National Health Interview Survey, COVID-19, and Online Data Collection Platforms: Adaptations, Tradeoffs, and New Directions

Comparative analysis of financial toxicity between SARS-CoV-2 infection and common comorbidities

Financial toxicity is common in individuals with COVID-19 and Long COVID. However, the extent of financial toxicity experienced, in comparison to other common comorbidities, is uncertain. Contributing factors exacerbating financial challenges in Long COVID are also unclear. These knowledge gaps are addressed via a cross-sectional analysis utilizing data from the 2022 National Health Interview Survey (NHIS), a representative sample drawn from the United States. COVID-19 cases were identified through self-reported positive testing or physician diagnoses. Long COVID was defined as experiencing COVID-19-related symptoms for more than three months. Comorbidity was assessed based on self-reported diagnoses of ten doctor-diagnosed conditions (Yes/No). Financial toxicity was defined as having difficulty paying medical bills, cost-related medication nonadherence, delaying healthcare due to cost, and/or not obtained healthcare due to cost. A total of 27,492 NHIS 2022 respondents were included in our analysis, representing 253 million U.S. adults. In multivariable logistic regression models, adults with Long COVID (excluding respondents with COVID-19 but not Long COVID), showed increased financial toxicity compared to those with other comorbidities, such as epilepsy (OR [95% CI]: 1.69 [1.22, 2.33]), dementia (1.51 [1.01, 2.25]), cancer (1.43 [1.19, 1.71]) or respiratory/cardiovascular conditions (1.18 [1.00, 1.40]/1.23 [1.02, 1.47]). Long COVID-related financial toxicity was associated with female sex, age <65 years, lack of medical insurance, current paid employment, residence region, food insecurity, fatigue, mild to severe depression symptoms experienced during the survey completion, visits to hospital emergency rooms, presence of arthritis, cardiovascular or respiratory conditions, and social activity limitations. In conclusion, American adults with Long COVID, but not those who had prior COVID-19 infection without Long COVID, exhibited a higher prevalence of financial toxicity compared to individuals with common comorbidities. Vulnerable populations were at greater risk for financial toxicity. These findings emphasize the importance of evaluating strategies to reduce economic burden and increase awareness of the effect of Long COVID-related financial toxicity on patient's healthcare and health status.

Adaptive Time-Location Sampling for COMPASS: A SARS-CoV-2 Prevalence Study in Fifteen Diverse Communities in the United States

BACKGROUND

COVID-19 has placed a disproportionate burden on underserved racial and ethnic groups, community members working in essential industries, those living in areas of high population density, and those reliant on in-person services such as transportation. The goal of this study was to estimate the cross-sectional prevalence of SARS-CoV-2 (active SARS-CoV-2 or prior SARS-CoV-2 infection) in children and adults attending public venues in 15 sociodemographically diverse communities in the United States and to develop a statistical design that could be rigorously implemented amidst unpredictable stay-at-home COVID-19 guidelines.

METHODS

We used time-location sampling with complex sampling involving stratification, clustering of units, and unequal probabilities of selection to recruit individuals from selected communities. We safely conducted informed consent, specimen collection, and face-to-face interviews outside of public venues immediately following recruitment.

RESULTS

We developed an innovative sampling design that adapted to constraints such as closure of venues, changing infection hotspots, and uncertain policies. We updated both the sampling frame and the selection probabilities over time using information acquired from prior weeks. We created site-specific survey weights that adjusted sampling probabilities for nonresponse and calibrated to county-level margins on age and sex at birth.

CONCLUSIONS

Although the study itself was specific to COVID-19, the strategies presented in this article could serve as a case study that can be adapted for performing population-level inferences in similar settings and could help inform rapid and effective responses to future global public health challenges.

Associations between COVID-19 incidence, weight status, and social participation restrictions in the U.S.: evidence from the national population, cross-sectional study

BACKGROUND

To explore the associations between coronavirus infection incidence and weight status and social participation restrictions among community-dwelling adults in the United States.

METHODS

We analyzed data from the 2021 National Health Interview Survey (NHIS), which included a representative sample of 29,394 individuals (Coronavirus disease 2019 (COVID-19): 3,205) and a weighted total of 252,461,316 individuals (COVID-19: 31,697,404), considering the complex sampling design used in the survey.

RESULTS

Age, race/ethnicity, education level, family income index, body mass index (BMI), and smoking status were significantly associated with COVID-19 infection. Weight status was significantly correlated with social participation restrictions and strongly associated with COVID-19 infection, particularly among individuals who were overweight or obese.

CONCLUSION

Weight status was shown to be associated not only with social participation restrictions but also with COVID-19 infection among U.S. adults. Understanding the complex interplay between weight status, social participation, and COVID-19 is crucial for developing effective preventive measures and promoting overall well-being in the community population.

Bias-Adjusted Predictions of County-Level Vaccination Coverage from the COVID-19 Trends and Impact Survey

BACKGROUND

The potential for selection bias in nonrepresentative, large-scale, low-cost survey data can limit their utility for population health measurement and public health decision making. We developed an approach to bias adjust county-level COVID-19 vaccination coverage predictions from the large-scale US COVID-19 Trends and Impact Survey.

DESIGN

We developed a multistep regression framework to adjust for selection bias in predicted county-level vaccination coverage plateaus. Our approach included poststratification to the American Community Survey, adjusting for differences in observed covariates, and secondary normalization to an unbiased reference indicator. As a case study, we prospectively applied this framework to predict county-level long-run vaccination coverage among children ages 5 to 11 y. We evaluated our approach against an interim observed measure of 3-mo coverage for children ages 5 to 11 y and used long-term coverage estimates to monitor equity in the pace of vaccination scale up.

RESULTS

Our predictions suggested a low ceiling on long-term national vaccination coverage (46%), detected substantial geographic heterogeneity (ranging from 11% to 91% across counties in the United States), and highlighted widespread disparities in the pace of scale up in the 3 mo following Emergency Use Authorization of COVID-19 vaccination for 5- to 11-y-olds.

LIMITATIONS

We relied on historical relationships between vaccination hesitancy and observed coverage, which may not capture rapid changes in the COVID-19 policy and epidemiologic landscape.

CONCLUSIONS

Our analysis demonstrates an approach to leverage differing strengths of multiple sources of information to produce estimates on the time scale and geographic scale necessary for proactive decision making.

IMPLICATIONS

Designing integrated health measurement systems that combine sources with different advantages across the spectrum of timeliness, spatial resolution, and representativeness can maximize the benefits of data collection relative to costs.

HIGHLIGHTS

The COVID-19 pandemic catalyzed massive survey data collection efforts that prioritized timeliness and sample size over population representativeness.The potential for selection bias in these large-scale, low-cost, nonrepresentative data has led to questions about their utility for population health measurement.We developed a multistep regression framework to bias adjust county-level vaccination coverage predictions from the largest public health survey conducted in the United States to date: the US COVID-19 Trends and Impact Survey.Our study demonstrates the value of leveraging differing strengths of multiple data sources to generate estimates on the time scale and geographic scale necessary for proactive public health decision making.

Prevalence of SARS-CoV-2 Infection among Children and Adults in 15 US Communities, 2021

During January-August 2021, the Community Prevalence of SARS-CoV-2 Study used time/location sampling to recruit a cross-sectional, population-based cohort to estimate SARS-CoV-2 seroprevalence and nasal swab sample PCR positivity across 15 US communities. Survey-weighted estimates of SARS-CoV-2 infection and vaccine willingness among participants at each site were compared within demographic groups by using linear regression models with inverse variance weighting. Among 22,284 persons >2 months of age and older, median prevalence of infection (prior, active, or both) was 12.9% across sites and similar across age groups. Within each site, average prevalence of infection was 3 percentage points higher for Black than White persons and average vaccine willingness was 10 percentage points lower for Black than White persons and 7 percentage points lower for Black persons than for persons in other racial groups. The higher prevalence of SARS-CoV-2 infection among groups with lower vaccine willingness highlights the disparate effect of COVID-19 and its complications.

Exploring preventive care practices among unvaccinated individuals in the United States during the COVID-19 pandemic

INTRODUCTION

Building on a Canadian study associating unvaccinated individuals to increased car accidents, we examined the relationship between COVID-19 vaccination status and US preventive care practices.

METHODS

We queried the 2021 National Health Interview Survey. First, we fitted a model to identify respondent-level factors associated with receipt of at least one COVID-19 vaccination. Second, we fitted a survey-weighted logistic regression model adjusted for respondent-level characteristics to examine whether the receipt of at least one COVID-19 vaccination predicted the receipt of preventive care services. Preventive care services assessed included serum cholesterol, glucose, and blood pressure measurements, as well as guideline-concordant cancer screening including breast, cervical, colorectal, and prostate cancer screening.

RESULTS

Factors predicting receipt of COVID-19 vaccination were age (adjusted Odds Ratio (aOR) 1.03; 95 % confidence interval (CI) [1.03-1.03]), Hispanic (aOR 1.25; 95 % CI [1.08-1.44]), and non-Hispanic Asian (aOR 3.52; 95 % CI [2.74-4.52]) ethnicity/race, and history of cancer (aOR 1.61; 95 % CI [1.13-2.30]). Unvaccinated respondents were less likely to have received serum cholesterol (aOR 0.69; 95 % CI [0.50-0.70), serum glucose (aOR 0.65; 95 % CI [0.56-0.75]), or blood pressure measurements (aOR 0.47; 95 % CI [0.33-0.66]); and were less likely to have received breast cancer (aOR 0.35; 95 % CI [0.25-0.48]), colorectal cancer (aOR 0.52; 95 % CI [0.46-0.60]) and prostate cancer screening (aOR 0.61; 95 % CI [0.48-0.76]). There was no significant association between unvaccinated respondents receiving cervical cancer screening (aOR 0.96; 95 % CI [0.81-1.13]; p = 0.616).

CONCLUSION

Non-receipt of COVID-19 vaccination was associated with non-receipt of preventive care services including cancer screening. Further studies are needed to assess if this association is due to system-level factors or reflects a general distrust of medical preventive care amongst this population.

Depression, Sleep Health & Sociodemographic Correlates in a Nationwide Survey: Implications for Depression Treatment During the COVID-19

Purpose

The COVID-19 pandemic has had a profound impact on mental health worldwide, with depression and sleep problems among the most common issues experienced by many individuals. Depression can lead to sleep problems, which can increase the risk of developing depressive symptoms. However, it is unclear which United States (US) sub-population was most affected by depression and sleep problems during the pandemic.

Methods

We conducted a secondary analysis using self-reported data from the 2021 National Health Interview Survey (NHIS), focusing on adults aged 18 years and above (n=29,763). We utilized self-reported responses to questions about prescription medication and frequency of depressive feelings to determine participants' depression status. Appropriate weights were applied to account for the sampling design of the surveys. Our analysis involved descriptive statistics and chi-squared tests to compare sociodemographic, clinical, behavioral, and sleep-related characteristics between US adults with and without depression. Additionally, logistic regression was used to examine the associations between sleep duration, sleep quality and depression.

Results

The overall prevalence of depression in our sample was 44.4%. It were higher in certain demographic groups, including younger adults (18-39 years, 47.7%), non-Hispanic whites (47.9%), females (50.1%), those at the lower income bracket (52.2%), those with no college or degree (48.7%) uninsured individuals (45.2%), and those reporting poor general health (71.9%). Individuals with depression had a 12% increased odds of experiencing short sleep (aOR: 1.12, 95% CI:1.04-1.20, p<0.001), 34% increased odds of experiencing long sleep (aOR: 1.34, 95% CI: 1.20-1.50, p < 0.001) and more than 2.5 fold increased odds of reporting poor sleep quality (aOR:2.57, 95% CI: 2.40-2.78; p<0.0001). In the multivariate analysis, all variables (sex, race/ethnicity, education, health insurance coverage, marital status, general health status and use of sleep medications, smoking and alcohol use status) were significantly predictors of poor sleep quality, with the exceptions of age and family income.

Conclusion

The findings emphasize the need to address sleep health in treating depression, especially during times of public health crises.

Artificial intelligence for healthcare and medical education: a systematic review

BACKGROUND

Human society has entered the age of artificial intelligence, medical practice and medical education are undergoing profound changes. Artificial intelligence (AI) is now applied in many industries, particularly in healthcare and medical education, where it deeply intersects. The purpose of this paper is to overview the current situation and problems of "AI+medicine/medical" education and to provide our own perspective on the current predicament.

METHODS

We searched PubMed, Embase, Cochrane and CNKI databases to assess the literature on AI+medical/medical education from 2017 to July 2022. The main inclusion criteria include literature describing the current situation or predicament of "AI+medical/medical education".

RESULTS

Studies have shown that the current application of AI in medical education is focused on clinical specialty training and continuing education, with the main application areas being radiology, diagnostics, surgery, cardiology, and dentistry. The main role is to assist physicians to improve their efficiency and accuracy. In addition, the field of combining AI with medicine/medical education is steadily expanding, and the most urgent need is for policy makers, experts in the medical field, AI and education, and experts in other fields to come together to reach consensus on ethical issues and develop regulatory standards. Our study also found that most medical students are positive about adding AI-related courses to the existing medical curriculum. Finally, the quality of research on "AI+medical/medical education" is poor.

CONCLUSION

In the context of the COVID-19 pandemic, our study provides an innovative systematic review of the latest "AI+medicine/medical curriculum". Since the AI+medicine curriculum is not yet regulated, we have made some suggestions.

Overview of the 2019 National Health Interview Survey Questionnaire Redesign

Data System. Federal health surveys, like the National Health Interview Survey (NHIS), represent important surveillance mechanisms for collecting timely, representative data that can be used to monitor the health and health care of the US population. Data Collection/Processing. Conducted by the National Center for Health Statistics (NCHS), NHIS uses an address-based, complex clustered sample of housing units, yielding data representative of the civilian noninstitutionalized US population. Survey redesigns that reduce survey length and eliminate proxy reporting may reduce respondent burden and increase participation. Such were goals in 2019, when NCHS implemented a redesigned NHIS questionnaire that also focused on topics most relevant and appropriate for surveillance of child and adult health. Data Analysis/Dissemination. Public-use microdata files and selected health estimates and detailed documentation are released online annually. Public Health Implications. Declining response rates may lead to biased estimates and weaken users' ability to make valid conclusions from the data, hindering public health efforts. The 2019 NHIS questionnaire redesign was associated with improvements in the survey's response rate, declines in respondent burden, and increases in data quality and survey relevancy. (Am J Public Health. 2023;113(4):408-415. https://doi.org/10.2105/AJPH.2022.307197).

Association between Cardiovascular Risk and Coronavirus Disease 2019: Findings from 2021 National Health Interview Survey

PURPOSE

To examine the association between pre-existing cardiovascular disorders and the risk of coronavirus disease 2019 (COVID-19) among community-dwelling adults in the United States (US).

METHODS

We analyzed data from the 2021 National Health Interview Survey (NHIS), encompassing 28,848 nationally representative participants aged ≥18. We examined the association by two age groups, younger adults (aged 18-59) and older adults (aged>60). Weighted analyses were conducted to consider the complex sampling design used in the NHIS.

RESULTS

Our results show that 13.9% of younger and 8.2% of older adults were infected with coronavirus, corresponding to a nationwide estimate of 23,701,358 COVID-19 cases in younger adults and 6,310,206 in older adults in 2021. Subjects who lived in the South region of the US had the highest COVID-19 rate (13.4%), followed by the Midwest (12.6%), West (10.9%), and Northeast (10.4%). Pre-existing cardiovascular risk factors (overweight, obesity, hypertension, and diabetes) were significantly associated with increased risk for COVID-19 infection in younger and older adults. Pre-existing cardiovascular diseases (angina, heart attack, and coronary heart disease) were significantly associated with COVID-19 in older adults but not significantly in younger adults. Significant dose-response relationships existed between the increased number of pre-existing cardiovascular risk factors and COVID-19 infection, with the strongest association in non-Hispanic (NH) Black and Hispanic ethnicities compared to NH White.

CONCLUSION

Pre-existing cardiovascular disorders are significantly associated with the risk of COVID-19 infection. The magnitudes of this risk association are stronger among the minority populations than NH White. Further studies are needed to determine the long-term impact of COVID-19 infection and its relationship to pre-existing cardiovascular disorders.

ADAPTIVE TIME LOCATION SAMPLING FOR COMPASS, A SARS-COV-2 PREVALENCE STUDY IN FIFTEEN DIVERSE COMMUNITIES IN THE UNITED STATES

The COVPN 5002 (COMPASS) study aimed to estimate the prevalence of SARS-CoV-2 (active SARS-CoV-2 or prior SARS-CoV-2 infection) in children and adults attending public venues in 15 socio-demographically diverse communities in the United States. To protect against potential challenges in implementing traditional sampling strategies, time-location sampling (TLS) using complex sampling involving stratification, clustering of units, and unequal probabilities of selection was used to recruit individuals from neighborhoods in selected communities. The innovative design adapted to constraints such as closure of venues; changing infection hotspots; and uncertain policies. Recruitment of children and the elderly raised additional challenges in sample selection and implementation. To address these challenges, the TLS design adaptively updated both the sampling frame and the selection probabilities over time using information acquired from prior weeks. Although the study itself was specific to COVID-19, the strategies presented in this paper could serve as a case study that can be adapted for performing rigorous population-level inferences in similar settings and could help inform rapid and effective responses to future global public health challenges.

Uptake and Equity in Influenza Vaccination Among Veterans with VA Coverage, Veterans Without VA Coverage, and Non-Veterans in the USA, 2019-2020

BACKGROUND

Vaccination is a primary method of reducing the burden of influenza, yet uptake is neither optimal nor equitable. Single-tier, primary care-oriented health systems may have an advantage in the efficiency and equity of vaccination.

OBJECTIVE

To assess the association of Veterans' Health Administration (VA) coverage with influenza vaccine uptake and disparities.

DESIGN

Cross-sectional.

PARTICIPANTS

Adult respondents to the 2019-2020 National Health Interview Survey.

MAIN MEASURES

We examined influenza vaccination rates, and racial/ethnic and income-based vaccination disparities, among veterans with VA coverage, veterans without VA coverage, and adult non-veterans. We performed multivariable logistic regressions adjusted for demographics and self-reported health, with interaction terms to examine differential effects by race/ethnicity and income.

KEY RESULTS

Our sample included n=2,277 veterans with VA coverage, n=2,821 veterans without VA coverage, and n=46,456 non-veterans. Veterans were more often White and male; among veterans, those with VA coverage had worse health and lower incomes. Veterans with VA coverage had a higher unadjusted vaccination rate (63.0%) than veterans without VA coverage (59.1%) and non-veterans (46.5%) (p<0.05 for each comparison). In our adjusted model, non-veterans were 11.4 percentage points (95% CI -14.3, -8.5) less likely than veterans with VA coverage to be vaccinated, and veterans without VA coverage were 6.7 percentage points (95% CI -10.3, -3.0) less likely to be vaccinated than those with VA coverage. VA coverage, compared with non-veteran status, was also associated with reduced racial/ethnic and income disparities in vaccination.

CONCLUSIONS

VA coverage is associated with higher and more equitable influenza vaccination rates. A single-tier health system that emphasizes primary care may improve the uptake and equity of vaccination for influenza, and possibly other pathogens, like SARS-CoV2.

Anxiety and depression symptoms among children before and during the COVID-19 pandemic

PURPOSE

The COVID-19 pandemic caused disruptions to children's daily lives due in part to stay-at-home orders and school closures, reducing interactions with both peers and extended family. Yet, few studies with nationally representative data have explored the potential association of the COVID-19 pandemic and children's mental health.

METHODS

The current study analyzed data from the 2019 and 2020 National Health Interview Survey (NHIS) to describe changes in the prevalence of symptoms of anxiety and depression before and during the first year of the pandemic among children aged 5-17 years. Changes in prevalence by child- and family-level characteristics were also examined.

RESULTS

During the COVID-19 pandemic, nearly one in six children aged 5-17 years had daily or weekly symptoms of anxiety or depression, a significant increase from before the COVID-pandemic (16.7% (95% CI:15.0-18.6) versus 14.4% (95% CI:13.4-15.3)). Males, children 5-11 years, non-Hispanic children, children living in families in large metropolitan areas, incomes at or below the federal poverty level, and whose highest educated parent had more than a HS education, also showed statistically significant increases in anxiety and depression symptoms.

CONCLUSIONS

NHIS data may be used to monitor this increase in mental health symptomatology and assist in identifying children at risk.

Nativity Status and Poly Tobacco Use among Young Adults in the United States

Young adults are the second largest segment of the immigrant population in the United States (US). Given recent trends in later age of initiation of tobacco use, we examined variation in use of tobacco products by nativity status for this population group. Methods: Our study included young adults 18–30 years of age sampled in the National Health Interview Survey (2015–2019), a nationally representative sample of the US population. We calculated prevalence of use of any and two or more tobacco products (cigarettes, cigars, pipes, e-cigarettes, and smokeless tobacco) for foreign-born (n = 3096) and US-born (n = 6811) young adults. Logistic regression models were adjusted for age, sex, race-ethnicity, education, and poverty, while accounting for the complex survey design. Results: Foreign-born young adults were significantly less likely to use any tobacco product (Cigarette = 7.3% vs. 10.7%; Cigar = 1.8% vs. 4.8%; E-cigarette = 2.3% vs. 4.5%, respectively; p < 0.01) or poly tobacco use (1.9% vs. 4.2%; p < 0.01) than US-born young adults. Adjusted regression models showed lower odds of poly tobacco use among the foreign-born than their US-born counterparts (Odds Ratio = 0.41, (95% Confidence Interval: 0.26–0.63)). Conclusions: The findings highlight the importance of targeted interventions by nativity status and further tobacco prevention efforts needed for the US-born.