Nevin Manimala

Biometrics. 2022 Jul 15. doi: 10.1111/biom.13719. Online ahead of print.

ABSTRACT

Inverse probability weighted estimators are the oldest and potentially most commonly used class of procedures for the estimation of causal effects. By adjusting for selection biases via a weighting mechanism, these procedures estimate an effect of interest by constructing a pseudo-population in which selection biases are eliminated. Despite their ease of use, these estimators require the correct specification of a model for the weighting mechanism, are known to be inefficient, and suffer from the curse of dimensionality. We propose a class of nonparametric inverse probability weighted estimators in which the weighting mechanism is estimated via undersmoothing of the highly adaptive lasso, a nonparametric regression function proven to converge at nearly -rate to the true weighting mechanism. We demonstrate that our estimators are asymptotically linear with variance converging to the nonparametric efficiency bound. Unlike doubly robust estimators, our procedures require neither derivation of the efficient influence function nor specification of the conditional outcome model. Our theoretical developments have broad implications for the construction of efficient inverse probability weighted estimators in large statistical models and a variety of problem settings. We assess the practical performance of our estimators in simulation studies and demonstrate use of our proposed methodology with data from a large-scale epidemiologic study.

PMID:35839293 | DOI:10.1111/biom.13719

Bipolar Disord. 2022 Jul 15. doi: 10.1111/bdi.13243. Online ahead of print.

ABSTRACT

BACKGROUND: The clinical effects of smartphone-based interventions for bipolar disorder (BD) have yet to be established.

OBJECTIVES: To examine the efficacy of smartphone-based interventions in BD and how the included studies reported user-engagement indicators.

METHODS: We conducted a systematic search on January 24, 2022, in PubMed, Scopus, Embase, APA PsycINFO, and Web of Science. We used random-effects meta-analysis to calculate the standardized difference (Hedges’ g) in pre-post change scores between smartphone intervention and control conditions. The study was pre-registered with PROSPERO (CRD42021226668).

RESULTS: The literature search identified 6,034 studies. Thirteen articles fulfilled the selection criteria. We included seven RCTs and performed meta-analyses comparing the pre-post change in depressive and (hypo)manic symptom severity, functioning, quality of life, and perceived stress between smartphone interventions and control conditions. There was significant heterogeneity among studies and no meta-analysis reached statistical significance. Results were also inconclusive regarding affective relapses and psychiatric readmissions. All studies reported positive user-engagement indicators.

CONCLUSION: We did not find evidence to support that smartphone interventions may reduce the severity of depressive or manic symptoms in BD. The high heterogeneity of studies supports the need for expert consensus to establish ideally how studies should be designed and the use of more sensitive outcomes, such as affective relapses and psychiatric hospitalizations, as well as the quantification of mood instability. The ISBD Big Data Task Force provides preliminary recommendations to reduce the heterogeneity and achieve more valid evidence in the field.

PMID:35839276 | DOI:10.1111/bdi.13243

PLoS One. 2022 Jul 15;17(7):e0271265. doi: 10.1371/journal.pone.0271265. eCollection 2022.

ABSTRACT

AIMS: The progesterone response of the nuclear progesterone receptor plays an important role in the female reproductive system. Changes in the function of the progesterone receptor gene may increase the risk of reproductive cancer. The present study performed a meta-analysis to examine whether the progesterone receptor gene PROGINS polymorphism was a susceptibility factor for female reproductive cancer.

MATERIALS AND METHODS: We searched the PubMed, Cochrane Library, Web of Science and EMBASE databases for literature on PROGINS polymorphisms and female reproductive cancer published before September 2020. We evaluated the risk using odds ratios [ORs] and 95% confidence intervals via fixed effects models and random-effects models, which were calculated for all five genetic models. We grouped the analyses by race, cancer, and HWE.

RESULTS: Thirty studies comprised of 25405 controls and 19253 female reproductive cancer cases were included in this meta-analysis. We observed that the Alu insertion polymorphism and the V660L polymorphism were significantly associated with female reproductive cancer in the allele and dominant genetic models. The allele genetic model and (Alu-insertion polymorphism: OR = 1.22, 95% CI = 1.02-1.45; V660L polymorphism: OR = 1.02, 95% CI = 1.00-1.13) dominant genetic model (Alu-insertion polymorphism: OR = 1.27, 95% CI = 1.03-1.58; V660L polymorphism: OR = 1.10, 95% CI = 1.011.19) demonstrated a significantly increased risk of female reproductive cancer. A subgroup analysis according to ethnicity found that the Alu insertion was associated with female reproductive cancer incidence in white (Allele model: OR = 1.21, 95% CI = 1.00-1.45; Heterozygous model: OR = 3.44, 95% CI = 1.30-9.09) and Asian (Dominant model: OR = 3.12, 95% CI = 1.25-7.79) populations, but the association disappeared for African and mixed racial groups. However, the V660L polymorphism was significantly associated with female reproductive cancer in the African (Allele model: OR = 2.52, 95% CI = 1.14-5.56; Heterozygous model: OR = 2.83, 95% CI = 1.26-6.35) and mixed racial groups (Dominant model: OR = 1.28, 95% CI = 1.01-1.62). Subgroup analysis by cancer showed that the PROGINS polymorphism increased the risk of cancer in the allele model, dominant mode and heterozygous model, but the confidence interval for this result spanned 1 and was not statistically significant. This sensitivity was verified in studies with HWE greater than 0.5.

CONCLUSION: Our meta-analysis showed that the progesterone receptor gene Alu insertion and the V660L polymorphism contained in the PROGINS polymorphism were susceptibility factors for female reproductive cancer.

PMID:35839271 | DOI:10.1371/journal.pone.0271265

PLoS Comput Biol. 2022 Jul 15;18(7):e1010212. doi: 10.1371/journal.pcbi.1010212. Online ahead of print.

ABSTRACT

Longitudinal ‘omics analytical methods are extensively used in the field of evolving precision medicine, by enabling ‘big data’ recording and high-resolution interpretation of complex datasets, driven by individual variations in response to perturbations such as disease pathogenesis, medical treatment or changes in lifestyle. However, inherent technical limitations in biomedical studies often result in the generation of feature-rich and sample-limited datasets. Analyzing such data using conventional modalities often proves to be challenging since the repeated, high-dimensional measurements overload the outlook with inconsequential variations that must be filtered from the data in order to find the true, biologically relevant signal. Tensor methods for the analysis and meaningful representation of multi-way data may prove useful to the biological research community by their advertised ability to tackle this challenge. In this study, we present tcam-a new unsupervised tensor factorization method for the analysis of multi-way data. Building on top of cutting-edge developments in the field of tensor-tensor algebra, we characterize the unique mathematical properties of our method, namely, 1) preservation of geometric and statistical traits of the data, which enables uncovering information beyond the inter-individual variation that often takes-over the focus, especially in human studies. 2) Natural and straightforward out-of-sample extension, making tcam amenable for integration in machine learning workflows. A series of re-analyses of real-world, human experimental datasets showcase these theoretical properties, while providing empirical confirmation of tcam’s utility in the analysis of longitudinal ‘omics data.

PMID:35839259 | DOI:10.1371/journal.pcbi.1010212

PLoS Comput Biol. 2022 Jul 15;18(7):e1010328. doi: 10.1371/journal.pcbi.1010328. Online ahead of print.

ABSTRACT

Building an accurate disease risk prediction model is an essential step in the modern quest for precision medicine. While high-dimensional genomic data provides valuable data resources for the investigations of disease risk, their huge amount of noise and complex relationships between predictors and outcomes have brought tremendous analytical challenges. Deep learning model is the state-of-the-art methods for many prediction tasks, and it is a promising framework for the analysis of genomic data. However, deep learning models generally suffer from the curse of dimensionality and the lack of biological interpretability, both of which have greatly limited their applications. In this work, we have developed a deep neural network (DNN) based prediction modeling framework. We first proposed a group-wise feature importance score for feature selection, where genes harboring genetic variants with both linear and non-linear effects are efficiently detected. We then designed an explainable transfer-learning based DNN method, which can directly incorporate information from feature selection and accurately capture complex predictive effects. The proposed DNN-framework is biologically interpretable, as it is built based on the selected predictive genes. It is also computationally efficient and can be applied to genome-wide data. Through extensive simulations and real data analyses, we have demonstrated that our proposed method can not only efficiently detect predictive features, but also accurately predict disease risk, as compared to many existing methods.

PMID:35839250 | DOI:10.1371/journal.pcbi.1010328

PLoS Genet. 2022 Jul 15;18(7):e1010281. doi: 10.1371/journal.pgen.1010281. Online ahead of print.

ABSTRACT

Estimating admixture histories is crucial for understanding the genetic diversity we see in present-day populations. Allele frequency or phylogeny-based methods are excellent for inferring the existence of admixture or its proportions. However, to estimate admixture times, spatial information from admixed chromosomes of local ancestry or the decay of admixture linkage disequilibrium (ALD) is used. One popular method, implemented in the programs ALDER and ROLLOFF, uses two-locus ALD to infer the time of a single admixture event, but is only able to estimate the time of the most recent admixture event based on this summary statistic. To address this limitation, we derive analytical expressions for the expected ALD in a three-locus system and provide a new statistical method based on these results that is able to resolve more complicated admixture histories. Using simulations, we evaluate the performance of this method on a range of different admixture histories. As an example, we apply the method to the Colombian and Mexican samples from the 1000 Genomes project. The implementation of our method is available at https://github.com/Genomics-HSE/LaNeta.

PMID:35839249 | DOI:10.1371/journal.pgen.1010281

PLoS One. 2022 Jul 15;17(7):e0271449. doi: 10.1371/journal.pone.0271449. eCollection 2022.

ABSTRACT

BACKGROUND: Community-acquired pneumonia (CAP) is a frequent cause of death worldwide, and in sub-Saharan Africa particularly. Human immunodeficiency virus infection (HIV) and tuberculosis (TB) influence pathogen distribution in patients with CAP. Previous studies in sub-Saharan Africa have shown different frequencies of respiratory pathogens and antibiotic susceptibility compared to studies outside Africa. This study aimed to investigate the aetiology, presentation, and treatment outcomes of community-acquired pneumonia in adults at the University Teaching Hospital in Lusaka, Zambia.

MATERIALS AND METHODS: Three-hundred-and-twenty-seven patients were enrolled at the University Teaching Hospital in Lusaka between March 2018 and December 2018. Clinical characteristics and laboratory data were collected. Sputum samples were tested by microscopy, other TB diagnostics, and bacterial cultures.

RESULTS: The commonest presenting complaint was cough (96%), followed by chest pain (60.6%), fever (59.3%), and breathlessness (58.4%). The most common finding on auscultation of the lungs was chest crackles (51.7%). Seventy percent of the study participants had complaints lasting at least a week before enrolment. The prevalence of HIV was 71%. Sputum samples were tested for 286 patients. The diagnostic yield was 59%. The most common isolate was Mycobacterium tuberculosis (20%), followed by Candida species (18%), Klebsiella pneumoniae (12%), and Pseudomonas aeruginosa (7%). Streptococcus pneumoniae was isolated in only four patients. There were no statistically significant differences between the rates of specific pathogens identified in HIV-infected patients compared with the HIV-uninfected. Thirty-day mortality was 30%. Patients with TB had higher 30-day mortality than patients without TB (p = 0.047).

CONCLUSION: Mycobacterium tuberculosis was the most common cause of CAP isolated in adults at the University Teaching Hospital in Lusaka, Zambia. Gram-negative organisms were frequently isolated. A high mortality rate was observed, as 30% of the followed-up study population had died after 30 days.

PMID:35839238 | DOI:10.1371/journal.pone.0271449

PLoS One. 2022 Jul 15;17(7):e0271561. doi: 10.1371/journal.pone.0271561. eCollection 2022.

ABSTRACT

BACKGROUND AND OBJECTIVES: We previously showed in a 6-month randomized controlled trial that resistance training and protein supplementation after bariatric surgery (Roux-en-Y gastric bypass, RYGB) improved muscle strength without significant effect on weight loss and body composition changes. We performed a 5-year follow-up study in these subjects with the aim 1) to assess the long-term effect of this exercise training intervention and 2) to analyze associations between habitual physical activity (PA) and weight regain at 5 years.

METHODS: Fifty-four out of 76 initial participants (follow-up rate of 71%) completed the 5-year follow-up examination (controls, n = 17; protein supplementation, n = 22; protein supplementation and resistance training, n = 15). We measured body weight and composition (DXA), lower-limb strength (leg-press one-repetition maximum) and habitual PA (Actigraph accelerometers and self-report). Weight regain at 5 years was considered low when <10% of 12-month weight loss.

RESULTS: Mean (SD) time elapse since RYGB was 5.7 (0.9) y. At 5 years, weight loss was 32.8 (10.1) kg, with a mean weight regain of 5.4 (SD 5.9) kg compared with the 12-month assessment. Moderate-to-vigorous PA (MVPA) assessed by accelerometry did not change significantly compared with pre-surgery values (+5.2 [SD 21.7] min/d, P = 0.059), and only 4 (8.2%) patients reported participation in resistance training. Muscle strength decreased over time (overall mean [SD]: -49.9 [53.5] kg, respectively, P<0.001), with no statistically significant difference between exercise training intervention groups. An interquartile increase in MVPA levels was positively associated with lower weight regain (OR [95% CI]: 3.27 [1.41;9.86]).

CONCLUSIONS: Early postoperative participation in a resistance training protocol after bariatric surgery was not associated with improved muscle strength after 5 years of follow-up; however, increasing physical activity of at least moderate intensity may promote weight maintenance after surgery. PA may therefore play an important role in the long-term management of patients with obesity after undergoing bariatric procedure.

PMID:35839214 | DOI:10.1371/journal.pone.0271561

Eur Arch Paediatr Dent. 2022 Jul 15. doi: 10.1007/s40368-022-00732-8. Online ahead of print.

ABSTRACT

BACKGROUND: This systematic review and meta-analysis aimed to answer the following question: Are children and adolescents with attention deficit hyperactivity disorder (ADHD) more likely to have gingival or periodontal disease-related outcomes than their non-ADHD peers?

METHODS: Searches were conducted in the following databases: Embase, Scopus, Web of Science, and PubMed. Google Scholar and OpenGrey were also verified. Observational studies were included in which children and adolescents with ADHD were compared with their healthy peers in terms of gingival and/or periodontal endpoints. Bias appraisal was performed using the Joann Briggs tool for case-control and cross-sectional studies. Meta-analysis was performed using R language. Results are reported as mean difference (MD) and odds ratio (OR). Statistical analyses were performed in RStudio.

RESULTS: A total of 149 records were identified in the searches. Seven studies were included. The meta-analysis showed that children and adolescents with ADHD had a higher mean gingival bleeding index (percentage) than their non-ADHD peers (MD = 11.25; CI = 0.08-22.41; I² = 73%). There was no difference between groups for plaque index (MD = 4.87; CI = – 2.56 to 12.30; I² = 63%) and gingivitis (OR = 1.42; CI = 0.22-9.21; I² = 76%). Regarding the assessment of risk of bias, the major issue found in the articles was the absence of analyses for the control of confounding factors.

CONCLUSION: Children and adolescents with ADHD had more gingival bleeding than their non-ADHD peers, but no difference regarding plaque or gingivitis was detected between groups.

CLINICAL REGISTRATION: CRD42021258404.

PMID:35838891 | DOI:10.1007/s40368-022-00732-8

Pharmacoeconomics. 2022 Jul 15. doi: 10.1007/s40273-022-01167-1. Online ahead of print.

ABSTRACT

BACKGROUND AND OBJECTIVE: Best-worst scaling is a theory-driven method that can be used to prioritize objects in health. We sought to characterize all studies of best-worst scaling to prioritize objects in health, to assess trends of using best-worst scaling in prioritization over time, and to assess the relationship between a legacy measure of quality (PREFS) and a novel assessment of subjective quality and policy relevance.

METHODS: A systematic review identified studies published through to the end of 2021 that applied best-worst scaling to study priorities in health (PROSPERO CRD42020209745), updating a prior review published in 2016. The PubMed, EBSCOhost, Embase, Scopus, APA PsychInfo, Web of Science, and Google Scholar databases were used and were supplemented by a hand search. Data describing the application, development, design, administration/analysis, quality, and policy relevance were summarized and we tested for trends by comparing articles before and after 1 January, 2017. Multivariate statistics were then used to assess the relationships between PREFS, subjective quality, policy relevance, and other possible indicators.

RESULTS: From a total of 2826 unique papers identified, 165 best-worst scaling studies were included in this review. Applications of best-worst scaling to study priorities in health have continued to grow (p < 0.01) and are now used in all regions of the world, most often to study the priorities of patients/consumers (67%). Several key trends can be observed over time: increased use of pretesting (p < 0.05); increased use of online administration (p < 0.01), and decreased use of paper self-administered surveys (p = 0.02); increased use of heterogeneity analysis (p = 0.02); an increase in having a clearly stated purpose (p < 0.01); and a decrease in comparing respondents to non-respondents (p = 0.01). The average sample size has more than doubled, from 228 to 472 respondents, but formal sample size justifications remain low (5.3%) and unchanged over time (p = 0.68). While the average PREFS score remained unchanged at 3.1/5, both subjective quality and policy relevance trended up, but changes were not statistically significant (p = 0.06 and p = 0.13). Most of the variation in subjective quality was driven by PREFS (R² = 0.42), but it was also positively assosciated with policy relevance, heterogeneity analysis, and using a balanced incomplete block design, and was negatively associated with not using developmental methods and an increasing sample size.

CONCLUSIONS: Using best-worst scaling to prioritize objects is now commonly used around the world to assess the priorities of patients and other stakeholders in health. Best practices are clearly emerging for best-worst scaling. Although legacy measures (PREFS) to measure study quality are reasonable, there may need to be new tools to assess both study quality and policy relevance.

PMID:35838889 | DOI:10.1007/s40273-022-01167-1