The Clustering of Health-Related Behaviours in Two British Birth Cohort Studies

Introduction

Modifiable negative health-related behaviours such as smoking, heavy alcohol consumption, physical inactivity and an unhealthy diet are leading causes of non-communicable disease globally. Research findings indicate that health-related behaviours do not co-occur within individuals by chance and therefore cluster. 

Methods

This study uses Latent Profile Analysis, to identify the clustered patterns of four health-related behaviours: Smoking, alcohol, diet and physical activity. We used data, collected when participants were in their early 30s, from two British cohorts born in 1958 and 1970 (N=21,019). Multi-group Latent Profile Analysis models were run separately for men and women testing for cohort differences in health-related behaviour cluster patterns.

Results

For both genders three clusters emerged: ‘Risky’ (1-9%), ‘Smokers’ (20-30%) and ‘Mainstream’ (68-77%). Members of the ‘Risky’ cluster smoked more cigarettes per day, drank more units of alcohol, ate fruit and vegetables less frequently, chips and fried food more frequently and were less physically active, compared to the other two clusters. In contrast, members of the ‘Mainstream’ cluster tended to be healthier than the other two clusters i.e. minimal cigarette smoking, eating fruit and vegetables more frequently, chips and fried food less frequently and being more physically active. A shift in membership to the ‘Mainstream’ cluster for men and women born in 1970 indicated a general improvement in HRBs overtime. However, behaviours in the ‘Mainstream’ cluster were not ideal i.e. the frequency of sweet food consumption was generally higher in the ‘Mainstream’ cluster. Moreover, the proportion of women born in 1970 drinking alcohol above the current UK recommended limits (21 units per week for men, 14 units per week for women) was almost double that of women born in 1958 across all three clusters.

Conclusions

Our findings provide further evidence of HRB clustering, identifying consistent HRBs cluster patterns across cohort and gender groups, implying some generalizability to mid-aged adults in the British population. Additionally, our findings suggest a general improvement in HRB patterns overtime, except for sweet food consumption and an increase in alcohol consumption amongst women. These findings can contribute to the development of individual and population level public health policies, interventions and campaigns that reflect the individual’s experience of health-related behaviour.