Generalized estimating equation modeling on correlated microbiome sequencing data with longitudinal measures

Trends and correlates in HIV viral load monitoring and viral suppression among adolescents and young adults in Dar es Salaam, Tanzania

BACKGROUND

Adolescents and young adults (AYA) living with HIV have been shown to have lower rates of viral load testing and viral suppression as compared to older adults. We examined trends over time and predictors of HIV viral load monitoring and viral suppression among AYA in a large HIV treatment programme in Dar es Salaam, Tanzania.

METHODS

We analysed longitudinal data of AYA aged 10-24 years initiated on antiretroviral therapy between January 2017 and October 2022. Trend models were used to assess changes in HIV viral load testing and viral suppression by calendar year. Generalised estimating equations were used to examine the relationship of sociodemographic and clinical factors with HIV viral load testing and viral suppression.

RESULTS

Out of 15,759 AYA, the percentage of those who received a 6-month HIV viral load testing increased from 40.6% in 2017 to 64.7% in 2022 and, a notable annual increase of 5.6% (p < 0.001). A higher HIV viral load testing uptake was observed among 20- to 24-year-olds (87.7%) compared to 10- to 19-year-olds (80.2%) (p < 0.001). The likelihood of not receiving an HIV viral load test within 12 months of antiretroviral therapy initiation was higher among 10- to 19-year-olds (adjusted odds ratio [aOR] = 1.7; 95% confidence interval [CI] = 1.4-2.0), advanced HIV disease (aOR = 1.3; 95% CI = 1.12-1.53), normal nutrition status at enrolment aOR 2.6 (95% CI = 1.59-4.26) and initiation of non-nucleoside reverse transcriptase inhibitors regimen aOR 1.2 (95% CI = 1.08-1.34). The proportion of AYA with viral suppression increased from 83.0% in 2017 to 94.6% in 2022. Notably, the overall trend in viral suppression increased significantly at 2.4% annually. The risk of not achieving viral suppression was greater among 10- to 14-year-olds (aOR = 2; 95% CI = 1.75-2.43) and 15- to 19-year-olds (aOR = 1.4; 95% CI = 1.24-1.58) as compared to 20-24 years; being male (aOR = 1.16; 95% CI = 1.02-1.32); undernourished (aOR = 1.53; 95% CI = 1.17-1.99); in WHO Stage II (aOR = 1.16; 95% CI = 1.02-1.33) and III (aOR = 1.21; 95% CI = 1.03-1.42) and being on an non-nucleoside reverse transcriptase inhibitors regimen (aOR = 1.32; 95% CI = 1.18-1.48).

CONCLUSION

HIV viral load testing uptake at 6 months of antiretroviral therapy initiation and viral suppression increased from 2017 to 2022; however, overall HIV viral load testing was suboptimal. Demographic and clinical characteristics can be used to identify AYA at greater risk for not having HIV viral load test and not achieving viral suppression.

Effect of continuous workshop training of the helping babies breathe program on the retention of midwives' knowledge and skills: A clinical trial study

AIM

To investigate the impact of ongoing workshop training of the "Helping Babies Breathe" program on the durability of midwives' knowledge and skills.

BACKGROUND

Implementing the Helping Babies Breathe (HBB) program is crucial as a simple protocol for neonatal resuscitation in low-resource healthcare settings to decrease the rate of asphyxia and perinatal mortality by the initial healthcare providers. In addition to training in this program, it is also essential to guarantee the retention of the acquired knowledge and skills.

DESIGN

A quasi-experimental clinical trial study with a single-group, pre-test-and-post-test design.

METHODS

This study was conducted throughout the year 2022, with a sample size of 61 midwives selected through a census sampling from those working in the delivery and operating rooms of X Hospital in x City. The midwives participated in 3-hour workshops. This study was performed in two stages: intervention and follow-up. The evaluation Instruments included the HBB educational package, which consisted of a questionnaire and 3 Objective Structured Clinical Exams. During the intervention phase, the HBB program training was conducted through a series of workshops held at four different time points over a span of six months. In the follow-up stage, the learners were not provided with any further training. The evaluation was done immediately after the initial training workshop of the HBB program, at the end of the final workshop in the sixth month and at the end of the follow-up period.

RESULTS

The mean knowledge score of the baseline, at six months and at twelve months after the initial workshop were documented as (17 SD1.2), (17.79 SD 0.4) and (17.73 SD 0.5), respectively. There was a statistically significant difference in the mean knowledge scores between the baseline and the six and twelve months (P<0.05), but no statistically significant difference was observed between six and twelve months (P>0.05). The mean skill scores showed a significant improvement and were maintained after six months compared with the initial assessment (P<0.05); however, there was a significant decrease in skill score twelve months later, in comparison to both the initial assessment and the first six months (P<0.05).

CONCLUSIONS

Healthcare workers can maintain their knowledge and skills by participating in ongoing training workshops. However, without continuous training, their skills may diminish. Therefore, it is essential to implement training programs that emphasize regular practice and repetition to ensure knowledge and skills retention.

REGISTRATION NUMBER

The present research was a part of the research work with the ethics ID IR.IRSHUMS.REC.1400.019.

Homogeneity pursuit and variable selection in regression models for multivariate abundance data

When building regression models for multivariate abundance data in ecology, it is important to allow for the fact that the species are correlated with each other. Moreover, there is often evidence species exhibit some degree of homogeneity in their responses to each environmental predictor, and that most species are informed by only a subset of predictors. We propose a generalized estimating equation (GEE) approach for simultaneous homogeneity pursuit (ie, grouping species with similar coefficient values while allowing differing groups for different covariates) and variable selection in regression models for multivariate abundance data. Using GEEs allows us to straightforwardly account for between-response correlations through a (reduced-rank) working correlation matrix. We augment the GEE with both adaptive fused lasso- and adaptive lasso-type penalties, which aim to cluster the species-specific coefficients within each covariate and encourage differing levels of sparsity across the covariates, respectively. Numerical studies demonstrate the strong finite sample performance of the proposed method relative to several existing approaches for modeling multivariate abundance data. Applying the proposed method to presence-absence records collected along the Great Barrier Reef in Australia reveals both a substantial degree of homogeneity and sparsity in species-environmental relationships. We show this leads to a more parsimonious model for understanding the environmental drivers of seabed biodiversity, and results in stronger out-of-sample predictive performance relative to methods that do not accommodate such features.

An analysis of socioeconomic factors on multiple chronic conditions and its economic burden: evidence from the National Health Service Survey in Yunnan Province, China

Background

The economic burden of multiple chronic conditions (MCCs) and its socio-economic influencing factors have widely raised public concerns. However, there are few large population-based studies on these problems in China. Our study aims at determining the economic burden of MCCs and associated factors specific to multimorbidity among middle-aged and older individuals.

Methods

As our study population, we extracted all 11,304 participants over 35 years old from the 2018 National Health Service Survey (NHSS) in Yunnan. Economic burden and socio-demographic characteristics were analyzed with descriptive statistics. Chi-square test and generalized estimating equations (GEE) regression models were used to identify influencing factors.

Results

The prevalence of chronic diseases was 35.93% in 11,304 participants and the prevalence of MCCs increased with age, was 10.12%. Residents who lived in rural areas were more likely to report MCCs than those who lived in urban areas (adjusted OR = 1.347, 97.5% CI: 1.116-1.626). Ethnic minority groups were less likely to report MCCs than those of Han (OR = 0.752, 97.5% CI: 0.601-0.942). Overweight or obese people were more likely to report MCCs than people with normal weight (OR = 1.317, 97.5% CI: 1.099-1.579). The per capita expenses of 2 weeks' illness, per capita hospitalization expenses, annual household income, annual household expenses, and annual household medical expenses of MCCs were ¥292.90 (±1427.80), ¥4804.22 (±11851.63), ¥51064.77 (±52158.76), ¥41933.50 (±39940.02) and ¥11724.94 (±11642.74), respectively. The per capita expenses of 2 weeks' illness, per capita hospitalization expenses, annual household income, annual household cost, and annual household medical expenses of hypertensive co-diabetic patients were more compared to those with other three comorbidity modes.

Conclusion

The prevalence of MCCs was relatively high among middle-aged and older individuals in Yunnan, China, which bought a heavy economic burden. This encourages policy makers and health providers to pay more attention to the behavioral/lifestyle factors, that contribute to multimorbidity to a great extent. Furthermore, health promotion and education in terms of MCCs need to be prioritized in Yunnan.

Detecting sparse microbial association signals adaptively from longitudinal microbiome data based on generalized estimating equations

The association between the compositions of microbial communities and various host phenotypes is an important research topic. Microbiome association research addresses multiple domains, such as human disease and diet. Statistical methods for testing microbiome-phenotype associations have been studied recently to determine their ability to assess longitudinal microbiome data. However, existing methods fail to detect sparse association signals in longitudinal microbiome data. In this paper, we developed a novel method, namely aGEEMIHC, which is a data-driven adaptive microbiome higher criticism analysis based on generalized estimating equations to detect sparse microbial association signals from longitudinal microbiome data. aGEEMiHC adopts generalized estimating equations framework that fully considers the correlation among different observations from the same subject in longitudinal data. To be robust to diverse correlation structures for longitudinal data, aGEEMiHC integrates multiple microbiome higher criticism analyses based on generalized estimating equations with different working correlation structures. Extensive simulation experiments demonstrate that aGEEMiHC can control the type I error correctly and achieve superior performance according to a statistical power comparison. We also applied it to longitudinal microbiome data with various types of host phenotypes to demonstrate the stability of our method. aGEEMiHC is also utilized for real longitudinal microbiome data, and we found a significant association between the gut microbiome and Crohn's disease. In addition, our method ranks the significant factors associated with the host phenotype to provide potential biomarkers.

Functional response regression model on correlated longitudinal microbiome sequencing data

Functional regression has been widely used on longitudinal data, but it is not clear how to apply functional regression to microbiome sequencing data. We propose a novel functional response regression model analyzing correlated longitudinal microbiome sequencing data, which extends the classic functional response regression model only working for independent functional responses. We derive the theory of generalized least squares estimators for predictors' effects when functional responses are correlated, and develop a data transformation technique to solve the computational challenge for analyzing correlated functional response data using existing functional regression method. We show by extensive simulations that our proposed method provides unbiased estimations for predictors' effect, and our model has accurate type I error and power performance for correlated functional response data, compared with classic functional response regression model. Finally we implement our method to a real infant gut microbiome study to evaluate the relationship of clinical factors to predominant taxa along time.

phyLoSTM: a novel deep learning model on disease prediction from longitudinal microbiome data

MOTIVATION

Research shows that human microbiome is highly dynamic on longitudinal timescales, changing dynamically with diet, or due to medical interventions. In this paper, we propose a novel deep learning framework "phyLoSTM", using a combination of Convolutional Neural Networks and Long Short Term Memory Networks (LSTM) for feature extraction and analysis of temporal dependency in longitudinal microbiome sequencing data along with host's environmental factors for disease prediction. Additional novelty in terms of handling variable timepoints in subjects through LSTMs, as well as, weight balancing between imbalanced cases and controls is proposed.

RESULTS

We simulated 100 datasets across multiple time points for model testing. To demonstrate the model's effectiveness, we also implemented this novel method into two real longitudinal human microbiome studies: (i) DIABIMMUNE three country cohort with food allergy outcomes (Milk, Egg, Peanut and Overall) (ii) DiGiulio study with preterm delivery as outcome. Extensive analysis and comparison of our approach yields encouraging performance with an AUC of 0.897 (increased by 5%) on simulated studies and AUCs of 0.762 (increased by 19%) and 0.713 (increased by 8%) on the two real longitudinal microbiome studies respectively, as compared to the next best performing method, Random Forest. The proposed methodology improves predictive accuracy on longitudinal human microbiome studies containing spatially correlated data, and evaluates the change of microbiome composition contributing to outcome prediction.

AVAILABILITY AND IMPLEMENTATION

https://github.com/divya031090/phyLoSTM.