Introduction

Since it was introduced a decade ago, smoke-free legislation in the UK has been successful in protecting the general public and workforce from harm arising from exposure to second-hand smoke (SHS).1–3 However, the legislation included an exemption for Her Majesty’s Prison and Probation Service (HMPPS, formally The National Offender Management Service (NOMS)) in England and Wales.4 The exemption allowed prisoners aged over 18 years to smoke in a single cell or in a cell shared with other smokers,5 staff smoking was prohibited within prison perimeter walls. Since around 80% of UK prisoners are smokers,6–9 and many of these are highly tobacco dependent,9 10 prisoners and prison staff remain at risk of high levels of SHS exposure.

In recent years, HMPPS has come under mounting pressure, from both the Prison Officers’ Association (the trade union representing prison officers throughout the UK)11 and from legal challenges by non-smoking prisoners citing poor health due to personal frequent exposure to SHS,12 to implement a smoke-free policy throughout the prison estate in England and Wales. In September 2015, in response to empirical research demonstrating high levels of SHS in English prisons,13 14 HMPPS announced the pilot implementation of a comprehensive smoke-free policy in four prisons in the South-West of England.15 This policy prohibited all staff members and prisoners from smoking tobacco and possessing tobacco or smoking paraphernalia (such as lighters and cigarette rolling paper) within the perimeter walls of the four prisons.

Prior to implementation, smoking cessation services were offered to prisoners free of charge (this included behavioural support and pharmacotherapy), and disposable electronic cigarettes were made available to purchase through the prison canteen. Prisoners were only permitted to use an electronic cigarette whilst in their cell. Tobacco and smoking paraphernalia were removed from the canteen list  two weeks before the smoke-free date at each establishment, to give prisoners the opportunity to smoke but not replace any remaining tobacco before the implementation date.

As part of an evaluation of the smoke-free policies introduced in four prisons in the South-West of England in 2016, we have compared indoor air quality, measured as concentrations of airborne particulate matter <2.5 microns in diameter (PM₂.₅), on wing landing locations three months before and three months after going smoke-free. By measuring concentrations of PM₂.₅, this study intends to determine to what extent the new policy reduces concentrations of SHS.

Methods

Results

A total of 200 datasets were collected from 32 wing landing locations throughout the four prisons. One SidePak monitor was destroyed during pre-implementation data collection, and on 12 occasions monitors were tampered with by prisoners (eg, by blocking the air inlet or turning off the monitor). The remaining 187 datasets included 113 collected before and 74 collected after policy implementation; the lower number after implementation arose primarily from restrictions on access to some prison wings. We therefore generated 74 paired sets of data for analysis (paired by prison, day of the week, time of the day, wing and monitor placement) which are presented in this paper. The 74 paired data sets were taken from across 29 wing landings (post-policy two wings at HMP 4 had been closed and one wing at HMP 3 could not be accessed by researchers due to security concerns), sampling particulate matter for an average of 5 hours and 8 minutes. Across all four prisons, monitors were placed on wing landings in the morning between 8:16 and 10:22 and in the afternoon between 14:38 and 18:00. (See table 2 for individual prison break down of mean sampling times and monitor placements times). On sampling days both pre-policy and post-policy implementation, all wings (apart from the Care and Separation Units) were at or near full capacity, with prisoner occupancy per wing ranging from 19 to 180.

Individual data from all four prisons, comparing PM2.5 concentrations collected pre-implementation and post-implementation

Data for the four prisons sampled (table 2) demonstrate that all but HMP 2 had mean PM₂.₅ concentrations above WHO 24-hour mean upper limit pre-policy implementation, and all had mean post-policy concentrations below this limit.25 All four prisons saw a statistically significant reduction in the PM₂.₅ concentration pre- to post- smoke-free policy (median percentage reductions, HMP 1=81%, HMP 2=45%, HMP 3=67%, HMP 4=72%, all four prisons, p<0.001). HMP 1, the local prison, had the highest mean and median PM₂.₅ concentrations pre-policy, and the largest percentage reduction post-policy for these samples. In HMP 2 (where monitors were not placed directly on the wing landings), the time spent over WHO 24-hour mean PM₂.₅ upper limit reduced from 7% to 0%; in the other three prisons, the reduction was from 53%–77% pre-policy to 13%–15% post-policy implementation. Figure 1 shows box plots of the distribution of PM₂.₅ concentrations measured in each prison before and after the smoke-free policy. An example of the difference in PM₂.₅ concentration profiles on a main residential wing at HMP 3, pre-implementation and post-implementation is presented in figure 2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Box plots of PM₂.₅ distributions in each of the four prisons pre- and post-smoke-free implementation. The horizontal line in each box represents the median value and the top and bottom of the box represent the 25th and 75th percentile, with the lines extending from the top and bottom of the boxes widening to the 5th and 95th percentile of the distribution. For ease of representation, figure 1 does not show PM₂.₅ concentrations over 500 µg/mᶟ (this only applies to samples taken from HMPs 1and 4). HMP, Her Majesty’s Prison.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Concentrations of PM₂.₅ sampled on a main residential wing at HMP 3 pre- and post- smoke-free implementation. HMP, Her Majesty’s Prison.

Discussion

This is the first study to compare particulate pollution before and after the implementation of smoke-free policy in English prisons. The air quality measures, which used concentrations of PM₂.₅ as a proxy for SHS, demonstrate that before the smoke-free policy was introduced PM₂.₅ levels were well in excess of WHO 24-hour mean PM₂.₅ upper limit,25 with half of all sampling time over this recommended guidance level. After introduction of the smoke-free policy, there was a substantial and statistically significant reduction in PM₂.₅ concentrations, to below WHO upper guidance limit of 25 µg/mᶟ per 24 hours. However, the range of concentrations sampled suggests that prisoners were still smoking on occasions under the smoke-free policy.

Our air quality measurements were not carried out in blind fashion, because researchers were obliged to answer questions from staff members and prisoners who enquired about the monitoring. However, while it is possible that prisoners or staff changed their behaviour in response to being monitored, we think that is unlikely to have occurred to any appreciable degree over the course of our measurements. SHS is not the only source of indoor PM₂.₅, which includes particulate matter released from sources such as open fires, toasters and microwaves. However, where toasters and microwaves were present on the wings, every effort was made to place the SidePak monitors as far away from these as possible. Safe locations for the SidePak monitors were limited, but researchers tried to collect data from all wings at each prison. Since security concerns and the design of the wings at HMP 2 and HMP 3 required us to place the SidePak monitors in wing offices (not directly on the wing landing) and in cupboards on several of the landings (which inhibited air flow), these measures are likely to have underestimated the true PM₂.₅ concentrations on these wing locations pre- and post-smoke-free policy. Nevertheless, reductions in PM₂.₅ concentrations were still observed after policy implementation in the majority of these samples. Similar issues with placement of SidePak monitors on wing locations were described in work carried out in a New Zealand prison, but that study also reported a significant reduction in PM₂.₅ concentration after going smoke-free.22

As an inevitable consequence of the smoke-free implementation dates in the four prisons, pre-policy air quality samples were taken during the winter months and post-policy during the summer months. It is possible that greater ventilation through open windows in the summer months may have contributed to the reduction in particulate levels between these two time points. However, to minimise this bias, SidePak monitors were placed towards the centre of the wings and away from any open windows during sampling. To examine whether outdoor air pollution (not only derived from SHS) could have contributed to indoor PM₂.₅ concentrations, a study which measured concentrations of particulate matter in 15 Scottish prisons, compared its indoor PM₂.₅ concentrations to outdoor measurements taken via the nearest static government monitoring station.21 Unfortunately, for this study, the nearest static government monitors were a considerable distance away (mean, 47 km) from the four prison sites sampled and were all placed in urban city centre locations (three of the four prisons sampled in this study were in remote rural locations). As PM₂.₅ is not specific to SHS and can also arise from traffic and industrial air pollution, researchers felt the comparison for this study was not suitable.

Pre-policy, researchers were able to work around any prison incidents (eg, regime changes, prisoner disturbances) in an attempt to sample all wing locations throughout the four prisons. Post-policy, researchers did not have the same flexibility as the sampling schedule was predetermined (in order to pair the samples for prison, day of the week, time, wing and monitor location), therefore fewer datasets were collected. We are unable to say whether this has significantly biased our findings. We recognise that our estimates of the proportion of time spent above WHO PM₂.₅ upper guidance limit of 25 µg/mᶟ as 24-hour mean are not truly representative because the maximum sampling time was determined by access to the wings, the battery life of the SidePak monitors used (around 11 hours), and only being able to place the monitors onto the wings during daytime hours. Since smoking does not occur during sleep, particulate levels are likely to have been considerably lower during the night. However, our data give a very good estimation, in view of the large amount of paired data collected pre-policy and post-policy (over 15 days pre-policy and post-policy), of SHS pollution during times when non-smokers would be exposed during waking hours.

In an earlier air quality monitoring study (which included two of the pilot smoke-free sites sampled here), we measured PM₂.₅ concentrations on wing landings where prisoners were permitted to smoke in their cells that were slightly higher than those three months prior to the smoke-free implementation in the present study (mean values 44 µg/mᶟ and 39 µg/mᶟ, respectively).14 A possible explanation for this is that the majority of samples taken in the current study were carried out on days leading up to the weekly delivery of tobacco to prisoners from the prison shop (data in this study were collected Monday to Friday, with canteen delivery typically occurring on Fridays) when many prisoners are running out of tobacco, whereas the earlier study included samples taken at the weekend (after tobacco delivery). This earlier study reported that PM₂.₅ concentrations were higher immediately after canteen delivery days.14 It is also possible however that three months before going smoke-free, prisoners were already starting to reduce their tobacco consumption or had been on a smoking cessation course at the prison in light of the impending policy. Further validation of SHS levels recorded in this study pre-policy comes from two further air quality monitoring studies carried out in Scottish prisons prior to their smoke-free policy13 21 which produced similar pre-policy PM₂.₅ concentrations.

Since 2005, the USA, New Zealand, Canada and Australia have all implemented smoke-free policies in their correctional facilities. International air quality studies from New Zealand and the USA have shown that comprehensive prison smoke-free policies are effective in substantially reducing SHS concentrations.20 22 26 All of these studies used markers of SHS, respirable particulate matter (eg, particulate matter less than 2.5 µg/m³ (PM₂.₅))20 22 and airborne nicotine26 to sample prison locations pre-policy and post-policy. The percentage reductions in PM₂.₅ concentrations in our study were very similar to those recorded in these other countries.20 22 26

Alongside reduced SHS concentrations, the potential health benefits of introducing a comprehensive smoke-free policy have been outlined in a study which examines the 10 years since the USA implemented its smoking ban in prisons. This study found that prisons which implemented a smoke-free policy had a 9% reduction in smoking-related deaths (particularly cardiovascular and pulmonary deaths), and that bans in place for longer than 9 years were associated with a reduction in cancer deaths.27 A study exploring natural deaths in male prisoners over 60 years of age in England and Wales reported diseases of the circulatory system (such as, coronary heart disease and cerebrovascular disease) and respiratory illnesses, all of which are substantially more common among smokers, as the most common cause of death.28 With these findings in mind, the roll-out of a comprehensive smoke-free policy across all 121 prisons in England and Wales set out by HMPPS15 has the potential to have the same positive health impact on the nearly 83 000 prisoners29 currently held, and the 32 000 staff members employed,30 in the prison estate.

Findings from this study suggest that prisoners were still smoking after the introduction of smoke-free, since PM₂.₅ concentrations post-policy ranged from 0 to 3073 µg/mᶟ, consistent with continued smoking in some areas. Hammond and Emmons measured nicotine concentrations before and after prisons in California, USA went smoke-free, and concluded that a smoking ban was effective in reducing SHS exposure but did not eliminate it.26 An ethnographic case study conducted in 10 prisons in the USA after implementing a complete smoking ban described the lengths prisoners would go to in order to acquire, exchange and smoke tobacco, and how tobacco had now become a more lucrative commodity to sell due to big demand and higher profit margin than illicit drugs.31 The study concluded that although prisoners smoked less post-policy, the emergent black market created by banning tobacco had a negative impact on prisoners. The emergence of a tobacco black market was also observed in New Zealand and the Northern Territories of Australia after their implementation of smoke-free.31–33

Conclusion

Smoke-free policies in these prisons have successfully reduced prisoner smoking, and both prisoner and staff exposure to SHS. Further work to reduce still further the occurrence of prisoner smoking is clearly required, and to assess the impact of the smoke-free policy on prisoner health. However, our data provide strong evidence in support of the continued implementation of the smoke-free policy throughout the English and Welsh prison estate and in other penal systems internationally.