Applied machine learning to identify differential risk groups underlying externalizing and internalizing problem behaviors trajectories: A case study using a cohort of Asian American children.

BACKGROUND: Internalizing and externalizing problems account for over 75% of the mental health burden in children and adolescents in the US, with higher burden among minority children. While complex interactions of multilevel factors are associated with these outcomes and may enable early identification of children in higher risk, prior research has been limited by data and application of traditional analysis methods. In this case example focused on Asian American children, we address the gap by applying data-driven statistical and machine learning methods to study clusters of mental health trajectories among children, investigate optimal predictions of children at high-risk cluster, and identify key early predictors. METHODS: Data from the US Early Childhood Longitudinal Study 2010-2011 were used. Multilevel information provided by children, families, teachers, schools, and care-providers were considered as predictors. Unsupervised machine learning algorithm was applied to identify groups of internalizing and externalizing problems trajectories. For prediction of high-risk group, ensemble algorithm, Superlearner, was implemented by combining several supervised machine learning algorithms. Performance of Superlearner and candidate algorithms, including logistic regression, was assessed using discrimination and calibration metrics via crossvalidation. Variable importance measures along with partial dependence plots were utilized to rank and visualize key predictors. FINDINGS: We found two clusters suggesting high- and low-risk groups for both externalizing and internalizing problems trajectories. While Superlearner had overall best discrimination performance, logistic regression had comparable performance for externalizing problems but worse for internalizing problems. Predictions from logistic regression were not well calibrated compared to those from Superlearner, however they were still better than few candidate algorithms. Important predictors identified were combination of test scores, child factors, teacher rated scores, and contextual factors, which showed non-linear associations with predicted probabilities. CONCLUSIONS: We demonstrated the application of data-driven analytical approach to predict mental health outcomes among Asian American children. Findings from the cluster analysis can inform critical age for early intervention, while prediction analysis has potential to inform intervention programing prioritization decisions. However, to better understand external validity, replicability, and value of machine learning in broader mental health research, more studies applying similar analytical approach is needed.

Lifetime quality-adjusted life years lost due to genital herpes acquired in the United States in 2018: a mathematical modeling study.

Background: Genital herpes (GH), caused by herpes simplex virus type 1 and type 2 (HSV-1, HSV-2), is a common sexually transmitted disease associated with adverse health outcomes. Symptoms associated with GH outbreaks can be reduced by antiviral medications, but the infection is incurable and lifelong. In this study, we estimate the long-term health impacts of GH in the United States using quality-adjusted life years (QALYs) lost. Methods: We used probability trees to model the natural history of GH secondary to infection with HSV-1 and HSV-2 among people aged 18-49 years. We modelled the following outcomes to quantify the major causes of health losses following infection: symptomatic herpes outbreaks, psychosocial impacts associated with diagnosis and recurrences, urinary retention caused by sacral radiculitis, aseptic meningitis, Mollaret's meningitis, and neonatal herpes. The model was parameterized based on published literature on the natural history of GH. We summarized losses of health by computing the lifetime number of QALYs lost per genital HSV-1 and HSV-2 infection, and we combined this information with incidence estimates to compute the total lifetime number of QALYs lost due to infections acquired in 2018 in the United States. Findings: We estimated 0.05 (95% uncertainty interval (UI) 0.02-0.08) lifetime QALYs lost per incident GH infection acquired in 2018, equivalent to losing 0.05 years or about 18 days of life for one person with perfect health. The average number of QALYs lost per GH infection due to genital HSV-1 and HSV-2 was 0.01 (95% UI 0.01-0.02) and 0.05 (95% UI 0.02-0.09), respectively. The burden of genital HSV-1 is higher among women, while the burden of HSV-2 is higher among men. QALYs lost per neonatal herpes infection was estimated to be 7.93 (95% UI 6.63-9.19). At the population level, the total estimated lifetime QALYs lost as a result of GH infections acquired in 2018 was 33,100 (95% UI 12,600-67,900) due to GH in adults and 3,140 (95% UI 2,260-4,140) due to neonatal herpes. Results were most sensitive to assumptions on the magnitude of the disutility associated with post-diagnosis psychosocial distress and symptomatic recurrences. Interpretation: GH is associated with substantial health losses in the United States. Results from this study can be used to compare the burden of GH to other diseases, and it provides inputs that may be used in studies on the health impact and cost-effectiveness of interventions that aim to reduce the burden of GH. Funding: The Center for Disease Control and Prevention.

The Estimated Lifetime Quality-Adjusted Life-Years Lost Due to Chlamydia, Gonorrhea, and Trichomoniasis in the United States in 2018.

BACKGROUND: Comprehensive evaluation of the quality-adjusted life-years (QALYs) lost attributable to chlamydia, gonorrhea, andtrichomoniasis in the United States is lacking. METHODS: We adapted a previous probability-tree model to estimate the average number of lifetime QALYs lost due to genital chlamydia, gonorrhea, and trichomoniasis, per incident infection and at the population level, by sex and age group. We conducted multivariate sensitivity analyses to address uncertainty around key parameter values. RESULTS: The estimated total discounted lifetime QALYs lost for men and women, respectively, due to infections acquired in 2018, were 1541 (95% uncertainty interval [UI], 186-6358) and 111 872 (95% UI, 29 777-267 404) for chlamydia, 989 (95% UI, 127-3720) and 12 112 (95% UI, 2 410-33 895) for gonorrhea, and 386 (95% UI, 30-1851) and 4576 (95% UI, 13-30 355) for trichomoniasis. Total QALYs lost were highest among women aged 15-24 years with chlamydia. QALYs lost estimates were highly sensitive to disutilities (health losses) of infections and sequelae, and to duration of infections and chronic sequelae for chlamydia and gonorrhea in women. CONCLUSIONS: The 3 sexually transmitted infections cause substantial health losses in the United States, particularly gonorrhea and chlamydia among women. The estimates of lifetime QALYs lost per infection help to prioritize prevention policies and inform cost-effectiveness analyses of sexually transmitted infection interventions.

Generating simple classification rules to predict local surges in COVID-19 hospitalizations.

Low rates of vaccination, emergence of novel variants of SARS-CoV-2, and increasing transmission relating to seasonal changes and relaxation of mitigation measures leave many US communities at risk for surges of COVID-19 that might strain hospital capacity, as in previous waves. The trajectories of COVID-19 hospitalizations differ across communities depending on their age distributions, vaccination coverage, cumulative incidence, and adoption of risk mitigating behaviors. Yet, existing predictive models of COVID-19 hospitalizations are almost exclusively focused on national- and state-level predictions. This leaves local policymakers in urgent need of tools that can provide early warnings about the possibility that COVID-19 hospitalizations may rise to levels that exceed local capacity. In this work, we develop a framework to generate simple classification rules to predict whether COVID-19 hospitalization will exceed the local hospitalization capacity within a 4- or 8-week period if no additional mitigating strategies are implemented during this time. This framework uses a simulation model of SARS-CoV-2 transmission and COVID-19 hospitalizations in the US to train classification decision trees that are robust to changes in the data-generating process and future uncertainties. These generated classification rules use real-time data related to hospital occupancy and new hospitalizations associated with COVID-19, and when available, genomic surveillance of SARS-CoV-2. We show that these classification rules present reasonable accuracy, sensitivity, and specificity (all ≥ 80%) in predicting local surges in hospitalizations under numerous simulated scenarios, which capture substantial uncertainties over the future trajectories of COVID-19. Our proposed classification rules are simple, visual, and straightforward to use in practice by local decision makers without the need to perform numerical computations.

Estimation of the Lifetime Quality-Adjusted Life Years (QALYs) Lost Due to Syphilis Acquired in the United States in 2018.

BACKGROUND: The purpose of this study was to estimate the health impact of syphilis in the United States in terms of the number of quality-adjusted life years (QALYs) lost attributable to infections in 2018. METHODS: We developed a Markov model that simulates the natural history and management of syphilis. The model was parameterized by sex and sexual orientation (women who have sex with men, men who have sex with women [MSW], and men who have sex with men [MSM]), and by age at primary infection. We developed a separate decision tree model to quantify health losses due to congenital syphilis. We estimated the average lifetime number of QALYs lost per infection, and the total expected lifetime number of QALYs lost due to syphilis acquired in 2018. RESULTS: We estimated the average number of discounted lifetime QALYs lost per infection as 0.09 (95% uncertainty interval [UI] .03-.19). The total expected number of QALYs lost due to syphilis acquired in 2018 was 13 349 (5071-31 360). Although per-case loss was the lowest among MSM (0.06), MSM accounted for 47.7% of the overall burden. For each case of congenital syphilis, we estimated 1.79 (1.43-2.16) and 0.06 (.01-.14) QALYs lost in the child and the mother, respectively. We projected 2332 (1871-28 250) and 79 (17-177) QALYs lost for children and mothers, respectively, due to congenital syphilis in 2018. CONCLUSIONS: Syphilis causes substantial health losses in adults and children. Quantifying these health losses in terms of QALYs can inform cost-effectiveness analyses and can facilitate comparisons of the burden of syphilis to that of other diseases.

Predicting resistance to fluoroquinolones among patients with rifampicin-resistant tuberculosis using machine learning methods.

Background: Limited access to drug-susceptibility tests (DSTs) and delays in receiving DST results are challenges for timely and appropriate treatment of multi-drug resistant tuberculosis (TB) in many low-resource settings. We investigated whether data collected as part of routine, national TB surveillance could be used to develop predictive models to identify additional resistance to fluoroquinolones (FLQs), a critical second-line class of anti-TB agents, at the time of diagnosis with rifampin-resistant TB. Methods and findings: We assessed three machine learning-based models (logistic regression, neural network, and random forest) using information from 540 patients with rifampicin-resistant TB, diagnosed using Xpert MTB/RIF and notified in the Republic of Moldova between January 2018 and December 2019. The models were trained to predict the resistance to FLQs based on demographic and TB clinical information of patients and the estimated district-level prevalence of resistance to FLQs. We compared these models based on the optimism-corrected area under the receiver operating characteristic curve (OC-AUC-ROC). The OC-AUC-ROC of all models were statistically greater than 0.5. The neural network model, which utilizes twelve features, performed best and had an estimated OC-AUC-ROC of 0.87 (0.83,0.91), which suggests reasonable discriminatory power. A limitation of our study is that our models are based only on data from the Republic of Moldova and since not externally validated, the generalizability of these models to other populations remains unknown. Conclusions: Models trained on data from phenotypic surveillance of drug-resistant TB can predict resistance to FLQs based on patient characteristics at the time of diagnosis with rifampin-resistant TB using Xpert MTB/RIF, and information about the local prevalence of resistance to FLQs. These models may be useful for informing the selection of antibiotics while awaiting results of DSTs.

Simple decision rules to predict local surges in COVID-19 hospitalizations during the winter and spring of 2022.

Low rates of vaccination, emergence of novel variants of SARS-CoV-2, and increasing transmission relating to seasonal changes leave many U.S. communities at risk for surges of COVID-19 during the winter and spring of 2022 that might strain hospital capacity, as in previous waves. The trajectories of COVID-19 hospitalizations during this period are expected to differ across communities depending on their age distributions, vaccination coverage, cumulative incidence, and adoption of risk mitigating behaviors. Yet, existing predictive models of COVID-19 hospitalizations are almost exclusively focused on national- and state-level predictions. This leaves local policymakers in urgent need of tools that can provide early warnings about the possibility that COVID-19 hospitalizations may rise to levels that exceed local capacity. In this work, we develop simple decision rules to predict whether COVID-19 hospitalization will exceed the local hospitalization capacity within a 4- or 8-week period if no additional mitigating strategies are implemented during this time. These decision rules use real-time data related to hospital occupancy and new hospitalizations associated with COVID-19, and when available, genomic surveillance of SARS-CoV-2. We showed that these decision rules present reasonable accuracy, sensitivity, and specificity (all ≥80%) in predicting local surges in hospitalizations under numerous simulated scenarios, which capture substantial uncertainties over the future trajectories of COVID-19 during the winter and spring of 2022. Our proposed decision rules are simple, visual, and straightforward to use in practice by local decision makers without the need to perform numerical computations. SIGNIFICANCE STATEMENT: In many U.S. communities, the risk of exceeding local healthcare capacity during the winter and spring of 2022 remains substantial since COVID-19 hospitalizations may rise due to seasonal changes, low vaccination coverage, and the emergence of new variants of SARS-CoV-2, such as the omicron variant. Here, we provide simple and easy-to-communicate decision rules to predict whether local hospital occupancy is expected to exceed capacity within a 4- or 8-week period if no additional mitigating measures are implemented. These decision rules can serve as an alert system for local policymakers to respond proactively to mitigate future surges in the COVID-19 hospitalization and minimize risk of overwhelming local healthcare capacity.

Improvement in post-partum uterine involution in rats treated with Apios americana.

OBJECTIVE: Apios americana, a plant used as a staple ingredient of native American diets, has various properties, including anti-cancer, anti-hyperglycemic, hypotensive, and anti-inflammatory activity. In Japan, Apios is used as a post-natal medication. After parturition, women undergo a period of recovery as they return to pre-pregnancy conditions. However, few health products that aid post-partum recovery are on the market. We explored whether Apios can accelerate the post-partum recovery process, in particular the involution of the uterus. METHODS: Female rats kept in individual cages were mated with two male rats, with the exception of the control group (female rats without mating, on basal diet; n=6). After delivery, rats were divided into five groups based on their diet: basal diet (model; n=6); basal diet+oral intake at 5.4 g/kg of Chanfukang granules (a Chinese patent medicine preparation for post-partum lochia) (positive; n=6); basal diet containing 10% Apios powder (low; n=6); basal diet containing 20% Apios powder (medium; n=6); basal diet containing 40% Apios powder (high; n=6). Five days later, uteri and spleens were weighed. Uterus and spleen indices for each rat were calculated by dividing visceral weight by the total weight. Hormone and cytokine concentrations were measured using enzyme-linked immunosorbent assay (ELISA). Histological analysis of uteri was completed using hematoxylin and eosin (H&E) staining. Expression of matrix metalloproteinases and inhibitors in uteri was measured by western blotting. RESULTS: Our results showed that Apios treatment reduced the post-partum uterus index and regulated the hormone concentrations. Moreover, we found that the process of uterine involution was accelerated, based on morphological changes in the uterus. In addition, our results indicated that Apios alleviated the inflammatory response induced by the involution process. Transforming growth factor ß was also found to be regulated by Apios. There were significant downregulation of matrix metalloproteinases and upregulation of their inhibitors by Apios, which suggested that Apios increased the rate of the collagen clearance process. CONCLUSIONS: These results, based on experimental observations at the molecular and protein levels, verified our hypothesis that Apios can improve uterine involution, and demonstrated the potential application of Apios in post-partum care.