Nicotine & Tobacco Research Advance Access published December 29, 2014 Nicotine & Tobacco Research, 2014, 1–7 doi:10.1093/ntr/ntu258 Original investigation
Original investigation
Cost-Effectiveness of Nicotine Patches for Smoking Cessation in Pregnancy: A Placebo Randomized Controlled Trial (SNAP) Holly N. Essex PhD1, Steve Parrott MSc1, Qi Wu MSc1, Jinshuo Li MPhil1, Sue Cooper PhD2, Tim Coleman MD2 Mental health and Addiction Research Group, Department of Health Sciences, University of York, York, UK; 2Division of Primary Care, NIHR School for Primary Care Research and UK Centre for Tobacco and Alcohol Studies, School of Medicine, University of Nottingham, Nottingham, UK
1
Corresponding Author: Holly N. Essex, PhD, Mental health and Addiction Research Group, Department of Health Sciences, University of York, York YO10 5DD, UK. Telephone: +441904-321817; E-mail: [email protected]
Abstract Introduction: Smoking during pregnancy is the most important, preventable cause of adverse pregnancy outcomes including miscarriage, premature birth, and low birth weight with huge financial costs to the National Health Service. However, there are very few published economic evaluations of smoking cessation interventions in pregnancy and previous studies are predominantly U.S.based and do not present incremental cost-effectiveness ratios (ICER). A number of studies have demonstrated cost-effectiveness of nicotine replacement therapy (NRT) in the general population, but this has yet to be tested among pregnant smokers. Methods: A cost-effectiveness analysis was undertaken alongside the smoking, nicotine, and pregnancy trial to compare NRT patches plus behavioral support to behavioral support alone, for pregnant women who smoked. Results: At delivery, biochemically verified quit rates were slightly higher at 9.4% in the NRT group compared to 7.6% in the control group (odds ratio = 1.26, 95% CI = 0.82–1.96), at an increased cost of around £90 per participant. Higher costs in the NRT group were mainly attributable to the cost of NRT patches (mean = £46.07). The incremental cost-effectiveness ratio associated with NRT was £4,926 per quitter and a sensitivity analysis including only singleton births yielded an ICER of £4,156 per quitter. However, wide confidence intervals indicated a high level of uncertainty. Conclusions: Without a specific willingness to pay threshold, and due to high levels of statistical uncertainty, it is hard to determine the cost-effectiveness of NRT in this population. Furthermore, future research should address compliance issues, as these may dilute any potential effects of NRT, thus reducing the cost-effectiveness.
Introduction Maternal smoking during pregnancy is the most important, preventable cause of adverse pregnancy outcomes including miscarriage, premature birth, and low birth weight.1 Preterm birth is the leading cause of neonatal death and morbidity and to which up to half of infants’ neurodevelopmental problems are attributable.2 Similarly, low birth weight births are associated with perinatal morbidity and
also the future development of coronary heart disease, type 2 diabetes, and obesity.3 Tobacco smoking in pregnancy is a worldwide public health problem. In high-income countries, maternal rates of smoking in pregnancy vary between 11% and 17% and prevalence appears to be falling.4–8 However, in low and middle income economies, rates are believed to be increasing9 such that the World Health Organization predicts a future epidemic there.10 Stopping smoking
© The Author 2014. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected].
1
2 in pregnancy prevents much morbidity and benefits infants as well as their mothers; effective cessation interventions which reduce maternal smoking concomitantly, also reduce the incidences of low birth weight infants and preterm births.11 Smoking cessation interventions delivered in pregnancy, therefore, could have even greater health benefits than those used by smokers who are not pregnant. The financial costs of smoking in pregnancy are high; in 2010, total annual maternal costs to the U.K. National Health Service (NHS) were estimated to be up to £64 million annually; infants’ costs were calculated as £12 to £23.5 million annually in addition to this.12 As smoking cessation interventions used in the general population are amongst the most cost-effective of all healthcare interventions,13,14 one would expect any effective cessation interventions used by pregnant smokers to have similar cost-efficacy. It has been estimated that spending between £13.60 and £37 per pregnant smoker would yield cost savings for the NHS,12 however, as there are very few published economic evaluations of cessation interventions in pregnancy, there are insufficient empirical data to verify this. A 2008 review identified only eight studies which tested cessation interventions in pregnancy and which also involved original economic analyses.15 Individual studies included in this review suggested favorable cost-benefit ratios for smoking cessation in pregnancy, but study designs were generally of low quality and none reported incremental cost effectiveness ratios (ICER). In the general population, nicotine replacement therapy (NRT) patches have been shown to be costeffective compared to behavioral counseling alone and other comparators.16 A few small trials have investigated the effect of nicotine replacement patches for use among pregnant smokers, but provided insufficient evidence on their efficacy or safety.17 Consequently, we report a cost-effectiveness analysis conducted alongside a large randomized controlled trial (RCT) comparing nicotine replacement therapy patches with visual identical ones when used for smoking cessation in pregnancy.18 Our objectives are to compare the costs associated with trial control and intervention strategies, estimate the consequences of these alternatives, and to assess the cost-effectiveness of NRT patches used in addition to behavioral support.
Methods The Smoking, Nicotine, and Pregnancy Trial The trial for which this economic evaluation was conducted was the smoking, nicotine, and pregnancy (SNAP) trial.18 Trial participants were aged 16–45 years; of 12–24 weeks gestation; smoked ≥10 cigarettes prior to pregnancy and smoked ≥5 cigarettes currently and had exhaled carbon monoxide (CO) readings of ≥8ppm. Between May 2007 and February 2010, 1,050 participants were recruited to the trial from seven English hospital antenatal clinics. Research midwives collected baseline data, prescribed trial patches and provided faceto-face behavioral support at enrollment, and collected follow-up data at contacts: 1 month and delivery. In both arms women received the same behavioral support to encourage cognitive and behavioral changes, with advice on preparation for quitting smoking and avoiding relapses, as well as how to use patches (placebo or NRT). Women received an initial behavioral support session lasting up to 1 hr at enrollment. A quit date was also set within 2 weeks of enrollment and follow-up points was measured from this. Women were offered additional behavioral support from the local NHS stop smoking services (SSS) and research midwives provided telephone support when women were contacted on their quit date, 3 days after this and at 1 month. Participants were randomized to receive either NRT
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00 (15 mg/16 hr) or identical placebo patches. The first 4 weeks supply of patches was issued on the quit date, with a second batch of 4 weeks of patches given to those women reported not smoking and who had CO validation at the 1 month follow-up. Full methods18 including the initial19 and final20 protocols for this study are published elsewhere.
Economic Evaluation A cost-effectiveness analysis was undertaken to compare NRT patches and behavioral support to behavioral support alone, for pregnant women who smoked. The outcome was biochemically validated smoking cessation, from quit date to delivery. As recommended by the National Institute for Health and Care Excellence (NICE),21 analyses were conducted from an NHS/personal social services viewpoint. The time horizon of the trial was up to 7 months (i.e., between 12 and 24 weeks gestation and delivery).
Cost Estimation There were two main costing components for the alternative strategies, firstly, the costs of the inputs required for the interventions and secondly, the resources used to care for each woman and her infant during the trial period. Intervention costs included training and staff time for midwives to deliver behavioral support, exhaled CO monitors, NRT patches and consumables associated with CO breath testing, and overheads for premises. Training costs were calculated as 2 days of training at an NHS SSS, for 10 midwives, provided by an NHS SSS Advisor, including salary costs, overheads, and the printing of 15-page manuals which were used both by midwives to guide intervention delivery and by participants to encourage their abstinence from smoking. At the baseline antenatal visit, all women provided a CO reading and after randomization, intervention group participants were given a 4-week supply of 15 mg/16 hr NRT patches; placebo arm women were given identical patches, without active ingredients. As placebo patches represent a research cost, these were excluded from the costing. At the baseline visit and prior to randomization, midwives also provided up to 1 hr of behavioral support. The time midwives spent providing face-to-face behavioral support was recorded and multiplied by salary and overhead costs to calculate a cost per session. Telephone behavioral support sessions were offered to women on the quit date and at 3 days and 1 month after this. Successful calls were logged; these varied in length and, although call times were not recorded, call lengths were estimated by a trial midwife reflecting on the average time that was likely spent on each type of call. Time estimates were: quit date call (2 min), + 3 days (2 min), 1 month calls to self-reported smokers (4 min) and, to self-reported non-smokers (2 min). At 1 month after randomization, the call for self-reported nonsmokers was shorter because, during this, midwives would arrange a home visit to validate smoking cessation and some items were asked at this subsequent visit, rather than on the phone. In some instances, multiple attempts were required to contact women, therefore, the cost of these attempts were also estimated. Failed calls were assumed to take 30 s, with three attempted for each woman who did not speak to the midwife on the appropriate day; for each successful call we also assumed a 30 s failed call attempt. Those women visited at home received a further session of face to face behavioral support and had exhaled CO readings to validate their reported abstinence at 1 month visit after their quit date. The cost of a home visit was estimated using the cost of a standard NHS antenatal home visit. Women who were biochemically verified as abstinent at 1 month were offered a second 4-week supply of
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00
3
patches. At delivery, midwives spent approximately 10 min using CO monitors to verify smoking status. To calculate the costs of CO monitoring, per use, at baseline, 1 month and delivery, the costs of the equipment and associated consumables were totaled and divided by the total number of uses. The cost of CO monitors was not depreciated as the life expectancy of these pieces of equipment was estimated to be around 5 years (the length of the trial). One calibration kit was required for every two monitors. Based on usage evidence from the trial we assumed semidisposable mouthpiece adaptors were changed after 60 uses, batteries were changed every 240 uses, and one disposable mouth piece and alcohol-free wipe was required for each use. Resource use costs: Resource use data were collected from trial participants at the primary outcome point and also from medical records. We asked participants about their use of NHS SSS, either face-to-face or by telephone. Furthermore, given that women who smoke during pregnancy are at an increased risk of maternal morbidity, miscarriage and low birth weight or preterm babies (who are more likely to be born via caesarean section and require neonatal special care), information about antenatal hospital admissions and mode of delivery were collected from maternal medical records, and data on admissions to neonatal special care were collected from infant medical records.
Valuation of Costs All data were valued in monetary terms and unit costs were reported in Pounds Sterling for the financial year 2009/2010 (representing the mid-point of the trial). Any costs occurring in prior or later price years were inflated or deflated using the Hospital and Community Health Services pay and prices index.22 As trial follow-up occurred 7 months post-randomization, no discounting was required, as the time frame was within 1 year. For standard NHS health care, U.K. unit costs were applied from national sources, increasing the generalizability of the results. Table 1 presents a summary of resource use and unit costs, with the calculation of the costs detailed below.
Calculating Costs To calculate the costs of face-to-face smoking cessation support provided outside of the trial by the NHS, a weighted average cost for individual and group sessions was calculated based on information from a NICE costing report.26 Salary and overheads information from the same report were used to calculate the cost per minute of a phone call (assumed to be the same length as the calls reported for providing behavioral support). The costs attributable to different of modes of delivery (caesarean section, assisted vaginal, or spontaneous vaginal birth) were established by calculating a weighted average of unit costs of different modes of delivery activities recorded in NHS reference costs. A similar method was used to calculate an average cost of a maternal antenatal admission, based on antenatal observations and investigations. To calculate the cost of an infant admission to neonatal care, a weighted average of bed day costs for neonatal critical care from NHS reference costs28 (£618) was multiplied by a weighted average length of stay for neonates with major diagnoses (12.2 days) according to Hospital Episode Statistics for 2009–2010.25 Quantities of services used were multiplied by the relevant unit costs to estimate overall cost profiles for trial participants.
Outcome Measures The measure of effectiveness for this economic evaluation was the primary outcome measure for the SNAP trial, biochemically-validated smoking cessation between a quit date and delivery; temporary lapses to smoking on up to five different occasions was permitted, provided no more than five cigarettes in total were smoked.29
Cost-Effectiveness Analysis A cost-effectiveness analysis was undertaken to combine the costs of the trial interventions with corresponding health outcomes. The ICER in terms of cost per additional quitter was calculated using the mean difference in cost between two trial arms divided by the mean difference in effectiveness (i.e., in quit rates). All analyses
Table 1. Unit Costs (2009/2010 Prices) Unit cost Interventions 15 mg/16 hr NRT patches Dispensing cost Band 7 midwife time (including overheads)a Antenatal midwife home visit CO monitors and consumables Printing Band 5/6 NHS SSS advisor (including overheads)a Resource use NHS SSS: Face-to-face (individual or group session) Telephone call (4 min call) Text message Maternal antenatal admission Mode of delivery: Unassisted vaginal delivery Assisted vaginal delivery Caesarean section Baby admission to neonatal unit
Unit
Source of unit cost
£35.84 £2.14 £35.28 £45.00 £0.47 £0.96 £19.12
4 week supply of 28 patches Prescription Per hour Visit Per use 15 page manual Per hour
The health and social care information centre23 www.psnc.org.uk/pages/archive.html The NHS staff council,24 Curtis22 Department of health25 Estimated from the SNAP trial Estimated from the SNAP trial NICE26
£11.69 £1.27 £0.16 £1,180.48
Session Call Message Admission Obstetric delivery
NICE26 NICE26 NHS connecting for health27 Department of health25 Department of health25
Admission
Department of health,25 The health and social care information centre28
£1,454.28 £2,095.06 £3,028.66 £7,532.31
Note. CO = carbon monoxide; NHS = National Health Service; NICE = National Institute for Health and Care Excellence; NRT = nicotine replacement therapy; SNAP = smoking, nicotine, and pregnancy; SSS = stop smoking services. a Band refers to the salary point on the NHS pay scale, with higher bands indicating a higher rate of pay.
4 were conducted on an intention to treat basis in which participants were analyzed within original randomization groups. Analyses were conducted in Excel. In the case of missing data on the effectiveness outcomes, these women were assumed to be continuing smokers at follow-up. Missing data for cost items were imputed using average costs for the appropriate arm of the trial, allowing the base case analysis to be completed for all women in the trial and, therefore, with the same quit rate as reported in the main trial.18 Cost data were bootstrapped to account for skewness, sampling with replacement observations 1,000 times to generate a new population of sample means with an approximate normal distribution. Bootstrap results were presented graphically using a cost-effectiveness plane,30 to show the uncertainty around the mean estimates of incremental costs and effects.
Results A total of 2,410 women expressed interest in the SNAP trial, of whom 1,050 (43%) were subsequently randomized and enrolled; 521 and 529 in the NRT and placebo arms, respectively.18 The groups were similar in terms of baseline demographic characteristics and were typically heavy smokers, smoking an average of around 14 cigarettes a day at baseline. Compliance rates were low in both arms of the trial; only 7.2% of women in the NRT group and 2.8% in the placebo arm reported using the trial patches for longer than 1 month. At delivery, biochemically verified quit rates were 9.4% (49/521 women) and 7.6% (40/529) in the NRT and placebo groups, respectively; a nonsignificant difference of 1.8% (odds ratio for abstinence with NRT = 1.26, 95% CI = 0.82–1.96).18
Costs Table 2 presents a breakdown of intervention costs for the trial arms. Costs for providing hospital-based and telephone behavioral support and CO breath testing were relatively equal in both arms, though costs of home visits at 4 weeks were slightly higher in the NRT group (£14.31 vs. £9.46). This was due to higher self-reported quit rates in the NRT group at 4 weeks (25.1% vs. 14.0%). The mean total intervention cost in the control group (providing behavioral support and CO testing) was £47.75. The comparable cost of
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00 behavioral support and CO testing was £52.24 in the NRT group, and the total mean cost including NRT patches was £98.31. Table 3 shows the use of healthcare resources by trial group. A higher proportion of women allocated to the NRT group had a caesarean section (20.9% vs. 16.1%), however, use of NHS SSS, maternal antenatal hospital admissions and admissions to neonatal units were similar in the two arms. Table 4 reports total mean costs for the trial groups (calculated by summing resource use and intervention costs for each participant). Total mean resource use costs were £40.26 higher in the NRT arm, and when added to intervention costs, overall mean costs were £90.81 higher for this group, however, these differences were not statistically significant.
Cost-Effectiveness An incremental cost of £4,926 per additional quitter was estimated, by combining the differential costs of the treatment groups, with the differential quit rates (bootstrapped 95% CI = −£114,128 to £126,747). NRT plus behavioral support was found to be slightly more costly than behavioral support alone, but with a higher quit rate. The cost-effectiveness plane in Figure 1 displays the variability around the mean differences in costs and effectiveness in the trial groups. The majority of the plots in the scatter plot fall in the northeast quadrant, indicating that NRT plus behavioral support is likely to be more effective, but more costly than behavioral support alone. However, the scatter plot shows the uncertainty around the cost estimates, with bootstrapped replications falling in all four quadrants of the cost-effectiveness plane.
Sensitivity Analysis We explored the impact of uncertainty in our estimates of costs and effectiveness by conducting a sensitivity analysis. As women who have multiple rather than singleton births are more likely to experience complicated pregnancies and deliveries we repeated the analysis, excluding women who had multiple births, which occurred more frequently in the control group (eight twin deliveries versus four in the NRT group). In the singleton only analysis, quit rates were very similar at 9.48% (+0.08%) in the NRT group and 7.10% (−0.46%) in the control group. Total costs were lower in the singleton only analysis, but by a similar amount in both groups (−£64.37
Table 2. Intervention Costs by Allocated Group £ in 2009/2010 prices Training cost (per face-to-face session) Training costs for midwives (per hospital behavioral support session) Treatment cost (per participant) NRT patches Behavioral support session at hospital (including CO monitoring) Telephone callsa Calls on quit date Calls on quit date + 3 days Calls on quit date + 4 weeks (self-reported smokers) Calls on quit date + 4 weeks (self-reported nonsmokers) Home visit at 4 weeks for behavioral support for self-reported nonsmokers (including CO monitoring) 10 min to monitor CO levels at delivery Average intervention costs (per participant) Note. CO = carbon monoxide; NRT = nicotine replacement therapy. a Calls include the cost of failed attempts.
NRT group, N = 521, Mean (SD) £4.18 £46.07 (£15.57) £21.72 (£4.18)
Control, N = 529, Mean (SD) £4.18 £0.00 £22.00 (£4.62)
£1.30 (£0.27) £1.97 (£0.64) £1.78 (£0.88) £1.07 (£0.27) £14.31 (£21.14)
£1.30 (£0.27) £1.94 (£0.66) £1.91 (£0.87) £1.00 (£0.24) £9.46 (£18.47)
£5.92 (£1.61) £98.31 (£35.21)
£5.96 (£1.54) £47.75 (£19.03)
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00
5
and −£72.32 in the NRT group and control group, respectively). This generated a similar ICER for singleton births of £4,156 per additional quitter (bootstrapped 95% CI = −£65,994 to £82,059).
Discussion To our knowledge, this is the first full economic evaluation alongside a RCT to have assessed, for pregnant women who smoke, the costeffectiveness of NRT in addition to behavioral support compared to behavioral support alone. The study demonstrated that the addition of NRT patches to behavioral support was on average only £90.81 more expensive to provide than behavioral support alone when considering intervention costs and health care resource use. This excess was largely attributable to the cost of NRT patches (mean cost of patches = £46.07) and patch costs were low as only a small proportion of women continued with their use; all 1,050 received 28 patches, but only 101 accepted a further 28 day supply. Furthermore, as quit rates were not significantly different at delivery, it is perhaps unsurprising that resource use (e.g., “non-trial” cessation support) and, hence costs were very similar in both trial arms. When combined with clinical effectiveness data from the trial, the introduction of NRT generated an ICER of £4,926 per additional quitter. This indicates that NRT may be considered cost-effective, if decision makers are willing to pay more than this amount for each additional quitter. However, due to the statistical insignificance of the quit rates between the two groups and high uncertainty around the estimate, this number should be interpreted with caution, as the low compliance level may have diluted the possible effects of NRT. Table 3. Healthcare Service Utilization NRT, N (%) NHS stop smoking service Missing data 36 (6.9%) Face-to-face session 0.18 (0.6%) Phone call 2.06 (2.3%) Text message 0.80 (1.8%) Maternal antenatal hospital admission Missing data 10 (1.9%) Admissions 79 (15.2%) Mode of birth Missing data 10 (1.9%) Unassisted vaginal birth 362 (69.5%) Assisted vaginal birth 40 (7.7%) Caesarean section 109 (20.9%) Baby admitted to neonatal unit Missing data 10 (1.9%) Admissionsa 37 (7.1%)
Control, N (%) 33 (6.2%) 0.16 (0.6%) 1.84 (2.0%) 0.71 (1.7%) 12 (2.3%) 82 (15.5%) 14 (2.7%) 386 (73.0%) 44 (8.3%) 85 (16.1%) 14 (2.7%) 39 (7.3%)
Note. NHS = National Health Service; NRT = nicotine replacement therapy. a Including multiple births (Four twin births in NRT group and eight in control group. Total N = 1,062 babies).
Our study had methodological limitations. Firstly, the outcome for the economic evaluation was smoking cessation, rather than a preference based measure such as a quality-adjusted life year (QALY), which would have increased the comparability of our results with other health care programmes, smoking-related or otherwise. Furthermore, the time horizon of the study was only around 7 months, with the primary endpoint of biochemically verified smoking cessation occurring at delivery. We considered the possibility of using published data to convert quit rates into QALYs gained to assess the cost and effectiveness occurring in the longer-term. However, this would have involved making unverifiable assumptions and, as the trial showed no significant differences in quit rates or birth outcomes,18 extrapolating our results in this way would probably not have added to our alongside-trial analyses. Unlike for QALYs, as there is no accepted “willingness to pay” threshold for achieving one additional pregnant woman who successfully stops smoking, this limits conclusions which can be drawn regarding the value for money of NRT use in this population. An overall limitation of the trial was that there was a very low compliance to either NRT or placebo patches, with only 7.2% of women in the NRT group and 2.8% in the placebo arm reporting using the trial patches for longer than 1 month. Low adherence was a problem in previous studies of NRT in pregnancy18 and in the general population most smokers will discontinue nicotine therapy within 1 month as they relapse or believe that the treatment is not working.31 However, it is not understood why compliance rates are so low, and further research is needed to explore the potential influence of adherence on smoking cessation rates, and if there is an increased benefit when using NRT for longer periods, how to improve compliance. Our analyses were conducted in accordance with guidelines produced by NICE21 and, therefore, followed an NHS perspective. However, it is unlikely any important costs were omitted by taking this viewpoint, rather than a wider societal viewpoint. Unlike other more intensive interventions found to be effective for pregnant smokers, such as counseling,32 providing NRT patches is not time intensive, and was incorporated into routine antenatal care, with minimal additional burden on the women or their families. Furthermore, the behavioral support element of the intervention included only a 1 hr face-to-face session, followed by the offer of short telephone calls for additional support, again, limiting the burden on women in terms of time and travel costs. The ICER was robust to sensitivity analysis excluding multiple births, which are likely to be more complex and therefore, more costly in the antenatal and perinatal periods. An analysis including only singleton births resulted in a slightly reduced ICER of £4,156 per quitter. Nevertheless, caution is required in interpreting the study results due to the substantial uncertainty around the estimates produced from both scenarios, indicated by the very wide confidence intervals. This uncertainty occurred as there were only small differences in total costs between the groups, but each groups’ total costs
Table 4. Total Costs £ in 2009/2010 prices Intervention costs Resource use costs Total costs
NRT group, N = 521, Mean (SD)
Control group, N = 529, Mean (SD)
Differencea (95% CI)
£98.31 (£35.21) £2,571.56 (£2,393.63) £2,669.87 (£2,394.09)
£47.75 (£19.03) £2,531.31 (£2,384.34) £2,579.06 (£2,385.68)
£50.56 (£47.13 to £53.99) £40.26 (−£248.76 to £329.27) £90.81 (−£198.31 to £379.94)
Note. NRT = nicotine replacement therapy. a Difference = cost in the intervention group − cost in the control group.
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00
6
Funding This work was supported by a grant from the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) Programme, (project number 06/07/01).
Declaration of Interests None declared.
Acknowledgments
Figure 1. Cost-effectiveness plane.
were affected by high within-group variability. This variability was particularly influenced by the high costs attributable to antenatal or neonatal admissions. Antenatal and neonatal admissions are an important consideration in studies with pregnant smoking populations, as smoking in pregnancy is associated with an increased risk of serious adverse pregnancy and neonatal outcomes, including placental abruption, miscarriage, preterm birth, and low birth weight.1 In the SNAP trial, around 7% of babies were admitted to a neonatal unit in both trial groups, with each admission costing over £7,000. This compares starkly with the average between-group difference in costs of only around £90. We cannot compare our findings with similar U.K. studies as, to our knowledge, there are currently no other economic evaluations of U.K. trials testing smoking cessation interventions in pregnancy; previous economic studies addressing this issue were predominantly U.S.-based, using data from the 1980s and 1990s and did not complete incremental cost-effectiveness analyses.15 The dearth of trials in this area also precludes synthesizing data in a meta-analysis focused on economic outcomes, however, the sample size of SNAP trial was large and permitted exploration of the cost-effectiveness of NRT. Two more recent studies have explored the incremental costeffectiveness ratio of smoking cessation among pregnant women and reported QALYs: one being a simple model based on an American trial of motivational interviewing (MI) versus usual care33 and another a hypothetical model constructed for NICE guidance.34 The American trial suggested that MI may be cost-effective for relapse prevention ($628/£367 per QALY saved), but not for smoking cessation (MI cost more but provided no additional benefit).33 The hypothetical NICE model suggested that smoking cessation interventions in pregnancy may be cost-effective, but due to inadequate data made a lot of assumptions (e.g., data on smoking heaviness was based on English data for nonpregnant female smokers and pregnancy-related maternal comorbidities were excluded).34 More high quality economic evaluations of smoking cessation interventions are required, especially from outside of the United States, to investigate the shortand longer-term cost-effectiveness of helping pregnant women to quit smoking. Furthermore, studies are needed exploring the costs and utility impacts of pregnancy-specific maternal comorbidities such as ectopic pregnancy and placental complications in order to inform future models of the full impact of smoking in pregnancy on maternal and infant outcomes.
Views and opinions expressed in this article are those of the authors and do not necessarily reflect those of the NIHR HTA Programme, the NIHR, the National Health Service or the English Department of Health. The SNAP trial was supported by a grant from the NIHR HTA Programme. The authors SC, SP, TC, and QW are members of the U.K. Centre for Tobacco and Alcohol Studies (UKCTAS). Funding from the British Heart Foundation, Cancer Research U.K., the Economic and Social Research Council, the Medical Research Council and the National Institute of Health Research, under the auspices of the U.K. Clinical Research Collaboration, is gratefully acknowledged. SC and TC are members of the NIHR School for Primary Care Research and TC acknowledges support from the East Midlands CLARHC. The SNAP trial is registered with controlled-trials.com (number ISRCTN07249128).
References 1. Royal College of Physicians. Smoking and the Young. A Report of a Working Party of the Royal College of Physicians. London, UK: RCP; 1992. 2. Green NS, Damus K, Simpson JL, et al.; March of Dimes Scientific Advisory Committee on Prematurity. Research agenda for preterm birth: recommendations from the March of Dimes. Am J Obstet Gynecol. 2005;193:626–635. 3. Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of in utero and early-life conditions on adult health and disease. N Engl J Med. 2008;359:61–73. 4. Centers for Disease Control and Prevention. Smoking during pregnancy— United States, 1990–2002. Morb Mortal Wkly Rep (MMWR). 2004;53:911– 915. http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5339a1.htm. Accessed July, 2014. 5. Jensen KE, Jensen A, Nohr B, Kjaer SK. Do pregnant women still smoke? A study of smoking patterns among 261,029 primiparous women in Denmark 1997–2005. Acta Obstet Gynecol Scand. 2008;87:760–767. 6. Mohsin M, Bauman AE. Socio-demographic factors associated with smoking and smoking cessation among 426,344 pregnant women in New South Wales, Australia. BMC Public Health. 2005;5. doi: 10.1186/1471-2458-5138. http://www.biomedcentral.com/1471-2458/5/138. 7. Schneider S, Maul H, Freerksen N, Pötschke-Langer M. Who smokes during pregnancy? An analysis of the German Perinatal Quality Survey 2005. Public Health. 2008;122:1210–1216. 8. The NHS Information Centre. Infant Feeding Survey 2010. Early Results. Leeds: The National Health Service Information Centre for Health and Social Care; 2011. http://www.hscic.gov.uk/catalogue/PUB00648. Accessed July, 2014. 9. Oncken CA, Dietz PM, Tong VT, et al. Prenatal tobacco prevention and cessation interventions for women in low- and middle-income countries. Acta Obstet Gynecol Scand. 2010;89:442–453. 10. World Health Organisation. WHO Report on the Global Tobacco Epidemic, 2008—the MPOWER Package. Geneva, Switzerland: World Health Organisation; 2008. http://www.who.int/tobacco/mpower/ mpower_report_full_2008.pdf. Accessed July, 2014. 11. Lumley J, Chamberlain C, Dowswell T, Oliver S, Oakley L, Watson L. Interventions for promoting smoking cessation during pregnancy.[update of Cochrane Database Syst Rev. 2004;(4):CD001055; PMID: 15495004]. Cochrane Database Syst Rev. 2009;(3):CD001055. doi: 10.1002/14651858.cd001055.pub3. 12. Godfrey C, Pickett KE, Parrott S, Eapen D. Estimating the costs to the NHS of smoking in pregnancy for pregnant women and infants. University
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00 of York: Public Health Research Consortium. 2010. http://phrc.lshtm. ac.uk/papers/PHRC_A3-06_Final_Report.pdf. Accessed July, 2014. 13. Parrott S, Godfrey C, Raw M, West R, McNeill A. Guidance for commissioners on the cost effectiveness of smoking cessation interventions. Health Educational Authority. Thorax. 1998;53(suppl 5, pt 2):S1–38. http://thorax.bmj.com/. Accessed July, 2014. 14. West R, McNeill A, Raw M. Smoking cessation guidelines for health professionals: an update. Health Education Authority. Thorax. 2000;55:987–999. 15. Ruger JP, Emmons KM. Economic evaluations of smoking cessation and relapse prevention programs for pregnant women: a systematic review. [Review]. Value Health. 2008;11:180–190. 16. Ruger JP, Lazar CM. Economic evaluation of pharmaco- and behavioral therapies for smoking cessation: a critical and systematic review of empirical research. Annu Rev Public Health. 2012;33:279–305. 17. Coleman T, Chamberlain C, Cooper S, Leonardi-Bee J. Efficacy and safety of nicotine replacement therapy for smoking cessation in pregnancy: systematic review and meta-analysis. Addiction. 2011;106:52–61. 18. Coleman T, Cooper S, Thornton JG, et al. A randomized trial of nicotine-replacement therapy patches in pregnancy. N Engl J Med. 2012;366:808–818. 19. Coleman T, Thornton J, Britton J, et al. Protocol for the smoking, nicotine and pregnancy (SNAP) trial: double-blind, placebo-randomised, controlled trial of nicotine replacement therapy in pregnancy. [Research Support, Non-U.S. Gov’t]. BMC Health Serv Res. 2007;7:2. 20. Coleman T, Thornton J, Britton J, et al. Protocol for the SNAP (Smoking Nicotine And Pregnancy) trial final version 7.0 11 June 2009. NIHR Health Technology Assessment Programme. 2009. http://www.nets.nihr. ac.uk/__data/assets/pdf_file/0017/51290/PRO-06-07-01.pdf. Accessed July, 2014. 21. NICE. Guide to the Methods of Technology Appraisal. London, UK: National Institute for Health and Clinical Excellence; 2008. 22. Curtis L. Unit Costs of Health and Social Care 2012. Canterbury, UK: Personal Social Services Research Unit; 2012. 23. The Health and Social Care Information Centre. Prescription cost analysis England 2009. Leeds, UK: The Health and Social Care Information Centre, Prescribing Support Unit; 2010. http://www.hscic.gov.uk/catalogue/PUB01414/pres-cost-anal-eng-2009-rep.pdf. Accessed July, 2014.
7 24. The NHS Staff Council. NHS Terms and Conditions of Service Handbook: Pay Circulars (AforC) 1 and 2/2013, Amendment Number 31. London, UK: The NHS Staff Council; 2014. http://www.nhsemployers.org/. Accessed July, 2014. 25. Department of Health. NHS Reference Costs 2009–2010: Appendix NSRC01: NHS Trust Reference Cost Schedules. London, UK: Department of Health; 2011. https://www.gov.uk/government/publications/nhs-reference-costs-2009–2010. Accessed July, 2014. 26. NICE. Smoking Cessation Services: Costing Report Implementing NICE Guidance. London, UK: National Institute for Health and Clinical Excellence; 2008. 27. NHS Connecting for Health. Savings Realised by Different Types of Organisations Across the NHS Through the Use of the NHSmail SMS Feature: GP Practice. Ayrshire, Scotland: NHS Connecting for Health; 2013. http://www.connectingforhealth.nhs.uk/systemsandservices/nhsmail/about/benefits. Accessed July, 2014. 28. The Health and Social Care Information Centre. Hospital Episode Statistics, Admitted Patient Care—England, 2009–10. Leeds, UK: The Health and Social Care Information Centre; 2010. http://www.hscic.gov. uk/pubs/hes0910. Accessed July, 2014. 29. West R, Hajek P, Stead L, et al. Outcome criteria in smoking cessation trials: proposal for a common standard. Addiction. 2005;100:299–303. 30. Black WC. The CE plane: a graphic representation of cost-effectiveness. Med Decis Making. 1990;10:212–214. 31. Balmford J, Borland R, Hammond D, Cummings KM. Adherence to and reasons for premature discontinuation from stop-smoking medications: data from the ITC Four-Country Survey. Nicotine Tob Res. 2011;13:94–102. 32. Chamberlain C, O’Mara-Eves A, Oliver S, et al. Psychosocial interventions for supporting women to stop smoking in pregnancy. Cochrane Database Syst Rev. 2013;10:CD001055. 33. Ruger JP, Weinstein MC, Hammond SK, Kearney MH, Emmons KM. Cost-effectiveness of motivational interviewing for smoking cessation and relapse prevention among low-income pregnant women: a randomized controlled trial. Value Health. 2008;11:191–198. 34. Taylor M. Economic Analysis of Interventions for Smoking Cessation Aimed at Pregnant Women. York, UK: York Health Economics Consortium; 2009. http://www.nice.org.uk/guidance/ph26/resources/quitting-smoking-in-pregnancy-and-following-childbirth-economic-analysisreview2. Accessed July, 2014.
Original investigation
Cost-Effectiveness of Nicotine Patches for Smoking Cessation in Pregnancy: A Placebo Randomized Controlled Trial (SNAP) Holly N. Essex PhD1, Steve Parrott MSc1, Qi Wu MSc1, Jinshuo Li MPhil1, Sue Cooper PhD2, Tim Coleman MD2 Mental health and Addiction Research Group, Department of Health Sciences, University of York, York, UK; 2Division of Primary Care, NIHR School for Primary Care Research and UK Centre for Tobacco and Alcohol Studies, School of Medicine, University of Nottingham, Nottingham, UK
1
Corresponding Author: Holly N. Essex, PhD, Mental health and Addiction Research Group, Department of Health Sciences, University of York, York YO10 5DD, UK. Telephone: +441904-321817; E-mail: [email protected]
Abstract Introduction: Smoking during pregnancy is the most important, preventable cause of adverse pregnancy outcomes including miscarriage, premature birth, and low birth weight with huge financial costs to the National Health Service. However, there are very few published economic evaluations of smoking cessation interventions in pregnancy and previous studies are predominantly U.S.based and do not present incremental cost-effectiveness ratios (ICER). A number of studies have demonstrated cost-effectiveness of nicotine replacement therapy (NRT) in the general population, but this has yet to be tested among pregnant smokers. Methods: A cost-effectiveness analysis was undertaken alongside the smoking, nicotine, and pregnancy trial to compare NRT patches plus behavioral support to behavioral support alone, for pregnant women who smoked. Results: At delivery, biochemically verified quit rates were slightly higher at 9.4% in the NRT group compared to 7.6% in the control group (odds ratio = 1.26, 95% CI = 0.82–1.96), at an increased cost of around £90 per participant. Higher costs in the NRT group were mainly attributable to the cost of NRT patches (mean = £46.07). The incremental cost-effectiveness ratio associated with NRT was £4,926 per quitter and a sensitivity analysis including only singleton births yielded an ICER of £4,156 per quitter. However, wide confidence intervals indicated a high level of uncertainty. Conclusions: Without a specific willingness to pay threshold, and due to high levels of statistical uncertainty, it is hard to determine the cost-effectiveness of NRT in this population. Furthermore, future research should address compliance issues, as these may dilute any potential effects of NRT, thus reducing the cost-effectiveness.
Introduction Maternal smoking during pregnancy is the most important, preventable cause of adverse pregnancy outcomes including miscarriage, premature birth, and low birth weight.1 Preterm birth is the leading cause of neonatal death and morbidity and to which up to half of infants’ neurodevelopmental problems are attributable.2 Similarly, low birth weight births are associated with perinatal morbidity and
also the future development of coronary heart disease, type 2 diabetes, and obesity.3 Tobacco smoking in pregnancy is a worldwide public health problem. In high-income countries, maternal rates of smoking in pregnancy vary between 11% and 17% and prevalence appears to be falling.4–8 However, in low and middle income economies, rates are believed to be increasing9 such that the World Health Organization predicts a future epidemic there.10 Stopping smoking
© The Author 2014. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected].
1
2 in pregnancy prevents much morbidity and benefits infants as well as their mothers; effective cessation interventions which reduce maternal smoking concomitantly, also reduce the incidences of low birth weight infants and preterm births.11 Smoking cessation interventions delivered in pregnancy, therefore, could have even greater health benefits than those used by smokers who are not pregnant. The financial costs of smoking in pregnancy are high; in 2010, total annual maternal costs to the U.K. National Health Service (NHS) were estimated to be up to £64 million annually; infants’ costs were calculated as £12 to £23.5 million annually in addition to this.12 As smoking cessation interventions used in the general population are amongst the most cost-effective of all healthcare interventions,13,14 one would expect any effective cessation interventions used by pregnant smokers to have similar cost-efficacy. It has been estimated that spending between £13.60 and £37 per pregnant smoker would yield cost savings for the NHS,12 however, as there are very few published economic evaluations of cessation interventions in pregnancy, there are insufficient empirical data to verify this. A 2008 review identified only eight studies which tested cessation interventions in pregnancy and which also involved original economic analyses.15 Individual studies included in this review suggested favorable cost-benefit ratios for smoking cessation in pregnancy, but study designs were generally of low quality and none reported incremental cost effectiveness ratios (ICER). In the general population, nicotine replacement therapy (NRT) patches have been shown to be costeffective compared to behavioral counseling alone and other comparators.16 A few small trials have investigated the effect of nicotine replacement patches for use among pregnant smokers, but provided insufficient evidence on their efficacy or safety.17 Consequently, we report a cost-effectiveness analysis conducted alongside a large randomized controlled trial (RCT) comparing nicotine replacement therapy patches with visual identical ones when used for smoking cessation in pregnancy.18 Our objectives are to compare the costs associated with trial control and intervention strategies, estimate the consequences of these alternatives, and to assess the cost-effectiveness of NRT patches used in addition to behavioral support.
Methods The Smoking, Nicotine, and Pregnancy Trial The trial for which this economic evaluation was conducted was the smoking, nicotine, and pregnancy (SNAP) trial.18 Trial participants were aged 16–45 years; of 12–24 weeks gestation; smoked ≥10 cigarettes prior to pregnancy and smoked ≥5 cigarettes currently and had exhaled carbon monoxide (CO) readings of ≥8ppm. Between May 2007 and February 2010, 1,050 participants were recruited to the trial from seven English hospital antenatal clinics. Research midwives collected baseline data, prescribed trial patches and provided faceto-face behavioral support at enrollment, and collected follow-up data at contacts: 1 month and delivery. In both arms women received the same behavioral support to encourage cognitive and behavioral changes, with advice on preparation for quitting smoking and avoiding relapses, as well as how to use patches (placebo or NRT). Women received an initial behavioral support session lasting up to 1 hr at enrollment. A quit date was also set within 2 weeks of enrollment and follow-up points was measured from this. Women were offered additional behavioral support from the local NHS stop smoking services (SSS) and research midwives provided telephone support when women were contacted on their quit date, 3 days after this and at 1 month. Participants were randomized to receive either NRT
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00 (15 mg/16 hr) or identical placebo patches. The first 4 weeks supply of patches was issued on the quit date, with a second batch of 4 weeks of patches given to those women reported not smoking and who had CO validation at the 1 month follow-up. Full methods18 including the initial19 and final20 protocols for this study are published elsewhere.
Economic Evaluation A cost-effectiveness analysis was undertaken to compare NRT patches and behavioral support to behavioral support alone, for pregnant women who smoked. The outcome was biochemically validated smoking cessation, from quit date to delivery. As recommended by the National Institute for Health and Care Excellence (NICE),21 analyses were conducted from an NHS/personal social services viewpoint. The time horizon of the trial was up to 7 months (i.e., between 12 and 24 weeks gestation and delivery).
Cost Estimation There were two main costing components for the alternative strategies, firstly, the costs of the inputs required for the interventions and secondly, the resources used to care for each woman and her infant during the trial period. Intervention costs included training and staff time for midwives to deliver behavioral support, exhaled CO monitors, NRT patches and consumables associated with CO breath testing, and overheads for premises. Training costs were calculated as 2 days of training at an NHS SSS, for 10 midwives, provided by an NHS SSS Advisor, including salary costs, overheads, and the printing of 15-page manuals which were used both by midwives to guide intervention delivery and by participants to encourage their abstinence from smoking. At the baseline antenatal visit, all women provided a CO reading and after randomization, intervention group participants were given a 4-week supply of 15 mg/16 hr NRT patches; placebo arm women were given identical patches, without active ingredients. As placebo patches represent a research cost, these were excluded from the costing. At the baseline visit and prior to randomization, midwives also provided up to 1 hr of behavioral support. The time midwives spent providing face-to-face behavioral support was recorded and multiplied by salary and overhead costs to calculate a cost per session. Telephone behavioral support sessions were offered to women on the quit date and at 3 days and 1 month after this. Successful calls were logged; these varied in length and, although call times were not recorded, call lengths were estimated by a trial midwife reflecting on the average time that was likely spent on each type of call. Time estimates were: quit date call (2 min), + 3 days (2 min), 1 month calls to self-reported smokers (4 min) and, to self-reported non-smokers (2 min). At 1 month after randomization, the call for self-reported nonsmokers was shorter because, during this, midwives would arrange a home visit to validate smoking cessation and some items were asked at this subsequent visit, rather than on the phone. In some instances, multiple attempts were required to contact women, therefore, the cost of these attempts were also estimated. Failed calls were assumed to take 30 s, with three attempted for each woman who did not speak to the midwife on the appropriate day; for each successful call we also assumed a 30 s failed call attempt. Those women visited at home received a further session of face to face behavioral support and had exhaled CO readings to validate their reported abstinence at 1 month visit after their quit date. The cost of a home visit was estimated using the cost of a standard NHS antenatal home visit. Women who were biochemically verified as abstinent at 1 month were offered a second 4-week supply of
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00
3
patches. At delivery, midwives spent approximately 10 min using CO monitors to verify smoking status. To calculate the costs of CO monitoring, per use, at baseline, 1 month and delivery, the costs of the equipment and associated consumables were totaled and divided by the total number of uses. The cost of CO monitors was not depreciated as the life expectancy of these pieces of equipment was estimated to be around 5 years (the length of the trial). One calibration kit was required for every two monitors. Based on usage evidence from the trial we assumed semidisposable mouthpiece adaptors were changed after 60 uses, batteries were changed every 240 uses, and one disposable mouth piece and alcohol-free wipe was required for each use. Resource use costs: Resource use data were collected from trial participants at the primary outcome point and also from medical records. We asked participants about their use of NHS SSS, either face-to-face or by telephone. Furthermore, given that women who smoke during pregnancy are at an increased risk of maternal morbidity, miscarriage and low birth weight or preterm babies (who are more likely to be born via caesarean section and require neonatal special care), information about antenatal hospital admissions and mode of delivery were collected from maternal medical records, and data on admissions to neonatal special care were collected from infant medical records.
Valuation of Costs All data were valued in monetary terms and unit costs were reported in Pounds Sterling for the financial year 2009/2010 (representing the mid-point of the trial). Any costs occurring in prior or later price years were inflated or deflated using the Hospital and Community Health Services pay and prices index.22 As trial follow-up occurred 7 months post-randomization, no discounting was required, as the time frame was within 1 year. For standard NHS health care, U.K. unit costs were applied from national sources, increasing the generalizability of the results. Table 1 presents a summary of resource use and unit costs, with the calculation of the costs detailed below.
Calculating Costs To calculate the costs of face-to-face smoking cessation support provided outside of the trial by the NHS, a weighted average cost for individual and group sessions was calculated based on information from a NICE costing report.26 Salary and overheads information from the same report were used to calculate the cost per minute of a phone call (assumed to be the same length as the calls reported for providing behavioral support). The costs attributable to different of modes of delivery (caesarean section, assisted vaginal, or spontaneous vaginal birth) were established by calculating a weighted average of unit costs of different modes of delivery activities recorded in NHS reference costs. A similar method was used to calculate an average cost of a maternal antenatal admission, based on antenatal observations and investigations. To calculate the cost of an infant admission to neonatal care, a weighted average of bed day costs for neonatal critical care from NHS reference costs28 (£618) was multiplied by a weighted average length of stay for neonates with major diagnoses (12.2 days) according to Hospital Episode Statistics for 2009–2010.25 Quantities of services used were multiplied by the relevant unit costs to estimate overall cost profiles for trial participants.
Outcome Measures The measure of effectiveness for this economic evaluation was the primary outcome measure for the SNAP trial, biochemically-validated smoking cessation between a quit date and delivery; temporary lapses to smoking on up to five different occasions was permitted, provided no more than five cigarettes in total were smoked.29
Cost-Effectiveness Analysis A cost-effectiveness analysis was undertaken to combine the costs of the trial interventions with corresponding health outcomes. The ICER in terms of cost per additional quitter was calculated using the mean difference in cost between two trial arms divided by the mean difference in effectiveness (i.e., in quit rates). All analyses
Table 1. Unit Costs (2009/2010 Prices) Unit cost Interventions 15 mg/16 hr NRT patches Dispensing cost Band 7 midwife time (including overheads)a Antenatal midwife home visit CO monitors and consumables Printing Band 5/6 NHS SSS advisor (including overheads)a Resource use NHS SSS: Face-to-face (individual or group session) Telephone call (4 min call) Text message Maternal antenatal admission Mode of delivery: Unassisted vaginal delivery Assisted vaginal delivery Caesarean section Baby admission to neonatal unit
Unit
Source of unit cost
£35.84 £2.14 £35.28 £45.00 £0.47 £0.96 £19.12
4 week supply of 28 patches Prescription Per hour Visit Per use 15 page manual Per hour
The health and social care information centre23 www.psnc.org.uk/pages/archive.html The NHS staff council,24 Curtis22 Department of health25 Estimated from the SNAP trial Estimated from the SNAP trial NICE26
£11.69 £1.27 £0.16 £1,180.48
Session Call Message Admission Obstetric delivery
NICE26 NICE26 NHS connecting for health27 Department of health25 Department of health25
Admission
Department of health,25 The health and social care information centre28
£1,454.28 £2,095.06 £3,028.66 £7,532.31
Note. CO = carbon monoxide; NHS = National Health Service; NICE = National Institute for Health and Care Excellence; NRT = nicotine replacement therapy; SNAP = smoking, nicotine, and pregnancy; SSS = stop smoking services. a Band refers to the salary point on the NHS pay scale, with higher bands indicating a higher rate of pay.
4 were conducted on an intention to treat basis in which participants were analyzed within original randomization groups. Analyses were conducted in Excel. In the case of missing data on the effectiveness outcomes, these women were assumed to be continuing smokers at follow-up. Missing data for cost items were imputed using average costs for the appropriate arm of the trial, allowing the base case analysis to be completed for all women in the trial and, therefore, with the same quit rate as reported in the main trial.18 Cost data were bootstrapped to account for skewness, sampling with replacement observations 1,000 times to generate a new population of sample means with an approximate normal distribution. Bootstrap results were presented graphically using a cost-effectiveness plane,30 to show the uncertainty around the mean estimates of incremental costs and effects.
Results A total of 2,410 women expressed interest in the SNAP trial, of whom 1,050 (43%) were subsequently randomized and enrolled; 521 and 529 in the NRT and placebo arms, respectively.18 The groups were similar in terms of baseline demographic characteristics and were typically heavy smokers, smoking an average of around 14 cigarettes a day at baseline. Compliance rates were low in both arms of the trial; only 7.2% of women in the NRT group and 2.8% in the placebo arm reported using the trial patches for longer than 1 month. At delivery, biochemically verified quit rates were 9.4% (49/521 women) and 7.6% (40/529) in the NRT and placebo groups, respectively; a nonsignificant difference of 1.8% (odds ratio for abstinence with NRT = 1.26, 95% CI = 0.82–1.96).18
Costs Table 2 presents a breakdown of intervention costs for the trial arms. Costs for providing hospital-based and telephone behavioral support and CO breath testing were relatively equal in both arms, though costs of home visits at 4 weeks were slightly higher in the NRT group (£14.31 vs. £9.46). This was due to higher self-reported quit rates in the NRT group at 4 weeks (25.1% vs. 14.0%). The mean total intervention cost in the control group (providing behavioral support and CO testing) was £47.75. The comparable cost of
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00 behavioral support and CO testing was £52.24 in the NRT group, and the total mean cost including NRT patches was £98.31. Table 3 shows the use of healthcare resources by trial group. A higher proportion of women allocated to the NRT group had a caesarean section (20.9% vs. 16.1%), however, use of NHS SSS, maternal antenatal hospital admissions and admissions to neonatal units were similar in the two arms. Table 4 reports total mean costs for the trial groups (calculated by summing resource use and intervention costs for each participant). Total mean resource use costs were £40.26 higher in the NRT arm, and when added to intervention costs, overall mean costs were £90.81 higher for this group, however, these differences were not statistically significant.
Cost-Effectiveness An incremental cost of £4,926 per additional quitter was estimated, by combining the differential costs of the treatment groups, with the differential quit rates (bootstrapped 95% CI = −£114,128 to £126,747). NRT plus behavioral support was found to be slightly more costly than behavioral support alone, but with a higher quit rate. The cost-effectiveness plane in Figure 1 displays the variability around the mean differences in costs and effectiveness in the trial groups. The majority of the plots in the scatter plot fall in the northeast quadrant, indicating that NRT plus behavioral support is likely to be more effective, but more costly than behavioral support alone. However, the scatter plot shows the uncertainty around the cost estimates, with bootstrapped replications falling in all four quadrants of the cost-effectiveness plane.
Sensitivity Analysis We explored the impact of uncertainty in our estimates of costs and effectiveness by conducting a sensitivity analysis. As women who have multiple rather than singleton births are more likely to experience complicated pregnancies and deliveries we repeated the analysis, excluding women who had multiple births, which occurred more frequently in the control group (eight twin deliveries versus four in the NRT group). In the singleton only analysis, quit rates were very similar at 9.48% (+0.08%) in the NRT group and 7.10% (−0.46%) in the control group. Total costs were lower in the singleton only analysis, but by a similar amount in both groups (−£64.37
Table 2. Intervention Costs by Allocated Group £ in 2009/2010 prices Training cost (per face-to-face session) Training costs for midwives (per hospital behavioral support session) Treatment cost (per participant) NRT patches Behavioral support session at hospital (including CO monitoring) Telephone callsa Calls on quit date Calls on quit date + 3 days Calls on quit date + 4 weeks (self-reported smokers) Calls on quit date + 4 weeks (self-reported nonsmokers) Home visit at 4 weeks for behavioral support for self-reported nonsmokers (including CO monitoring) 10 min to monitor CO levels at delivery Average intervention costs (per participant) Note. CO = carbon monoxide; NRT = nicotine replacement therapy. a Calls include the cost of failed attempts.
NRT group, N = 521, Mean (SD) £4.18 £46.07 (£15.57) £21.72 (£4.18)
Control, N = 529, Mean (SD) £4.18 £0.00 £22.00 (£4.62)
£1.30 (£0.27) £1.97 (£0.64) £1.78 (£0.88) £1.07 (£0.27) £14.31 (£21.14)
£1.30 (£0.27) £1.94 (£0.66) £1.91 (£0.87) £1.00 (£0.24) £9.46 (£18.47)
£5.92 (£1.61) £98.31 (£35.21)
£5.96 (£1.54) £47.75 (£19.03)
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00
5
and −£72.32 in the NRT group and control group, respectively). This generated a similar ICER for singleton births of £4,156 per additional quitter (bootstrapped 95% CI = −£65,994 to £82,059).
Discussion To our knowledge, this is the first full economic evaluation alongside a RCT to have assessed, for pregnant women who smoke, the costeffectiveness of NRT in addition to behavioral support compared to behavioral support alone. The study demonstrated that the addition of NRT patches to behavioral support was on average only £90.81 more expensive to provide than behavioral support alone when considering intervention costs and health care resource use. This excess was largely attributable to the cost of NRT patches (mean cost of patches = £46.07) and patch costs were low as only a small proportion of women continued with their use; all 1,050 received 28 patches, but only 101 accepted a further 28 day supply. Furthermore, as quit rates were not significantly different at delivery, it is perhaps unsurprising that resource use (e.g., “non-trial” cessation support) and, hence costs were very similar in both trial arms. When combined with clinical effectiveness data from the trial, the introduction of NRT generated an ICER of £4,926 per additional quitter. This indicates that NRT may be considered cost-effective, if decision makers are willing to pay more than this amount for each additional quitter. However, due to the statistical insignificance of the quit rates between the two groups and high uncertainty around the estimate, this number should be interpreted with caution, as the low compliance level may have diluted the possible effects of NRT. Table 3. Healthcare Service Utilization NRT, N (%) NHS stop smoking service Missing data 36 (6.9%) Face-to-face session 0.18 (0.6%) Phone call 2.06 (2.3%) Text message 0.80 (1.8%) Maternal antenatal hospital admission Missing data 10 (1.9%) Admissions 79 (15.2%) Mode of birth Missing data 10 (1.9%) Unassisted vaginal birth 362 (69.5%) Assisted vaginal birth 40 (7.7%) Caesarean section 109 (20.9%) Baby admitted to neonatal unit Missing data 10 (1.9%) Admissionsa 37 (7.1%)
Control, N (%) 33 (6.2%) 0.16 (0.6%) 1.84 (2.0%) 0.71 (1.7%) 12 (2.3%) 82 (15.5%) 14 (2.7%) 386 (73.0%) 44 (8.3%) 85 (16.1%) 14 (2.7%) 39 (7.3%)
Note. NHS = National Health Service; NRT = nicotine replacement therapy. a Including multiple births (Four twin births in NRT group and eight in control group. Total N = 1,062 babies).
Our study had methodological limitations. Firstly, the outcome for the economic evaluation was smoking cessation, rather than a preference based measure such as a quality-adjusted life year (QALY), which would have increased the comparability of our results with other health care programmes, smoking-related or otherwise. Furthermore, the time horizon of the study was only around 7 months, with the primary endpoint of biochemically verified smoking cessation occurring at delivery. We considered the possibility of using published data to convert quit rates into QALYs gained to assess the cost and effectiveness occurring in the longer-term. However, this would have involved making unverifiable assumptions and, as the trial showed no significant differences in quit rates or birth outcomes,18 extrapolating our results in this way would probably not have added to our alongside-trial analyses. Unlike for QALYs, as there is no accepted “willingness to pay” threshold for achieving one additional pregnant woman who successfully stops smoking, this limits conclusions which can be drawn regarding the value for money of NRT use in this population. An overall limitation of the trial was that there was a very low compliance to either NRT or placebo patches, with only 7.2% of women in the NRT group and 2.8% in the placebo arm reporting using the trial patches for longer than 1 month. Low adherence was a problem in previous studies of NRT in pregnancy18 and in the general population most smokers will discontinue nicotine therapy within 1 month as they relapse or believe that the treatment is not working.31 However, it is not understood why compliance rates are so low, and further research is needed to explore the potential influence of adherence on smoking cessation rates, and if there is an increased benefit when using NRT for longer periods, how to improve compliance. Our analyses were conducted in accordance with guidelines produced by NICE21 and, therefore, followed an NHS perspective. However, it is unlikely any important costs were omitted by taking this viewpoint, rather than a wider societal viewpoint. Unlike other more intensive interventions found to be effective for pregnant smokers, such as counseling,32 providing NRT patches is not time intensive, and was incorporated into routine antenatal care, with minimal additional burden on the women or their families. Furthermore, the behavioral support element of the intervention included only a 1 hr face-to-face session, followed by the offer of short telephone calls for additional support, again, limiting the burden on women in terms of time and travel costs. The ICER was robust to sensitivity analysis excluding multiple births, which are likely to be more complex and therefore, more costly in the antenatal and perinatal periods. An analysis including only singleton births resulted in a slightly reduced ICER of £4,156 per quitter. Nevertheless, caution is required in interpreting the study results due to the substantial uncertainty around the estimates produced from both scenarios, indicated by the very wide confidence intervals. This uncertainty occurred as there were only small differences in total costs between the groups, but each groups’ total costs
Table 4. Total Costs £ in 2009/2010 prices Intervention costs Resource use costs Total costs
NRT group, N = 521, Mean (SD)
Control group, N = 529, Mean (SD)
Differencea (95% CI)
£98.31 (£35.21) £2,571.56 (£2,393.63) £2,669.87 (£2,394.09)
£47.75 (£19.03) £2,531.31 (£2,384.34) £2,579.06 (£2,385.68)
£50.56 (£47.13 to £53.99) £40.26 (−£248.76 to £329.27) £90.81 (−£198.31 to £379.94)
Note. NRT = nicotine replacement therapy. a Difference = cost in the intervention group − cost in the control group.
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00
6
Funding This work was supported by a grant from the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) Programme, (project number 06/07/01).
Declaration of Interests None declared.
Acknowledgments
Figure 1. Cost-effectiveness plane.
were affected by high within-group variability. This variability was particularly influenced by the high costs attributable to antenatal or neonatal admissions. Antenatal and neonatal admissions are an important consideration in studies with pregnant smoking populations, as smoking in pregnancy is associated with an increased risk of serious adverse pregnancy and neonatal outcomes, including placental abruption, miscarriage, preterm birth, and low birth weight.1 In the SNAP trial, around 7% of babies were admitted to a neonatal unit in both trial groups, with each admission costing over £7,000. This compares starkly with the average between-group difference in costs of only around £90. We cannot compare our findings with similar U.K. studies as, to our knowledge, there are currently no other economic evaluations of U.K. trials testing smoking cessation interventions in pregnancy; previous economic studies addressing this issue were predominantly U.S.-based, using data from the 1980s and 1990s and did not complete incremental cost-effectiveness analyses.15 The dearth of trials in this area also precludes synthesizing data in a meta-analysis focused on economic outcomes, however, the sample size of SNAP trial was large and permitted exploration of the cost-effectiveness of NRT. Two more recent studies have explored the incremental costeffectiveness ratio of smoking cessation among pregnant women and reported QALYs: one being a simple model based on an American trial of motivational interviewing (MI) versus usual care33 and another a hypothetical model constructed for NICE guidance.34 The American trial suggested that MI may be cost-effective for relapse prevention ($628/£367 per QALY saved), but not for smoking cessation (MI cost more but provided no additional benefit).33 The hypothetical NICE model suggested that smoking cessation interventions in pregnancy may be cost-effective, but due to inadequate data made a lot of assumptions (e.g., data on smoking heaviness was based on English data for nonpregnant female smokers and pregnancy-related maternal comorbidities were excluded).34 More high quality economic evaluations of smoking cessation interventions are required, especially from outside of the United States, to investigate the shortand longer-term cost-effectiveness of helping pregnant women to quit smoking. Furthermore, studies are needed exploring the costs and utility impacts of pregnancy-specific maternal comorbidities such as ectopic pregnancy and placental complications in order to inform future models of the full impact of smoking in pregnancy on maternal and infant outcomes.
Views and opinions expressed in this article are those of the authors and do not necessarily reflect those of the NIHR HTA Programme, the NIHR, the National Health Service or the English Department of Health. The SNAP trial was supported by a grant from the NIHR HTA Programme. The authors SC, SP, TC, and QW are members of the U.K. Centre for Tobacco and Alcohol Studies (UKCTAS). Funding from the British Heart Foundation, Cancer Research U.K., the Economic and Social Research Council, the Medical Research Council and the National Institute of Health Research, under the auspices of the U.K. Clinical Research Collaboration, is gratefully acknowledged. SC and TC are members of the NIHR School for Primary Care Research and TC acknowledges support from the East Midlands CLARHC. The SNAP trial is registered with controlled-trials.com (number ISRCTN07249128).
References 1. Royal College of Physicians. Smoking and the Young. A Report of a Working Party of the Royal College of Physicians. London, UK: RCP; 1992. 2. Green NS, Damus K, Simpson JL, et al.; March of Dimes Scientific Advisory Committee on Prematurity. Research agenda for preterm birth: recommendations from the March of Dimes. Am J Obstet Gynecol. 2005;193:626–635. 3. Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of in utero and early-life conditions on adult health and disease. N Engl J Med. 2008;359:61–73. 4. Centers for Disease Control and Prevention. Smoking during pregnancy— United States, 1990–2002. Morb Mortal Wkly Rep (MMWR). 2004;53:911– 915. http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5339a1.htm. Accessed July, 2014. 5. Jensen KE, Jensen A, Nohr B, Kjaer SK. Do pregnant women still smoke? A study of smoking patterns among 261,029 primiparous women in Denmark 1997–2005. Acta Obstet Gynecol Scand. 2008;87:760–767. 6. Mohsin M, Bauman AE. Socio-demographic factors associated with smoking and smoking cessation among 426,344 pregnant women in New South Wales, Australia. BMC Public Health. 2005;5. doi: 10.1186/1471-2458-5138. http://www.biomedcentral.com/1471-2458/5/138. 7. Schneider S, Maul H, Freerksen N, Pötschke-Langer M. Who smokes during pregnancy? An analysis of the German Perinatal Quality Survey 2005. Public Health. 2008;122:1210–1216. 8. The NHS Information Centre. Infant Feeding Survey 2010. Early Results. Leeds: The National Health Service Information Centre for Health and Social Care; 2011. http://www.hscic.gov.uk/catalogue/PUB00648. Accessed July, 2014. 9. Oncken CA, Dietz PM, Tong VT, et al. Prenatal tobacco prevention and cessation interventions for women in low- and middle-income countries. Acta Obstet Gynecol Scand. 2010;89:442–453. 10. World Health Organisation. WHO Report on the Global Tobacco Epidemic, 2008—the MPOWER Package. Geneva, Switzerland: World Health Organisation; 2008. http://www.who.int/tobacco/mpower/ mpower_report_full_2008.pdf. Accessed July, 2014. 11. Lumley J, Chamberlain C, Dowswell T, Oliver S, Oakley L, Watson L. Interventions for promoting smoking cessation during pregnancy.[update of Cochrane Database Syst Rev. 2004;(4):CD001055; PMID: 15495004]. Cochrane Database Syst Rev. 2009;(3):CD001055. doi: 10.1002/14651858.cd001055.pub3. 12. Godfrey C, Pickett KE, Parrott S, Eapen D. Estimating the costs to the NHS of smoking in pregnancy for pregnant women and infants. University
Nicotine & Tobacco Research, 2014, Vol. 00, No. 00 of York: Public Health Research Consortium. 2010. http://phrc.lshtm. ac.uk/papers/PHRC_A3-06_Final_Report.pdf. Accessed July, 2014. 13. Parrott S, Godfrey C, Raw M, West R, McNeill A. Guidance for commissioners on the cost effectiveness of smoking cessation interventions. Health Educational Authority. Thorax. 1998;53(suppl 5, pt 2):S1–38. http://thorax.bmj.com/. Accessed July, 2014. 14. West R, McNeill A, Raw M. Smoking cessation guidelines for health professionals: an update. Health Education Authority. Thorax. 2000;55:987–999. 15. Ruger JP, Emmons KM. Economic evaluations of smoking cessation and relapse prevention programs for pregnant women: a systematic review. [Review]. Value Health. 2008;11:180–190. 16. Ruger JP, Lazar CM. Economic evaluation of pharmaco- and behavioral therapies for smoking cessation: a critical and systematic review of empirical research. Annu Rev Public Health. 2012;33:279–305. 17. Coleman T, Chamberlain C, Cooper S, Leonardi-Bee J. Efficacy and safety of nicotine replacement therapy for smoking cessation in pregnancy: systematic review and meta-analysis. Addiction. 2011;106:52–61. 18. Coleman T, Cooper S, Thornton JG, et al. A randomized trial of nicotine-replacement therapy patches in pregnancy. N Engl J Med. 2012;366:808–818. 19. Coleman T, Thornton J, Britton J, et al. Protocol for the smoking, nicotine and pregnancy (SNAP) trial: double-blind, placebo-randomised, controlled trial of nicotine replacement therapy in pregnancy. [Research Support, Non-U.S. Gov’t]. BMC Health Serv Res. 2007;7:2. 20. Coleman T, Thornton J, Britton J, et al. Protocol for the SNAP (Smoking Nicotine And Pregnancy) trial final version 7.0 11 June 2009. NIHR Health Technology Assessment Programme. 2009. http://www.nets.nihr. ac.uk/__data/assets/pdf_file/0017/51290/PRO-06-07-01.pdf. Accessed July, 2014. 21. NICE. Guide to the Methods of Technology Appraisal. London, UK: National Institute for Health and Clinical Excellence; 2008. 22. Curtis L. Unit Costs of Health and Social Care 2012. Canterbury, UK: Personal Social Services Research Unit; 2012. 23. The Health and Social Care Information Centre. Prescription cost analysis England 2009. Leeds, UK: The Health and Social Care Information Centre, Prescribing Support Unit; 2010. http://www.hscic.gov.uk/catalogue/PUB01414/pres-cost-anal-eng-2009-rep.pdf. Accessed July, 2014.
7 24. The NHS Staff Council. NHS Terms and Conditions of Service Handbook: Pay Circulars (AforC) 1 and 2/2013, Amendment Number 31. London, UK: The NHS Staff Council; 2014. http://www.nhsemployers.org/. Accessed July, 2014. 25. Department of Health. NHS Reference Costs 2009–2010: Appendix NSRC01: NHS Trust Reference Cost Schedules. London, UK: Department of Health; 2011. https://www.gov.uk/government/publications/nhs-reference-costs-2009–2010. Accessed July, 2014. 26. NICE. Smoking Cessation Services: Costing Report Implementing NICE Guidance. London, UK: National Institute for Health and Clinical Excellence; 2008. 27. NHS Connecting for Health. Savings Realised by Different Types of Organisations Across the NHS Through the Use of the NHSmail SMS Feature: GP Practice. Ayrshire, Scotland: NHS Connecting for Health; 2013. http://www.connectingforhealth.nhs.uk/systemsandservices/nhsmail/about/benefits. Accessed July, 2014. 28. The Health and Social Care Information Centre. Hospital Episode Statistics, Admitted Patient Care—England, 2009–10. Leeds, UK: The Health and Social Care Information Centre; 2010. http://www.hscic.gov. uk/pubs/hes0910. Accessed July, 2014. 29. West R, Hajek P, Stead L, et al. Outcome criteria in smoking cessation trials: proposal for a common standard. Addiction. 2005;100:299–303. 30. Black WC. The CE plane: a graphic representation of cost-effectiveness. Med Decis Making. 1990;10:212–214. 31. Balmford J, Borland R, Hammond D, Cummings KM. Adherence to and reasons for premature discontinuation from stop-smoking medications: data from the ITC Four-Country Survey. Nicotine Tob Res. 2011;13:94–102. 32. Chamberlain C, O’Mara-Eves A, Oliver S, et al. Psychosocial interventions for supporting women to stop smoking in pregnancy. Cochrane Database Syst Rev. 2013;10:CD001055. 33. Ruger JP, Weinstein MC, Hammond SK, Kearney MH, Emmons KM. Cost-effectiveness of motivational interviewing for smoking cessation and relapse prevention among low-income pregnant women: a randomized controlled trial. Value Health. 2008;11:191–198. 34. Taylor M. Economic Analysis of Interventions for Smoking Cessation Aimed at Pregnant Women. York, UK: York Health Economics Consortium; 2009. http://www.nice.org.uk/guidance/ph26/resources/quitting-smoking-in-pregnancy-and-following-childbirth-economic-analysisreview2. Accessed July, 2014.
