Dataset Information

Effect of imbalanced sampling and missing data on associations between gender norms and risk of adolescent HIV.

ABSTRACT:

Background

Despite strides towards gender equality, inequalities persist or remain unstudied, due potentially to data gaps. Although mapped, the effects of key data gaps remain unknown. This study provides a framework to measure effects of gender- and age-imbalanced and missing covariate data on gender-health research. The framework is demonstrated using a previously studied pathway for effects of pre-marital sex norms among adults on adolescent HIV risk.

Methods

After identifying gender-age-imbalanced Demographic and Health Survey (DHS) datasets, we resampled responses and restricted covariate data from a relatively complete, balanced dataset derived from the 2007 Zambian DHS to replicate imbalanced gender-age sampling and covariate missingness. Differences in model outcomes due to sampling were measured using tests for interaction. Missing covariate effects were measured by comparing fully-adjusted and reduced model fitness.

Findings

We simulated data from 25 DHS surveys across 20 countries from 2005-2014 on four sex-stratified models for pathways of adult attitude-behaviour discordance regarding pre-marital sex and adolescent risk of HIV. On average, across gender-age-imbalanced surveys, males comprised 29.6% of responses compared to 45.3% in the gender-balanced dataset. Gender-age-imbalanced sampling significantly affected regression coefficients in 40% of model-scenarios (N = 40 of 100) and biased relative-risk estimates away from gender-age-balanced sampling outcomes in 46% (N = 46) of model-scenarios. Model fitness was robust to covariate removal with minor effects on male HIV models. No consistent trends were observed between sampling distribution and risk of biased outcomes.

Interpretation

Gender-health model outcomes may be affected by sampling gender-age-imbalanced data and less-so by missing covariates. Although occasionally attenuated, the effect magnitude of gender-age-imbalanced sampling is variable and may mask true associations, thus misinforming policy dialogue. We recommend future surveys improve balanced gender-age sampling to promote research reliability.

Funding

Bill & Melinda Gates Foundation grant OPP1140262 to Stanford University.

SUBMITTER: Gupta R

PROVIDER: S-EPMC9241092 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:In numerous classification problems, class distribution is not balanced. For example, positive examples are rare in the fields of disease diagnosis and credit card fraud detection. General machine learning methods are known to be suboptimal for such imbalanced classification. One popular solution is to balance training data by oversampling the underrepresented (or undersampling the overrepresented) classes before applying machine learning algorithms. However, despite its popularity, the effectiveness of sampling has not been rigorously and comprehensively evaluated. This study assessed combinations of seven sampling methods and eight machine learning classifiers (56 varieties in total) using 31 datasets with varying degrees of imbalance. We used the areas under the precision-recall curve (AUPRC) and receiver operating characteristics curve (AUROC) as the performance measures. The AUPRC is known to be more informative for imbalanced classification than the AUROC. We observed that sampling significantly changed the performance of the classifier (paired t-tests P < 0.05) only for few cases (12.2% in AUPRC and 10.0% in AUROC). Surprisingly, sampling was more likely to reduce rather than improve the classification performance. Moreover, the adverse effects of sampling were more pronounced in AUPRC than in AUROC. Among the sampling methods, undersampling performed worse than others. Also, sampling was more effective for improving linear classifiers. Most importantly, we did not need sampling to obtain the optimal classifier for most of the 31 datasets. In addition, we found two interesting examples in which sampling significantly reduced AUPRC while significantly improving AUROC (paired t-tests P < 0.05). In conclusion, the applicability of sampling is limited because it could be ineffective or even harmful. Furthermore, the choice of the performance measure is crucial for decision making. Our results provide valuable insights into the effect and characteristics of sampling for imbalanced classification.

Project description:ImportanceEngaging adolescent boys and young men in preventing violence against women is a potentially impactful public health strategy.ObjectiveTo evaluate the effectiveness of a community-based, gender-transformative program (ie, Manhood 2.0) on perpetration of gender-based violence by adolescent boys and young men.Design, setting, and participantsIn this unblinded cluster randomized clinical trial, neighborhoods were designated as the unit of clustering (1:1 allocation). Three-month (ie, time point 2 [T2]) and 9-month (ie, time point 3 [T3]) follow-ups were conducted. The trial took place in 20 Pittsburgh, Pennsylvania, neighborhoods and 1 centrally located site with concentrated disadvantage. Pittsburgh-based adolescent boys and young men (ages 13 to 19 years) were recruited between July 27, 2015, and June 5, 2017, through youth-serving organizations and community-based alternatives to residential placement for juvenile justice-involved youth. Intention-to-treat analysis was conducted from June 2018 to November 2019.InterventionsManhood 2.0, an international program adapted for adolescent boys and young men in US urban communities, encourages these individuals to challenge gender norms that foster violence against women and unhealthy sexual relationships. Individuals in the control population received job-readiness training. Each program was 18 hours.Main outcomes and measuresThe primary outcome was change in participant-level perpetration of sexual violence (SV) or adolescent relationship abuse (ARA) at T3.ResultsAmong 866 participants, 465 individuals (54%) enrolled in 11 intervention clusters and 401 individuals (46%) enrolled in 10 control clusters. In the intervention group, 325 participants (70%) were analyzed at T2 and 334 participants (72%) were analyzed at T3; in the control group, 262 participants (65%) were analyzed at T2 and 301 participants (75%) were analyzed at T3. Mean (SD) age was 15.5 (1.6) years; 609 participants (70%) self-identified as non-Hispanic Black, and 178 (20%) self-identified as Hispanic, multiracial, or other race/ethnicity other than White. Among individuals in the intervention group, 296 participants (64%) reported any SV or ARA perpetration at baseline, and 173 participants (52%) reported any SV or ARA perpetration at T3. Among individuals in the control group, 213 participants (53%) reported any SV or ARA perpetration at baseline, and 124 participants (41%) reported any SV or ARA perpetration at T3). The difference in reduction between groups was not significant. There was no evidence of an intervention effect for the primary outcome (adjusted odds ratio [OR], 1.32; 95% CI, 0.86-2.01; P?=?.20).Conclusions and relevanceThe findings from this evaluation of a community-based gender-transformative program for adolescent boys and young men did not show a significant intervention effect in reducing SV or ARA perpetration between Manhood 2.0 and a job-readiness control program. Combining gender-transformative approaches with job-readiness programs may be relevant for violence prevention in low-resource urban settings. Attention to improving implementation and strategies to sustain such community-based efforts are needed.Trial registrationClinicalTrials.gov Identifier: NCT02427061.