 593 (25·3%) of 358 741 during control periods to 40 380 (22·6%) of 178 523 in the first 6 months, and then 52 288 (21·5%) of 243 350 at 6 months or later (aRD for control vs the intervention -4·2 percentage points, 95% CI -4·8 to -3·7). 15-month retention improved from 33 668 (80·2%) of 41 998 in control to 35 959 (83·6%) of 43 005 during intervention (aRD 5·9 percentage points, 95% CI 0·6 to 11·2), with larger effects in clients newly starting treatment (aRD 12·7 percentage points, 1·4 to 23·9). We found no effect on treatment success (based on viral load) in a nested subcohort (379 [83·7%] of 453 in the control phase vs 402 [83·8%] of 480 in the intervention phase; aRD 0·9 percentage points, -5·4 to 7·2). Improving the caring aspects of health-care worker behaviour is feasible in public health settings, enhances client experience, reduces missed appointments, and increases retention. The Bill & Melinda Gates Foundation.

Cardiovascular, Kidney, and Safety Outcomes With GLP-1 Receptor Agonists Alone and in Combination With SGLT2 Inhibitors in Type 2 Diabetes: A Systematic Review and Meta-Analysis.

GLP-1 (glucagon-like peptide-1) receptor agonists and SGLT2 (sodium-glucose cotransporter 2) inhibitors both improve cardiovascular and kidney outcomes in people with type 2 diabetes. We conducted a systematic review and meta-analysis to assess the effects of GLP-1 receptor agonists on clinical outcomes with and without SGLT2 inhibitors. We searched MEDLINE and Embase databases from inception until July 12, 2024, for randomized, double-blind, placebo-controlled outcome trials of GLP-1 receptor agonists in type 2 diabetes that reported treatment effects by baseline use of SGLT2 inhibitors, with findings supplemented by unpublished data. We estimated treatment effects by baseline SGLT2 inhibitor use using inverse variance-weighted meta-analysis. The main cardiovascular outcomes were major adverse cardiovascular events (nonfatal myocardial infarction, stroke, or cardiovascular death) and hospitalization for heart failure. Kidney outcomes included a composite of ≥50% reduction in estimated glomerular filtration rate, kidney failure or death caused by kidney failure, and annualized rate of decline in estimated glomerular filtration rate (estimated glomerular filtration rate slope). Serious adverse events and severe hypoglycemia were also evaluated. This meta-analysis was registered on the International Prospective Register of Systematic Reviews (PROSPERO; CRD42024565765). We identified 3 trials with 1743 of 17 072 (10.2%) participants with type 2 diabetes receiving an SGLT2 inhibitor at baseline. GLP-1 receptor agonists reduced the risk of major adverse cardiovascular events by 21% (hazard ratio [HR], 0.79 [95% CI, 0.71-0.87]), with consistent effects in those receiving and not receiving SGLT2 inhibitors at baseline (HR, 0.77 [95% CI, 0.54-1.09] and HR, 0.79 [95% CI, 0.71-0.87], respectively; P-heterogeneity=0.78). The effect on hospitalization for heart failure was similarly consistent regardless of SGLT2 inhibitor use (HR, 0.58 [95% CI, 0.36-0.93] and HR, 0.73 [95% CI, 0.63-0.85]; P-heterogeneity=0.26). Effects on the composite kidney outcome (risk ratio, 0.79 [95% CI, 0.66-0.95]) and estimated glomerular filtration rate slope (0.78 mL/min/1.73 m2/y [95% CI, 0.57-0.98]) also did not vary according to SGLT2 inhibitor use (P-heterogeneity=0.53 and 0.94, respectively). Serious adverse effects and severe hypoglycemia were also similar regardless of SGLT2 inhibitor use (P-heterogeneity=0.29 and 0.50, respectively). In people with type 2 diabetes, the cardiovascular and kidney benefits of GLP-1 receptor agonists are consistent regardless of SGLT2 inhibitor use.

Designing observational studies to estimate treatment effects in geriatric oncology: A closer look at confounding and its mitigation.

Designing observational studies to estimate treatment effects in geriatric oncology: A closer look at confounding and its mitigation.

Covariate balancing for high-dimensional samples in controlled experiments

In controlled experiments, achieving covariate balancing across all groups is crucial as it ensures that the estimated treatment effects are not confounded by the effects of covariates. This study proposes a mixed-integer nonlinear programming model to address the covariate balancing problem. Specifically, we introduce a new covariate imbalance measure, which is the maximum discrepancy in both the first and second central moments between any two groups. The second central moment can effectively capture the correlation of covariates in a physical sense, which is crucial for partitioning high-dimensional samples. A mixed-integer nonlinear programming model is constructed to minimize the proposed measure to obtain the optimal partitioning results. The nonlinear model is then linearized to accelerate the optimization process. We conduct computational experiments based on simulated datasets, including one-dimensional, two-dimensional, and three-dimensional Gaussian distributed samples, and a real clinic trial dataset. Compared to the conventional discrepancy-based method, our method achieves a 54.81% and a 40.6% reduction in the maximum discrepancy of partitioning results in the two-dimensional simulated Gaussian samples and the real clinic trial dataset, respectively. These results demonstrate the superiority of the proposed model in partitioning high-dimensional samples with correlated covariates compared with the conventional discrepancy-based method.

Possible Glycemic Effects of Vagus Nerve Stimulation Evaluated by Continuous Glucose Monitoring in People with Diabetes and Autonomic Neuropathy: A Randomized, Sham-Controlled Trial.

Objective: Autonomic neuropathy is associated with dysglycemia that is difficult to control. We investigated if transcutaneous vagus nerve stimulation (tVNS) could improve glycemic levels. Methods: We randomized 145 individuals with type 1 diabetes (T1D) (n = 70) or type 2 diabetes (T2D) (n = 75) and diabetic autonomic neuropathy (DAN) to self-administered treatment with active cervical tVNS (n = 68) or sham (n = 77) for 1 week (4 daily stimulations) and 8 weeks (2 daily stimulations), separated by a wash-out period of at least 2 weeks. Continuous glucose monitoring (CGM) indices were measured for 104 participants starting 5 days prior to intervention periods, during the 1-week period, and at end of the 8-week period. Primary outcomes were between-group differences in changes in coefficient of variation (CV) and in time in range (TIR 3.9-10 mmol/L). Secondary outcomes were other metrics of CGM and HbA1c. Results: For the 1-week period, median [interquartile range] changes of CV from baseline to follow-up were -1.1 [-4.3;2.0] % in active and -1.5 [-4.4;2.5] % in sham, with no significance between groups (P = 0.54). For TIR, the corresponding changes were 2.4 [-2.1;7.4] % in active and 5.1 [-2.6;8.8] in sham group (P = 0.84). For the 8-week treatment period, changes in CV and TIR between groups were also nonsignificant. However, in the subgroup analysis, persons with T1D receiving active tVNS for 8 weeks had a significant reduction in CV compared with the T1D group receiving sham stimulation (estimated treatment effect: -11.6 [95% confidence interval -20.2;-2.0] %, P = 0.009). None of the changes in secondary outcomes between treatment groups were significantly different. Conclusions: Overall, no significant changes were observed in CGM metrics between treatment arms, while individuals with T1D and DAN decreased their CV after 8 weeks of tVNS treatment.

Comparison of statistical methods for the analysis of patient-reported outcomes in randomised controlled trials: A simulation study.

Patient-reported outcomes(PROs) that aim to measure patients' subjective attitudes towards their health or health-related conditions in various fields have been increasingly used in randomised controlled trials(RCTs). PRO data is likely to be bounded, discrete, and skewed. Although various statistical methods are available for the analysis of PROs in RCT settings, there is no consensus on what statistical methods are the most appropriate for use. This study aims to use simulation methods to compare the performance (in terms of bias, empirical standard error, coverage of the confidence interval, Type I error, and power) of three different statistical methods, multiple linear regression(MLR), Tobit regression (Tobit), and median regression (Median), to estimate a range of predefined treatment effects for a PRO in a two-arm balanced RCT. We assumed there was an underlying latent continuous outcome that the PRO was measuring, but the actual scores observed were equally spaced and discrete. This study found that MLR was associated with little bias of the estimated treatment effect, small standard errors, and appropriate coverage of the confidence interval under most scenarios. Tobit performed worse than MLR for analysing PROs with a small number of levels, but it had better performance when analysing PROs with more discrete values. Median showed extremely large bias and errors, associated with low power and coverage for most scenarios especially when the number of possible discrete values was small. We recommend MLR as a simple and universal statistical method for the analysis of PROs inRCT settings.

The application of Bayesian method in estimating heterogeneity of treatment effect

It is well recognized that the benefit of medical interventions may not be distributed evenly in the target population due to patient heterogeneity, and conclusions based on traditional randomized clinical trials do not apply to everyone. Considering the increasing cost of randomized trials and difficulties in recruiting patients, there is an urgent need to develop analytical methods to estimate treatment effects in subgroups. In particular, due to limited sample sizes of subgroups, standard analysis tends to yield wide confidence intervals for the treatment effect, which are often uninformative. In this article, the superiority of the empirical Bayes estimate of the conditional average treatment effect is demonstrated by using a linear regression model to specifically represent treatment-covariate interactions. To obtain more accurate prior distributions and posterior estimates of treatment effects in subgroups, we utilize the covariate-dependent precision parameter across subgroups. Based on subgroup data obtained from clinical trials, the covariate-dependent precision parameter is used to assess the similarity of treatment effects across subgroups, which is an approach to obtain more reasonable prior information. To verify the validity of the empirical Bayes method, we evaluate it by a real clinical trial comparing insulin and metformin.

Model Averaging for Estimating Treatment Effects With Binary Responses

ABSTRACTIn this article, we present a novel approach for estimating the conditional average treatment effect in models with binary responses. Our proposed method involves model averaging, and we establish a weight choice criterion based on jackknife model averaging. We analyze the theoretical properties of this approach, including its asymptotic optimality in achieving the lowest possible squared error and the convergence rate of the weights assigned to correctly specified models. Additionally, we introduce a new matching method that combines partition and nearest neighbor pairing, leveraging the strengths of both techniques. To evaluate the performance of our method, we conduct comparisons with existing approaches via a Monte Carlo study and a real data analysis. Overall, our results demonstrate the effectiveness and practicality of our proposed approach for estimating the conditional average treatment effect in binary response models.

Analyzing missingness patterns in real-world data using the SMDI toolkit: application to a linked EHR-claims pharmacoepidemiology study

BackgroundMissing data in confounding variables present a frequent challenge in generating evidence using real-world data, including electronic health records (EHR). Our objective was to apply a recently published toolkit for characterizing missing data patterns and based on the toolkit results about likely missingness mechanisms, illustrate the decision-making process for analyses in an empirical case example.MethodsWe utilized the Structural Missing Data Investigations (SMDI) toolkit to characterize missing data patterns in the context of a pharmacoepidemiology study comparing cardiovascular outcomes of initiating sodium-glucose-cotransporter-2 inhibitors (SGLT2i) and dipeptidyl peptidase‐4 inhibitors (DPP‐4i) among older adults. The study used a linked EHR-Medicare claims dataset from Duke Health patients (2015–2017), focusing on partially observed confounders from EHR data (HbA1c lab and body mass index [BMI] values). Our analysis incorporated SMDI's descriptive functions and diagnostic tests to explore missingness patterns and determine missingness mitigation approaches. We used findings from these investigations to inform estimation of adjusted hazard ratios comparing the two classes of medications.ResultsHigh levels of missingness were noted for important confounding variables including HbA1c (63.6%) and BMI (16.5%). Diagnostic tests resulted in output that described: 1) the distributions of patient characteristics, exposure, and outcome between patients with or without an observed value of the partially observed covariate, 2) the ability to predict missingness based on observed covariates, and 3) estimate if the missingness of a partially observed covariate is differential with respect to the outcome. There was evidence that missingness could be sufficiently described using observed data, which allowed multiple imputation by chained equations using random forests to address missing confounder data in estimating treatment effects. Multiple imputation resulted in improved alignment of effect estimates with previous studies.ConclusionsWe were able to demonstrate the practical application of the SMDI toolkit in a real-world setting. Application of the SMDI toolkit and the resulting insights of potential missingness patterns can inform the choice of appropriate analytic methods and increase transparency of research methods in handling missing data. This type of approach can inform analytic decision making and may increase our ability to generate evidence from real-world data.

Disentangled latent factors for muti-cause treatment effect estimation

Abstract Existing methods estimate treatment effects from observational data and assume that covariates are all confounders. However, observed covariates may not directly represent confounding variables that influence both treatment and outcome. They always include variables that only affect the treatment or the outcome. In addition, for multi-dimensional binary treatments, disentangled methods are mainly designed for binary variables and ignore the impact of multi-cause treatment variables on the inference of latent factors. To address these two issues, based on variable decomposition and proxy inference, we propose the Disentangled Latent Factors for Multi-cause Treatment Estimation (DEMTE) algorithm. It utilizes an identifiable autoencoder to infer and disentangle latent factors based on the joint distribution of variables in observational data. DEMTE evaluates the treatment effect on the disentangled factors. Synthetic experiments and semi-synthetic experiments demonstrate the effectiveness of the inference and disentanglement techniques and our method achieves more accurate treatment effect estimation.

Estimating Treatment Effects Using Observational Data and Experimental Data with Non-Overlapping Support

Estimating Treatment Effects Using Observational Data and Experimental Data with Non-Overlapping Support

Modelling wound area in studies of wound healing interventions

BackgroundExperimental studies of wound healing often use survival analysis and time to event outcomes or differences in wound area at a specific time point. However, these methods do not use a potentially large number of observations made over the course of a trial and may be inefficient. A model-based approach can leverage all trial data, but there is little guidance on appropriate models and functional forms to describe wound healing.MethodsWe derive a general statistical model and review a wide range of plausible mathematical models to describe wound healing. We identify a range of possible derived estimands and their derivation from the models. Using data from a trial of an intervention to promote ulcer healing in patients affected by leprosy that included three measurement methods repeated across the course of the study, we compare the goodness-of-fit of the models using a range of methods and estimate treatment effects and healing rate functions with the best-fitting models.ResultsOverall, we included 5,581 ulcer measurements of 1,578 unique images from 130 patients. We examined the performance of a range of models. The square root, log square root, and log quadratic models were the best fitting models across all outcome measurement methods. The estimated treatment effects magnitude and sign varied by time post-randomisation, model type, and outcome type, but across all models there was little evidence of effectiveness. The estimated effects were significantly more precise than non-parametric alternatives. For example, estimated differences from the three outcome measurements at 42-days post-randomisation were − 0.01 cm2 (-0.77, 0.74), -0.44 cm2 (-1.64, 0.76), and 0.11 cm2 (-0.87, 1.08) using a non-parametric method versus − 0.03 cm2 (-0.14, 0.06), 0.06 cm2 (-0.05, 0.17), and 0.03 cm2 (-0.07, 0.17) using a square-root model.ConclusionsModel-based analyses can dramatically improve the precision of estimates but care must be taken to carefully compare and select the best fitting models. The (log) square-root model is strongly recommended reflecting advice from a century ago.

Covariate-constrained randomization in cluster randomized 2 × 2 factorial trials: application to a diabetes prevention study

BackgroundCluster randomized trials (CRTs) are randomized trials where randomization takes place at an administrative level (e.g., hospitals, clinics, or schools) rather than at the individual level. When the number of available clusters is small, researchers may not be able to rely on simple randomization to achieve balance on cluster-level covariates across treatment conditions. If these cluster-level covariates are predictive of the outcome, covariate imbalance may distort treatment effects, threaten internal validity, lead to a loss of power, and increase the variability of treatment effects. Covariate-constrained randomization (CR) is a randomization strategy designed to reduce the risk of imbalance in cluster-level covariates when performing a CRT. Existing methods for CR have been developed and evaluated for two- and multi-arm CRTs but not for factorial CRTs.MethodsMotivated by the BEGIN study—a CRT for weight loss among patients with pre-diabetes—we develop methods for performing CR in 2 × 2 factorial cluster randomized trials with a continuous outcome and continuous cluster-level covariates. We apply our methods to the BEGIN study and use simulation to assess the performance of CR versus simple randomization for estimating treatment effects by varying the number of clusters, the degree to which clusters are associated with the outcome, the distribution of cluster level covariates, the size of the constrained randomization space, and analysis strategies.ResultsCompared to simple randomization of clusters, CR in the factorial setting is effective at achieving balance across cluster-level covariates between treatment conditions and provides more precise inferences. When cluster-level covariates are included in the analyses model, CR also results in greater power to detect treatment effects, but power is low compared to unadjusted analyses when the number of clusters is small.ConclusionsCR should be used instead of simple randomization when performing factorial CRTs to avoid highly imbalanced designs and to obtain more precise inferences. Except when there are a small number of clusters, cluster-level covariates should be included in the analysis model to increase power and maintain coverage and type 1 error rates at their nominal levels.

Positive Airway Pressure, Mortality, and Cardiovascular Risk in Older Adults With Sleep Apnea

Positive airway pressure (PAP) is the first-line treatment for obstructive sleep apnea (OSA), but evidence on its beneficial effect on major adverse cardiovascular events (MACE) and mortality prevention is limited. To determine whether PAP initiation and utilization are associated with lower mortality and incidence of MACE among older adults with OSA living in the central US. This retrospective clinical cohort study included Medicare beneficiaries with 2 or more distinct OSA claims identified from multistate, statewide, multiyear (2011-2020) Medicare fee-for-service claims data. Individuals were followed up until death or censoring on December 31, 2020. Analyses were performed between December 2021 and December 2023. Evidence of PAP initiation and utilization based on PAP claims after OSA diagnosis. All-cause mortality and MACE, defined as a composite of myocardial infarction, heart failure, stroke, or coronary revascularization. Doubly robust Cox proportional hazards models with inverse probability of treatment weights were used to estimate treatment effect sizes controlling for sociodemographic and clinical factors. Among 888 835 beneficiaries with OSA included in the analyses (median [IQR] age, 73 [69-78] years; 390 598 women [43.9%]; 8115 Asian [0.9%], 47 122 Black [5.3%], and 760 324 White [85.5%] participants; median [IQR] follow-up, 3.1 [1.5-5.1] years), those with evidence of PAP initiation (290 015 [32.6%]) had significantly lower all-cause mortality (hazard ratio [HR], 0.53; 95% CI, 0.52-0.54) and MACE incidence risk (HR, 0.90; 95% CI, 0.89-0.91). Higher quartiles (Q) of annual PAP claims were progressively associated with lower mortality (Q2 HR, 0.84; 95% CI, 0.81-0.87; Q3 HR, 0.76; 95% CI, 0.74-0.79; Q4 HR, 0.74; 95% CI, 0.72-0.77) and MACE incidence risk (Q2 HR, 0.92; 95% CI, 0.89-0.95; Q3 HR, 0.89; 95% CI, 0.86-0.91; Q4 HR, 0.87; 95% CI, 0.85-0.90). In this cohort study of Medicare beneficiaries with OSA, PAP utilization was associated with lower all-cause mortality and MACE incidence. Results might inform trials assessing the importance of OSA therapy toward minimizing cardiovascular risk and mortality in older adults.

Synthesis estimators for transportability with positivity violations by a continuous covariate

Abstract Studies intended to estimate the effect of a treatment, like randomized trials, may not be sampled from the desired target population. To correct for this discrepancy, estimates can be transported to the target population. Methods for transporting between populations are often premised on a positivity assumption, such that all relevant covariate patterns in one population are also present in the other. However, eligibility criteria, particularly in the case of trials, can result in violations of positivity when transporting to external populations. To address nonpositivity, a synthesis of statistical and mathematical models can be considered. This approach integrates multiple data sources (e.g. trials, observational, pharmacokinetic studies) to estimate treatment effects, leveraging mathematical models to handle positivity violations. This approach was previously demonstrated for positivity violations by a single binary covariate. Here, we extend the synthesis approach for positivity violations with a continuous covariate. For estimation, two novel augmented inverse probability weighting estimators are proposed. Both estimators are contrasted with other common approaches for addressing nonpositivity. Empirical performance is compared via Monte Carlo simulation. Finally, the competing approaches are illustrated with an example in the context of two-drug vs. one-drug antiretroviral therapy on CD4 T cell counts among women with HIV.

SOX-5 Transcription Factor: a Novel Psoriatic Autoantigen Preferentially Found in Women.

Adaptive immunity mediates psoriatic disease pathogenesis. We aimed to identify novel psoriatic autoantigens and their phenotypic associations in deeply characterized patient cohorts. Sera from psoriatic arthritis (PsA) patients were used for autoantibody discovery. Immunoprecipitations performed with cell lysates were on-bead digested, and autoantigens were identified by mass spectrometry. Prevalence and clinical features associated with anti-SRY-Box transcription factor-D (SOX-D) antibodies were determined by screening discovery cohorts of patients with PsA (n = 135), patients with psoriasis without PsA (n = 24), and healthy controls (n = 41). A PsA validation cohort (n = 325) and disease control samples of individuals with rheumatoid arthritis (RA; n = 66) and systemic lupus erythematosus (SLE, n = 66) were assayed for anti-SOX5 antibodies. Disease characteristics were compared by antibody status. Longitudinal data were analyzed using linear mixed-effects models with patient-specific intercept to ascertain associations. We also tested PsA sera for the recently described anti-ADAMTS-L5 autoantibody in PsA. The novel autoantigens identified were SOX-D transcription factors, with SOX-5 being the focus of this analysis. Anti-SOX5 antibodies were present in 8.9% (12 of 135) and 4.3% (14 of 323) of patients in the PsA discovery and validation cohorts, respectively, 12.5% of patients (3 of 24) in the psoriasis group, 2.4% (1 of 41) of healthy controls, and 7.6% (5 of 66) each of patients in the RA and SLE groups. Anti-SOX5 were associated with female sex in both PsA cohorts (discovery: 15.7% women, 2.6% men, P = 0.006; validation: 6.3% women, 1.4% men, P = 0.049). In a longitudinal analysis adjusted for sex, anti-SOX5 associated with biologic disease-modifying antirheumatic drug treatment (95% vs 61%; P = 0.001; n = 96) and with differences in estimated treatment effects by mechanism of action. Anti-ADAMTS-L5 autoantibodies were identified in 8 of 124 patients (6.5%) in the PsA group. SOX-D transcription factors are novel psoriatic autoantigens. Anti-SOX5 antibodies were preferentially found in women with PsA and associated with specific clinical and treatment characteristics, suggesting that anti-SOX5 antibodies may identify mechanistic subgroups. We independently validated anti-ADAMTS-L5 autoantibodies in PsA.

Surrogate endpoint metaregression: useful statistics for regulators and trialists

ObjectivesThe main purpose of using a surrogate endpoint is to estimate the treatment effect on the true endpoint sooner than with a true endpoint. Based on a metaregression of historical randomized trials with surrogate and true endpoints, we discuss statistics for applying and evaluating surrogate endpoints. MethodsWe computed statistics from 2 types of linear metaregressions for trial-level data: simple random effects and novel random effects with correlations among estimated treatment effects in trials with more than 2 arms. A key statistic is the estimated intercept of the metaregression line. An intercept that is small or not statistically significant increases confidence when extrapolating to a new treatment because of consistency with a single causal pathway and invariance to labeling of treatments as controls. For a regulator applying the metaregression to a new treatment, a useful statistic is the 95% prediction interval. For a clinical trialist planning a trial of a new treatment, useful statistics are the surrogate threshold effect proportion, the sample size multiplier adjusted for dropouts, and the novel true endpoint advantage. ResultsWe illustrate these statistics with surrogate endpoint metaregressions involving antihypertension treatment, breast cancer screening, and colorectal cancer treatment. ConclusionRegulators and trialists should consider using these statistics when applying and evaluating surrogate endpoints.

Tissue-source effect on mesenchymal stem cells as living biodrugs for heart failure: Systematic review and meta-analysis.

Mesenchymal stem cells (MSCs), as living biodrugs, have entered advanced phases of clinical assessment for cardiac function restoration in patients with myocardial infarction and heart failure. While MSCs are available from diverse tissue sources, bone-marrow-derived MSCs (BM-MSCs) remain the most well-studied cell type, besides umbilical-cord-derived MSCs (UC-MSCs). The latter offers advantages, including noninvasive availability without ethical considerations. To compare the safety and efficacy of BM-MSCs and UC-MSCs in terms of left ventricular ejection fraction (LVEF), 6-min walking distance (6MWD), and major adverse cardiac events (MACEs). Five databases were systematically searched to identify randomized controlled trials (RCTs). Thirteen RCTs (693 patients) were included using predefined eligibility criteria. Weighted mean differences and odds ratio (OR) for the changes in the estimated treatment effects. UC-MSCs significantly improved LVEF vs controls by 5.08% [95% confidence interval (CI): 2.20%-7.95%] at 6 mo and 2.78% (95%CI: 0.86%-4.70%) at 12 mo. However, no significant effect was observed for BM-MSCs vs controls. No significant changes were observed in the 6MWD with either of the two cell types. Also, no differences were observed for MACEs, except rehospitalization rates, which were lower only with BM-MSCs (odds ratio 0.48, 95%CI: 0.24-0.97) vs controls. UC-MSCs significantly improved LVEF compared with BM-MSCs. Their advantageous characteristics position them as a promising alternative to MSC-based therapy.

Treatment of hidradenitis suppurativa with adalimumab in the PIONEER I and II trials reduced indices of systemic inflammation, recognised risk factors for cardiovascular disease.

Hidradenitis suppurativa (HS) is associated with increased cardiovascular disease (CVD) risk. Systemic immune inflammation index (SII), neutrophil/lymphocyte (NLR), platelet/lymphocyte (PLR) and monocyte/lymphocyte ratios (MLR) are biomarkers of systemic inflammation and CVD. One small study identified a lower NLR and PLR in patients treated with adalimumab. Our aim was to assess changes in SII, NLR, PLR and MLR in a larger cohort, and to evaluate their association with disease severity and treatment response. This was a post-hoc analysis of PIONEER I and PIONEER II, two phase 3 randomised placebo-controlled clinical trials of adalimumab for HS. SII, NLR, PLR and MLR were log10-transformed and linear mixed model was used to estimate treatment effect. SII, NLR, PLR and MLR reduced from baseline levels with adalimumab treatment by week 12, when the primary response endpoint was assessed. Significant changes first appeared at week 4 and were maintained to week 36. In contrast, no significant changes were observed in placebo-treated patients. In patients re-randomised at week 12 from placebo to adalimumab, SII, NLR, PLR and MLR also reduced within 4 weeks. In patients re-randomised from adalimumab to placebo, these biomarkers returned to baseline by week 36. In addition, SII, NLR and PLR correlated with draining fistula count (r=0.26-0.43, p<0.001). Adalimumab non-responders in PIONEER I had a higher SII, NLR and PLR at baseline and week 12, but this was not significant when adjusted for draining fistulae. Treatment of HS patients with adalimumab results in rapid sustained reduction in systemic inflammation measured by SII, NLR, PLR and MLR which correlate with CVD risk. SII, NLR and PLR may predict adalimumab response, although dependent on their interaction with the number of draining fistulae.