MRC/CSO Social and Public Health Sciences Unit

Published 26th April 2022

By Dr Emily Tweed, Clinical Lecturer, MRC/CSO Social and Public Health Sciences Unit

There’s always a moment of anticipation, midway through a project, when you press ‘run’ on the code for what you think is going to be the key graph or statistical test. In this case, I gave a small happy gasp when the plot popped up.

The project was the Tobacco in Prisons Study (TIPs), and I was looking at a graph of respiratory medication use by people in prison in Scotland, before and after the introduction of a comprehensive smoke-free policy across all prisons. I had joined the interdisciplinary team for this large mixed-methods evaluation to explore how we might use routinely collected anonymised data from healthcare services to understand potential impacts of this policy on the health of people in prison custody.

Why evaluate smoke-free prisons?

Surprisingly, although a number of countries worldwide have introduced smoke-free prison policies, there is little evidence of their impacts on health. Rates of smoking in people in prison tend to be extremely high – for instance, in Scotland when smoking was still permitted within prisons in 2017, 68% of people in prison smoked compared to 18% of adults at liberty. You’d have to go back to the 1950s to find smoking rates that high in the general population: a time when tobacco advertising was widespread, the risks of smoking were poorly understood, and life was lived through a haze of second-hand smoke, Mad Men-style.

In 2017, it was announced that all Scottish prisons would go completely smoke-free by November 2018. Smoke-free policies in the general population have been shown to have major health benefits. For people in prison, whether smokers or not, smoke-free policies could have similar benefits – but the centrality of tobacco to prison life meant that there might also be risks particular to the prison setting. In particular, there were concerns that taking away tobacco might make coping with everyday life much harder among a population already at high risk of poor mental health. 

As one part of the TIPs study, we therefore wanted to investigate how Scotland’s prisons going smoke-free might impact – positively or negatively – on the health of people in their care. Alongside the self-reported measures of health collected through surveys and detailed one-to-one qualitative interviews by other TIPs team members, we were keen to look at objective indicators that might tell us about people’s health status or use of health services. Prison pharmacy data on medication dispensing was a natural choice: it was routinely collected already, as part of a single national contract, and thanks to the sterling work of our colleague Tom Byrne, national prisons pharmacy adviser, relatively easy to obtain.

What did we do?

We identified four groups of medications that we were interested in: those used to support people in their attempts to quit or abstain from smoking (like nicotine replacement therapy, or varenicline); those for illnesses caused or exacerbated by smoking (including angina, respiratory conditions like asthma and COPD, gastro-oesophageal reflux, and conjunctivitis); treatments for depression and anxiety (as potential unintended consequences of the policy); and a group of medications that we didn’t expect to change, that we could use as a control group (in this case, treatments for epilepsy). To account for long-term trends, including seasonal variation in medication use, we used a statistical technique called auto-regressive integrated moving average models. Thanks to the qualitative strands of TIPs, we knew that even the announcement that the policy was going to be introduced had had an impact on smoking practices within prisons, so we checked for changes associated with both policy announcement and implementation.

And that graph?

Well, at the date of policy implementation, there was a noticeable drop-off in the dispensing of inhalers and antibiotics, which was sustained over time – so the policy seemed to have had a positive impact on these indicators of short-term respiratory health. In keeping with what we knew from our work with prison health stakeholders, there were also big increases in dispensing of nicotine replacement therapy to support people in custody and a smooth roll-out of the policy. The lack of change in our control group of medications (anti-epilepsy medications) also made us more confident that the changes we did see were due to the intervention. Reassuringly, we saw no change in dispensing of medications for mental health, although this is a relatively crude measure of people’s mental state, and doesn’t completely rule out the possibility of some negative impacts, especially for those most at risk.

Together with the other results from TIPs, and the broader literature on tobacco control in community settings, we concluded that smoke-free prison policies can have beneficial effects on physical health, at least over the short term. This is important, as a number of countries worldwide – such as Finland and Northern Ireland – are considering whether to introduce such policies.

However, it’s yet to be seen whether these translate into long-term benefits, and whether smoke-free habits are kept up when people return to their communities after release. That’s the next project – and the next set of anxious moments, waiting for that code to run.

This blog has also been published on the Society for the Study of Addiction website, and can be found here.

Find out more

Read our paper in the Lancet Public Health here (open access): Tweed et al (2021) Evaluation of a national smoke-free prisons policy using medication dispensing: an interrupted time-series analysis

Read our plain-language summary: How did a complete smoke-free policy affect the health of people in prison in Scotland? Findings from the Tobacco in Prisons study

Read the full report from TIPs (open access): https://www.journalslibrary.nihr.ac.uk/phr/wglf1204/

---

First published: 11 April 2022