Comparison of methods for analyzing longitudinal binary outcomes: cognitive status as an example

A longitudinal study of transitions between smoking and smokeless tobacco use from the ITC Bangladesh Surveys: implications for tobacco control in the Southeast Asia region

Background

In Southeast Asia, tobacco use is a major public health threat. Tobacco users in this region may switch between or concurrently use smoked tobacco and smokeless tobacco (SLT), which makes effective tobacco control challenging. This study tracks transitions of use among different product users (cigarettes, bidis, and SLT) in Bangladesh, one of the largest consumers of tobacco in the region, and examines factors related to transitions and cessation.

Methods

Four waves (2009-2015) of the International Tobacco Control (ITC) Bangladesh Survey with a cohort sample of 3245 tobacco users were analysed. Generalized Estimating Equations (GEE) models were used to explore the socioeconomic correlates of transitions from the exclusive use of cigarettes, bidis, or SLT to the use of other tobacco products or quitting over time.

Findings

Among exclusive cigarette users, most remained as exclusive cigarette users (68.1%). However, rural smokers were more likely than urban smokers to transition to bidi use (odds ratio [OR] = 3.02, 95% confidence interval [CI] = 1.45-6.29); to SLT use (OR = 2.68, 95% CI = 1.79-4.02) and to quit tobacco (OR = 1.57, 95% CI = 1.06-2.33). Among exclusive bidi users, transitional patterns were more volatile. Fewer than half (43.3%) of the exclusive bidi users maintained their status throughout the waves. Those with higher socio-economic status (SES) were more likely to quit (OR = 4.16, 95% CI = 1.08-13.12) compared to low SES smokers. Exclusive SLT users either continued using SLT or quit with minimal transitions to other products (≤2%). Nevertheless, males were more likely to switch to other tobacco products; younger (OR = 2.94, 95% CI = 1.23-6.90 vs. older), more educated (OR = 1.55, 95% CI = 1.77-3.12 vs. less educated), and urban SLT users (OR = 0.52, 95% CI = 0.30-0.86 for rural vs. urban users) were more likely to quit.

Interpretation

Complex transitional patterns were found among different types of tobacco product users over time in Bangladesh. These findings can inform more comprehensive and multi-faceted approaches to tackle diversified tobacco use in Bangladesh and neighbouring countries in the Southeast Asia region with similar tobacco user profiles of smoked tobacco and SLT products.

Funding

This is an unfunded observational study with the use the ITC Bangladesh datasets. The ITC Bangladesh Surveys were supported by grants from the US National Cancer Institute (P01 CA138389), the International Development Research Centre (IDRC Grant 104831-003), and Canadian Institutes of Health Research (MOP-79551, MOP-115016).

Perceived risk of infection while traveling during the COVID-19 pandemic: Insights from Columbus, OH

The COVID-19 outbreak caused major disruptions on individuals' out-of-home activities. Worldwide mandates to slow down the spread of the disease resulted in significant reductions in travel. This study analyzes the changes in individuals' travel outcomes and their risk perceptions related to exposure and specific travel modes during the COVID-19 pandemic. We use data collected through an online survey with residents of Columbus, OH from April 30 to May 7, 2020. Employing multiple generalized estimating equations (GEEs) with a logit link function, we analyze the perceived risk of infection while traveling with different modes controlling for socio-demographics. The findings show that on average individuals are more likely to find shared modes (i.e., transit, ride-hailing, carsharing) riskier as compared to individual ones (i.e., walking, autos) when it comes to COVID-19 exposure. This study also suggests that the associations between perceptions related to exposure and various travel modes vary across groups with (1) different primary mode preferences (auto users vs non-auto users (e.g., transit users, bicyclists, etc.)), and (2) different socio-demographics. For example, auto users are more likely to find shared modes such as ride-hailing or transit riskier as compared to personal car. The conclusions present recommendations for future transportation policies in the post-COVID era. These include building upon the emerging positive perceptions towards non-motorized modes as an opportunity to promote sustainable transportation as well as formulating viable solutions to address the high-risk perceptions associated with transit.

Obesity and socioeconomic disadvantage in midlife female public sector employees: a cohort study

BACKGROUND

The two-way relationship between obesity and socioeconomic disadvantage is well established but previous studies on social and economic consequences of obesity have primarily focused on relatively young study populations. We examined whether obesity is associated with socioeconomic disadvantage through the 10-12-year follow-up, and how obesity-related socioeconomic inequalities develop during midlife among women.

METHODS

Baseline data were derived from the female population of the Helsinki Health Study cohort, comprising 40-60 -year-old employees of the City of Helsinki, Finland in 2000-2002 (n = 6913, response rate 69%). The follow-up surveys were carried out in 2007 (n = 5810) and 2012 (n = 5400). Socioeconomic disadvantage was measured by five dichotomous measures. Repeated logistic regression analyses utilising generalized estimating equations (GEE) were used to test the association between baseline self-reported obesity and the likelihood of socioeconomic disadvantage through all phases. The effect of time on the development of inequalities was examined by time interaction terms in random effect logistic regression models.

RESULTS

After adjustment for educational level, baseline obesity was associated with repeated poverty (OR = 1.23; 95% CI; 1.05-1.44), frequent economic difficulties (OR = 1.74; 95% CI; 1.52-1.99), low household net income (OR = 1.23; 95% CI; 1.07-1.41), low household wealth (OR = 1.90; 95% CI; 1.59-2.26) and low personal income (OR = 1.22; 95% CI; 1.03-1.44). The differences in poverty rate and low personal income between the participants with obesity and participants with normal weight widened during the follow-up. Living without a partner and early exit from paid employment explained the widening of inequalities.

CONCLUSIONS

Weight status inequalities in socioeconomic disadvantage persisted or widened during the late adulthood.

Relationship characteristics associated with anal sex among female drug users

BACKGROUND

Anal sex is an important yet little studied HIV risk behavior for women.

METHODS

Using information collected on recent sexual encounters, we examined the influence of sex partner and relationship characteristics on the likelihood of engaging in anal sex among women with a high risk of HIV infection.

RESULTS

Anal sex was nearly 3 times more common among actively bisexual women (OR = 2.96, 95% CI: 2.17-4.03). Women were more likely to have anal sex with partners who injected drugs (OR = 2.32, 95% CI: 1.44-3.75), were not heterosexual (OR = 1.85, 95% CI: 1.18-2.90), and with whom they exchanged money or drugs for sex (OR = 1.79, 95% CI: 1.10-2.90). The likelihood of anal sex also increased with the number of nights sleeping together (OR = 1.15, 95% CI: 1.06-1.24). In contrast, emotional closeness and social closeness were not associated with anal sex. Condom use during anal sex was uncommon, and did not vary according to partner or relationship characteristics.

CONCLUSIONS

Our findings support the need for HIV prevention interventions that target anal sex among heterosexuals, particularly in drug-using populations residing in neighborhoods with elevated levels of HIV prevalence.

Influenza vaccination concerns among older blacks: a randomized controlled trial

CONTEXT

Vaccination can decrease the morbidity and mortality caused by influenza, yet vaccination rates remain low, especially among minority groups. Previous studies have found that important barriers to vaccination include the fear of adverse reactions and concern that the vaccine causes influenza.

BACKGROUND

This research aimed to assess the effects of messages designed to address concerns about the safety and effectiveness of vaccination among blacks aged >or=50 years.

DESIGN

In a randomized controlled trial conducted in 2007 with pre-exposure and post-exposure measurements, participants were randomly assigned to read either the vaccine safety messages (VSM) developed for the study (treatment condition) or the vaccine information statement (VIS) currently given to those getting the vaccine (control condition).

SETTING/PARTICIPANTS

A total of 108 participants participated in the study. Data were collected in either participants' residences, community settings, or university conference rooms.

MAIN OUTCOME MEASURES

Influenza vaccine-related beliefs and intention to receive vaccination were assessed.

RESULTS

Data analysis in 2007 and 2008 found that the randomization hypotheses of equal groups were retained. Participants exposed to the VSM showed greater improvement in post-exposure beliefs about how the vaccine works (p=0.0006) and the transmission of the flu (p=0.0034) as well as greater post-exposure disagreement with the belief that the vaccine causes influenza (p=0.0411).

CONCLUSIONS

The VSM affected beliefs about vaccine safety and effectiveness to a greater degree than did the VIS. These findings show that VSM are effective in changing beliefs empirically linked with influenza vaccination and vaccination disparities. The dissemination of VSM to older blacks may increase vaccination rates and reduce vaccination disparities.

Determinants of injection drug user (IDU) syringe sharing: the relationship between availability of syringes and risk network member characteristics in Winnipeg, Canada

AIMS

Despite the establishment of syringe exchange programmes, syringe-sharing behaviour remains common among some injection drug users (IDU). Previous studies have identified several individual- and social network-level variables associated with syringe sharing. We examine the extent to which each of these variables is related independently to this behaviour within a diverse study population.

DESIGN, SETTING AND PARTICIPANTS

A cross-sectional survey of 435 IDU conducted between December 2003 and September 2004 in Winnipeg, Canada.

MEASUREMENTS

Individual and social-network variables were obtained from a survey instrument administered through a personal interview. Syringe sharing was defined as receptive syringe sharing in the last 6 months. Logistic regression analysis with generalized estimating equations was used to determine simultaneously the role of individual-level and risk network member-level variables on the odds of syringe sharing.

FINDINGS

Individuals' relationship to a risk network member (sex partner, OR: 15.3 95% CI: 7.6-30.8; family member, OR: 3.4 95% CI: 1.3-9.0) and difficulty of access to syringes (OR: 3.6 95% CI: 1.3-9.9) were predictive of syringe sharing. Dyads who 'often' pooled resources to obtain drugs were at 4.9 times (95% CI: 2.1-11.6) the odds of syringe sharing, while those who 'sometimes' pooled resources were at 2.8 times (95% CI: 1.1-6.7) the odds, compared to those who 'never' pooled resources together.

CONCLUSIONS

Syringe sharing in this population depended on both the availability of clean syringes and social network relationships. Adopting interventions that take into account relationships and behaviours that shape social norms present in networks/dyads would be a necessary prevention strategy alongside the provision of clean syringes.

Modeling association in longitudinal binary outcomes: a brief review

In order to better understand aging, longitudinal studies are run in which participants are evaluated repeatedly and selected end-points (e.g., score on a cognitive screen, falls, occurrence/reoccurrence of a condition) are examined. The objective of the present paper is primarily to describe the methods available that take into account correlation between binary outcomes, and in particular to model the association of binary outcomes after controlling for covariates by using an implementation of generalized estimating equations (GEE) called 'alternating logistic regression' (ALR). In GEE, association within longitudinal outcomes is accounted for but not estimated. Alternating logistic regression, however, basically enables simultaneous estimation of pair-wise odds ratios of outcomes within a cluster, while accounting for the dependence of the outcome on covariates. A sub-sample (n=2458) from a community-based sample of Duke Established Populations for Epidemiologic Studies of the Elderly is used. In the example used here, logistic regression using GEE and ALR is used to model binary outcomes at three time points (baseline, three and six years later) and to control for covariates in a representative community-based sample 65 years of age and older (n=2458). The outcomes indicate any problem versus no problem on a five-item activities of daily living (ADL) scale in a community sample. The ALR model, however, provides insight into decline in ADL from baseline to each of the time-points whereas GEE does not. In both controlled and uncontrolled analyses, decline in ADL over three and six-year intervals (baseline to three years later, baseline to six years and three years post-baseline to six years post-baseline) is significant.