Can primary care data be used to evaluate the effectiveness of tobacco control policies?: data quality, method development and assessment of the impact of smokefree legislation using data from the Health Improvement Network

Szatkowski, Lisa Catherine (2011) Can primary care data be used to evaluate the effectiveness of tobacco control policies?: data quality, method development and assessment of the impact of smokefree legislation using data from the Health Improvement Network. PhD thesis, University of Nottingham.

[img]
Preview
PDF
4Mb

Abstract

Background: Smokefree legislation is just one of a number of tobacco control policies introduced in the UK in the last decade in an attempt to curb the harm caused by smoking. Whilst such legislation is known to have reduced non-smokers’ exposure to environmental tobacco smoke, less is known about whether the introduction of a smoking ban encourages existing smokers to attempt to quit and to seek support to do so from appropriate sources such as their general practitioner. High quality data are needed to evaluate the effectiveness of legislation in prompting smokers to change their smoking behaviour, and data collected routinely in primary care may provide such an opportunity. However, there is little contemporary evidence about the quality of the smoking data recorded in primary care, nor how best to analyse these data, which must be addressed before the resource can be used to evaluate the effectiveness of tobacco control policies.

Methods: Initially, a systematic review was undertaken to assess the impact of national comprehensive smokefree legislation on population smoking prevalence, cigarette consumption and quitting behaviour. Then, the quality of smoking status and cessation intervention recording in The Health Improvement Network (THIN) database, a large database of UK primary care records, was investigated using indirect standardisation to compare rates of recording with external data sources. Having identified Autoregressive Integrated Moving Average (ARIMA) interrupted time series analysis as an appropriate method to assess the impact of smokefree legislation on measures of smoking-related clinical activity recorded in THIN data, several sensitivity analyses were untaken to assess the impact of decisions that must be taken during the data analysis process. In the light of this knowledge, ARIMA models were used to investigate changes in the rate of recording of patients’ smoking status, delivery of cessation advice, referral of smokers to specialist cessation services and prescribing of smoking cessation medications in the months leading up to, and after, the introduction of smokefree legislation.

Results: The findings of the systematic review provide some evidence that in populations where well-enforced, comprehensive smokefree policies have been implemented quitting activity increased in the run up to, and/or following, the introduction of the legislation. Assessment of the quality of the smoking information recorded in THIN showed that the data have improved in recent years, such that the recorded prevalence of smoking is now similar to that reported in national surveys. Some uncertainty does, however, remain about the quality of recording of the delivery of cessation advice or referral of smokers to cessation services. ARIMA modelling highlighted a 6.2% increase in Nicotine Replacement Therapy (NRT) prescribing in the six months before smokefree legislation was introduced in England, and a 13.2% increase in bupropion prescribing in the three months pre-ban. A 5.5% decline in NRT prescribing and a 13.7% decline in bupropion prescribing were seen in the nine months post-legislation, declines which were offset to an extent, but not completely, by prescribing of varenicline which was first available on prescription in December 2006. Similar, though non-statistically significant, patterns were seen in Scotland, Wales and Northern Ireland, where the smaller number of practices in THIN in these countries reduced the power to detect small changes in prescribing. In England, the patterns of change in prescribing did not differ with patient sex, age group, medical history or social class.

Conclusions: The improved quality of the smoking data recorded in the THIN dataset suggests that primary care data may be a valuable resource with which to evaluate the effectiveness of tobacco control policies such as smokefree legislation. The significant increases in prescribing of NRT and bupropion in the run-up to the introduction of smokefree legislation in the UK suggest that smokers looking to quit may seek support to do so from primary care, though the decline in rates of prescribing post-legislation suggests that this positive change may not be sustained. This may represent a missed opportunity to maximise the impact of smoking bans by ensuring that smokers are aware of, and indeed access, cessation support available through primary care both before and after legislation is enacted, and should be noted by policy makers planning the introduction of smokefree legislation elsewhere. Ensuring that smokers are aware of, and indeed access, the effective support that is available through primary care to help them quit may be one way to maximise the positive impacts of smokefree legislation and reduce the health and economic burdens of continued tobacco use.

Item Type:Thesis (PhD)
Supervisors:Coleman, T.
Lewis, S.
McNeill, A.
Uncontrolled Keywords:smoking cessation, primary care data, tobacco control, data quality, smokefree legislation, thin, health improvement network
Faculties/Schools:UK Campuses > Faculty of Medicine and Health Sciences > School of Community Health Sciences
ID Code:1902
Deposited By:Dr Lisa Szatkowski
Deposited On:17 Oct 2011 11:10
Last Modified:17 Oct 2011 11:10

Archive Staff Only: item control page