Article Text
Abstract
Objective To estimate the carbon footprint of behavioural support services for smoking cessation: text message support, telephone counselling, group counselling and individual counselling.
Design Carbon footprint analysis.
Data source Publicly available data on National Health Service Stop Smoking Services and per unit carbon emissions; published effectiveness data from the txt2stop trial and systematic reviews of smoking cessation services.
Main outcome measures Carbon dioxide equivalents (CO2e) per 1000 smokers, per lifetime quitter, and per quality-adjusted life year gained, and cost-effectiveness, including social cost of carbon, of smoking cessation services.
Results Emissions per 1000 participants were 8143 kg CO2e for text message support, 8619 kg CO2e for telephone counselling, 16 114 kg CO2e for group counselling and 16 372 kg CO2e for individual counselling. Emissions per intervention lifetime quitter were 636 (95% CI 455 to 958) kg CO2e for text message support, 1051 (95% CI 560 to 2873) kg CO2e for telephone counselling, 1143 (95% CI 695 to 2270) kg CO2e for group counselling and 2823 (95% CI 1688 to 6549) kg CO2e for individual counselling. Text message, telephone and group counselling remained cost-effective when cost-effectiveness analysis was revised to include the environmental and economic cost of damage from carbon emissions.
Conclusions All smoking cessation services had low emissions compared to the health gains produced. Text message support had the lowest emissions of the services evaluated. Smoking cessation services have small carbon footprints and were cost-effective after accounting for the societal costs of greenhouse gas emissions.
- Environment
- Health Services
- Cessation
- Economics
- Primary Health Care
Statistics from Altmetric.com
Introduction
Health systems are significant sources of greenhouse gas emissions in developed countries.1 Patient and staff travel, building energy consumption, and other healthcare components produce emissions that contribute to climate change. The National Health Service (NHS) in England emits around 20 billion kilograms of carbon dioxide equivalents (CO2e) annually, accounting for around 25% of public sector carbon emissions and 3% of total emissions in the UK.2–4
Reducing carbon emissions from healthcare is an important part of reducing overall emissions and mitigating climate change.3 In 2009, the NHS became one of the first health systems to pledge to reduce its carbon footprint, following the UK government's introduction of legally binding emissions targets.2 ,5 To meet the goal of at least 10% reduction by 2015, the NHS Sustainable Development Unit has called for prioritisation of health services that are clinically effective and produce low carbon emissions.6
However, identifying services that are clinically effective and produce low carbon emissions has been difficult.7 The NHS reports that around 50 kg carbon is emitted per appointment, but has not compared emissions of different health services.8 ,9 This is largely due to lack of appropriate measures: existing measures for carbon emissions do not account for clinical effectiveness and cost-effectiveness of services.6 ,10
In this paper, we estimate the carbon footprint of different smoking cessation services: text message support, telephone counselling, group counselling and individual counselling. To identify effective services with low-carbon emissions, we combine existing data on clinical and cost-effectiveness with the carbon footprint estimates for each service. We develop and use new measures to compare the carbon-effectiveness—carbon emissions per health gain produced—and cost-effectiveness, including the cost of carbon, of smoking cessation services.
Methods
We used national guidelines from the UK Department of Health and National Centre for Health and Clinical Excellence (NICE) to describe telephone, group and individual counselling.11 As limited data are available on text message support in practice, we used the protocol from the txt2stop trial.12
Carbon emissions
Text message support, telephone, group and individual counselling involve different amounts of patient and staff travel, clinic and office space, and technology (eg, telephones, computers, text messages). For each method of cessation support, we estimated the necessary staff numbers and space of a typical service (see online supplementary appendix 1). Staff numbers and space requirements were used to calculate carbon emissions from staff commuting, computers and building emissions.
Travel, building space and technology produce typical amounts of carbon emissions per unit, such as emissions per kilometre travelled.13 We multiplied each component of smoking cessation services by their standard carbon emissions per unit. The UK government defines standard carbon emissions through estimation of greenhouse gases involved in producing, using and recycling a product: for example, carbon emissions for driving include petrol and materials used in car manufacture.14
To obtain the total carbon footprint of a service, we summed emissions from patient and staff travel, clinic and office space, and technology usage. All carbon emissions are expressed in kilograms of CO2e, which approximate the contribution of different greenhouse gases to global warming. All prices are from 2010.
Travel
The NHS estimates that patient and staff travel generates 0.25 kg CO2e per kilometre.5 This accounts for differences in carbon emissions by mode of transport and distance.9
Patient travel
A carbon emissions audit by the National Institute for Health Research (NIHR) estimated the average distance travelled to primary care settings to be 4.8 km roundtrip.13
Building emissions and energy consumption
For different smoking cessation behavioural support services, we used the number of employees and per person clinic and office space to estimate building emissions and energy consumption.17 We assumed that smoking cessation services were delivered full-time, so space was reserved for smoking cessation within multipurpose clinics.
Staff numbers
Based on Department of Health recommendations, we assume two staff members share an office of 12 m2.18
Primary care setting
The Department of Health recommends rooms of 16 m2 for group counselling and 8 m2 for individual counselling.18 An audit by the NHS NIHR estimated emissions of 65.79 kg CO2e per m2 of primary care clinic space per year.13
Air-conditioned office
Based on the NIHR audit, we estimated emissions of 154.30 kg CO2e per m2 for air-conditioned office space. Telephone counselling and administrative staff are predominantly based in air-conditioned offices.13
Telephone usage
Telephone call
Data from the telephone industry estimates emissions of 0.0075 kg CO2e/min of telephone calling.19 ,20 This includes emissions from phone manufacture, network operation and communication infrastructure.
Text message
Based on available industry data, we estimated emissions of 4.17×10−6 kg CO2e per text message.21
Information technologies
Manufacture and delivery
The UK Department of Environment, Food and Rural Affairs estimates emissions of 0.58 kg CO2e per pound (£) spent on computers and office machinery.22 Information technologies (IT) spending data was not available for NHS Stop Smoking Services, so we estimated one computer per counsellor or administrative staff. We assumed that computers are used exclusively for smoking cessation services, replaced every 5 years, and cost £500 on average. For text message support, we calculated carbon emissions from txt2stop trial expenditures of £700/year. For telephone counselling, we estimated £100/year spent on needed equipment.
Usage
Carbon emissions from IT usage are included in energy consumption at service sites.
Carbon-effectiveness
We compared carbon emissions in terms of two measures of smoking-related health gain: intervention lifetime quitters and quality-adjusted life years gained (QALY). We used pooled relative risks of biochemically verified smoking abstinence from the most recent trials of text message support and Cochrane systematic reviews of telephone, group and individual counselling to calculate lifetime quitters per 1000 participants.12 ,23–25 We assumed a 2% background quit rate, a 21% relapse rate between 6 months and 12 months, and a 30% lifetime relapse rate.26–28
We calculated carbon emissions per QALY using an estimate from the cost-effectiveness analysis of txt2stop. This model estimated 2.27 QALYs gained per lifetime quitter, predominantly due to a reduction in smoking-related diseases.29 It used a lifetime horizon to calculate QALYs, a 3.5% discount rate and a comparator of current practice as per NICE guidance.30 ,31 Estimates of lifetime quitters and QALYs gained for each counselling method are shown in table 1. Group counselling produced the greatest gain in lifetime quitters and QALYs, followed by text-message support. As CIs for carbon emissions have not been described, we used the Box method to calculate CIs for carbon-effectiveness, focusing on the uncertainty in lifetime quitters and QALYs.32
Cost-effectiveness including carbon emissions
The societal impact of carbon emissions can be integrated into cost-effectiveness analysis by calculating the intervention cost, including the costs of carbon emissions, per QALY. Our comparator was no intervention with zero cost and gain in QALYs. The UK Department of Energy and Climate Change estimates the environmental, economic and social cost of damages from carbon emissions to be £0.028/kg CO2e emitted.33 Carbon emissions were multiplied by this figure and added to base intervention costs for each smoking cessation service.
We obtained base intervention costs per 1000 participants, excluding the cost of stop smoking medications, from recent UK studies. All prices were adjusted for inflation using the UK hospital and community services inflation index.34 The estimated costs of running a text message service were £16 120/1000 participants based on data from the txt2stop trial.29 Based on modelling by NICE, the estimated cost of telephone counselling was £59 000/1000 participants.31 An analysis of NHS smoking cessation clinics estimated the cost of group counselling and individual counselling to be £73 000/1000 participants.35
Sensitivity analysis
As telemedicine is increasingly encouraged, staff commuting patterns may change.36 To account for this, we conducted a sensitivity analysis excluding staff commuting.
Results
Table 2 shows carbon emissions for each component of smoking cessation services.
Carbon emissions
Overall, emissions per 1000 participants were 8143 kg CO2e for text message support, 8619 kg CO2e for telephone counselling, 16 114 kg CO2e for group counselling and 16 372 kg CO2e for individual counselling.
Patient travel was the largest source of emissions for group counselling (60%) and individual counselling (44%). Staff commuting accounted for the largest source of carbon emissions for text message support (59%) and telephone counselling (56%) and the second largest source for group (30%) and individual counselling (39%). Building emissions were the third largest source of emission for all services. Owing to different intervention activities, telephone usage and IT accounted for 7% of emissions for text message support, 12% for telephone counselling, 2% for group counselling and 2% for individual counselling.
Carbon-effectiveness
Carbon emissions per intervention lifetime quitter were 636 (95% CI 455 to 958) kg CO2e for text message support, 1051 (95% CI 560 to 2873) kg CO2e for telephone counselling, 1143 (95% CI 695 to 2270) kg CO2e for group counselling and 2823 (95% CI 1688 to 6549) kg CO2e for individual counselling. Carbon emissions per QALY gained were 281 (95% CI 201 to 422) kg CO2e for text-message support, 466 (95% CI 248 to 1268) kg CO2e for telephone counselling, 504 (95% CI 306 to 995) kg CO2e for group counselling, and 1259 (95% CI 741 to 2823) kg CO2e for individual counselling.
Cost-effectiveness
Figure 1 shows the cost per QALY for text message, telephone and group counselling support after including the cost of damages from carbon emissions. The cost of damages from carbon emissions per QALY was £7.9 for text message support, £13.0 for telephone support, £14.1 for group counselling and £35.3 for individual counselling. The total cost per QALY was £564 (95% CI 403 to 847) for text message support, £3202 (95% CI 1702 to 8712) for telephone support, £2295 (95% CI 1394 to 4534) for group counselling and £5651 (95% CI 3324 to 12 665) for individual counselling.
Sensitivity analysis
As shown in table 2, excluding emissions generated by staff, commuting decreased the carbon footprint for all smoking cessation services. Emissions decreased to a greater extent for telephone and text message support as proportionally more of their emissions are due to staff travel.
Discussion
All smoking cessation services had small carbon footprints and remained highly cost-effective after accounting for the societal impact of carbon emissions. Text message support had the smallest carbon footprint per quitter and per QALY.
Strengths and weaknesses
To our knowledge, this is the first study to estimate the carbon footprint of smoking cessation behavioural support services and develop an integrated measure of carbon emissions and clinical effectiveness. Carbon emissions are a social cost of health services provision, and our findings indicate that they can be incorporated into cost-effectiveness analysis.
In comparing smoking cessation methods, variability in population and use of medications may influence the relative risks of quitting The systematic reviews of counselling effectiveness included trials with a range of control groups from self-help to brief interventions, while the txt2stop trial encouraged controls to use existing NHS smoking cessation support services.12 In all included trials, medication (bupropion or nicotine replacement therapy) was offered to the treatment and control groups or to neither; the majority reported similar usage of stop-smoking medications across groups.12 ,23–25 In practice, many people receiving inperson counselling may be prescribed concurrent stop-smoking medication, increasing the likelihood of quitting and the carbon emissions of smoking cessation interventions.
Our estimate of carbon emissions from patient travel for group counselling may underestimate travel as group counselling services are usually provided for a wider geographical area than local clinics.37 Administrative staff and enrolment time may be underestimates because we did not include reception staffing or call-holding. The NHS and telephone industry are working to increase energy efficiency, so we may overestimate building and telephone emissions.3 ,19
We excluded recruitment costs (initial general practitioner visits), water usage and clinical waste from our analysis, although they are likely to generate a large amount of emissions. We did not calculate the carbon cost of stop smoking medications, which should be subject to a separate analysis. Pharmaceuticals account for a fifth of NHS England's overall carbon footprint.5 In smoking cessation, carbon emissions from stop-smoking medications and the additional medical monitoring necessary may be partially offset by increased likelihood of quitting among those using medication.11 Our CIs for carbon-effectiveness may be imprecise. We could not account for uncertainty in carbon emissions because carbon emission data are available as point estimates without CIs.
Our analysis focused on the health system-dependent carbon emissions of interventions. We did not take into account potential long-term savings in emissions, due to decreased tobacco consumption and reduced need for treatment of smoking-related diseases.29 ,31 Manufacture and distribution of tobacco products produces an estimated 0.93 kg CO2e per pound spent.22 On average, British smokers spend £1500 on tobacco per year.38 ,39 Carbon emissions from the manufacture and distribution of this tobacco are estimated to 1400 kg CO2e per smoker per year or 1.4 million kg CO2e per 1000 smokers. This estimate does not include carbon emissions from smoking the tobacco, which have not been published and likely increase carbon emissions from smoking. Per year, the carbon emissions of producing the tobacco used by the average smoker are greater than the carbon emissions of enrolling 1000 smokers in text message, telephone or group counselling.
We also did not take into account potential increases in carbon emissions from increased life expectancy or substituted products, although all emissions are important to climate change. The NHS aims to provide services which save lives, treat and prevent disease, while minimising the environmental impact of those services. Since average carbon emissions per capita in the UK are over 8000 kg/year, any healthcare intervention that prolongs life will increase societal carbon costs.40 Yet it would be counter to the goals of the NHS or other health services to contemplate reducing societal carbon emissions by reducing access to life-saving interventions.
Comparison with other studies
There have been few studies comparing the carbon footprints of different methods of delivering healthcare. Three previous studies of NHS secondary care reported higher per patient carbon footprints as might be expected from drug-intensive and surgical procedures. These studies of renal dialysis and reflux control reported high proportions of emissions from patient and staff travel, which is consistent with our findings.41–43 For smoking cessation services, carbon emissions per participant, 8.1–16.3 kg CO2e, are significantly lower than the NHS outpatient average of 50 kg CO2e per appointment.8
Implications for research and practice
Based on our findings, the NHS and other health systems should consider including text message support in the mix of smoking cessation services. For group and individual counselling, careful consideration of the number and location of sessions could reduce carbon emissions from travel. The measures developed—carbon-effectiveness and cost-effectiveness including the societal costs of carbon emissions—could be used to prioritise health services that are clinically effective and have small carbon footprints.6 Text message, telephone and group counselling interventions remain highly cost-effective after accounting for the societal cost of carbon emissions (in a UK context the cut-off for cost-effectiveness of services is around £20 000/QALY).31
Further research could analyse the carbon-effectiveness of stop-smoking medications and other primary care services. The long-term social cost of carbon is debated, so we urge further research on the societal impact of carbon emissions.33
Meaning of study
Smoking cessation behavioural support services have small carbon footprints, are carbon-effective and cost-effective. Text message, telephone, group and individual counselling remained highly cost-effective after accounting for the environmental cost of carbon emissions. Measures of carbon-effectiveness could help health systems meet targets for reducing carbon emissions while maximising patient benefit.
What is already known
-
Health services are major sources of greenhouse gas emissions, and reducing healthcare-related carbon emission can help mitigate climate change.
-
Existing methods for assessing the environmental impact of health services do not account for clinical effectiveness and cost-effectiveness.
What this paper adds
-
All smoking cessation services had low carbon emissions relative to the health gains produced and remained cost-effective after accounting for the environmental cost of carbon emissions.
-
Text message support had the lowest carbon emissions of services evaluated.
-
Behavioural support services for smoking cessation have small carbon footprints, are carbon-effective and cost-effective.
Acknowledgments
We thank the MRC for funding the txt2stop trial. AJBS thanks the Marshall Aid Commemoration Commission for her studentship.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online appendix
Footnotes
-
Contributors AJBS conducted the analysis of carbon emissions and drafted the protocol and manuscript. IT provided carbon cost estimates and informed methodology. IR suggested analysing carbon cost for different smoking cessation support services and revised the article critically. JC provided guidance on the cost-effectiveness analysis. CF conceived carbon cost per QALY analysis, provided guidance on the protocol, informed specifics of NHS smoking cessation services and text message support, and edited each draft.
-
Competing interests None.
-
Patient consent Obtained.
-
Provenance and peer review Not commissioned; externally peer reviewed.