The landscape of liver transplantation for patients with Alcohol-associated liver disease in USA.

BACKGROUND/AIMS: Indications for liver transplants have expanded to include patients with alcohol-associated liver disease (ALD) over the last decade. Concurrently, the liver allocation policy was updated in February 2020 replacing the Donor Service Area with Acuity Circles (AC). The aim is to compare the transplantation rate, waitlist outcomes, and post-transplant survival of candidates with ALD to non-ALD and assess differences in that effect after the implementation of the AC policy. APPROACH: Scientific Registry for Transplant Recipients data for adult liver transplant (LT) candidates was reviewed from the post-AC era (2/4/2020 - 3/1/2022) and compared with an equivalent length of time before AC were implemented. RESULTS: The adjusted transplant rates were significantly higher for those with ALD pre-AC, and this difference increased after AC implementation (transplant rate ratio comparing ALD to non-ALD=1.20, 1.13, 1.61, and 1.32 for MELD categories 37-40, 33-36, 29-32, 25-28, respectively, in the post-AC era, p <0.05 for all). The adjusted likelihood of death/removal from the waitlist was lower for ALD patients across all lower MELD categories (aSHR=0.70, 0.81, 0.84, 0.70 for MELD categories at list 25-28, 20-24, 15-19, 6-14, respectively, p <0.05). Adjusted post-transplant survival was better for those with ALD (aHR=0.81, p <0.05). Waiting list and post-transplant mortality tended to improve more for those with ALD since the implementation of AC but not significantly. CONCLUSIONS: ALD is a growing indication for liver transplantation. Although ALD patients continue to have excellent post-transplant outcomes and lower wait list mortality, candidates with ALD have higher adjusted transplant rates, and these differences have increased after AC implementation.

SMART use of medications for the treatment of adolescent severe obesity: A sequential multiple assignment randomized trial protocol.

BACKGROUND: Severe obesity is a complex, chronic disease affecting nearly 9% of adolescents in the U.S. Although the current mainstay of treatment is lifestyle therapy, pediatric clinical practice guidelines recommend the addition of adjunct anti-obesity medication (AOM), such as phentermine and topiramate. However, guidance regarding when adjunct AOM should be started and how AOM should be used is unclear. Furthermore, an inherent limitation of current treatment guidelines is their "one-size-fits-all" approach, which does not account for the heterogeneous nature of obesity and high degree of patient variability in response to all interventions. METHODS: This paper describes the study design and methods of a sequential multiple assignment randomized trial (SMART), "SMART Use of Medications for the Treatment of Adolescent Severe Obesity." The trial will examine 1) when to start AOM (specifically phentermine) in adolescents who are not responding to lifestyle therapy and 2) how to modify AOM when there is a sub-optimal response to the initial pharmacological intervention (specifically, for phentermine non-responders, is it better to add topiramate to phentermine or switch to topiramate monotherapy). Critically, participant characteristics that may differentially affect response to treatment will be assessed and evaluated as potential moderators of intervention efficacy. CONCLUSION: Data from this study will be used to inform the development of an adaptive intervention for the treatment of adolescent severe obesity that includes empirically-derived decision rules regarding when and how to use AOM. Future research will test this adaptive intervention against standard "one-size-fits-all" treatments.

Appetitive and psychological phenotypes of pediatric patients with obesity.

BACKGROUND: Obesity is a heterogeneous disease with variable treatment response. Identification of the unique constellation of contributors to obesity may allow for targeted interventions and improved outcomes. OBJECTIVE: Identify empirically derived phenotypes of pediatric patients with obesity based on appetitive and psychological correlates of obesity. METHODS: This cross-sectional study included patients aged 5-12 years who were treated in a weight management clinic and completed standard intake questionnaires including Child Eating Behavior Questionnaire (CEBQ), Vanderbilt ADHD Scale and Pediatric Symptom Checklist. Phenotypes were elicited using latent profile analysis of 12 indicators: eight CEBQ subscales, inattention, hyperactivity/impulsivity, internalizing and externalizing symptoms. RESULTS: Parents/guardians of 384 patients (mean age 9.8 years, mean BMI 30.3 kg/m2 ) completed the intake questionnaires. A 4-phenotype model best fits the data. Hedonic Impulsive phenotype (42.5%) exhibited high food enjoyment and hyperactivity/impulsivity. Inattentive Impulsive phenotype (27.4%) exhibited overall low food approach and high food avoid behaviours, and highest inattention. Hedonic Emotional phenotype (20.8%) scored the highest on food enjoyment, internalizing and externalizing symptoms. Picky Eating phenotype (9.3%) scored the lowest on food approach, inattention, hyperactivity/impulsivity, internalizing and externalizing symptoms. CONCLUSION: Appetitive traits and psychological symptoms appear to cluster in distinct patterns, giving rise to four unique phenotypic profiles, which, if replicated, may help inform the development of tailored treatment plans.

School Connectedness and Adolescent E-cigarette Susceptibility in an Urban Sample of Middle and High School Students.

Adolescent school connectedness generally protects from risk behaviors such as tobacco use; however, its relationship to e-cigarette use is unclear. This study examines the relationship between adolescent school connectedness and e-cigarette susceptibility in a diverse longitudinal sample. This secondary analysis of a school-based intervention surveyed 608 middle (66%) and high school (34%) students from 10 schools at 3 time points over 1 year. At baseline, respondents had a mean age of 14 years, 54% were female, and 71% were BIPOC (Black, Indigenous, People of Color). Logistic regression models examined unadjusted and adjusted associations between school connectedness (both baseline and concurrent) and e-cigarette susceptibility over time. E-cigarettes represented the most prevalent form of current nicotine-containing product use in spring 2019 (2.3%), and most respondents reported no e-cigarette susceptibility (69%). E-cigarette susceptibility remained relatively stable during the study. Higher baseline school connectedness levels were associated with lower odds of e-cigarette susceptibility over time. Similarly, higher concurrent school connectedness scores were associated with lower odds of e-cigarette susceptibility over time: spring 2019 (OR, 0.39; 95% CI, 0.32, 0.47), fall 2019 (OR, 0.49; 95% CI, 0.34, 0.72), and spring 2020 (OR, 0.64; 95% CI, 0.47, 0.87). Findings were similar for middle and high school students and did not differ significantly after adjusting for other covariates. Adolescents' school connectedness appears to protect from e-cigarette susceptibility over time, underscoring the importance of promoting positive school experiences to reduce adolescent risk e-cigarette use.

Should children receive a kidney transplant before 2 years of age?

BACKGROUND: The optimal age of kidney transplantation for infants and toddlers with kidney failure is unclear. We aimed to evaluate the patient survival associated with kidney transplantation before 2 years of age versus remaining on the waitlist until ≥2 years. METHOD: We used the Scientific Registry of Transplant Recipients to identify all children added to the deceased-donor waitlist before 2 years of age between 1/1/2000 and 4/30/2020. For each case aged <2 years at transplant, we created a control group comprising all candidates on the waitlist on the case's transplant date. Patient survival was evaluated using sequential Cox regression. Dialysis-free time was defined as graft survival time for cases and the sum of dialysis-free time on the waitlist and graft survival time for controls. RESULTS: We observed similar patient survival for posttransplant periods 0-3 and 4-12 months but higher survival for period >12 months for <2-year decreased-donor recipients (aHR: 0.32; 95% CI: 0.13-0.78; p = .01) compared with controls. Similarly, patient survival was higher for <2-year living-donor recipients for posttransplant period >12 months (aHR: 0.21; 95% CI: 0.06-0.73; p = .01). The 5-year dialysis-free survival was higher for <2-year deceased- (difference: 0.59 years; 95% CI: 0.23-0.93) and living-donor (difference: 1.84 years; 95% CI: 1.31-2.25) recipients. CONCLUSION: Kidney transplantation in children <2 years of age is associated with improved patient survival and reduced dialysis exposure compared with remaining on the waitlist until ≥2 years.

Modified interactive Q-learning for attenuating the impact of model misspecification with treatment effect heterogeneity.

A sequential multiple assignment randomized trial, which incorporates multiple stages of randomization, is a popular approach for collecting data to inform personalized and adaptive treatments. There is an extensive literature on statistical methods to analyze data collected in sequential multiple assignment randomized trials and estimate the optimal dynamic treatment regime. Q-learning with linear regression is widely used for this purpose due to its ease of implementation. However, model misspecification is a common problem with this approach, and little attention has been given to the impact of model misspecification when treatment effects are heterogeneous across subjects. This article describes the integrative impact of two possible types of model misspecification related to treatment effect heterogeneity: omitted early-stage treatment effects in late-stage main effect model, and violated linearity assumption between pseudo-outcomes and predictors despite non-linearity arising from the optimization operation. The proposed method, aiming to deal with both types of misspecification concomitantly, builds interactive models into modified parametric Q-learning with Murphy's regret function. Simulations show that the proposed method is robust to both sources of model misspecification. The proposed method is applied to a two-stage sequential multiple assignment randomized trial with embedded tailoring aimed at reducing binge drinking in first-year college students.

Outcome trajectory estimation for optimal dynamic treatment regimes with repeated measures.

In recent sequential multiple assignment randomized trials, outcomes were assessed multiple times to evaluate longer-term impacts of the dynamic treatment regimes (DTRs). Q-learning requires a scalar response to identify the optimal DTR. Inverse probability weighting may be used to estimate the optimal outcome trajectory, but it is inefficient, susceptible to model mis-specification, and unable to characterize how treatment effects manifest over time. We propose modified Q-learning with generalized estimating equations to address these limitations and apply it to the M-bridge trial, which evaluates adaptive interventions to prevent problematic drinking among college freshmen. Simulation studies demonstrate our proposed method improves efficiency and robustness.

A Microsimulation of Well-Being and Literacy Interventions to Reduce Scam Susceptibility in Older Adults.

Poor financial and health literacy and poor psychological well-being are significant correlates of scam susceptibility in older adults; yet, no research has examined whether interventions that target these factors may effectively reduce susceptibility. Using longitudinal data from older adults in the Rush Memory and Aging Project (MAP) (N = 1,231), we used microsimulations to estimate the causal effect of hypothetical well-being and literacy interventions on scam susceptibility over six years. Microsimulations can simulate a randomized trial to estimate intervention effects using observational data. We simulated hypotheticalinterventions that improved well-being or literacy scores by either 10% or 30% from baseline, or to the maximum scores, for an older adult population and for income and education subgroups. Simulations suggest thathypotheticalinterventions that increase well-being or literacy cause statistically significant reductions in scam susceptibility of older adults over time, but improving well-being caused a greater-albeit not significantly different-reduction compared to improving literacy.

Optimizing Longitudinal Tobacco Cessation Treatment in Lung Cancer Screening: A Sequential, Multiple Assignment, Randomized Trial.

Importance: Nearly half of the 14.8 million US adults eligible for lung cancer screening (LCS) smoke cigarettes. The optimal smoking cessation program components for the LCS setting are unclear. Objective: To assess the effect of adding a referral to prescription medication therapy management (MTM) to the tobacco longitudinal care (TLC) program among patients eligible for LCS who smoke and do not respond to early tobacco treatment and to assess the effect of decreasing the intensity of TLC among participants who do respond to early treatment. Design, Setting, and Participants: This randomized clinical trial included patients who currently smoked cigarettes daily and were eligible for LCS. Recruitment took place at primary care centers and LCS programs at 3 large health systems in the US and began in October 2016, and 18-month follow-up was completed April 2021. Interventions: (1) TLC comprising intensive telephone coaching and combination nicotine replacement therapy for 1 year with at least monthly contact; (2) TLC with MTM, MTM offered pharmacist-referral for prescription medications; and (3) Quarterly TLC, intensity of TLC was decreased to quarterly contact. Intervention assignments were based on early response to tobacco treatment (abstinence) that was assessed either 4 weeks or 8 weeks after treatment initiation. Main outcomes and Measures: Self-reported, 6-month prolonged abstinence at 18-month. Results: Of 636 participants, 228 (35.9%) were female, 564 (89.4%) were White individuals, and the median (IQR) age was 64.3 (59.6-68.8) years. Four weeks or 8 weeks after treatment initiation, 510 participants (80.2%) continued to smoke (ie, early treatment nonresponders) and 126 participants (19.8%) had quit (ie, early treatment responders). The 18 month follow-up survey response rate was 83.2% (529 of 636). Across TLC groups at 18 months follow-up, the overall 6-month prolonged abstinence rate was 24.4% (129 of 529). Among the 416 early treatment nonresponders, 6-month prolonged abstinence for TLC with MTM vs TLC was 17.8% vs 16.4% (adjusted odds ratio [aOR] 1.13; 95% CI, 0.67-1.89). In TLC with MTM, 98 of 254 participants (39%) completed at least 1 MTM visit. Among 113 early treatment responders, 6-month prolonged abstinence for Quarterly TLC vs TLC was 24 of 55 (43.6%) vs 34 of 58 (58.6%) (aOR, 0.54; 95% CI, 0.25-1.17). Conclusions and Relevance: In this randomized clinical trial, adding referral to MTM with TLC for participants who did not respond to early treatment did not improve smoking abstinence. Stepping down to Quarterly TLC among early treatment responders is not recommended. Integrating longitudinal tobacco cessation care with LCS is feasible and associated with clinically meaningful quit rates. Trial Registration: ClinicalTrials.gov Identifier: NCT02597491.

Examining engagement effects in an adaptive preventive intervention for college student drinking.

OBJECTIVE: This study determined the characteristics of engagement and whether engagement in an adaptive preventive intervention (API) was associated with reduced binge drinking and alcohol-related consequences. METHOD: Incoming students were recruited for a sequential multiple assignment randomized trial (SMART; N = 891, 62.4% female, 76.8% non-Hispanic White) with an assessment-only control group. The API occurred during the first semester of college, with outcomes assessed at the end of the semester. The API involved two stages. Stage 1 included universal intervention components (personalized normative feedback [PNF] and self-monitoring). Stage 2 bridged heavy drinkers to access additional resources. We estimated the effect of engagement in Stage 1 only and in the whole API (Stages 1 and 2) among the intervention group, and the effect of the API versus control had all students assigned an API engaged, on alcohol-related outcomes. RESULTS: Precollege binge drinking, intention to pledge a fraternity/sorority, and higher conformity motives were most associated with lower odds of Stage 1 engagement. Action (readiness to change) and PNF engagement were associated with Stage 2 engagement. API engagement was associated with significant reductions in alcohol-related consequences among heavy drinkers. Compared to the control, we estimated the API would reduce the relative increase in alcohol-related consequences from baseline to follow-up by 25%, had all API students engaged. CONCLUSIONS: Even partial engagement in each component of the "light-touch" API rendered benefits. Analyses suggested that had all students in the intervention group engaged, the API would significantly reduce the change in alcohol-related consequences over the first semester in college. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Transportability of causal inference under random dynamic treatment regimes for kidney-pancreas transplantation.

A difficult decision for patients in need of kidney-pancreas transplant is whether to seek a living kidney donor or wait to receive both organs from one deceased donor. The framework of dynamic treatment regimes (DTRs) can inform this choice, but a patient-relevant strategy such as "wait for deceased-donor transplant" is ill-defined because there are multiple versions of treatment (i.e., wait times, organ qualities). Existing DTR methods average over the distribution of treatment versions in the data, estimating survival under a "representative intervention." This is undesirable if transporting inferences to a target population such as patients today, who experience shorter wait times thanks to evolutions in allocation policy. We, therefore, propose the concept of a generalized representative intervention (GRI): a random DTR that assigns treatment version by drawing from the distribution among strategy compliers in the target population (e.g., patients today). We describe an inverse-probability-weighted product-limit estimator of survival under a GRI that performs well in simulations and can be implemented in standard statistical software. For continuous treatments (e.g., organ quality), weights are reformulated to depend on probabilities only, not densities. We apply our method to a national database of kidney-pancreas transplant candidates from 2001-2020 to illustrate that variability in transplant rate across years and centers results in qualitative differences in the optimal strategy for patient survival.

Spatially-aware clustering improves AJCC-8 risk stratification performance in oropharyngeal carcinomas.

OBJECTIVE: Evaluate the effectiveness of machine learning tools that incorporate spatial information such as disease location and lymph node metastatic patterns-of-spread, for prediction of survival and toxicity in HPV+ oropharyngeal cancer (OPC). MATERIALS & METHODS: 675 HPV+ OPC patients that were treated at MD Anderson Cancer Center between 2005 and 2013 with curative intent IMRT were retrospectively collected under IRB approval. Risk stratifications incorporating patient radiometric data and lymph node metastasis patterns via an anatomically-adjacent representation with hierarchical clustering were identified. These clusterings were combined into a 3-level patient stratification and included along with other known clinical features in a Cox model for predicting survival outcomes, and logistic regression for toxicity, using independent subsets for training and validation. RESULTS: Four groups were identified and combined into a 3-level stratification. The inclusion of patient stratifications in predictive models for 5-yr Overall survival (OS), 5-year recurrence free survival, (RFS) and Radiation-associated dysphagia (RAD) consistently improved model performance measured using the area under the curve (AUC). Test set AUC improvements over models with clinical covariates, was 9 % for predicting OS, and 18 % for predicting RFS, and 7 % for predicting RAD. For models with both clinical and AJCC covariates, AUC improvement was 7 %, 9 %, and 2 % for OS, RFS, and RAD, respectively. CONCLUSION: Including data-driven patient stratifications considerably improve prognosis for survival and toxicity outcomes over the performance achieved by clinical staging and clinical covariates alone. These stratifications generalize well to across cohorts, and sufficient information for reproducing these clusters is included.

Diagnosed Gonorrhea Among Privately Insured Women: Analysis of United States Claims Data.

Background: Gonorrhea incidence in the United States has risen by nearly 50% in the last decade, while screening rates have increased. Gonorrhea sequelae rates could indicate whether increased gonorrhea incidence is due to better screening. We estimated the association of gonorrhea diagnosis with pelvic inflammatory disease (PID), ectopic pregnancy (EP), and tubal factor infertility (TFI) in women and detected changes in associations over time. Materials and Methods: This retrospective cohort study included 5,553,506 women aged 18-49 tested for gonorrhea in the IBM MarketScan claims administrative database from 2013-2018 in the United States. We estimated incidence rates and hazard ratios (HRs) of gonorrhea diagnosis for each outcome, adjusting for potential confounders using Cox proportional hazards models. We tested the interaction between gonorrhea diagnosis and the initial gonorrhea test year to identify changes in associations over time. Results: We identified 32,729 women with a gonorrhea diagnosis (mean follow-up time in years: PID = 1.73, EP = 1.75, TFI = 1.76). A total of 131,500 women were diagnosed with PID, 64,225 had EP, and 41,507 had TFI. Women with gonorrhea diagnoses had greater incidence per 1000 person-years for all outcomes (PID = 33.5, EP = 9.4, TFI = 5.3) compared to women without gonorrhea diagnoses (PID = 13.9, EP = 6.7, TFI = 4.3). After adjustment, HRs were higher in women with a gonorrhea diagnosis vs. those without [PID = 2.29 (95% confidence interval, CI: 2.15-2.44), EP = 1.57, (95% CI: 1.41-1.76), TFI = 1.70 (95% CI: 1.47-1.97)]. The interaction of gonorrhea diagnosis and test year was not significant, indicating no change in relationship by initial test year. Conclusion: The relationship between gonorrhea and reproductive outcomes has persisted, suggesting a higher disease burden.

The Minnesota attributable risk of kidney donation (MARKD) study: a retrospective cohort study of long-term (> 50 year) outcomes after kidney donation compared to well-matched healthy controls.

BACKGROUND: There is uncertainty about the long-term risks of living kidney donation. Well-designed studies with controls well-matched on risk factors for kidney disease are needed to understand the attributable risks of kidney donation. METHODS: The goal of the Minnesota Attributable Risk of Kidney Donation (MARKD) study is to compare the long-term (> 50 years) outcomes of living donors (LDs) to contemporary and geographically similar controls that are well-matched on health status. University of Minnesota (n = 4022; 1st transplant: 1963) and Mayo Clinic LDs (n = 3035; 1st transplant: 1963) will be matched to Rochester Epidemiology Project (REP) controls (approximately 4 controls to 1 donor) on the basis of age, sex, and race/ethnicity. The REP controls are a well-defined population, with detailed medical record data linked between all providers in Olmsted and surrounding counties, that come from the same geographic region and era (early 1960s to present) as the donors. Controls will be carefully selected to have health status acceptable for donation on the index date (date their matched donor donated). Further refinement of the control group will include confirmed kidney health (e.g., normal serum creatinine and/or no proteinuria) and matching (on index date) of body mass index, smoking history, family history of chronic kidney disease, and blood pressure. Outcomes will be ascertained from national registries (National Death Index and United States Renal Data System) and a new survey administered to both donors and controls; the data will be supplemented by prior surveys and medical record review of donors and REP controls. The outcomes to be compared are all-cause mortality, end-stage kidney disease, cardiovascular disease and mortality, estimated glomerular filtration rate (eGFR) trajectory and chronic kidney disease, pregnancy risks, and development of diseases that frequently lead to chronic kidney disease (e.g. hypertension, diabetes, and obesity). We will additionally evaluate whether the risk of donation differs based on baseline characteristics. DISCUSSION: Our study will provide a comprehensive assessment of long-term living donor risk to inform candidate living donors, and to inform the follow-up and care of current living donors.

Practical guidance on modeling choices for the virtual twins method.

Individuals can vary drastically in their response to the same treatment, and this heterogeneity has driven the push for more personalized medicine. Accurate and interpretable methods to identify subgroups that respond to the treatment differently from the population average are necessary to achieving this goal. The Virtual Twins (VT) method is a highly cited and implemented method for subgroup identification because of its intuitive framework. However, since its initial publication, many researchers still rely heavily on the authors' initial modeling suggestions without examining newer and more powerful alternatives. This leaves much of the potential of the method untapped. We comprehensively evaluate the performance of VT with different combinations of methods in each of its component steps, under a collection of linear and nonlinear problem settings. Our simulations show that the method choice for Step 1 of VT, in which dense models with high predictive performance are fit for the potential outcomes, is highly influential in the overall accuracy of the method, and Superlearner is a promising choice. We illustrate our findings by using VT to identify subgroups with heterogeneous treatment effects in a randomized, double-blind trial of very low nicotine content cigarettes.

Long-term Outcomes Associated With Post-kidney Donation Pregnancy Complications.

BACKGROUND: Kidney donors have increased risk of postdonation gestational hypertension (gHTN) and preeclampsia. In the general population, pregnancy complications are associated with long-term maternal risk. However, little data exist on whether donors with postdonation pregnancy-related complications have similar increased long-term risks. We studied whether postdonation gHTN, preeclampsia/eclampsia, or gestational diabetes (gDM) was associated with increased risk of developing hypertension, DM, cardiovascular disease, or estimated glomerular filtration rate <45 mL/min/1.73 m 2 . METHODS: Postdonation pregnancies with complications were matched to pregnancies without complications based on time from donation. Incidence of outcomes was compared using sequential Cox regression with robust standard errors. Donors with predonation pregnancy complications were excluded. Models were adjusted for age at pregnancy, gravidity, year of donation, and family history of hypertension, DM, and heart disease. RESULTS: Of the 384 donors with postdonation pregnancies (median [quartiles] follow-up of 27.0 [14.2-36.2] y after donation), 39 experienced preeclampsia/eclampsia, 29 gHTN without preeclampsia, and 17 gDM. Median interval from donation to first pregnancy with preeclampsia was 5.1 (2.9-8.6) y; for gHTN, 3.7 (1.9-7.8) y; and for gDM, 7.3 (3.7-10.3) y. Preeclampsia/eclampsia (hazard ratio [HR] 2.70; 95% confidence interval [CI], 1.53-4.77) and gHTN (HR 2.39; 95% CI, 1.24-4.60) were associated with development of hypertension. Preeclampsia/eclampsia (HR 2.15; 95% CI, 1.11-4.16) and gDM (HR 5.60; 95% CI, 1.41-22.15) were associated with development of DM. Pregnancy-related complications were not associated with increased risk of cardiovascular disease or estimated glomerular filtration rate <45 mL/min/1.73 m 2 . CONCLUSIONS: In our single-center study, postdonation preeclampsia, gHTN, or gDM was associated with long-term risk of hypertension or DM.

Can a Single Variable Predict Early Dropout From Digital Health Interventions? Comparison of Predictive Models From Two Large Randomized Trials.

BACKGROUND: A single generalizable metric that accurately predicts early dropout from digital health interventions has the potential to readily inform intervention targets and treatment augmentations that could boost retention and intervention outcomes. We recently identified a type of early dropout from digital health interventions for smoking cessation, specifically, users who logged in during the first week of the intervention and had little to no activity thereafter. These users also had a substantially lower smoking cessation rate with our iCanQuit smoking cessation app compared with users who used the app for longer periods. OBJECTIVE: This study aimed to explore whether log-in count data, using standard statistical methods, can precisely predict whether an individual will become an iCanQuit early dropout while validating the approach using other statistical methods and randomized trial data from 3 other digital interventions for smoking cessation (combined randomized N=4529). METHODS: Standard logistic regression models were used to predict early dropouts for individuals receiving the iCanQuit smoking cessation intervention app, the National Cancer Institute QuitGuide smoking cessation intervention app, the WebQuit.org smoking cessation intervention website, and the Smokefree.gov smoking cessation intervention website. The main predictors were the number of times a participant logged in per day during the first 7 days following randomization. The area under the curve (AUC) assessed the performance of the logistic regression models, which were compared with decision trees, support vector machine, and neural network models. We also examined whether 13 baseline variables that included a variety of demographics (eg, race and ethnicity, gender, and age) and smoking characteristics (eg, use of e-cigarettes and confidence in being smoke free) might improve this prediction. RESULTS: The AUC for each logistic regression model using only the first 7 days of log-in count variables was 0.94 (95% CI 0.90-0.97) for iCanQuit, 0.88 (95% CI 0.83-0.93) for QuitGuide, 0.85 (95% CI 0.80-0.88) for WebQuit.org, and 0.60 (95% CI 0.54-0.66) for Smokefree.gov. Replacing logistic regression models with more complex decision trees, support vector machines, or neural network models did not significantly increase the AUC, nor did including additional baseline variables as predictors. The sensitivity and specificity were generally good, and they were excellent for iCanQuit (ie, 0.91 and 0.85, respectively, at the 0.5 classification threshold). CONCLUSIONS: Logistic regression models using only the first 7 days of log-in count data were generally good at predicting early dropouts. These models performed well when using simple, automated, and readily available log-in count data, whereas including self-reported baseline variables did not improve the prediction. The results will inform the early identification of people at risk of early dropout from digital health interventions with the goal of intervening further by providing them with augmented treatments to increase their retention and, ultimately, their intervention outcomes.

Feature selection for support vector regression using a genetic algorithm.

Support vector regression (SVR) is particularly beneficial when the outcome and predictors are nonlinearly related. However, when many covariates are available, the method's flexibility can lead to overfitting and an overall loss in predictive accuracy. To overcome this drawback, we develop a feature selection method for SVR based on a genetic algorithm that iteratively searches across potential subsets of covariates to find those that yield the best performance according to a user-defined fitness function. We evaluate the performance of our feature selection method for SVR, comparing it to alternate methods including LASSO and random forest, in a simulation study. We find that our method yields higher predictive accuracy than SVR without feature selection. Our method outperforms LASSO when the relationship between covariates and outcome is nonlinear. Random forest performs equivalently to our method in some scenarios, but more poorly when covariates are correlated. We apply our method to predict donor kidney function 1 year after transplant using data from the United Network for Organ Sharing national registry.

Design considerations and analytical framework for reliably identifying a beneficial individualized treatment rule.

An individualized treatment rule (ITR) formalizes personalized medicine by assigning treatment as a function of patients' clinical information, which contrasts with a static treatment rule that assigns everyone the same treatment. ITR identification has become a common aim in randomized clinical trials but sample size considerations for this aim are lacking. One approach is to select a sample size that will reliably identify an ITR with a performance close to the theoretical optimal rule. However, this approach could still lead to identifying ITRs that perform worse than the optimal static rule, particularly in the absence of substantial effect heterogeneity. This limitation motivates sample size considerations aimed at reliable identification of a beneficial ITR, which outperforms the optimal static rule, and analysis methods that identify the estimated optimal static rule when there is substantial uncertainty about whether an ITR will improve outcomes. To address these limitations, we propose a sample size approach based on the probability of identifying a beneficial ITR and introduce an approach for selecting the LASSO penalty parameter such that in the absence of treatment effect heterogeneity the estimated optimal static rule is identified with high probability. We apply these approaches to the PLUTO trial aimed at developing methods to assist with smoking cessation.