Challenges of COVID-19 Case Forecasting in the US, 2020-2021.

During the COVID-19 pandemic, forecasting COVID-19 trends to support planning and response was a priority for scientists and decision makers alike. In the United States, COVID-19 forecasting was coordinated by a large group of universities, companies, and government entities led by the Centers for Disease Control and Prevention and the US COVID-19 Forecast Hub (https://covid19forecasthub.org). We evaluated approximately 9.7 million forecasts of weekly state-level COVID-19 cases for predictions 1-4 weeks into the future submitted by 24 teams from August 2020 to December 2021. We assessed coverage of central prediction intervals and weighted interval scores (WIS), adjusting for missing forecasts relative to a baseline forecast, and used a Gaussian generalized estimating equation (GEE) model to evaluate differences in skill across epidemic phases that were defined by the effective reproduction number. Overall, we found high variation in skill across individual models, with ensemble-based forecasts outperforming other approaches. Forecast skill relative to the baseline was generally higher for larger jurisdictions (e.g., states compared to counties). Over time, forecasts generally performed worst in periods of rapid changes in reported cases (either in increasing or decreasing epidemic phases) with 95% prediction interval coverage dropping below 50% during the growth phases of the winter 2020, Delta, and Omicron waves. Ideally, case forecasts could serve as a leading indicator of changes in transmission dynamics. However, while most COVID-19 case forecasts outperformed a naïve baseline model, even the most accurate case forecasts were unreliable in key phases. Further research could improve forecasts of leading indicators, like COVID-19 cases, by leveraging additional real-time data, addressing performance across phases, improving the characterization of forecast confidence, and ensuring that forecasts were coherent across spatial scales. In the meantime, it is critical for forecast users to appreciate current limitations and use a broad set of indicators to inform pandemic-related decision making.

Evaluation of individual and ensemble probabilistic forecasts of COVID-19 mortality in the United States.

Short-term probabilistic forecasts of the trajectory of the COVID-19 pandemic in the United States have served as a visible and important communication channel between the scientific modeling community and both the general public and decision-makers. Forecasting models provide specific, quantitative, and evaluable predictions that inform short-term decisions such as healthcare staffing needs, school closures, and allocation of medical supplies. Starting in April 2020, the US COVID-19 Forecast Hub (https://covid19forecasthub.org/) collected, disseminated, and synthesized tens of millions of specific predictions from more than 90 different academic, industry, and independent research groups. A multimodel ensemble forecast that combined predictions from dozens of groups every week provided the most consistently accurate probabilistic forecasts of incident deaths due to COVID-19 at the state and national level from April 2020 through October 2021. The performance of 27 individual models that submitted complete forecasts of COVID-19 deaths consistently throughout this year showed high variability in forecast skill across time, geospatial units, and forecast horizons. Two-thirds of the models evaluated showed better accuracy than a naïve baseline model. Forecast accuracy degraded as models made predictions further into the future, with probabilistic error at a 20-wk horizon three to five times larger than when predicting at a 1-wk horizon. This project underscores the role that collaboration and active coordination between governmental public-health agencies, academic modeling teams, and industry partners can play in developing modern modeling capabilities to support local, state, and federal response to outbreaks.

Comparing trained and untrained probabilistic ensemble forecasts of COVID-19 cases and deaths in the United States.

The U.S. COVID-19 Forecast Hub aggregates forecasts of the short-term burden of COVID-19 in the United States from many contributing teams. We study methods for building an ensemble that combines forecasts from these teams. These experiments have informed the ensemble methods used by the Hub. To be most useful to policymakers, ensemble forecasts must have stable performance in the presence of two key characteristics of the component forecasts: (1) occasional misalignment with the reported data, and (2) instability in the relative performance of component forecasters over time. Our results indicate that in the presence of these challenges, an untrained and robust approach to ensembling using an equally weighted median of all component forecasts is a good choice to support public health decision-makers. In settings where some contributing forecasters have a stable record of good performance, trained ensembles that give those forecasters higher weight can also be helpful.

Gambling formats, involvement, and problem gambling: which types of gambling are more risky?

BACKGROUND: The recognition of problem gambling as a public health issue has increased as the availability of gambling expands. Research has found that some formats of gambling are more closely linked to problem gambling than others. Conflicting evidence, however, has emerged, suggesting that the most important consideration is involvement (i.e., number of gambling formats an individual participates in). This debate has important implications for the regulation of gambling formats and for the allocation of problem gambling prevention and treatment services. METHOD: Analyses utilized the Baseline General Population Survey (BGPS) and the Baseline Online Panel Survey (BOPS) of Massachusettscollected in 2013-2014. The BGPS contains a representative sample of 9523 Massachusetts adults and the BOPS contains a sample of 5046 Massachusetts adults. All participants were administered the same comprehensive survey of their past year gambling behavior and problem gambling symptomology. Only those who gambled regularly in the past 12 months (n = 5852) were included. The Problem and Pathological Gambling Measure was used to classify gambling behavior. Within the sample, there were 446 problem gamblers. We assessed: 1) whether some gambling formats are more related to problem gambling; 2) whether problem gambling is positively related to high involvement in gambling; 3) the relationship between involvement in gambling and intensity of gambling; and 4) whether gambling formats mediate the relationship between gambling involvement and problem gambling. RESULTS: Groups of monthly gamblers participating in casino gambling, bingo, and sports betting contained a higher proportion of problem gamblers. High gambling involvement was also positively associated with problem gambling; however, a large minority of gamblers experienced problems when engaging in only one or two forms of gambling. Gambling involvement was also positively associated with intensity of gambling. Therefore, intensity of gambling may be partly driving the relationship between involvement and problem gambling. Specific gambling formats mediated the relationship between involvement and problem gambling. CONCLUSIONS: The gambling format an individual participates in is connected to whether an individual is likely to experience problem gambling. We also found that the level of involvement (and its relationship to intensity) may affect the likelihood that an individual will experience problematic gambling behavior. Ultimately, the type of gambling format an individual partakes in does mediate the relationship between problem gambling and involvement. In Massachusetts, participating in casino gambling was more closely associated with problem gambling than other formats across all levels of involvement.

Correlates of At-Risk and Problem Gambling Among Veterans in Massachusetts.

Few studies have examined problem gambling among veterans and, of those studies, there are conflicting conclusions surrounding correlates of problem gambling in veterans. Our study aims to assess problem gambling prevalence among veterans using non-Veterans Affairs data and to evaluate correlates of problem gambling among veterans in a general population sample. We obtained a probability sample of adult Massachusetts residents using address based sampling in 2013-2014. Participants completed a questionnaire on demographics, veteran status, and gambling behaviors and motivations. We identified n = 129 problem gamblers from a sample of n = 9578 participants. Of the problem gamblers who had veteran status information, 20.6% were veterans. Due to sample size limitations, we analyzed veteran problem and at-risk gamblers compared to veteran recreational gamblers. Having friends and family members engaged in gambling and engaging in more gambling formats were significantly, positively associated with veteran problem and at-risk gambler status. Participating in raffles in the past year was associated with lower odds of being a veteran problem and at-risk gambler compared to veteran recreational gamblers (OR 0.31, 95% CI 0.18-0.52). These discriminators of at-risk and problem gambling may be useful in developing clinical treatment approaches for veteran problem gamblers. Future studies should focus on changes in the prevalence of veteran problem gambling and additional correlates that may better capture social support domains and gambling activity among veterans.

The importance of friends and family to recreational gambling, at-risk gambling, and problem gambling.

BACKGROUND: The variables correlated with problem gambling are routinely assessed and fairly well established. However, problem gamblers were all 'at-risk' and 'recreational' gamblers at some point. Thus, it is instructive from a prevention perspective to also understand the variables which discriminate between recreational gambling and at-risk gambling and whether they are similar or different to the ones correlated with problem gambling. This is the purpose of the present study. METHOD: Between September 2013 to May 2014, a representative sample of 9,523 Massachusetts adults was administered a comprehensive survey of their past year gambling behavior and problem gambling symptomatology. Based on responses to the Problem and Pathological Gambling Measure, respondents were categorized as Non-Gamblers (2,523), Recreational Gamblers (6,271), At-Risk Gamblers (600), or Problem/Pathological Gamblers (129). With the reference category of Recreational Gambler, a series of binary logistic regressions were conducted to identify the demographic, health, and gambling related variables that differentiated Recreational Gamblers from Non-Gamblers, At-Risk-Gamblers, and Problem/Pathological Gamblers. RESULTS: The strongest discriminator of being a Non-Gambler rather than a Recreational Gambler was having a lower portion of friends and family that were regular gamblers. Compared to Recreational Gamblers, At-Risk Gamblers were more likely to: gamble at casinos; play the instant and daily lottery; be male; gamble online; and be born outside the United States. Compared to Recreational Gamblers, Problem and Pathological Gamblers were more likely to: play the daily lottery; be Black; gamble at casinos; be male; gamble online; and play the instant lottery. Importantly, having a greater portion of friends and family who were regular gamblers was the second strongest correlate of being both an At-Risk Gambler and Problem/Pathological Gambler. CONCLUSIONS: These analyses offer an examination of the similarities and differences between gambling subtypes. An important finding throughout the analyses is that the gambling involvement of family and friends is strongly related to Recreational Gambling, At-Risk Gambling, and Problem/Pathological Gambling. This suggests that targeting the social networks of heavily involved Recreational Gamblers and At-Risk Gamblers (in addition to Problem/Pathological Gamblers) could be an important focus of efforts in problem gambling prevention.

Triggering interventions for influenza: the ALERT algorithm.

BACKGROUND: Early, accurate predictions of the onset of influenza season enable targeted implementation of control efforts. Our objective was to develop a tool to assist public health practitioners, researchers, and clinicians in defining the community-level onset of seasonal influenza epidemics. METHODS: Using recent surveillance data on virologically confirmed infections of influenza, we developed the Above Local Elevated Respiratory Illness Threshold (ALERT) algorithm, a method to identify the period of highest seasonal influenza activity. We used data from 2 large hospitals that serve Baltimore, Maryland and Denver, Colorado, and the surrounding geographic areas. The data used by ALERT are routinely collected surveillance data: weekly case counts of laboratory-confirmed influenza A virus. The main outcome is the percentage of prospective seasonal influenza cases identified by the ALERT algorithm. RESULTS: When ALERT thresholds designed to capture 90% of all cases were applied prospectively to the 2011-2012 and 2012-2013 influenza seasons in both hospitals, 71%-91% of all reported cases fell within the ALERT period. CONCLUSIONS: The ALERT algorithm provides a simple, robust, and accurate metric for determining the onset of elevated influenza activity at the community level. This new algorithm provides valuable information that can impact infection prevention recommendations, public health practice, and healthcare delivery.

Marginal structural models for the estimation of the risk of Diabetes Mellitus in the presence of elevated depressive symptoms and antidepressant medication use in the Women's Health Initiative observational and clinical trial cohorts.

BACKGROUND: We evaluate the combined effect of the presence of elevated depressive symptoms and antidepressant medication use with respect to risk of type 2 diabetes among approximately 120,000 women enrolled in the Women's Health Initiative (WHI), and compare several different statistical models appropriate for causal inference in non-randomized settings. METHODS: Data were analyzed for 52,326 women in the Women's Health Initiative Clinical Trials (CT) Cohort and 68,169 women in the Observational Study (OS) Cohort after exclusions. We included follow-up to 2005, resulting in a median duration of 7.6 years of follow up after enrollment. Results from three multivariable Cox models were compared to those from marginal structural models that included time varying measures of antidepressant medication use, presence of elevated depressive symptoms and BMI, while adjusting for potential confounders including age, ethnicity, education, minutes of recreational physical activity per week, total energy intake, hormone therapy use, family history of diabetes and smoking status. RESULTS: Our results are consistent with previous studies examining the relationship of antidepressant medication use and risk of type 2 diabetes. All models showed a significant increase in diabetes risk for those taking antidepressants. The Cox Proportional Hazards models using baseline covariates showed the lowest increase in risk , with hazard ratios of 1.19 (95 % CI 1.06 - 1.35) and 1.14 (95 % CI 1.01 - 1.30) in the OS and CT, respectively. Hazard ratios from marginal structural models comparing antidepressant users to non-users were 1.35 (95 % CI 1.21 - 1.51) and 1.27 (95 % CI 1.13 - 1.43) in the WHI OS and CT, respectively - however, differences among estimates from traditional Cox models and marginal structural models were not statistically significant in both cohorts. One explanation suggests that time-dependent confounding was not a substantial factor in these data, however other explanations exist. Unadjusted Cox Proportional Hazards models showed that women with elevated depressive symptoms had a significant increase in diabetes risk that remained after adjustment for confounders. However, this association missed the threshold for statistical significance in propensity score adjusted and marginal structural models. CONCLUSIONS: Results from the multiple approaches provide further evidence of an increase in risk of type 2 diabetes for those on antidepressants.

Relations of depressive symptoms and antidepressant use to body mass index and selected biomarkers for diabetes and cardiovascular disease.

OBJECTIVES: We investigated whether depressive symptoms and antidepressant use are associated with biomarkers for glucose dysregulation and inflammation, body mass index (BMI), and waist circumference. METHODS: Postmenopausal women were recruited into the Women's Health Initiative from 1993 to 1998, and data were collected at regular intervals through 2005. We used multiple linear regression models to examine whether depressive symptoms and antidepressant use are associated with BMI, waist circumference, and biomarkers. RESULTS: Analysis of data from 71, 809 women who completed all relevant baseline and year 3 assessments showed that both elevated depressive symptoms and antidepressant use were significantly associated with higher BMI and waist circumference. Among 1950 women, elevated depressive symptoms were significantly associated with increased insulin levels and measures of insulin resistance. Analyses of baseline data from 2242 women showed that both elevated depressive symptoms and antidepressant use were associated with higher C-reactive protein levels. CONCLUSIONS: Monitoring body habitus and other biomarkers among women with elevated depression symptoms or taking antidepressant medication may be prudent to prevent diabetes and cardiovascular disease.

The United States COVID-19 Forecast Hub dataset.

Academic researchers, government agencies, industry groups, and individuals have produced forecasts at an unprecedented scale during the COVID-19 pandemic. To leverage these forecasts, the United States Centers for Disease Control and Prevention (CDC) partnered with an academic research lab at the University of Massachusetts Amherst to create the US COVID-19 Forecast Hub. Launched in April 2020, the Forecast Hub is a dataset with point and probabilistic forecasts of incident cases, incident hospitalizations, incident deaths, and cumulative deaths due to COVID-19 at county, state, and national, levels in the United States. Included forecasts represent a variety of modeling approaches, data sources, and assumptions regarding the spread of COVID-19. The goal of this dataset is to establish a standardized and comparable set of short-term forecasts from modeling teams. These data can be used to develop ensemble models, communicate forecasts to the public, create visualizations, compare models, and inform policies regarding COVID-19 mitigation. These open-source data are available via download from GitHub, through an online API, and through R packages.

DNA methylation in paired breast epithelial and white blood cells from women undergoing reduction mammoplasty.

BACKGROUND: The extent to which white blood cell (WBC) DNA methylation provides information on the status of breast epithelial cell DNA is unknown. PATIENTS AND METHODS: We examined the correlation between methylation in Ras-association domain family-1 gene (RASSF1), a tumor-suppressor gene, and methylation in repetitive elements in paired sets of DNA from WBC and breast epithelial cells collected from 32 women undergoing reduction mammoplasty. RESULTS: We observed no evidence of correlation in methylation levels for ALU, long interspersed nuclear element-1 (LINE1) or juxtacentromeric satellite-2 (SAT2) (r=0.02 for LINE1, p=0.98; r=0.28 for ALU, p=0.12; r=0.26 for SAT2, p=0.17) for matched sets of DNA from WBC and breast epithelial cells. Variability in these markers across individuals and in the same tissue was low. Five women had an average methylation level above 5% for RASSF1 in breast epithelial cell DNA; however, average methylation levels in WBC DNA for these women were all below 1%. CONCLUSION: Methylation patterns in WBC DNA did not reflect methylation patterns in the breast.

Elevated depressive symptoms, antidepressant use, and diabetes in a large multiethnic national sample of postmenopausal women.

OBJECTIVE: To examine elevated depressive symptoms and antidepressant use in relation to diabetes incidence in the Women's Health Initiative. RESEARCH DESIGN AND METHODS: A total of 161,808 postmenopausal women were followed for over an average of 7.6 years. Hazard ratios (HRs) estimating the effects of elevated depressive symptoms and antidepressant use on newly diagnosed incident diabetes were obtained using Cox proportional hazards models adjusted for known diabetes risk factors. RESULTS: Multivariable-adjusted HRs indicated an increased risk of incident diabetes with elevated baseline depressive symptoms (HR 1.13 [95% CI 1.07-1.20]) and antidepressant use (1.18 [1.10-1.28]). These associations persisted through year 3 data, in which respective adjusted HRs were 1.23 (1.09-1.39) and 1.31 (1.14-1.50). CONCLUSIONS: Postmenopausal women with elevated depressive symptoms who also use antidepressants have a greater risk of developing incident diabetes. In addition, longstanding elevated depressive symptoms and recent antidepressant medication use increase the risk of incident diabetes.

Insulin therapy to improve BMI in cystic fibrosis-related diabetes without fasting hyperglycemia: results of the cystic fibrosis related diabetes therapy trial.

OBJECTIVE: Cystic fibrosis-related diabetes (CFRD) without fasting hyperglycemia (CFRD FH-) is not associated with microvascular or macrovascular complications, leading to controversy about the need for treatment. The Cystic Fibrosis Related Diabetes Therapy (CFRDT) Trial sought to determine whether diabetes therapy improves BMI in these patients. RESEARCH DESIGN AND METHODS: A three-arm multicenter randomized trial compared 1 year of therapy with premeal insulin aspart, repaglinide, or oral placebo in subjects with cystic fibrosis who had abnormal glucose tolerance. RESULTS: One hundred adult patients were enrolled. Eighty-one completed the study, including 61 with CFRD FH- and 20 with severly impaired glucose tolerance (IGT). During the year before therapy, BMI declined in all groups. Among the group with CFRD FH-, insulin-treated patients lost 0.30 +/- 0.21 BMI units the year before therapy. After 1 year of insulin therapy, this pattern reversed, and they gained 0.39 +/- 21 BMI units (P = 0.02). No significant change in the rate of BMI decline was seen in placebo-treated patients (P = 0.45). Repaglinide-treated patients had an initial significant BMI gain (0.53 +/- 0.19 BMI units, P = 0.01), but this effect was not sustained. After 6 months of therapy they lost weight so that by 12 months there was no difference in the rate of BMI change during the study year compared with the year before (P = 0.33). Among patients with IGT, neither insulin nor repaglinide affected the rate of BMI decline. No significant differences were seen in the rate of lung function decline or the number of hospitalizations in any group. CONCLUSIONS: Insulin therapy safely reversed chronic weight loss in patients with CFRD FH-.

A process evaluation of a telephone-based peer-delivered smoking cessation intervention for adult survivors of childhood cancer: the partnership for health study.

OBJECTIVE: We report on the process evaluation of an efficacious national smoking cessation intervention for adult survivors of childhood cancer. We examine associations between intervention implementation characteristics and study outcomes, as well as participant characteristics related to level of involvement in the intervention. METHODS: The study was conducted at the Dana-Farber Cancer Institute in Boston, Massachusetts, from 1999-2001. Participants (n = 398) were randomly assigned to receive a proactive telephone-based peer counseling intervention. They received up to 6 counseling calls, individually tailored and survivor-targeted materials, and nicotine replacement therapy (NRT) patches if they were prepared to quit smoking. RESULTS: Forty-two percent of survivors participated in the maximum number of calls (5-6), and 29% of participants requested and received NRT. Total counseling time was an average of 51 min. Quit status at follow-up was related to intervention dose, and participants who received NRT were significantly more likely to make a 24-h quit attempt. Demographic variables (females, White), higher daily smoking rate, poorer perceived health and moderate perceived risk of smoking were significantly related to greater intervention involvement. CONCLUSIONS: A brief peer-delivered, telephone counseling intervention is an effective way to intervene with adult survivors of childhood cancer who are smoking. Findings from the process evaluation data (call length and number, frequency, and spacing) will inform future telephone counseling cessation programs.