Study designs and populations

These two sources used different study designs and different kinds of samples. Beard et al. 2017 used repeated monthly surveys to study population trends in England from 2014 to 2016. Fa-Binefa et al. used a population-based cohort design, where participants were recruited and then followed over time to examine later cardiovascular disease and mortality.

1. Beard et al. 2017: monthly representative household surveys

According to Beard et al. 2017, the data came from the Alcohol and Smoking Toolkit Studies in England between April 2014 and June 2016. The paper describes these as monthly household face-to-face surveys of representative samples of about 1,700 adults in England. This means the study design was not a cohort following the same people over time. Instead, it repeatedly sampled new representative groups each month and analyzed trends across months.

Who was included?

• Adults aged 16 years and older in England.
• Participants surveyed between April 2014 and June 2016.
• People who answered questions on smoking status and alcohol consumption.
• Total included: 45,414 adults.

Beard et al. 2017 also reports the sample characteristics: 51% were female, mean age was 46.9 years, 45.5% were in routine or manual jobs or unemployed, and 87% were of white ethnicity. In short, this source used repeated cross-sectional, representative population samples rather than a single followed cohort.

2. Fa-Binefa et al.: REGICOR population-based cohort with follow-up

Fa-Binefa et al. used a different approach. The source identifies the study as the REGICOR population-based cohort study. Participants were recruited from 1995 to 2005, were 25 to 79 years old, and were then followed over time. The figure caption states that the cohort had a median follow-up of 7.02 years. This makes it a population-based cohort study, because the same participants were observed after baseline to assess later outcomes such as cardiovascular disease events and mortality.

Who was included?

• People from the REGICOR population-based cohort.
• Adults aged 25 to 79 years at recruitment.
• Participants recruited between 1995 and 2005.
• Initial sample: 11,158 individuals.
• Main analysis sample: 4,499 individuals.

The source also notes that, for one analysis, smoking onset age between 9 and 30 years was considered, with 4,171 individuals mentioned for that range. Participants with smoking onset before 9 years or after 30 years were excluded from that analysis. Baseline data collection included fasting before the visit, a blood sample, standardized questionnaires, and a physical examination.

3. Main differences between the study designs

• Beard et al. 2017: repeated monthly representative household surveys in England.
• Fa-Binefa et al.: population-based cohort study with follow-up.
• Beard et al. 2017: focused on monthly population trends and associations in smoking and alcohol-related behaviors.
• Fa-Binefa et al.: focused on future risk of cardiovascular disease and mortality after baseline exposure measurement.
• Beard et al. 2017: sampled adults aged 16+ in England.
• Fa-Binefa et al.: included adults aged 25–79 in the REGICOR cohort.

4. Summary

In summary, Beard et al. 2017 used monthly repeated cross-sectional representative household surveys of adults in England from 2014 to 2016, including 45,414 adults aged 16 and older who answered smoking and alcohol questions. By contrast, Fa-Binefa et al. used the REGICOR population-based cohort, recruiting adults aged 25 to 79 years from 1995 to 2005 and following them over time, with 11,158 initially available and 4,499 included in the main analysis. The first study measured changing population patterns month by month; the second study followed a defined population over time to study later health outcomes.