#4Corners4Health Social Media Cancer Prevention Campaign for Emerging Adults: Protocol for a Randomized Stepped-Wedge Trial

BACKGROUND

Many emerging adults (EAs) are prone to making unhealthy choices, which increase their risk of premature cancer morbidity and mortality. In the era of social media, rigorous research on interventions to promote health behaviors for cancer risk reduction among EAs delivered over social media is limited. Cancer prevention information and recommendations may reach EAs more effectively over social media than in settings such as health care, schools, and workplaces, particularly for EAs residing in rural areas.

OBJECTIVE

This pragmatic randomized trial aims to evaluate a multirisk factor intervention using a social media campaign designed with community advisers aimed at decreasing cancer risk factors among EAs. The trial will target EAs from diverse backgrounds living in rural counties in the Four Corners states of Arizona, Colorado, New Mexico, and Utah.

METHODS

We will recruit a sample of EAs (n=1000) aged 18 to 26 years residing in rural counties (Rural-Urban Continuum Codes 4 to 9) in the Four Corners states from the Qualtrics' research panel and enroll them in a randomized stepped-wedge, quasi-experimental design. The inclusion criteria include English proficiency and regular social media engagement. A social media intervention will promote guideline-related goals for increased physical activity, healthy eating, and human papillomavirus vaccination and reduced nicotine product use, alcohol intake, and solar UV radiation exposure. Campaign posts will cover digital and media literacy skills, responses to misinformation, communication with family and friends, and referral to community resources. The intervention will be delivered over 12 months in Facebook private groups and will be guided by advisory groups of community stakeholders and EAs and focus groups with EAs. The EAs will complete assessments at baseline and at 12, 26, 39, 52, and 104 weeks after randomization. Assessments will measure 6 cancer risk behaviors, theoretical mediators, and participants' engagement with the social media campaign.

RESULTS

The trial is in its start-up phase. It is being led by a steering committee. Team members are working in 3 subcommittees to optimize community engagement, the social media intervention, and the measures to be used. The Stakeholder Organization Advisory Board and Emerging Adult Advisory Board were formed and provided initial input on the priority of cancer risk factors to target, social media use by EAs, and community resources available. A framework for the social media campaign with topics, format, and theoretical mediators has been created, along with protocols for campaign management.

CONCLUSIONS

Social media can be used as a platform to counter misinformation and improve reliable health information to promote health behaviors that reduce cancer risks among EAs. Because of the popularity of web-based information sources among EAs, an innovative, multirisk factor intervention using a social media campaign has the potential to reduce their cancer risk behaviors.

TRIAL REGISTRATION

ClinicalTrials.gov NCT05618158; https://classic.clinicaltrials.gov/ct2/show/NCT05618158.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

PRR1-10.2196/50392.

Discriminative machine learning for maximal representative subsampling

Biased population samples pose a prevalent problem in the social sciences. Therefore, we present two novel methods that are based on positive-unlabeled learning to mitigate bias. Both methods leverage auxiliary information from a representative data set and train machine learning classifiers to determine the sample weights. The first method, named maximum representative subsampling (MRS), uses a classifier to iteratively remove instances, by assigning a sample weight of 0, from the biased data set until it aligns with the representative one. The second method is a variant of MRS - Soft-MRS - that iteratively adapts sample weights instead of removing samples completely. To assess the effectiveness of our approach, we induced artificial bias in a public census data set and examined the corrected estimates. We compare the performance of our methods against existing techniques, evaluating the ability of sample weights created with Soft-MRS or MRS to minimize differences and improve downstream classification tasks. Lastly, we demonstrate the applicability of the proposed methods in a real-world study of resilience research, exploring the influence of resilience on voting behavior. Through our work, we address the issue of bias in social science, amongst others, and provide a versatile methodology for bias reduction based on machine learning. Based on our experiments, we recommend to use MRS for downstream classification tasks and Soft-MRS for downstream tasks where the relative bias of the dependent variable is relevant.

Scoping review of the methodology of large health surveys conducted in Spain early on in the COVID-19 pandemic

Background

The use of health surveys has been key in the scientific community to promptly communicate results about the health impact of COVID-19. But what information was collected, where, when and how, and who was the study population?

Objective

To describe the methodological characteristics used in large health surveys conducted in Spain early on in the COVID-19 pandemic.

Methods

Scoping review. Inclusion criteria: observational studies published between January 2020 and December 2021, with sample sizes of over 2,000 persons resident in Spain. Databases consulted: PubMed, CINAHL, Literatura Latinoamericana y del Caribe en CC de la Salud, Scopus, PsycINFO, Embase, Sociological Abstracts, Dialnet and Web of Science Core Collection. We analyzed the characteristics of the literature references, methodologies and information gathered in the surveys selected. Fifty five studies were included.

Results

Sixty percentage of the studies included had mental health as their main topic and 75% were conducted on the general adult population. Thirteen percentage had a longitudinal design, 93% used the internet to gather information and the same percentage used non-probability sampling. Thirty percentage made some type of sampling correction to reduce coverage or non-response biases, but not selection biases. Sixty seven percentage did not state the availability of their data.

Conclusions

Consistent with the extensive use of non-probability sampling without any bias correction in the extraordinary setting created by COVID-19, quality population frameworks are required so that probability and representative samples can be extracted quickly to promptly address other health crises, as well as to reduce potential coverage, non-response and particularly selection biases by utilizing reweighting techniques. The low data accessibility despite the huge opportunity that COVID-19 provided for Open Science-based research is striking.

Participation bias in the UK Biobank distorts genetic associations and downstream analyses

While volunteer-based studies such as the UK Biobank have become the cornerstone of genetic epidemiology, the participating individuals are rarely representative of their target population. To evaluate the impact of selective participation, here we derived UK Biobank participation probabilities on the basis of 14 variables harmonized across the UK Biobank and a representative sample. We then conducted weighted genome-wide association analyses on 19 traits. Comparing the output from weighted genome-wide association analyses (neffective = 94,643 to 102,215) with that from standard genome-wide association analyses (n = 263,464 to 283,749), we found that increasing representativeness led to changes in SNP effect sizes and identified novel SNP associations for 12 traits. While heritability estimates were less impacted by weighting (maximum change in h2, 5%), we found substantial discrepancies for genetic correlations (maximum change in rg, 0.31) and Mendelian randomization estimates (maximum change in βSTD, 0.15) for socio-behavioural traits. We urge the field to increase representativeness in biobank samples, especially when studying genetic correlates of behaviour, lifestyles and social outcomes.

Weight smoothing for nonprobability surveys

Adjustment techniques to mitigate selection bias in nonprobability samples often involve modelling the propensity to participate in the nonprobability sample along with inverse propensity weighting. It is well known that procedures for estimating weights are effective if the covariates selected in the propensity model are related to both the variable of interest and the participation indicator. In most surveys, there are many variables of interest, making weight adjustments difficult to determine as a suitable weight for one variable may be unsuitable for other variables. The standard compromise is to include a large number of covariates in the propensity model but this may increase the variability of the estimates, especially when some covariates are weakly related to the variables of interest. Weight smoothing, developed for probability surveys, could be helpful in these situations. It aims to remove the variability caused by overfit propensity models by replacing the inverse propensity weights with predicted weights obtained using a smoothing model. In this article, we study weight smoothing in the nonprobability survey context, both theoretically and empirically, to understand its effectiveness at improving the efficiency of estimates.

Use of Random Domain Intercept Technology to Track COVID-19 Vaccination Rates in Real-Time Across the United States: Survey Study

Background

Accurate and timely COVID-19 vaccination coverage data are vital for informing targeted, effective messaging and outreach and identifying barriers to equitable health service access. However, gathering vaccination rate data is challenging, and efforts often result in information that is either limited in scope (eg, limited to administrative data) or delayed (impeding the ability to rapidly respond). The evaluation of innovative technologies and approaches that can assist in addressing these limitations globally are needed.

Objective

The objective of this survey study was to assess the validity of Random Domain Intercept Technology (RDIT; RIWI Corp) for tracking self-reported vaccination rates in real time at the US national and state levels. RDIT—a form of online intercept sampling—has the potential to address the limitations of current vaccination tracking systems by allowing for the measurement of additional data (eg, attitudinal data) and real-time, rapid data collection anywhere there is web access.

Methods

We used RDIT from June 30 to July 26, 2021, to reach a broad sample of US adult (aged ≥18 years) web users and asked questions related to COVID-19 vaccination. Self-reported vaccination status was used as the focus of this validation exercise. National- and state-level RDIT-based vaccination rates were compared to Centers for Disease Control and Prevention (CDC)–reported national and state vaccination rates. Johns Hopkins University’s and Emory University’s institutional review boards designated this project as public health practice to inform message development (not human subjects research).

Results

By using RDIT, 63,853 adult web users reported their vaccination status (6.2% of the entire 1,026,850 American web-using population that was exposed to the survey). At the national level, the RDIT-based estimate of adult COVID-19 vaccine coverage was slightly higher (44,524/63,853, 69.7%; 95% CI 69.4%-70.1%) than the CDC-reported estimate (67.9%) on July 15, 2021 (ie, midway through data collection; t_63,852=10.06; P<.001). The RDIT-based and CDC-reported state-level estimates were strongly and positively correlated (r=0.90; P<.001). RDIT-based estimates were within 5 percentage points of the CDC’s estimates for 29 states.

Conclusions

This broad-reaching, real-time data stream may provide unique advantages for tracking the use of a range of vaccines and for the timely evaluation of vaccination interventions. Moreover, RDIT could be harnessed to rapidly assess demographic, attitudinal, and behavioral constructs that are not available in administrative data, which could allow for deeper insights into the real-time predictors of vaccine uptake–enabling targeted and timely interventions.

Building Infrastructure for Surveillance of Adverse and Positive Childhood Experiences: Integrated, Multimethod Approaches to Generate Data for Prevention Action

Adverse and positive childhood experiences have a profound impact on lifespan health and well-being. However, their incorporation into ongoing population-based surveillance systems has been limited. This paper outlines critical steps in building a comprehensive approach to adverse and positive childhood experiences surveillance, provides examples from the Preventing Adverse Childhood Experiences: Data to Action cooperative agreement, and describes improvements needed to optimize surveillance data for action. Components of a comprehensive approach to adverse and positive childhood experiences surveillance include revisiting definitions and measurement, including generating and using uniform definitions for adverse and positive childhood experiences across data collection efforts; conducting youth-based surveillance of adverse and positive childhood experiences; using innovative methods to gather and analyze near real-time data; leveraging available data, including from administrative sources; and integrating data on community- and societal-level risk and protective factors for adverse childhood experiences, including social and health inequities such as racism and poverty, as well as policies and conditions that create healthy environments for children and families. Comprehensive surveillance data on adverse and positive childhood experiences can inform data-driven prevention and intervention efforts, including focusing prevention programming and services to populations in greatest need. Data can be used to evaluate progress in reducing the occurrence of adverse childhood experiences and bolstering the occurrence of positive childhood experiences. Through expansion and improvement in adverse and positive childhood experiences surveillance-including at federal, state, territorial, tribal, and local levels-data-driven action can reduce children's exposure to violence and other adversities and improve lifelong health and well-being.

What is the adult experience of Perthes' disease? : initial findings from an international web-based survey

AIMS

Perthes' disease is an uncommon hip disorder with limited data on the long-term outcomes in adulthood. We partnered with community-based foundations and utilized web-based survey methodology to develop the Adult Perthes Survey, which includes demographics, childhood and adult Perthes' disease history, the University of California Los Angeles (UCLA) Activity Scale item, Short Form-36, the Hip disability and Osteoarthritis Outcome Score, and a body pain diagram. Here we investigate the following questions: 1) what is the feasibility of obtaining > 1,000 survey responses from adults who had Perthes' disease using a web-based platform?; and 2) what are the baseline characteristics and demographic composition of our sample?

METHODS

The survey link was available publicly for 15 months and advertised among support groups. Of 1,505 participants who attempted the Adult Perthes survey, 1,182 completed it with a median timeframe of 11 minutes (IQR 8.633 to 14.72). Participants who dropped out were similar to those who completed the survey on several fixed variables. Participants represented 45 countries including the USA (n = 570; 48%), UK (n = 295; 25%), Australia (n = 133; 11%), and Canada (n = 46; 4%). Of the 1,182 respondents, 58% were female and the mean age was 39 years (SD 12.6).

RESULTS

Ages at onset of Perthes' disease were < six years (n = 512; 43%), six to seven years (n = 321; 27%), eight to 11 years (n = 261; 22%), and > 11 years (n = 76; 6%), similar to the known age distribution of Perthes' disease. During childhood, 40% (n = 476) of respondents had at least one surgery. Bracing, weightbearing restriction, and absence of any treatment varied significantly between USA and non-USA respondents (p < 0.001, p = 0.002, and p < 0.001, respectively). As adults, 22% (n = 261) had at least one total hip arthroplasty, and 30% (n = 347) had any type of surgery; both more commonly reported among women (p = 0.002).

CONCLUSION

While there are limitations due to self-sampling, our study shows the feasibility of obtaining a large set of patient-reported data from adults who had childhood Perthes' from multiple countries. Cite this article: Bone Jt Open 2022;3(5):404-414.

Masks, money, and mandates: A national survey on efforts to increase COVID-19 vaccination intentions in the United States

Various efforts to increase COVID-19 vaccination rates have been employed in the United States. We sought to rapidly investigate public reactions to these efforts to increase vaccination, including self-reported responses to widespread reduced masking behavior, monetary incentive programs to get vaccinated, and work vaccination requirements. Using a unique method for data collection (Random Domain Intercept Technology), we captured a large (N = 14,152), broad-based sample of the United States Web-using population (data collected from June 30 -July 26, 2021). About 3/4 of respondents reported being vaccinated. The likelihood of vaccination and vaccination intention differed across various demographic indicators (e.g., gender, age, income, political leaning). We observed mixed reactions to efforts aimed at increasing vaccination rates among unvaccinated respondents. While some reported that specific efforts would increase their likelihood of getting vaccinated (between 16% and 32%), others reported that efforts would decrease their likelihood of getting vaccinated (between 17% and 42%). Reactions differed by general vaccination intention, as well as other demographic indicators (e.g., race, education). Our results highlight the need to fully understand reactions to policy changes, programs, and mandates before they are communicated to the public and employed. Moreover, the results emphasize the importance of understanding how reactions differ across groups, as this information can assist in targeting intervention efforts and minimizing potentially differential negative impact.

Variable selection in Propensity Score Adjustment to mitigate selection bias in online surveys

The development of new survey data collection methods such as online surveys has been particularly advantageous for social studies in terms of reduced costs, immediacy and enhanced questionnaire possibilities. However, many such methods are strongly affected by selection bias, leading to unreliable estimates. Calibration and Propensity Score Adjustment (PSA) have been proposed as methods to remove selection bias in online nonprobability surveys. Calibration requires population totals to be known for the auxiliary variables used in the procedure, while PSA estimates the volunteering propensity of an individual using predictive modelling. The variables included in these models must be carefully selected in order to maximise the accuracy of the final estimates. This study presents an application, using synthetic and real data, of variable selection techniques developed for knowledge discovery in data to choose the best subset of variables for propensity estimation. We also compare the performance of PSA using different classification algorithms, after which calibration is applied. We also present an application of this methodology in a real-world situation, using it to obtain estimates of population parameters. The results obtained show that variable selection using appropriate methods can provide less biased and more efficient estimates than using all available covariates.

What's to Like? Facebook as a Tool for Survey Data Collection

In this paper, we explore the use of Facebook targeted advertisements for the collection of survey data. We illustrate the potential of survey sampling and recruitment on Facebook through the example of building a large employee-employer linked dataset as part of The Shift Project. We describe the workflow process of targeting, creating, and purchasing survey recruitment advertisements on Facebook. We address concerns about sample selectivity, and apply post-stratification weighting techniques to adjust for differences between our sample and that of "gold-standard" data sources. We then compare univariate and multi-variate relationships in the Shift data against the Current Population Survey and the National Longitudinal Survey of Youth-1997. Finally, we provide an example of the utility of the firm-level nature of the data by showing how firm-level gender composition is related to wages. We conclude by discussing some important remaining limitations of the Facebook approach, as well as highlighting some unique strengths of the Facebook targeting advertisement approach, including the ability for rapid data collection in response to research opportunities, rich and flexible sample targeting capabilities, and low cost, and we suggest broader applications of this technique.

Immediate and informative feedback during a pandemic: Using stated preference analysis to predict vaccine uptake rates

In response to an emerging pandemic, there is urgent need for information regarding individual evaluation of risk and preferences toward mitigation strategies such as vaccinations. However, with social distancing policies and financial stress during an outbreak, traditional robust survey methodologies of face-to-face, probabilistic sampling, may not be feasible to deploy quickly, especially in developing countries. We recommend a protocol that calls for a sensitive survey design, acceptance of a web-based approach and adjustments for uncertainty of respondents, to deliver urgently needed information to policymakers as the public health crisis unfolds, rather than in its aftermath. This information is critical to tailor comprehensive vaccination campaigns that reach critical immunity thresholds. We apply our recommendations in a regional study of 16 Latin American countries in the month following index cases of COVID-19. We use a split-sample, contingent valuation approach to evaluate the effects of cost, duration of immunity and effectiveness of the vaccine. Our results show that cost and duration of immunity are significant factors in the decision to vaccinate, while the degree of effectiveness is insignificant, unless the vaccine is 100% effective. Income as well as perceived risk and severity of the virus are important determinants also.

Ensuring survey research data integrity in the era of internet bots

We used an internet-based survey platform to conduct a cross-sectional survey regarding the impact of COVID-19 on the LGBTQ + population in the United States. While this method of data collection was quick and inexpensive, the data collected required extensive cleaning due to the infiltration of bots. Based on this experience, we provide recommendations for ensuring data integrity. Recruitment conducted between May 7 and 8, 2020 resulted in an initial sample of 1251 responses. The Qualtrics survey was disseminated via social media and professional association listservs. After noticing data discrepancies, research staff developed a rigorous data cleaning protocol. A second wave of recruitment was conducted on June 11–12, 2020 using the original recruitment methods. The five-step data cleaning protocol led to the removal of 773 (61.8%) surveys from the initial dataset, resulting in a sample of 478 participants in the first wave of data collection. The protocol led to the removal of 46 (31.9%) surveys from the second two-day wave of data collection, resulting in a sample of 98 participants in the second wave of data collection. After verifying the two-day pilot process was effective at screening for bots, the survey was reopened for a third wave of data collection resulting in a total of 709 responses, which were identified as an additional 514 (72.5%) valid participants and led to the removal of an additional 194 (27.4%) possible bots. The final analytic sample consists of 1090 participants. Although a useful and efficient research tool, especially among hard-to-reach populations, internet-based research is vulnerable to bots and mischievous responders, despite survey platforms’ built-in protections. Beyond the depletion of research funds, bot infiltration threatens data integrity and may disproportionately harm research with marginalized populations. Based on our experience, we recommend the use of strategies such as qualitative questions, duplicate demographic questions, and incentive raffles to reduce likelihood of mischievous respondents. These protections can be undertaken to ensure data integrity and facilitate research on vulnerable populations.

An Empirical Approach to Differences in Flexible Electricity Consumption Behaviour of Urban and Rural Populations—Lessons Learned in Germany

This article analyses two major trends of the 21st century. Firstly, the transition from fossil fuel-based energy production to renewable energy sources. Secondly, the inexorable urbanisation which can be witnessed all over the globe. The most promising renewable energy production technologies for the near future, i.e., wind and solar energy, are volatile by nature which makes matching supply and demand essential for a successful transition. Therefore, the aspects that determine the willingness of consumers to flexibilise their demand has gained growing attention. Initial research shows that different settings for (co-)ownership in terms of available prosumption options and used production technologies have a varying impact on demand flexibility. However, existing research has analysed flexibility drivers solely for the general population as an aggregate without any distinction regarding spatial, economic, or social factors. In this article, the authors go one step further and analyse whether those drivers for flexible consumption behaviour differ in rural or urban areas acknowledging differences in day-to-day life in both cases. This study is based on 2074 completed questionnaires from German consumers which were analysed using propensity score matching. The results show that people from rural and urban areas do not significantly differ in their willingness to be demand flexible in general. However, (co-)owners of RE installations from rural areas are generally significantly more demand flexible than (co-)owners of RE installations from urban areas. Further, when looking at different RE technologies, the results show that (co-)owners of solar installations are significantly more demand flexible if they are from rural areas. Lastly, when looking at usage options, people who solely consume produced electricity are more demand flexible if they are from rural areas as well.

Depressive symptoms and perception of risk during the first wave of the COVID-19 pandemic: A web-based cross-country comparative survey

Evidence is accumulating of the negative impact of the COVID-19 pandemic and related public health measures on mental health. In this emergent field, there has been little research into the role of risk perception on depressive symptoms and the contribution of health-care resources to model risk perception and mental health. The aim of this paper is to describe the relationship between individual-level perception of risk and depression, controlling for a set of confounders and for country-level heterogeneity. A cross-sectional and observational online survey was conducted using a non-probability snowball sampling technique. We use data on 11,340 respondents, living in six European countries (Italy, Sweden, United Kingdom, France, Poland, Czech Republic) who completed survey questionnaires during the first months of the pandemic. We used a fixed-effect approach, which included individual and macro-level variables. The findings suggest that a high proportion of people suffering from depression and heightened risk perception is positively associated with reporting depressive symptoms, even if this relationship varies significantly between countries. Moreover, the association is moderated by contextual factors including health-care expenditure as a percentage of Gross Domestic Product, hospital beds for acute care, and number of medical specialists per head of population. Investment in health care offers a concrete means of protecting the mental health of a population living under pandemic restrictions.

An Empirical Study of How Household Energy Consumption Is Affected by Co-Owning Different Technological Means to Produce Renewable Energy and the Production Purpose

The transition from fossil fuel-based to renewable energy sources is one of the main economic and social challenges of the early 21st century. Due to the volatile character of wind and solar power production, matching supply and demand is essential for this transition to be successful. In this context, the willingness of private consumers to use energy flexibly has gained growing attention. Research indicates that a viable driver to motivate consumers to be demand flexible is to make them (co-)owners of renewable energy production facilities. However, existing research has only analyzed this question from an aggregated perspective. This article analyses whether behavioral changes triggered by (co-)ownership in renewables differ according to the type of installation; be it solar, wind, or bioenergy. In addition, the prosumption options self-consumption/self-consumption and sale/sale are considered. To do so, we collected 2074 completed questionnaires on energy consumption that entered an econometric model using propensity score matching to control for estimation biases. We find significant differences in the willingness to consume electricity in a flexible manner for (co-)owners of solar installations. However, only the usage of household appliances proves to be statistically significant (p-value = 0.04). Furthermore, the results show that within the group of (co-)owners of solar installation, the choice between self-consumption and sale of the produced energy has a significant effect on the inclination to become demand flexible (p-value ≤ 0.001; p-value = 0.003).

Level of confidence in and endorsement of the health system among internet users in 12 low-income and middle-income countries

INTRODUCTION

People's confidence in and endorsement of the health system are key measures of system performance, yet are undermeasured in low-income and middle-income countries (LMICs). We explored the prevalence and predictors of these measures in 12 countries.

METHODS

We conducted an internet survey in Argentina, China, Ghana, India, Indonesia, Kenya, Lebanon, Mexico, Morocco, Nigeria, Senegal and South Africa collecting demographics, ratings of quality, and confidence in and endorsement of the health system. We used multivariable logistic regression to assess the association between confidence/endorsement and self-reported quality of recent healthcare.

RESULTS

Of 13 489 respondents, 62% reported a health visit in the past year. Applying population weights, 32% of these users were very confident that they could receive effective care if they were to 'become very sick tomorrow'; 30% endorsed the health system, that is, agreed that it 'works pretty well and only needs minor changes'. Reporting high quality in the last visit was associated with 4.48 and 2.69 greater odds of confidence (95% CI 3.64 to 5.52) and endorsement (95% CI 2.33 to 3.11). Having health insurance was positively associated with confidence and endorsement (adjusted odds ratio (AOR) 1.68, 95% CI 1.49 to 1.90 and AOR 1.34, 95% CI 1.22 to 1.48), while experiencing discrimination in healthcare was negatively associated (AOR 0.67, 95% CI 0.56 to 0.80 and AOR 0.63, 95% CI 0.53 to 0.76).

CONCLUSION

Confidence and endorsement of the health system were low across 12 LMICs. This may hinder efforts to gain support for universal health coverage. Positive patient experience was strongly associated with confidence in and endorsement of the health system.

Explaining Conspiracy Beliefs and Scepticism around the COVID-19 Pandemic

Public opinion on COVID-19 provides new empirical evidence for the debate on the ideological contours of conspiracy theories. I report findings from a web survey in Greece where participants were recruited via paid advertising on Facebook and the study sample was adjusted for age, gender, education, domicile, and region of residence using a nationally representative reference sample. I find that beliefs about conspiracy theories are more correlated than the values associated with established political ideologies, and that conspiracy beliefs and scepticism about the pandemic are best explained by belief in unrelated political and medical conspiracy theories. No other demographic or attitudinal variable has such a strong influence, and the results are robust to different statistical specifications. In comparison, the effect of ideology measured by left-right self-placement is rather negligible and further moderated by trust in government. The results have implications for the strategies aimed at fighting disinformation during public health emergencies.

Estimating General Parameters from Non-Probability Surveys Using Propensity Score Adjustment

This study introduces a general framework on inference for a general parameter using nonprobability survey data when a probability sample with auxiliary variables, common to both samples, is available. The proposed framework covers parameters from inequality measures and distribution function estimates but the scope of the paper is broader. We develop a rigorous framework for general parameter estimation by solving survey weighted estimating equations which involve propensity score estimation for units in the non-probability sample. This development includes the expression of the variance estimator, as well as some alternatives which are discussed under the proposed framework. We carried a simulation study using data from a real-world survey, on which the application of the estimation methods showed the effectiveness of the proposed design-based inference on several general parameters.

An evaluation of whether propensity score adjustment can remove the self-selection bias inherent to web panel surveys addressing sensitive health behaviours

Background

In health research, population estimates are generally obtained from probability-based surveys. In market research surveys are frequently conducted from volunteer web panels. Propensity score adjustment (PSA) is often used at analysis to try to remove bias in the web survey, but empirical evidence of its effectiveness is mixed. We assess the ability of PSA to remove bias in the context of sensitive sexual health research and the potential of web panel surveys to replace or supplement probability surveys.

Methods

Four web panel surveys asked a subset of questions from the third British National Survey of Sexual Attitudes and Lifestyles (Natsal-3). Five propensity scores were generated for each web survey. The scores were developed from progressively larger sets of variables, beginning with demographic variables only and ending with demographic, sexual identity, lifestyle, attitudinal and sexual behaviour variables together. The surveys were weighted to match Natsal-3 based on propensity score quintiles. The performance of each survey and weighting was assessed by calculating the average ‘absolute’ odds ratio (inverse of the odds ratio if less than 1) across 22 pre-specified sexual behaviour outcomes of interest comparing the weighted web survey with Natsal-3. The average standard error across odds ratios was examined to assess the impact of weighting upon variance.

Results

Propensity weighting reduced bias relative to Natsal-3 as more variables were added for males, but had little effect for females, and variance increased for some surveys. Surveys with more biased estimates before propensity weighting showed greater reduction in bias from adjustment. Inconsistencies in performance were evident across surveys and outcomes. For most surveys and outcomes any reduction in bias was only partial and for some outcomes the bias increased.

Conclusions

Even after propensity weighting using a rich range of information, including some sexual behaviour variables, some bias remained and variance increased for some web surveys. Whilst our findings support the use of PSA for web panel surveys, the reduction in bias is likely to be partial and unpredictable, consistent with the findings from market research. Our results do not support the use of volunteer web panels to generate unbiased population health estimates.

Inference from Non-Probability Surveys with Statistical Matching and Propensity Score Adjustment Using Modern Prediction Techniques

Online surveys are increasingly common in social and health studies, as they provide fast and inexpensive results in comparison to traditional ones. However, these surveys often work with biased samples, as the data collection is often non-probabilistic because of the lack of internet coverage in certain population groups and the self-selection procedure that many online surveys rely on. Some procedures have been proposed to mitigate the bias, such as propensity score adjustment (PSA) and statistical matching. In PSA, propensity to participate in a nonprobability survey is estimated using a probability reference survey, and then used to obtain weighted estimates. In statistical matching, the nonprobability sample is used to train models to predict the values of the target variable, and the predictions of the models for the probability sample can be used to estimate population values. In this study, both methods are compared using three datasets to simulate pseudopopulations from which nonprobability and probability samples are drawn and used to estimate population parameters. In addition, the study compares the use of linear models and Machine Learning prediction algorithms in propensity estimation in PSA and predictive modeling in Statistical Matching. The results show that statistical matching outperforms PSA in terms of bias reduction and Root Mean Square Error (RMSE), and that simpler prediction models, such as linear and k-Nearest Neighbors, provide better outcomes than bagging algorithms.

Propensity score adjustment using machine learning classification algorithms to control selection bias in online surveys

Modern survey methods may be subject to non-observable bias, from various sources. Among online surveys, for example, selection bias is prevalent, due to the sampling mechanism commonly used, whereby participants self-select from a subgroup whose characteristics differ from those of the target population. Several techniques have been proposed to tackle this issue. One such is Propensity Score Adjustment (PSA), which is widely used and has been analysed in various studies. The usual method of estimating the propensity score is logistic regression, which requires a reference probability sample in addition to the online nonprobability sample. The predicted propensities can be used for reweighting using various estimators. However, in the online survey context, there are alternatives that might outperform logistic regression regarding propensity estimation. The aim of the present study is to determine the efficiency of some of these alternatives, involving Machine Learning (ML) classification algorithms. PSA is applied in two simulation scenarios, representing situations commonly found in online surveys, using logistic regression and ML models for propensity estimation. The results obtained show that ML algorithms remove selection bias more effectively than logistic regression when used for PSA, but that their efficacy depends largely on the selection mechanism employed and the dimensionality of the data.

Smokers' behavioral intentions in response to a low-nicotine cigarette policy

INTRODUCTION

Evidence suggests that reducing the nicotine concentration in cigarettes to sub-addictive levels would reduce use. Until a low-nicotine cigarette policy is enacted, population-level effects are unknown. This study examines the behavioral intentions of current U.S. cigarette smokers if a low-nicotine policy were implemented.

METHODS

Data were drawn from a nationally representative probability-based panel and opt-in panel. Weighted logistic regressions examined likelihood to (1) smoke lower nicotine cigarettes, (2) quit using tobacco, (3) use e-cigarettes, (4) illegally buy high-nicotine cigarettes, and (5) smoke cigars, cigarillos, or little cigars (CLCCs) among smokers, controlling for demographics, tobacco products used, dependence, and intentions to quit cigarettes. Latent class analyses (LCA) characterized patterns of behavioral intentions.

RESULTS

If a low-nicotine policy were implemented, most participants indicated a likelihood to smoke low-nicotine cigarettes (78.4%) or quit tobacco (61.9%), followed by use e-cigarettes (46.5%). Individuals with greater dependence had greater odds of intending to smoke low-nicotine cigarettes, use e-cigarettes, and illegally buy high-nicotine cigarettes. Current e-cigarette or CLCCs users had higher odds of intending to use these products. LCA revealed that individuals would 1) use low-nicotine cigarettes with low intentions to use other tobacco products or 2) use multiple tobacco products, including low-nicotine cigarettes.

CONCLUSIONS

A reduced nicotine standard for all combustible tobacco products is needed given that many tobacco users would likely intend to continue to use tobacco products. Differences in intentions by tobacco use and demographic characteristics indicate a need for additional cessation support and education around the harms of continued use of combustible tobacco.

Consumer beliefs about healthy foods and diets

Background

The U.S. Food and Drug Administration has begun a public process to redefine how companies are allowed to use the term “healthy” on food packages. Although the definition is formulated based on the latest consensus in nutrition and epidemiological research, it is also important to understand how consumers define and understand the term if it is to be behaviorally relevant. This research is an exploratory study designed to provide a descriptive account of consumers’ perceptions of and beliefs about the meaning of “healthy” food.

Methods

A nationwide U.S. sample of 1,290 food consumers was surveyed in December 2018. Respondents answered 15 questions designed to gauge perceptions of healthy food and to elicit preference for policies surrounding healthy food definitions. Responses are weighted to demographically match the population. Categorical variables have a sampling error of ±2.7%. Exploratory factor analysis is used to determine latent dimensions of health perceptions related to food type.

Results

Consumers were about evenly split on whether a food can be deemed healthy based solely on the foods’ nutritional content (52.1% believing as such) or whether there were other factors that affect whether a food is healthy (47.9% believing as such). Consumers were also about evenly split on whether an individual food can be considered healthy (believed by 47.9%) or whether this healthiness is instead a characteristic of one’s overall diet (believed by 52.1%). Ratings of individual food products revealed that “healthy” perceptions are comprised of at least three underlying latent dimensions related to animal origin, preservation, and freshness/processing. Focusing on individual macronutrients, perceived healthiness was generally decreasing in a food’s fat, sodium, and carbohydrate content and increasing in protein content. About 40% of consumers thought a healthy label implied they should increase consumption of the type of food bearing the label and about 15% thought the label meant they could eat all they wanted.

Conclusions

Results suggest consumer’s perceptions of “healthy,” which is primarily based on fat content, partially aligns with the FDA definition but also suggest consumers perceive the word as a broader and more nuanced concept that defies easy, uniform definition. Results highlight areas where nutrition education may be needed and suggest disclosures may need to accompany health claims so that consumers know what, precisely, is being communicated.

JUUL use and reasons for initiation among adult tobacco users

Background

The electronic nicotine delivery system (ENDS) JUUL has quickly captured the ENDS market, representing 74.6% of the total dollar share for this category as of November 2018. Although JUUL is marketed as an alternative to cigarettes intended for current adult smokers, evidence suggests that a majority of ENDS users are concurrently current cigarette smokers. Little is known about the dual use of JUUL and cigarettes, as well as the reasons for trying JUUL among adult tobacco users.

Methods

A survey fielded via web and phone of 1332 current cigarette, cigar, little cigar or cigarillo (CLCC), and ENDS users aged 18–54 years was conducted from March to May 2018. Weighted descriptive and bivariate analyses examined JUUL use and reasons for trying JUUL by demographics, combustible tobacco use, ENDS use and intention to quit.

Results

Approximately 81% of our sample reported current use of two or more tobacco products, including cigarettes, CLCCs or ENDS. Among current tobacco users, 15% ever used JUUL and 12% used JUUL in the past 30 days. Most individuals (74%) reporting ever JUUL use indicated using it for 5 days or fewer in the past 30 days. The most common reason for trying JUUL was trying to quit smoking cigarettes (37%), followed by family, friends or colleagues using the product (32%).

Conclusion

Although some tobacco users may be initiating JUUL to decrease combustible use, most were using the product infrequently and concurrently with other products. Findings have significant implications for cessation intervention efforts and policy development to help smokers quit.