The Lancet Public Health · 2022 · 87 citations · 21 references
Public HealthHealth BehaviorPopulation Health SciencesEpidemiologic MethodsEpidemiologySocial Determinants Of HealthChronic DiseaseChronic Disease EpidemiologyEpidemiology Of Noncommunicable DiseasesHealth Risk BehaviorsSocial EpidemiologySocioeconomic Determinants Of HealthHealth Behavior ChangeHealth-related Behavior ChangeHousehold Socioeconomic StatusLongitudinal Cohort StudiesNoncommunicable Disease RiskPooled Odds RatiosProspective Cohort DataSmoking Epidemiology
BACKGROUND: Behavioural risk factors of non-communicable diseases (NCDs) are socially patterned. However, the direction and the extent to which socioeconomic status (SES) influences behaviour changes before and after the diagnosis of NCDs is not clearly understood. We aimed to investigate the influence of SES on behaviour changes (physical inactivity and smoking) before and after the diagnosis of major NCDs. METHODS: In this multicohort study, we pooled individual-level data from six prospective cohort studies across 17 countries. We included participants who were diagnosed with either diabetes, cardiovascular disease, chronic lung disease, or cancer after recruitment. Participants were surveyed every 2 years. Education and total household wealth were used to construct SES. We measured behaviour changes as whether or not participants continued or initiated physical inactivity or smoking after NCD diagnosis. We used multivariable logistic regression models to estimate odds ratios (ORs), prevalence ratios (PRs), and 95% CIs for the associations between SES and continuation or initiation of unfavourable behaviours. FINDINGS: We included 8107 individuals recruited between March, 2002, and January, 2016. Over the 4-year period before and after NCD diagnosis, 886 (60·4%) of 1466 individuals continued physical inactivity and 1018 (68·8%) of 1480 participants continued smoking; 1047 (15·8%) of 6641 participants with physical activity before diagnosis initiated physical inactivity after diagnosis and 132 (2·0%) of 6627 non-smokers before diagnosis initiated smoking after diagnosis. Compared with participants with high SES, those with low SES were more likely to continue physical inactivity (244 [70·3%] of 347 vs 23 [50.0%] of 46; PR 1·41 [95% CI 1·05-1·99]; OR 2·28 [1·18-4·41]), continue smoking (214 [75·4%] of 284 vs 39 [60·9%] of 64; PR 1·27 [1·03-1·59]; OR 2·08 [1·14-3·80]), but also to initiate physical inactivity (188 [26·1%] of 720 vs 47 [7·4%] of 639; PR 3·59 [2·58-4·85]; OR 4·31 [3·02 - 6·14]). INTERPRETATION: Low SES was associated with continuing or initiating physical inactivity and continuing smoking after NCD diagnosis. Reducing socioeconomic inequality in health behaviour changes should be prioritised and integrated into NCD-prevention programmes. FUNDING: Zhejiang University and Fundamental Research Funds for the Central Universities.
21
Closing the gap in a generation: health equity through action on the social determinants of health
Michael Marmot, Sharon Friel, Ruth Bell et al. · The Lancet · 2008 · 7.9K citations
Statistical Analysis with Missing Data
Subir Ghosh, Roderick J. A. Little, Donald B. Rubin · Technometrics · 1988 · 3.1K citations