FHBF: Federated hybrid boosted forests with dropout rates for supervised learning tasks across highly imbalanced clinical datasets

Although several studies have deployed gradient boosting trees (GBT) as a robust classifier for federated learning tasks (federated GBT [FGBT]), even with dropout rates (federated gradient boosting trees with dropout rate [FDART]), none of them have investigated the overfitting effects of FGBT across heterogeneous and highly imbalanced datasets within federated environments nor the effect of dropouts in the loss function. In this work, we present the federated hybrid boosted forests (FHBF) algorithm, which incorporates a hybrid weight update approach to overcome ill-posed problems that arise from overfitting effects during the training across highly imbalanced datasets in the cloud. Eight case studies were conducted to stress the performance of FHBF against existing algorithms toward the development of robust AI models for lymphoma development across 18 European federated databases. Our results highlight the robustness of FHBF, yielding an average loss of 0.527 compared with FGBT (0.611) and FDART (0.584) with increased classification performance (0.938 sensitivity, 0.732 specificity).

Professional strategies in upper secondary school dropout management among youth in the Sami areas of Norway: a focus group study

The upper secondary school dropout rate is a challenge in many western countries, and measures have been taken to prevent dropout. The dropout rate in Norway is stable but is the highest among the northernmost counties. The aim of this study is to explore the strategies employed by upper secondary school teachers and their collaborators to prevent dropout from upper secondary school among Sami youth in northern Norway. This study is based on three focus group interviews with teachers, advisers, nurses, and counsellors in the Sami areas of northern Norway. The thematic analysis identified two main strategies, namely tracking the student and giving the student time. A transparent environments, cultural competence, and interdisciplinary collaboration were identified as prerequisites for successfully implementing the two strategies to prevent dropout from upper secondary school.

Civic engagement training at a school for youth with a history of dropping out

Community psychologists have made significant contributions to the study of civic engagement, yet scarce studies have examined the impact of civic engagement training among youth with a history of dropping out. We describe an effort to promote civic education and action through a curriculum implemented at an alternative school that focuses on (a) developing awareness of the importance of engaging in social/political issues; (b) increasing civic participation; and (c) acquiring political advocacy and organizing experience. This evaluation of the civic engagement training summarizes the issues students reported in a public presentation as having had an impact in their lives; their historical, political, and social understanding of the issues; the ways in which they used a variety of social media to communicate information to different audiences; and their engagement in civic actions to impact their selected issues. Overall, students became more aware of their role as citizens and voters and wanted to share their experiences with their peers, friends, and families. The implications of promoting civic engagement among youth with a history of dropping out of school are discussed, as well as the challenges of the training implementation.

Gene representation in scRNA-seq is correlated with common motifs at the 3' end of transcripts

One important characteristic of single-cell RNA sequencing (scRNA-seq) data is its high sparsity, where the gene-cell count data matrix contains high proportion of zeros. The sparsity has motivated widespread discussions on dropouts and missing data, as well as imputation algorithms of scRNA-seq analysis. Here, we aim to investigate whether there exist genes that are more prone to be under-detected in scRNA-seq, and if yes, what commonalities those genes may share. From public data sources, we gathered paired bulk RNA-seq and scRNA-seq data from 53 human samples, which were generated in diverse biological contexts. We derived pseudo-bulk gene expression by averaging the scRNA-seq data across cells. Comparisons of the paired bulk and pseudo-bulk gene expression profiles revealed that there indeed exists a collection of genes that are frequently under-detected in scRNA-seq compared to bulk RNA-seq. This result was robust to randomization when unpaired bulk and pseudo-bulk gene expression profiles were compared. We performed motif search to the last 350 bp of the identified genes, and observed an enrichment of poly(T) motif. The poly(T) motif toward the tails of those genes may be able to form hairpin structures with the poly(A) tails of their mRNA transcripts, making it difficult for their mRNA transcripts to be captured during scRNA-seq library preparation, which is a mechanistic conjecture of why certain genes may be more prone to be under-detected in scRNA-seq.

Bubble: a fast single-cell RNA-seq imputation using an autoencoder constrained by bulk RNA-seq data

Single-cell RNA-sequencing technology (scRNA-seq) brings research to single-cell resolution. However, a major drawback of scRNA-seq is large sparsity, i.e. expressed genes with no reads due to technical noise or limited sequence depth during the scRNA-seq protocol. This phenomenon is also called 'dropout' events, which likely affect downstream analyses such as differential expression analysis, the clustering and visualization of cell subpopulations, cellular trajectory inference, etc. Therefore, there is a need to develop a method to identify and impute these dropout events. We propose Bubble, which first identifies dropout events from all zeros based on expression rate and coefficient of variation of genes within cell subpopulation, and then leverages an autoencoder constrained by bulk RNA-seq data to only impute those values. Unlike other deep learning-based imputation methods, Bubble fuses the matched bulk RNA-seq data as a constraint to reduce the introduction of false positive signals. Using simulated and several real scRNA-seq datasets, we demonstrate that Bubble enhances the recovery of missing values, gene-to-gene and cell-to-cell correlations, and reduces the introduction of false positive signals. Regarding some crucial downstream analyses of scRNA-seq data, Bubble facilitates the identification of differentially expressed genes, improves the performance of clustering and visualization, and aids the construction of cellular trajectory. More importantly, Bubble provides fast and scalable imputation with minimal memory usage.

Improving Cavalieri volume estimation based on non-equidistant planar sections: The trapezoidal estimator

The Cavalieri estimator allows one to infer the volume of an object from area measurements in equidistant planar sections. It is known that applying this estimator in the non-equidistant case may inflate the coefficient of error considerably. We therefore consider a newly introduced variant, the trapezoidal estimator, and make it available to practitioners. Its typical variance behaviour for natural objects is comparable to the equidistant case. We state this unbiased estimator, describe variance estimates and explain how the latter can be simplified under rather general but realistic models for the gaps between sections. Simulations and an application to a synthetic area function based on parietal lobes of 18 monkeys illustrate the new methods.

Multiply robust subgroup analysis based on a single-index threshold linear marginal model for longitudinal data with dropouts

Identifying subpopulations that may be sensitive to the specific treatment is an important step toward precision medicine. On the other hand, longitudinal data with dropouts is common in medical research, and subgroup analysis for this data type is still limited. In this paper, we consider a single-index threshold linear marginal model, which can be used simultaneously to identify subgroups with differential treatment effects based on linear combination of the selected biomarkers, estimate the treatment effects in different subgroups based on regression coefficients, and test the significance of the difference in treatment effects based on treatment-subgroup interaction. The regression parameters are estimated by solving a penalized smoothed generalized estimating equation and the selection bias caused by missingness is corrected by a multiply robust weighting matrix, which allows multiple missingness models to be taken account into estimation. The proposed estimator remains consistent when any model for missingness is correctly specified. Under regularity conditions, the asymptotic normality of the estimator is established. Simulation studies confirm the desirable finite-sample performance of the proposed method. As an application, we analyze the data from a clinical trial on pancreatic cancer.

Dropout associated with osteopathic manual treatment for chronic noncancerous pain in randomized controlled trials

CONTEXT

Reviews exploring harm outcomes such as adverse effects (AE), all cause dropouts (ACD), dropouts due to inefficacy, and dropouts due to AE associated with osteopathic manipulative treatment (OMT) or osteopathic manual therapy (OMTh) are scant.

OBJECTIVES

To explore the overall AE, ACD, dropouts due to inefficacy, and AE in chronic noncancerous pain (CNCP) patients receiving OMTh through a systematic review of previous literature.

METHODS

For this systematic review and meta-analysis, the authors searched MEDLINE, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), Physiotherapy Evidence Database (PEDro), EMCare, and Allied and Complementary Medicine Database (AMED), and Ostmed.Dr, as well as the bibliographical references of previous systematic reviews evaluating OMTh for pain severity, disability, quality of life, and return to work outcomes. Randomized controlled trials with CNCP patients 18 years or older with OMTh as an active or combination intervention and the presence of a control or combination group were eligible for inclusion. In this sub-study of a previous, larger systematic review, 11 studies (n=1,015) reported data that allowed the authors to perform meta-analyses on ACD and dropouts due to AE. The risk of bias (ROB) was assessed with the Cochrane ROB tool and the quality of evidence was determined with the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach.

RESULTS

The pooled analysis showed that ACD was not significantly different for visceral OMTh (vOMTh) vs. OMTh control (odds ratio [OR]=2.66 [95% confidence interval [[CI]], 0.28, 24.93]) or for OMTh vs. standard care (OR=1.26 [95% CI, 0.84, 1.89]; I²=0%). Single study analysis showed that OMTh results were nonsignificant in comparison with chemonucleolysis, gabapentin, and exercise. OMTh in combination with gabapentin (vs. gabapentin alone) and OMTh in combination with exercise (vs. exercise alone) showed nonsignificant ACD. Dropouts due to AE were not significantly different, but the results could not be pooled due to an insufficient number of studies.

CONCLUSIONS

Most articles did not explicitly report AEs, ACD rates, or dropouts due to AEs and inefficacy. The limited data available on dropouts showed that OMTh was well tolerated compared with control interventions, and that the ACD and dropouts due to AEs were not significantly different than comparators. Future trials should focus on explicit reporting of dropouts along with beneficial outcomes to provide a better understanding of OMTh efficacy.

How Do Military Family Caregivers Who Completed a Supportive Intervention Differ From Those Who Dropped Out?

Background. To create efficacious interventions for military family caregivers (MFCs), it is important to understand the characteristics and predictors of completers and dropouts of newly developed supportive interventions. Aim. The purpose of this study was to examine completion patterns in MFCs enrolled in an educational intervention feasibility study. Method. Baseline data are presented from MFC completers (n = 64) and dropouts (n = 60) of a national feasibility study for an innovative intervention. Measures include depression (Patient Health Questionnaire-2), anxiety (Generalized Anxiety Disorder-7), somatic symptoms (Patient Health Questionnaire-15), quality of life (World Health Organization Quality of Life-Brief), relationship satisfaction (Relationship Assessment Scale), and military care recipient number of injuries. Analysis of variance was used to evaluate differences between completers and dropouts and logistic regression was used to identify predictors of intervention completion. Results. Results indicated that MFCs with greater anxiety, χ²(3) = 10.33, p = .02; depression, χ²(1) = 8.18, p = .004; somatic symptoms, F(1, 106) = 6.26, p = .01; care recipient number of injuries, F(1, 118) = 16.31, p < .001; lower general satisfaction with treatment, F(1, 96) = 4.34, p = .04; and lower satisfaction with accessibility and convenience with treatment, F(1, 89) = 4.18, p = .04, were significantly more likely to complete the intervention. After multivariate analysis, the sole predictor of intervention completion was the number of care recipients' injuries, χ²(6) = 14.89, N = 77, p < .05. Conclusions. Overall, findings indicate that MFCs who were more "at risk" were more likely to complete the intervention. Findings present patterns of intervention completion and provide insight on areas in need of further investigation on intervention development supporting the needs of MFCs.

Tracking coverage, dropout and multidimensional equity gaps in immunisation systems in West Africa, 2000-2017

Background

Several West African countries are unlikely to achieve the recommended Global Vaccine Action Plan (GVAP) immunisation coverage and dropout targets in a landscape beset with entrenched intra-country equity gaps in immunisation. Our aim was to assess and compare the immunisation coverage, dropout and equity gaps across 15 West African countries between 2000 and 2017.

Methods

We compared Bacille Calmette Guerin (BCG) and the third dose of diphtheria–tetanus–pertussis (DTP3) containing vaccine coverage between 2000 and 2017 using the WHO and Unicef Estimates of National Immunisation Coverage for 15 West African countries. Estimated subregional median and weighted average coverages, and dropout (DTP1–DTP3) were tracked against the GVAP targets of ≥90% coverage (BCG and DTP3), and ≤10% dropouts. Equity gaps in immunisation were assessed using the latest disaggregated national health survey immunisation data.

Results

The weighted average subregional BCG coverage was 60.7% in 2000, peaked at 83.2% in 2009 and was 65.7% in 2017. The weighted average DTP3 coverage was 42.3% in 2000, peaked at 70.3% in 2009 and was 61.5% in 2017. As of 2017, 46.7% of countries (7/15) had met the GVAP targets on DTP3 coverage. Average weighted subregional immunisation dropouts consistently reduced from 16.4% in 2000 to 7.4% in 2017, meeting the GVAP target in 2008. In most countries, inequalities in BCG, and DTP3 coverage and dropouts were mainly related to equity gaps of more than 20% points between the wealthiest and the poorest, high coverage regions and low coverage regions, and between children of mothers with at least secondary education and those with no formal education. A child’s sex and place of residence (urban or rural) minimally determined equity gaps.

Conclusions

The West African subregion made progress between 2000 and 2017 in ensuring that its children utilised immunisation services, however, wide equity gaps persist.

Systematic review and network meta-analysis of approved medicines for the treatment of idiopathic pulmonary fibrosis

Background: Clinical practice guidelines for the treatment of idiopathic pulmonary fibrosis (IPF) currently recommend pirfenidone and nintedanib. However, there is a lack of evidence from head-to-head comparisons.

Objectives: To perform a systematic review and network meta-analysis (NMA) to access the efficacy and tolerability of two new treatments for IPF, pirfenidone and nintedanib.

Methods: Randomized controlled trials (RCTs) selection (CENTRAL, MEDLINE, Embase), data extraction, risk of bias analysis, and GRADE assessment were carried out by two authors separately. Direct estimates were calculated using standard pairwise meta-analysis. A Bayesian mixed treatment comparison approach for NMA estimates, with 95% confidence intervals (CI), was used to compare the treatments, calculating odds ratios (OR) and number needed to treat (NNTB) or harm (NNTH).

Results: The NMA on 10 randomized controlled trials showed that each drug had a positive effect on percentage of forced vital capacity (FVC) decline ≥ 10% (pirfenidone OR = 0.54 [95% CI = 0.37–0.80], NNTB = 9 [95% CI = 7–22]; nintedanib OR = 0.59 [95% CI = 0.41–0.84], NNTB = 9 [95% CI = 6–23]), but no significant differences were noted when comparing pirfenidone and nintedanib with respect to acute exacerbations, mortality, and serious adverse events (FVC decline OR = 0.91 [95% CI = 0.45–2.03]) or dropouts (OR = 0.75 [95% CI = 0.33–1.27]). Nintedanib showed an effect on dropouts, OR = 1.61 (1.13–2.28) and NNTH = 14 (8–61).

Conclusions: Based on RCTs of 12 month duration in patients with IPF, a positive effect on FVC decline was noted for both treatments and on dropouts for nintedanib, but no significant differences were noted between treatments.

McImpute: Matrix Completion Based Imputation for Single Cell RNA-seq Data

Motivation: Single-cell RNA sequencing has been proved to be revolutionary for its potential of zooming into complex biological systems. Genome-wide expression analysis at single-cell resolution provides a window into dynamics of cellular phenotypes. This facilitates the characterization of transcriptional heterogeneity in normal and diseased tissues under various conditions. It also sheds light on the development or emergence of specific cell populations and phenotypes. However, owing to the paucity of input RNA, a typical single cell RNA sequencing data features a high number of dropout events where transcripts fail to get amplified. Results: We introduce mcImpute, a low-rank matrix completion based technique to impute dropouts in single cell expression data. On a number of real datasets, application of mcImpute yields significant improvements in the separation of true zeros from dropouts, cell-clustering, differential expression analysis, cell type separability, the performance of dimensionality reduction techniques for cell visualization, and gene distribution. Availability and Implementation: https://github.com/aanchalMongia/McImpute_scRNAseq.

Outpatient Addiction Treatment for Problematic Alcohol Use: What Makes Patients Who Dropped Out Different from Those Who Did Not?

BACKGROUND

A minority of individuals with problematic alcohol use effectively seek help. Moreover, dropouts from care are not uncommon. It remains a major concern for health professionals, as adherence to treatment is significantly associated with better physical and psychological outcomes.

OBJECTIVES

The main aim of this research was to assess what factors could distinguish patients with problematic alcohol use who dropped out from those who did not.

METHODS

The sample included 150 patients followed-up in an outpatient treatment center in France for a problematic alcohol use. Two measurement times were planned: at the first appointment and after six month of treatment. A large set of individual, environmental and institutional variables were considered to compare both subgroups.

RESULTS

Patients who dropped out mostly differ from patients who did not with a higher level of alcohol-related problems, ambivalence, inclinations to use the substance, number of missed appointments. Significant results were also observed regarding a lower time gap between the first contact with the center and the first appointment, as well as the season of the last appointment.

CONCLUSIONS

Tailored motivational interventions could be offered to ambivalent patients, especially during the beginning of the treatment and some significant periods of the year. A particular focus should be brought on patients presenting such profiles in terms of level of alcohol problems, inclinations to drink and motivation to change. Overall, the study provides elements to better understand what may bring one patient to drop out of the treatment, and to improve the continuity of care.

Using Predictive Analytics to Identify Children at High Risk of Defaulting From a Routine Immunization Program: Feasibility Study

BACKGROUND

Despite the availability of free routine immunizations in low- and middle-income countries, many children are not completely vaccinated, vaccinated late for age, or drop out from the course of the immunization schedule. Without the technology to model and visualize risk of large datasets, vaccinators and policy makers are unable to identify target groups and individuals at high risk of dropping out; thus default rates remain high, preventing universal immunization coverage. Predictive analytics algorithm leverages artificial intelligence and uses statistical modeling, machine learning, and multidimensional data mining to accurately identify children who are most likely to delay or miss their follow-up immunization visits.

OBJECTIVE

This study aimed to conduct feasibility testing and validation of a predictive analytics algorithm to identify the children who are likely to default on subsequent immunization visits for any vaccine included in the routine immunization schedule.

METHODS

The algorithm was developed using 47,554 longitudinal immunization records, which were classified into the training and validation cohorts. Four machine learning models (random forest; recursive partitioning; support vector machines, SVMs; and C-forest) were used to generate the algorithm that predicts the likelihood of each child defaulting from the follow-up immunization visit. The following variables were used in the models as predictors of defaulting: gender of the child, language spoken at the child's house, place of residence of the child (town or city), enrollment vaccine, timeliness of vaccination, enrolling staff (vaccinator or others), date of birth (accurate or estimated), and age group of the child. The models were encapsulated in the predictive engine, which identified the most appropriate method to use in a given case. Each of the models was assessed in terms of accuracy, precision (positive predictive value), sensitivity, specificity and negative predictive value, and area under the curve (AUC).

RESULTS

Out of 11,889 cases in the validation dataset, the random forest model correctly predicted 8994 cases, yielding 94.9% sensitivity and 54.9% specificity. The C-forest model, SVMs, and recursive partitioning models improved prediction by achieving 352, 376, and 389 correctly predicted cases, respectively, above the predictions made by the random forest model. All models had a C-statistic of 0.750 or above, whereas the highest statistic (AUC 0.791, 95% CI 0.784-0.798) was observed in the recursive partitioning algorithm.

CONCLUSIONS

This feasibility study demonstrates that predictive analytics can accurately identify children who are at a higher risk for defaulting on follow-up immunization visits. Correct identification of potential defaulters opens a window for evidence-based targeted interventions in resource limited settings to achieve optimal immunization coverage and timeliness.

Mixed results: the protective role of schooling in the HIV epidemic in Swaziland

Swaziland has the highest HIV prevalence in the world. It is recognised that young women, especially adolescents, are particularly vulnerable to HIV infection and bear a disproportionate burden of HIV incidence. The HIV data from Swaziland show the location of the epidemic, which is particularly high among adolescent girls and young women. This paper is based on research in Swaziland, commissioned because of the perception that large numbers of children were dropping out of the school. It was assumed that these "dropouts" had increased risk of HIV exposure. This study carried out a detailed analysis using the Annual Education Census Reports from 2012 to 2014 produced by the Ministry of Education. In addition, this topic was explored, during fieldwork with key informants in the country. While HIV prevalence rises rapidly among young women in Swaziland, as is the case across most of Southern Africa, the data showed there were few dropouts. This was the case at all levels of education - primary, junior secondary and senior secondary. The major reason for dropping out of primary school was family reasons; and in junior and senior secondary, pregnancy was the leading cause. Swaziland is doing well in terms of getting its children into school, and, for the most part, keeping them there. This paper identifies the students who face increased vulnerability: the limited number of dropouts; repeaters who consequently were "out-of-age for grade"; and orphans and vulnerable children (OVC). The learners who were classified as repeaters and OVC greatly outnumbered the dropouts. We argue, on the basis of these data, for re-focussed attention and the need to develop a method for tracking children as they move across the vulnerable groups. We acknowledge schooling is protective in reducing children's vulnerability to HIV, and Swaziland is on the right track in education, although there are challenges.

"Open mesh" or "strictly selected population" recruitment? The experience of the randomized controlled MeMeMe trial

Among randomized controlled trials (RCTs), trials for primary prevention require large samples and long follow-up to obtain a high-quality outcome; therefore the recruitment process and the drop-out rates largely dictate the adequacy of the results. We are conducting a Phase III trial on persons with metabolic syndrome to test the hypothesis that comprehensive lifestyle changes and/or metformin treatment prevents age-related chronic diseases (the MeMeMe trial, EudraCT number: 2012-005427-32, also registered on ClinicalTrials.gov [NCT02960711]). Here, we briefly analyze and discuss the reasons which may lead to participants dropping out from trials. In our experience, participants may back out of a trial for different reasons. Drug-induced side effects are certainly the most compelling reason. But what are the other reasons, relating to the participants' perception of the progress of the trial which led them to withdraw after randomization? What about the time-dependent drop-out rate in primary prevention trials? The primary outcome of this analysis is the point of drop-out from trial, defined as the time from the randomization date to the withdrawal date. Survival functions were non-parametrically estimated using the product-limit estimator. The curves were statistically compared using the log-rank test (P=0.64, not significant). Researchers involved in primary prevention RCTs seem to have to deal with the paradox of the proverbial "short blanket syndrome". Recruiting only highly motivated candidates might be useful for the smooth progress of the trial but it may lead to a very low enrollment rate. On the other hand, what about enrolling all the eligible subjects without considering their motivation? This might boost the enrollment rate, but it can lead to biased results on account of large proportions of drop-outs. Our experience suggests that participants do not change their mind depending on the allocation group (intervention or control). There is no single answer to sort out the short blanket syndrome.

Robust estimation of partially linear models for longitudinal data with dropouts and measurement error

Outliers, measurement error, and missing data are commonly seen in longitudinal data because of its data collection process. However, no method can address all three of these issues simultaneously. This paper focuses on the robust estimation of partially linear models for longitudinal data with dropouts and measurement error. A new robust estimating equation, simultaneously tackling outliers, measurement error, and missingness, is proposed. The asymptotic properties of the proposed estimator are established under some regularity conditions. The proposed method is easy to implement in practice by utilizing the existing standard generalized estimating equations algorithms. The comprehensive simulation studies show the strength of the proposed method in dealing with longitudinal data with all three features. Finally, the proposed method is applied to data from the Lifestyle Education for Activity and Nutrition study and confirms the effectiveness of the intervention in producing weight loss at month 9. Copyright © 2016 John Wiley & Sons, Ltd.

Child Abuse and Neglect Among Children Who Drop Out of School: A Study in Izmir, Turkey

Child abuse and neglect (CAN), and dropping out of school have long been recognized as pervasive social problems globally, and Turkey is no exception. This study aims to explore the prevalence and incidence of CAN in children who drop out of school of Turkey, using the ISPCAN Child abuse Screening Tool, Children's Version, which is an appropriate tool for multinational comparisons. Data from a convenience sample of children who drop out of school age 11, 13, and 16 from Izmir were collected either by interviews or by self-completion. The results show that, compared to children who do not drop out of school, children who drop out of school have higher rates of psychological and physical abuse and neglect within the family. This study not only highlights the need for preventive laws for CAN and dropping out of school, but also points to direction for future research.

Intention-to-treat analysis with treatment discontinuation and missing data in clinical trials

Motivated by a recent National Research Council study, we discuss three aspects of the analysis of clinical trials when participants prematurely discontinue treatments. First, we distinguish treatment discontinuation from missing outcome data. Data collection is often stopped after treatment discontinuation, but outcome data could be recorded on individuals after they discontinue treatment, as the National Research Council study recommends. Conversely, outcome data may be missing for individuals who do not discontinue treatment, as when there is loss to follow up or missed clinic visits. Missing outcome data is a standard missing data problem, but treatment discontinuation is better viewed as a form of noncompliance and treated using ideas from the causal literature on noncompliance. Second, the standard intention to treat estimand, the average effect of randomization to treatment, is compared with three alternative estimands for the intention to treat population: the average effect when individuals continue on the assigned treatment after discontinuation, the average effect when individuals take a control treatment after treatment discontinuation, and a summary measure of the effect of treatment prior to discontinuation. We argue that the latter choice of estimand has advantages and should receive more consideration. Third, we consider when follow-up measures after discontinuation are needed for valid measures of treatment effects. The answer depends on the choice of primary estimand and the plausibility of assumptions needed to address the missing data. Ideas are motivated and illustrated by a reanalysis of a past study of inhaled insulin treatments for diabetes, sponsored by Eli Lilly.

Intended and Unintended Effects of State-Mandated High School Science and Mathematics Course Graduation Requirements on Educational Attainment

Mathematics and science course graduation requirement (CGR) increases in the 1980s and 1990s might have had both intended and unintended consequences. Using logistic regression with Census and American Community Survey (ACS) data (n = 2,892,444), we modeled CGR exposure on (a) high school dropout, (b) beginning college, and (c) obtaining any college degree. Possible between-groups differences were also assessed. We found that higher CGRs were associated with higher odds to drop out of high school, but results for the college-level outcomes varied by group. Some were less likely to enroll, whereas others who began college were more likely to obtain a degree. Increased high school dropout was consistent across the population, but some potential benefit was also observed, primarily for those reporting Hispanic ethnicity.

Use of the 'patient journey' model in the internet-based pre-fitting counseling of a person with hearing disability: lessons from a failed clinical trial

BACKGROUND

Persons with a hearing impairment have various experiences during their 'journey' through hearing loss. In our previous studies we have developed 'patient journey' models of person with hearing impairment and their communication partners (CPs). The study was aimed to evaluate the effectiveness of using the patient journey model in the internet-based pre-fitting counseling of a person with hearing disability (ClinicalTrials.gov Protocol Registration System: NCT01611129, registered 2012 May 14).

METHOD

The study employed a randomized controlled trial (RCT) with waiting list control (WLC) design. Even though we had intended to recruit 158 participants, we only managed to recruit 80 participants who were assigned to one of two groups: (1) Intervention group; and (2) WLC. Participants from both groups completed a 30 day internet-based counseling program (group 2 waited for a month before intervention) based on the 'patient journey' model. Various outcome measures which focus on self-reported hearing disability, self-reported depression and anxiety, readiness to change and self-reported hearing disability acceptance were administered pre- and post-intervention.

RESULTS

The trial results suggest that the intervention was not feasible. Treatment compliancy was one of the main problems with a high number of dropouts. Only 18 participants completed both pre- and post-intervention outcome measures. Their results were included in the analysis. Results suggest no statistically significant differences among groups over time in all four measures.

CONCLUSIONS

Due to the limited sample size, no concrete conclusions can be drawn about the hypotheses from the current study. Furthermore, possible reasons for failure of this trial and directions for future research are discussed.

Choosing profile double-sampling designs for survival estimation with application to President's Emergency Plan for AIDS Relief evaluation

Most studies that follow subjects over time are challenged by having some subjects who dropout. Double sampling is a design that selects and devotes resources to intensively pursue and find a subset of these dropouts, then uses data obtained from these to adjust naïve estimates, which are potentially biased by the dropout. Existing methods to estimate survival from double sampling assume a random sample. In limited-resource settings, however, generating accurate estimates using a minimum of resources is important. We propose using double-sampling designs that oversample certain profiles of dropouts as more efficient alternatives to random designs. First, we develop a framework to estimate the survival function under these profile double-sampling designs. We then derive the precision of these designs as a function of the rule for selecting different profiles, in order to identify more efficient designs. We illustrate using data from the United States President's Emergency Plan for AIDS Relief-funded HIV care and treatment program in western Kenya. Our results show why and how more efficient designs should oversample patients with shorter dropout times. Further, our work suggests generalizable practice for more efficient double-sampling designs, which can help maximize efficiency in resource-limited settings.

Predictors of participant retention in a guided online self-help program for university students: prospective cohort study

Background

Attrition is a persistent issue in online self-help programs, but limited research is available on reasons for attrition or successful methods for improving participant retention. One potential approach to understanding attrition and retention in such programs is to examine person-related variables (eg, beliefs and attitudes) that influence behavior. Theoretical models, such as the Theory of Planned Behavior, that describe conditions influencing human behavior may provide a useful framework for predicting participant retention in online-based program.

Objective

We examined predictors of participant retention in a guided online anxiety, depression, and stress self-help program for university students using the theory of planned behavior. We also explored whether age, symptom severity, and type of coaching (ie, email vs phone) affected participant retention.

Methods

65 university students with mild to moderate depression, anxiety, and stress were enrolled in this prospective cohort study. Participants completed a questionnaire based on the theory of planned behavior prior to commencing the online-based program and the Depression Anxiety and Stress Scale (DASS) during the assessment module of the program. Participant retention was operationalized as the number of program modules completed.

Results

Perceived control over completing the online program significantly predicted intention to complete the program (F_3,62=6.7; P=.001; adjusted R²=.2; standardized beta=.436, P=.001). Age (standardized beta=.319, P=.03) and perceived behavioral control (standardized beta=.295, P=.05) predicted the number of program modules completed (F_3,61=3.20, P=.03, adjusted R² =.11). Initial level of distress (ie, symptom severity) did not predict participant retention (P=.55). Participants who chose phone-based coaching completed more program modules than participants who chose email-based coaching (Mann-Whitney’s U=137; P=.004).

Conclusions

Participants’ age, level of perceived behavioral control, and choice of interaction (ie, phone-based or email-based coaching) were found to influence retention in this online-based program.

The design and conduct of clinical trials to limit missing data

This article summarizes recommendations on the design and conduct of clinical trials of a National Research Council study on missing data in clinical trials. Key findings of the study are that (a) substantial missing data is a serious problem that undermines the scientific credibility of causal conclusions from clinical trials; (b) the assumption that analysis methods can compensate for substantial missing data is not justified; hence (c) clinical trial design, including the choice of key causal estimands, the target population, and the length of the study, should include limiting missing data as one of its goals; (d) missing-data procedures should be discussed explicitly in the clinical trial protocol; (e) clinical trial conduct should take steps to limit the extent of missing data; (f) there is no universal method for handling missing data in the analysis of clinical trials - methods should be justified on the plausibility of the underlying scientific assumptions; and (g) when alternative assumptions are plausible, sensitivity analysis should be conducted to assess robustness of findings to these alternatives. This article focuses on the panel's recommendations on the design and conduct of clinical trials to limit missing data. A companion paper addresses the panel's findings on analysis methods.

Psychological and sociodemographic predictors of premature discontinuation of a 1-year multimodal outpatient weight-reduction program: an attrition analysis

OBJECTIVE

Attrition rates of up to 77% have been reported in conservative weight-reduction programs for the treatment of obesity. In view of the cost of such programs to the health system, there is a need to identify the variables that predict premature discontinuation of treatment. Previous studies have focused mainly on somatic and sociodemographic parameters. The prospective influence of psychological factors has not been systematically investigated to date.

METHODS

A total of 164 patients (138 of whom were women) with a mean age of 45 years and a mean body mass index of 39.57 participated in a 1-year outpatient weight-reduction program at the Charité - Universitätsmedizin Berlin University Hospital. The program included movement therapy, dietary advice, psychoeducational and behavioral interventions, relaxation procedures, and consultations with a specialist in internal medicine and a psychologist. Patients also underwent regular laboratory and psychological testing. The results were evaluated using a t-test, χ(2)-test, and logistic regression analysis.

RESULTS

Seventy-one of the 164 patients (61 women, mean age = 43 years, mean body mass index = 39.53) withdrew before the end of the program (attrition rate = 43.3%). While there were no differences between the somatic and metabolic characteristics of those who withdrew and those who remained, the sociodemographic and psychological factors had some relevance. In particular, "expectation of self-efficacy" (Fragebogen zu Selbstwirksamkeit, Optimismus und Pessimismus [SWOP]), "not working," "tiredness" (Berliner Stimmungsfragebogen [BSF]), "pessimism" (SWOP) and "positive reframing" (Brief-COPE) were found to play a role in whether participants subsequently dropped out of the treatment. "Support coping" (Brief-COPE) and "older age" prior to the start of treatment were identified as variables that promoted treatment adherence.

CONCLUSION

The results are discussed in light of previous findings and with regard to whether the modules of the weight-reduction program should be adapted.

An activity-based intervention for obese and physically inactive children organized in primary care: feasibility and impact on fitness and BMI A one-year follow-up study

OBJECTIVE

To investigate the feasibility and impact on BMI and physical fitness of an intervention for obese and inactive children, based on physical activity and carried out in primary health care.

DESIGN

A prospective, longitudinal one-year follow-up study.

SETTING

The community of Kristiansand, Norway (80 000 inhabitants).

INTERVENTION

A 40-week structured intervention based on physical training with some lifestyle advice for the obese child and one parent. Subjects. A total of 62 physically inactive children aged 6-14 years with iso-BMI ≥ 30 kg/m².

MAIN OUTCOME MEASURES

Body mass index (BMI), maximum oxygen uptake, and physical fitness in tests of running, jumping, throwing, and climbing assessed at baseline and after six and 12 months as well as number of dropouts and predicting factors.

RESULTS

A total of 49 out of 62 children completed the first six months and 37 children completed 12 months. Dropout rate was higher when parents reported being physically inactive at baseline or avoided physical participation in the intervention. The children's maximum oxygen uptake increased significantly after 12 months from 27.0 to 32.0 ml/kg/min (means), as did physical fitness (endurance, speed, agility, coordination, balance, strength) and BMI was significantly reduced.

CONCLUSION/IMPLICATIONS

This one-year activity-based intervention for obese and inactive children performed in primary health care succeeded by increasing cardiovascular capacity and physical fitness combined with reduced BMI in those who completed. Dropout was substantial and depended on the attendance and compliance with physical activity by the parents.

Reporting practices of dropouts in psychological research using a wait-list control: current state and suggestions for improvement

Reporting practices regarding dropouts in wait-list control studies hold great importance for the ability to replicate, generalize, and draw conclusions from research. This concern is applicable to all psychological research utilizing wait-list controls, regardless of purpose of research (e.g., treatment outcome). The current study assessed the present state of reporting practices in this type of experimental design and discussed the limitations and implications of the insufficient reporting found. 171 articles from psychology journals utilizing wait-list control design were surveyed regarding the reporting of the number of dropouts from the wait-list control and experimental conditions, characteristics and assessment scores of the dropouts, and total dropouts. Variables that are crucial to interpreting research findings are not consistently reported. Additionally, journal impact factor and year of publication were positively correlated with the adequacy of reporting. Consistencies with previous findings were noted, and suggestions for remedying the reporting inadequacies were discussed.

Rule-of-thumb adjustment of sample sizes to accommodate dropouts in a two-stage analysis of repeated measurements

Recent contributions to the statistical literature have provided elegant model-based solutions to the problem of estimating sample sizes for testing the significance of differences in mean rates of change across repeated measures in controlled longitudinal studies with differentially correlated error and missing data due to dropouts. However, the mathematical complexity and model specificity of these solutions make them generally inaccessible to most applied researchers who actually design and undertake treatment evaluation research in psychiatry. In contrast, this article relies on a simple two-stage analysis in which dropout-weighted slope coefficients fitted to the available repeated measurements for each subject separately serve as the dependent variable for a familiar ANCOVA test of significance for differences in mean rates of change. This article is about how a sample of size that is estimated or calculated to provide desired power for testing that hypothesis without considering dropouts can be adjusted appropriately to take dropouts into account. Empirical results support the conclusion that, whatever reasonable level of power would be provided by a given sample size in the absence of dropouts, essentially the same power can be realized in the presence of dropouts simply by adding to the original dropout-free sample size the number of subjects who would be expected to drop from a sample of that original size under conditions of the proposed study.