Home » Additional Features

Surveys, Dashboards, and Innovations in Response to COVID-19

1 September 2021 692 views No Comment
Dan Liao, Survey Research Methods Section Publication Officer
    In human history, innovations have often come rapidly in response to a crisis, rather than resulting from continuous efforts. This is true for the statistical world in response to COVID-19—we are witnessing emerging innovations and technological shifts. National, state, and local governments need statisticians to collect, analyze, and interpret relevant data more rapidly than ever before, so innovators have jumped up to meet this need. Some of these innovations have been given great attention by the media, while others have quietly had a profound impact on the lives of millions, perhaps billions, of people.

    Here, ASA members share their experiences and offer insights into how the pandemic has changed and will continue to change statistics and data science.

    ICF Covid Monitor Poll, United States

    John Boyle, Thomas Brassell, and James Dayton, ICF

    Topics covered in the ICF COVID-19 Monitor Survey of US adults include public health, economic and personal finance impact, mitigation, and mental health. The survey questions remain largely consistent from wave to wave, allowing us to track how Americans’ attitudes about COVID-19 are changing over time. The ICF COVID-19 Monitor Survey is a cross-sectional, census-balanced, nonprobability mobile panel survey with approximately 1,000 interviews per iteration. ICF completed the initial poll in March 2020 and the 11th iteration in April 2021.

    Predictors of Early Uptake of the COVIDSafe App, Australia

    Benjamin Phillips, Social Research Centre

    The COVIDSafe app was developed and deployed by the Australian government to enable rapid contact tracing for people who are diagnosed with COVID-19. To provide an early picture of who had and had not downloaded the COVIDSafe app, the Social Research Centre surveyed a sample of 542 Australian adults from the Australian probability online panel, Life in Australia. This data set was supplemented with an extensive range of longitudinal information by linking at an individual level to research by the Australian National University conducted in previous waves of data collection.

    Although Life in Australia is representative of Australians, the study sample was a subset of panel members and over-represented those who have higher levels of digital aptitude. For this reason, the Heckman model was used for the analysis to address the problem where some results are unobserved due to a selection effect (in this case, selection into the sample).

    At the time of data collection, about 31 percent of Australian smartphone users were estimated to have downloaded and installed COVIDSafe. Several important factors were found that either motivated people to download and install or acted as a barrier to installation.

    Research on the Epidemiology of SARS-CoV-2 in Essential Response Personnel (RECOVER) Study

    Mark Thompson, Centers for Disease Control and Prevention; Lauren Olsho, Abt Associates

    The RECOVER study performed by Abt Associates under contract to the US Centers for Disease Control and Prevention (CDC) is a prospective, longitudinal cohort study of approximately 3,000 health care personnel, first responders, and essential workers from six US geographies. This research offers CDC timely information about COVID-19 illness characterization and correlates of protection in this highly exposed cohort. The research will also provide rates of asymptomatic infection and re-infection, as well as details about viral shedding. Since this population has been prioritized for vaccination with novel COVID-19 vaccine products and has thus been vaccinated for several months, the study also accommodates an early real-world evaluation of the effectiveness of vaccines. The data from the study was underlying the first assessments of the real-world COVID-19 vaccine effectiveness by CDC.

    The CDC was able to estimate that COVID-19 symptomatic illness and asymptomatic infection was reduced by 94 percent among health care personnel who were fully vaccinated, defined in this study as seven or more days after receipt of a second vaccine dose, and by 82 percent among those who were partially vaccinated, defined in this study as 14 days after receipt of dose one through six days after dose two.

    These findings support the CDC’s recommendation that everyone should get both doses of an mRNA COVID-19 vaccine to get the most protection. The study also demonstrated shorter illness duration and less severe symptoms among “breakthrough” cases, those infected after vaccination.

    The Ohio COVID-19 Survey

    Timothy Sahr and Robert Ashmead, Ohio Colleges of Medicine Government Resource Center; Bo Lu, The Ohio State University; and Marcus Berzofsky and Caroline Scruggs, RTI International

    The Ohio COVID-19 Survey (OCS) is a 10-minute web-/telephone-based survey of Ohio residents that uses a rotating panel design to estimate changes over time in regional health and economic indicators related to the COVID-19 pandemic. The OCS takes advantage of an existing large statewide representative survey, the 2019 Ohio Medicaid Assessment Survey (OMAS), which it uses as a sampling frame to create representative sub-samples in a timely fashion.

    Over the course of the COVID-19 pandemic, the survey has changed focus with topics ranging from COVID-19 symptoms and social distancing behaviors to vaccine uptake and hesitancy. The OCS dashboard displays summary results of the survey and is currently updated biweekly as new data is made available.

    The OCS is sponsored by a partnership between the Ohio Department of Health, Ohio Department of Medicaid’s Medicaid Technical Assistance and Policy Program, and The Ohio State University. It is a subproject of the OMAS, with research contributions from the faculty of The Ohio State University’s College of Public Health and Department of Sociology, Ohio University, Ohio Department of Health, Ohio Department of Medicaid, RTI International, and the Ohio Colleges of Medicine Government Resource Center.

    Health District of Northern Larimer County Triennial Community Health Survey 2020 Covid Follow-Up

    Sue Hewitt and Suman Mathur, Health District of Northern Larimer County, and David Brown and Jay Breidt, Colorado State University

    Graduate student David Brown and Jay Breidt, a professor in the department of statistics at Colorado State University, collaborated with Sue Hewitt and Suman Mathur of the Health District of Northern Larimer County (HDNLC) on the design and analysis of the 2019 Triennial Community Health Survey, which examined the health status, needs, and behaviors of the adult population in Larimer County, Colorado. In 2020, HDNLC assessed the impacts of COVID-19 by resurveying the respondents from the 2019 study. Brown and Breidt assisted with the design and analysis of the updated study, which allowed HDNLC to ask new questions (e.g., adherence to public health guidelines, COVID-19 testing results, etc.) and look for changes in previous responses (e.g., mental health status, insurance coverage, etc.) after the onset of the COVID-19 pandemic.

    Dallas-Fort Worth COVID-19 Prevalence Study

    Jill A. Dever, Nicole Mack, and Jamie Ridenhour, RTI International; Amit Singal and Jasmin Tiro, University of Texas Southwestern Medical Center; and Andrew Masica, Texas Health Resources

    The Dallas-Fort Worth COVID Prevalence Study, sponsored by The University of Texas Southwestern Medical Center (UTSW) and Texas Health Resources (THR), included two protocols, each designed to produce prevalence estimates of active and past COVID-19 infections separately for Dallas and Tarrant counties.

    In Protocol 1, one eligible adult aged 18–89 years was identified within the households randomly chosen from RTI’s Enhanced Address-Based Sampling frame. Protocol 2 was initiated for study-eligible county residents who voluntarily enrolled in the study to address low enrollment rates in Protocol 1.

    RTI International collected questionnaire responses via web, telephone, and paper, followed by biospecimen collection (i.e., nasal swab and blood draw) at a UTSW / THR-directed site.

    NYC Neighborhoods COVID-19 Dashboard

    Qixuan Chen, Columbia University

    New York City (NYC) was an early epicenter of the COVID-19 pandemic in the United States. Although the NYC Department of Health has collected and made publicly available data on COVID-19, there was no layperson-friendly website that provided COVID tracking and development information in local neighborhoods. Motivated by this, Qixuan Chen and her students in the department of biostatistics at Columbia University developed the NYC Neighborhoods COVID-19 Dashboard.

    The dashboard was first published in August 2020 and has been updated daily since then. It tracks daily new cases, hospitalizations, deaths, and tests for every NYC neighborhood by ZIP code and provides data visualizations of distributions and time trends for COVID cases, hospitalizations, deaths, and tests by neighborhoods and demographics. Interactive maps of new cases and incidence rate by ZIP code enable identification of clusters of neighborhoods with the most virus cases emerging on a daily basis. Projections plots visualize the projected trends on new infections, new cases, new hospitalizations, and new deaths in the next eight weeks.

    Statistical Adjustments of Sample Representation in Community-Level Prevalence Estimates of COVID-19 Transmission and Immunity

    Yajuan Si, University of Michigan; Andrew Gelman, Columbia University

    Throughout the COVID-19 pandemic, government policy and health care implementation responses have been guided by reported positivity rates and counts of positive cases in the community. The selection bias of this data calls into question its validity as a measure of the actual viral incidence in the community and predictor of clinical burden. Yajuan Si, a professor at the University of Michigan, and Andrew Gelman, a professor at Columbia University, have developed a statistical adjustment approach under multilevel regression and poststratification for viral RNA testing data collected from asymptomatic patients who present for elective procedures within a hospital system to estimate true viral incidence and immunity prevalence in the community.

    Empirical studies found this model predicted the clinical burden of SARS-CoV-2 earlier and more accurately than currently accepted metrics. This method can be implemented easily in a wide variety of hospital settings.

    COVID-19 Survey Participation and Well-Being: A Survey Experiment

    Kate Sollis, Nicholas Biddle, and Ben Edwards, Australian National University, and Diane Herz, Social Research Centre

    Individuals throughout the world are being recruited into studies to examine the social impacts of COVID-19. While previous literature has illustrated how research participation can affect distress and well-being, no study has examined this in the COVID-19 context to the authors’ best knowledge. Using an innovative approach, this study analyzes the impacts of participation in a COVID-19 survey in Australia on subjective well-being through a survey experiment.

    At a population level, the authors found no evidence that participation affects subjective well-being. However, this may not hold for those with mental health concerns and those living in financial insecurity.

    These findings provide the research community with a deeper understanding of the potential well-being effects from COVID-19–related research participation.

    Application of Superpopulation Model to Weighting of VicHealth Coronavirus Victorian Well-Being Impact Study

    VicHealth Coronavirus Victorian Wellbeing Impact Study (2020) and Victorian Health Promotion Foundation, Melbourne in collaboration with the Social Research Centre

    VicHealth undertook a survey to understand the impact on Victorians of the first coronavirus-related restrictions, which took place March–May 2020 and are now known as the first lockdown. A series of questions was asked covering a range of health and lifestyle areas to establish whether the lockdown changed people’s healthy lifestyles compared to life in February 2020 and to understand factors that may have influenced these changes.

    A survey of 2,000 Victorians was conducted via a nonprobability online panel. Data was calibrated with probability-based data collected by Life in Australia to overcome some of the biases associated with data collection via nonprobability panels using a super-population approach to derive weights for each respondent in the nonprobability sample. Population distributions for demographic characteristics were obtained from the Australian Bureau of Statistics, and those for lifestyle characteristics and key survey outcomes were obtained from Life in Australia.

    The final adopted solution reduced the average bias by more than 50 percent, compared to the unweighted solution, while still achieving an acceptable level of variability in the weights.

    1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
    Loading...

    Comments are closed.