More Like this

1. Development and evaluation of a machine learning model to predict acute care for opioid use disorder among Medicaid enrollees engaged in a community‐based treatment program

   https://doi.org/10.1111/add.70079  

   Xue, Lingshu; Yin, Ruofei; Cole, Evan S; Lo‐Ciganic, Wei‐Hsuan; Gellad, Walid F; Donohue, Julie; Tang, Lu (April 2025, Addiction)

   Abstract AimsTo develop machine‐learning algorithms for predicting the risk of a hospitalization or emergency department (ED) visit for opioid use disorder (OUD) (i.e. OUD acute events) in Pennsylvania Medicaid enrollees in the Opioid Use Disorder Centers of Excellence (COE) program and to evaluate the fairness of model performance across racial groups. MethodsWe studied 20 983 United States Medicaid enrollees aged 18 years or older who had COE visits between April 2019 and March 2021. We applied multivariate logistic regression, least absolute shrinkage and selection operator models, random forests, and eXtreme Gradient Boosting (XGB), to predict OUD acute events following the initial COE visit. Our models included predictors at the system, patient, and regional levels. We assessed model performance using multiple metrics by racial groups. Individuals were divided into a low, medium and high‐risk group based on predicted risk scores. ResultsThe training (n = 13 990) and testing (n = 6993) samples displayed similar characteristics (mean age 38.1 ± 9.3 years, 58% male, 80% White enrollees) with 4% experiencing OUD acute events at baseline. XGB demonstrated the best prediction performance (C‐statistic = 76.6% [95% confidence interval = 75.6%–77.7%] vs. 72.8%–74.7% for other methods). At the balanced cutoff, XGB achieved a sensitivity of 68.2%, specificity of 70.0%, and positive predictive value of 8.3%. The XGB model classified the testing sample into high‐risk (6%), medium‐risk (30%), and low‐risk (63%) groups. In the high‐risk group, 40.7% had OUD acute events vs. 16.5% and 5.0% in the medium‐ and low‐risk groups. The high‐ and medium‐risk groups captured 44% and 26% of individuals with OUD events. The XGB model exhibited lower false negative rates and higher false positive rates in racial/ethnic minority groups than White enrollees. ConclusionsNew machine‐learning algorithms perform well to predict risks of opioid use disorder (OUD) acute care use among United States Medicaid enrollees and improve fairness of prediction across racial and ethnic groups compared with previous OUD‐related models. 

   more » « less
2. Opioid Treatment Deserts: Concept development and application in a US Midwestern urban county

   https://doi.org/10.1371/journal.pone.0250324  

   Hyder, Ayaz; Lee, Jinhyung; Dundon, Ashley; Southerland, Lauren T.; All, David; Hammond, Gretchen; Miller, Harvey J. (May 2021, PLOS ONE)

   West, Brooke (Ed.)

   Objectives An Opioid Treatment Desert is an area with limited accessibility to medication-assisted treatment and recovery facilities for Opioid Use Disorder. We explored the concept of Opioid Treatment Deserts including racial differences in potential spatial accessibility and applied it to one Midwestern urban county using high resolution spatiotemporal data. Methods We obtained individual-level data from one Emergency Medical Services (EMS) agency (Columbus Fire Department) in Franklin County, Ohio. Opioid overdose events were based on EMS runs where naloxone was administered from 1/1/2013 to 12/31/2017. Potential spatial accessibility was measured as the time (in minutes) it would take an individual, who may decide to seek treatment after an opioid overdose, to travel from where they had the overdose event, which was a proxy measure of their residential location, to the nearest opioid use disorder (OUD) treatment provider that provided medically-assisted treatment (MAT). We estimated accessibility measures overall, by race and by four types of treatment providers (any type of MAT for OUD, Buprenorphine, Methadone, or Naltrexone). Areas were classified as an Opioid Treatment Desert if the estimate travel time to treatment provider (any type of MAT for OUD) was greater than a given threshold. We performed sensitivity analysis using a range of threshold values based on multiple modes of transportation (car and public transit) and using only EMS runs to home/residential location types. Results A total of 6,929 geocoded opioid overdose events based on data from EMS agencies were used in the final analysis. Most events occurred among 26–35 years old (34%), identified as White adults (56%) and male (62%). Median travel times and interquartile range (IQR) to closest treatment provider by car and public transit was 2 minutes (IQR: 3 minutes) and 17 minutes (IQR: 17 minutes), respectively. Several neighborhoods in the study area had limited accessibility to OUD treatment facilities and were classified as Opioid Treatment Deserts. Travel time by public transit for most treatment provider types and by car for Methadone-based treatment was significantly different between individuals who were identified as Black adults and White adults based on their race. Conclusions Disparities in access to opioid treatment exist at the sub-county level in specific neighborhoods and across racial groups in Columbus, Ohio and can be quantified and visualized using local public safety data (e.g., EMS runs). Identification of Opioid Treatment Deserts can aid multiple stakeholders better plan and allocate resources for more equitable access to MAT for OUD and, therefore, reduce the burden of the opioid epidemic while making better use of real-time public safety data to address a public health epidemic that has turned into a public safety crisis. 

   more » « less
3. Can past variants of SARS-CoV-2 predict the impact of future variants? Machine learning for early warning of US counties at risk

   https://doi.org/10.1007/s10729-025-09728-4  

   Smith, Kevin B; Shen, Siqian; Denton, Brian T (December 2025, Health Care Management Science)

   AbstractIn this paper, we determine whether machine learning (ML) models created using data from the novel SARS-CoV-2 Alpha variant can prospectively predict county-level incidence of emerging variants, validated using data of the Omicron variant. We first select publicly-available sociodemographic, economic, and health-related characteristics of 3140 United States (US) counties at the time of the confirmed early US outbreak of the novel SARS-CoV-2 virus in March 2020 for analysis. Our primary result is the set of US counties that experienced the upper quartile of population-adjusted Omicron variant incidence at certain period (e.g., 100 days) after Omicron variant’s appearance in the US. We show more predictive results by incorporating additional data and features (e.g., human mobility) that can be acquired dynamically after the outbreak, to improve prediction accuracy at the cost of additional waiting time and effort. Towards the goal of decision support, we aim to prospectively evaluate our models’ ability to classify and rank US counties at risk. We measure their classification performance using the area under receiver operating characteristic curve score with 95% confidence intervals. We further calculate the proportion of the top counties by Omicron incidence that our models correctly identify, and compare their score with those of individual county-level features that can serve as a heuristic predictive performance baseline. Our results show that ML algorithms predict county-level Omicron variant incidence with better performance than natural heuristics that decision makers might otherwise use. More generally, historical data from the first wave of a novel pandemic can help predict the incidence of future variants and strengthen state or federal pandemic response interventions. HighlightsData-driven predictive models that capture patterns from early viral variants can support policymaking related to emerging viral variants.County-level sociodemographic, health, and economic characteristics are predictive of early COVID-19 outcomes in the United States (US).Machine learning models trained on early US county-level COVID-19 outcomes are additionally predictive of county-level SARS-CoV-2 Omicron variant outcomes.County-level machine learning models can be used as a critical policymaking tool given the inevitability of novel emerging viruses. 

   more » « less
4. US Abortion Bans and Fertility

   https://doi.org/10.1001/jama.2024.28527  

   Bell, Suzanne O; Franks, Alexander M; Arbour, David; Anjur-Dietrich, Selena; Stuart, Elizabeth A; Ben-Michael, Eli; Feller, Avi; Gemmill, Alison (April 2025, JAMA)

   ImportanceAbortion bans may lead to births among those who are unable to overcome barriers to abortion. The population-level effects of these policies, particularly their unequal impacts across subpopulations in the US, remain unclear. ObjectiveTo assess heterogeneity in the association of abortion bans with changes in fertility in the US, within and across states. Design, Setting, and ParticipantsDrawing from birth certificate and US Census Bureau data from 2012 through 2023 for all 50 states and the District of Columbia, this study used a bayesian panel data model to evaluate state-by-subgroup-specific changes in fertility associated with complete or 6-week abortion bans in 14 US states. The average percent and absolute change in the fertility rate among females aged 15 through 44 years was estimated overall and by state, and within and across states by age, race and ethnicity, marital status, education, and insurance payer. ExposureComplete or 6-week abortion ban. Main outcome and MeasuresFertility rate (births per 1000 reproductive-aged females) overall and by subgroups. ResultsThere were an estimated 1.01 (95% credible interval [CrI], 0.45-1.64) additional births above expectation per 1000 females aged 15 through 44 years (reproductive age) in states following adoption of abortion bans (60.55 observed vs 59.54 expected; 1.70% increase; 95% CrI, 0.75%-2.78%), equivalent to 22 180 excess births, with evidence of variation by state and subgroup. Estimated differences above expectation were largest for racially minoritized individuals (≈2.0%), unmarried individuals (1.79%), individuals younger than 35 years (≈2.0%), Medicaid beneficiaries (2.41%), and those without college degrees (high school diploma, 2.36%; some college, 1.58%), particularly in southern states. Differences in race and ethnicity and education across states explain most of the variability in the state-level association between abortion bans and fertility rates. Conclusion and RelevanceThese findings provide evidence that fertility rates in states with abortion bans were higher than would have been expected in the absence of these policies, with the largest estimated differences among subpopulations experiencing the greatest structural disadvantages and in states with among the worst maternal and child health and well-being outcomes. 

   more » « less
5. Prevalence and Correlates of Long COVID Symptoms Among US Adults

   https://doi.org/10.1001/jamanetworkopen.2022.38804  

   Perlis, Roy H.; Santillana, Mauricio; Ognyanova, Katherine; Safarpour, Alauna; Lunz Trujillo, Kristin; Simonson, Matthew D.; Green, Jon; Quintana, Alexi; Druckman, James; Baum, Matthew A.; et al (October 2022, JAMA Network Open)

   ImportancePersistence of COVID-19 symptoms beyond 2 months, or long COVID, is increasingly recognized as a common sequela of acute infection. ObjectivesTo estimate the prevalence of and sociodemographic factors associated with long COVID and to identify whether the predominant variant at the time of infection and prior vaccination status are associated with differential risk. Design, Setting, and ParticipantsThis cross-sectional study comprised 8 waves of a nonprobability internet survey conducted between February 5, 2021, and July 6, 2022, among individuals aged 18 years or older, inclusive of all 50 states and the District of Columbia. Main Outcomes and MeasuresLong COVID, defined as reporting continued COVID-19 symptoms beyond 2 months after the initial month of symptoms, among individuals with self-reported positive results of a polymerase chain reaction test or antigen test. ResultsThe 16 091 survey respondents reporting test-confirmed COVID-19 illness at least 2 months prior had a mean age of 40.5 (15.2) years; 10 075 (62.6%) were women, and 6016 (37.4%) were men; 817 (5.1%) were Asian, 1826 (11.3%) were Black, 1546 (9.6%) were Hispanic, and 11 425 (71.0%) were White. From this cohort, 2359 individuals (14.7%) reported continued COVID-19 symptoms more than 2 months after acute illness. Reweighted to reflect national sociodemographic distributions, these individuals represented 13.9% of those who had tested positive for COVID-19, or 1.7% of US adults. In logistic regression models, older age per decade above 40 years (adjusted odds ratio [OR], 1.15; 95% CI, 1.12-1.19) and female gender (adjusted OR, 1.91; 95% CI, 1.73-2.13) were associated with greater risk of persistence of long COVID; individuals with a graduate education vs high school or less (adjusted OR, 0.67; 95% CI, 0.56-0.79) and urban vs rural residence (adjusted OR, 0.74; 95% CI, 0.64-0.86) were less likely to report persistence of long COVID. Compared with ancestral COVID-19, infection during periods when the Epsilon variant (OR, 0.81; 95% CI, 0.69-0.95) or the Omicron variant (OR, 0.77; 95% CI, 0.64-0.92) predominated in the US was associated with diminished likelihood of long COVID. Completion of the primary vaccine series prior to acute illness was associated with diminished risk for long COVID (OR, 0.72; 95% CI, 0.60-0.86). Conclusions and RelevanceThis study suggests that long COVID is prevalent and associated with female gender and older age, while risk may be diminished by completion of primary vaccination series prior to infection. 

   more » « less