New UCL research using harmonised data from four UK cohort studies shows the extent of the decline in cigarette smoking over the past five decades. While one in three people born in 1946 smoked cigarettes in their early 40s, only around one in five adults born in 1970 smoked when they were the same age.
Researchers from the UCL Centre for Longitudinal Studies and Department of Behavioural Science and Health undertook the analyses to highlight the research opportunities available using harmonised data on smoking.
Code to generate the harmonised variables so researchers can use the data in their own analyses is publicly available on the CLS GitHub website.
For the harmonisation, the research team used data reported by more than 50,000 people born in 1946, 1958, 1970 and 2000-02 who are taking part in the UK’s cohort studies, and also information collected from parents. They recoded and standardised variables about smoking that were collected in different ways so that survey data were comparable across the studies, and across multiple sweeps of the same study.
This included information about study participants’ current and former smoking habits and information about their mothers’ smoking during pregnancy and parental smoking during when they were growing up.
The code that has been released will allow researchers to combine and compare data from the four studies, enhancing cross-cohort research to more accurately compare smoking trends within and across generations.
Rates of daily cigarette smoking declined across generations. By age 42/43, daily smoking rates were:
In addition, the average amount of cigarettes smoked by those who reported smoking on a daily basis in midlife declined across generations:
Within each group, smoking rates increased during adolescence, reached a peak in early adulthood (ages 20–30 years) and then declined. The average number of cigarettes people smoked daily peaked in early adulthood and then dropped. People who continued to smoke daily smoked fewer cigarettes as they grew older. Males were generally more likely to smoke and tended to smoke more than females, though differences were smaller in the later-born cohorts.
The report explored the smoking habits of participants’ parents. Fathers of 1946 cohort participants were almost four times more likely to smoke when their children were growing up compared to the fathers of the generation born at the turn of the century (80.5% v 24.2%).
Unsurprisingly, rates of smoking in pregnancy were lower among mothers of the 2000-02 generation compared to those of their older counterparts. However, rates were higher among mothers of the 1970 group than those from the 1958 cohort (42.3% v 33.5%).
Lead author, Dr Liam Wright (UCL Centre for Longitudinal Studies) said: “Previous studies that have mapped smoking trends across the UK have relied on repeat cross-sectional or retrospectively measured smoking data, which have limitations regarding accurate measurement. Here, we introduce a new resource detailing harmonisable smoking data in four longitudinal UK birth cohort studies.
“The cohorts represent a unique and largely underutilised resource for investigating trends in smoking across life. This analysis only scratches the surface of questions that can be answered with these data. Future research can combine these harmonised data with rich socioeconomic information, including on social class, education and income, to enable investigation of changing inequalities in smoking between cohorts and over the lifecourse. Detailed health data could be used for investigating the impact of smoking on public health.”
‘Cigarette smoking across life from 1946 to 2018: Harmonisation of four British birth cohort studies,’ by Liam Wright, Loren Kock, Harry Tattan-Birch, and David Bann is available on the Addiction website.
Ryan Bradshaw
Editorial Content Manager
Phone: 020 7612 6516
Email: r.bradshaw@ucl.ac.uk