RESEARCH NEWS & VIEWS surface of Mars to analyse the composition of soil and rocks, looking for clues to the planet’s past environment and whether conditions may have been favourable for microbial life6. Opti cal spectrometers routinely support activities that underpin our daily lives, such as biomedi cal research, drug discovery, renewable energy, forensic science, environmental monitoring and chemical detection. The optical spectrometers used in these applications tend to be complex and costly because of their numerous high-precision optical and mechanical components and the stringent requirements for the align ment of these parts.They can also suffer from poor throughput, because much of the input light is scattered or absorbed as it passes through many components before reaching the detector for analysis. Finally, to be able to distinguish, or resolve, two nearby wavelengths, the instruments must typically be large. Bao and Bawendi have overcome many of these limitations through an elegant integration of nanotechnology with the image sensors used in digital cameras. Their spectrometer is based on a design involv ing broadband absorptive filters, which are similar to the coatings applied to sunglasses to block ultraviolet light. The filters are made from a series of CQDs, each with a specific particle size. This design concept can be explained by considering the case of a single sensor and a broadband absorptive optical filter. The sensor detects visible light and the filter has a specific, known (measured independently) cut-off wavelength, below which light is totally absorbed. Ideally, the filter should efficiently transmit light above this cut-off wavelength. Now consider an additional sensor and an absorptive filter tuned to another, slightly dif ferent, cut-off wavelength. Both pairs of sen sors and filters are illuminated with light of unknown colour content. The difference in the signals registered by the two sensors is a measure of the incident light power between the different cut-off wavelengths. In principle, extending this approach to greater numbers of sensors with filters tuned to different cut-off wavelengths will increase the range of colours that can be measured and the ability to resolve two adjacent colours. Bao and Bawendi achieved this scaling by applying 195 different broadband absorp tive CQD filters to hundreds of locations on the pixelated image sensor and by using an extended spectrum-reconstruction method to handle the large set of sensor readings. Commercial nanotechnology developments have led to simplified quantum-dot synthesis and precise control of dot size. CQDs can be used as tailorable broadband absorptive filters, because both the spectral absorption and the emission properties vary with particle
size. It is now feasible to produce CQDs with continuously and finely varying absorption cut-off colours, from the deep violet to nearinfrared wavelengths. These solutions can be directly patterned on the image-sensor pixels using inkjet or direct-contact printing meth ods. The long-term stability of the patterned quantum dots has also improved to provide reasonable device lifetime. It is these key fac tors that enabled Bao and Bawendi to develop their CQD spectrometer, and the simplicity of their design overcomes the constraints usually seen in conventional spectrometers. Future developments in nanotechnology and their potential for commercialization may provide the full complement of CQD materi als needed for spectral measurements beyond the visible range. One promising research area concerns chalcogenide CQDs, which have optical emission and absorption properties that extend to longer, infrared wavelengths7. Further technical challenges must be overcome to improve CQD materials and reduce optical losses. If practical automated quantum-dot
patterning on image sensors can be realized, then the costs that limit widespread integration of this technology into consumer electronics will be reduced. In the future, we may see tiny, high-resolution CQD spectrometers used on space missions or as ubiquitous sensing ele ments in household devices connected to the Internet. ■ Norm C. Anheier is at the Pacific Northwest National Laboratory, Richland, Washington 99352, USA. e-mail: [email protected] 1. Faraday, M. Phil. Trans. R. Soc. Lond. 147, 145–181 (1857). 2. Bao, J. & Bawendi, M. G. Nature 523, 67–70 (2015). 3. Ekimov, A. I. & Onushchenko, A. A. JETP Lett. 34, 345–349 (1981). 4. Newton, I. Opticks (Smith & Walford, 1704). 5. Kaltenegger, L. et al. Astrobiology 10, 89–102 (2010). 6. Wiens, R. C., Maurice, S. & the ChemCam Team. Geochem. News 145, 41–48 (2011). 7. Kershaw, S. V. et al. Chem. Soc. Rev. 42, 3033–3087 (2013).
PU BL I C H EA LT H
The case for pay to quit A randomized controlled trial of four financial-incentive programmes for smoking cessation finds that reward-based schemes lead to sustained abstinence, but low public acceptability of such schemes threatens their adoption. THERESA M. MARTEAU & ELENI MANTZARI
T
obacco remains the most lethal legal product1, killing up to half of all users2 and accounting for more than half of the difference in life expectancy between rich and poor members of society3. Higher pricing, which equates to a financial penalty, is thought to be the most successful interven tion for reducing smoking. A 50% increase in inflation-adjusted prices is estimated to cut consumption by 20% in high-, middle- and low-income countries, with the largest impacts on the young and on poor people4. For those wanting to quit, behavioural interventions and pharmacotherapy can help, although few attempts result in sustained quitting5. Financial incentives are a relatively new addition to the repertoire of behavioural interventions. Writ ing in the New England Journal of Medicine, Halpern et al.6 present a trial of four incentive schemes that adds to the growing evidence base (Fig. 1). The trial is the first to compare four incen tive schemes for smoking cessation against usual care. It involved 2,538 employees of a US company, their relatives and friends. Two of the schemes targeted individuals and two
4 0 | N AT U R E | V O L 5 2 3 | 2 J U LY 2 0 1 5
© 2015 Macmillan Publishers Limited. All rights reserved
targeted groups of six participants who were incentivized through the collective perfor mance of the group. One individual and one group-based scheme provided rewards of US$800 for smoking cessation, and the other two schemes required refundable $150 depos its, together with a $650 reward, for successful participants. Acceptance of the reward-based schemes — assessed as the proportion of participants enrolling in the scheme to which they were assigned — was 90%, much higher than for the deposit schemes, at 13.7%. There were no differences in the acceptance of individual and group-based schemes. When comparing the entire cohort of participants, regardless of whether or not they enrolled on the scheme they were offered (with those not enrolling in the scheme assumed to have remained smokers), the quit rates in the intervention groups at 6 months ranged from 9.4% to 16% — all higher than the 6% quit rate in the usual-care group. Quitting was higher for the reward- than for the deposit-based schemes (15.7% versus 10.2%) and similar for individ ual and group-based schemes. Among the small proportion of participants who accepted the deposit schemes, 52% had quit at 6 months, compared with 17% of those
PETER DAZELEY/GETTY
NEWS & VIEWS RESEARCH
Figure 1 | Carrots work for smokers. Halpern et al.6 show that reward-based financial-incentive schemes are more commonly accepted by smokers than are deposit-based schemes, but that both approaches lead to higher rates of quitting than usual care.
on reward schemes. This outcome led Halpern et al. to conclude that use of deposits is better. However, the lack of interest in these schemes limits their usefulness, as does the absence of sustained effects. At 12 months, 6 months after the incentives were stopped, about 50% of quit ters were smoking again. Only those who had been enrolled in reward-based schemes, who retained quit rates of 7.5% for the individual and 8.7% for the group schemes, showed an advantage over the 3.4% quit rate at 12 months achieved through usual care. The problem of ‘gaming’ — faking being a smoker to qualify for enrolment on a scheme or being a non-smoker to remain on a scheme — is of particular concern when using financial incentives for smoking cessation. Although reports of quitting by participants in Halpern and colleagues’ trial were validated by checking cotinine levels (a component of tobacco) in their saliva, smoking at enrolment was confirmed in only a small minority of par ticipants. An estimated 20% of those recruited were non-smokers, although this did not affect the study’s findings. By contrast, a study of pregnant smokers who were offered participa tion in a financial-incentive scheme revealed that none of the 239 enrolled, all of whom were biochemically tested at baseline, were non-smokers7. This raises questions about the contextual features of incentive schemes that can be used to minimize gaming. Halpern and colleagues’ well-designed study, the largest known of its kind so far, makes a considerable contribution to the growing evidence about financial incentives
for smoking cessation. Two recent systematic reviews8,9, one of which8 includes the results of this trial, suggest that financial-incentive schemes can be effective in achieving sustained quitting, particularly when substantial incen tives are used and when offered to pregnant smokers. The effect of incentives is also dou bled in more-deprived populations9, suggest ing that this approach motivates greater change in people on low incomes, among whom rates of smoking are not only markedly higher but also resistant to reduction. This finding high lights the potential contribution of financial incentives for reducing the health inequalities related to smoking, including the significantly higher death rates from tobacco among poor people. Were such schemes a pill or a behavioural intervention not involving money, they would doubtless be included in the range of interventions on offer to smokers want ing to quit. We are, however, “funny about money” and in particular about “money out of place”10. Given their equal effectiveness to other interventions, financial incentives are less acceptable to the public, and in turn potentially to health professionals and policymakers, than other interventions for changing behaviours11. Why is there not greater acceptability of a seemingly cost-effective intervention for reducing the leading cause of preventable premature death worldwide? The low accept ability of paying people to stop smoking (or to lose weight or take medication) reflects sev eral concerns about ideas of fairness, including
‘coercing the vulnerable’, ‘rewarding the feckless’ and ‘not rewarding the responsible’. The first of these often arises in the context of incentivizing the disadvantaged, because it is assumed that such individuals are less free to resist monetary temptations. This com promises the use of incentives for addressing health inequalities. An example of the effect of the latter two concerns is a situation in the United States12, in which objections from non-smoking employees led to the rejection by management of a $750 reward scheme for smoking cessation (for which there was strong evidence of effectiveness) and the adoption of a $625 penalty deducted from smokers’ salaries — an intervention for which at the time there was no evidence of effectiveness. The acceptability of financial incentives can change: it increases with information about effectiveness and varies with incentive type (acceptability is lower for cash incentives and higher for voucher schemes)13. Nonetheless, financial interventions to change behav iour — whether they involve price increases through taxation or reward schemes — are much less acceptable than other, often less effective, interventions14. Studies are needed that explore the scale and duration of finan cial-incentive schemes in different popula tions of smokers, and that aim to identify the characteristics of schemes producing the greatest changes. In parallel with this, research is needed to develop ways of increas ing the acceptability of interventions that could significantly help to reduce the death toll from tobacco, particularly among poor individuals. ■ Theresa M. Marteau and Eleni Mantzari are in the Behaviour and Health Research Unit, University of Cambridge, Cambridge CB2 0SR, UK. e-mail: [email protected] 1. www.vox.com/2014/5/19/5727712/the-threedeadliest-drugs-in-america-are-all-totally-legal 2. Peto, R. et al. Br. Med. Bull. 52, 12-21 (1996). 3. Jarvis, M. J. & Wardle, J. in Social Determinants of Health (eds Marmot, M. & Wilkinson, R.) 240–255 (Oxford Univ. Press, 1999). 4. Jha, P. & Peto, R. N. Engl. J. Med. 370, 60–68 (2014). 5. US Department of Health & Human Services. The Health Consequences of Smoking — 50 Years of Progress: A Report of the Surgeon General, 2014 (HHS, 2014). 6. Halpern, S. D. et al. N. Engl. J. Med. 372, 2108–2117 (2015). 7. Ierfino, D. et al. Addiction 110, 680–688 (2015). 8. Cahill, K., Hartmann-Boyce, J. & Perera, R. Incentives for Smoking Cessation (Cochrane Library, 2015). 9. Mantzari, E. et al. Prevent. Med. 75, 75–85 (2015). 10. Sandel, M. J. What Money Can’t Buy: The Moral Limits of Markets (Farrar, Straus & Giroux, 2012). 11. Promberger, M., Brown, R. C. H., Ashcroft, R. E. & Marteau, T. M. J. Med. Ethics 37, 682–687 (2011). 12. Volpp, K. G. & Galvin, R. J. Am. Med. Assoc. 311, 909–910 (2014). 13. Promberger, M., Dolan, P. & Marteau, T. M. Soc. Sci. Med. 75, 2509–2514 (2012). 14. Diepeveen, S., Ling, T., Suhrcke, M., Roland, M. & Marteau, T. M. BMC Public Health 13, 756 (2013). 2 J U LY 2 0 1 5 | V O L 5 2 3 | N AT U R E | 4 1
© 2015 Macmillan Publishers Limited. All rights reserved
size. It is now feasible to produce CQDs with continuously and finely varying absorption cut-off colours, from the deep violet to nearinfrared wavelengths. These solutions can be directly patterned on the image-sensor pixels using inkjet or direct-contact printing meth ods. The long-term stability of the patterned quantum dots has also improved to provide reasonable device lifetime. It is these key fac tors that enabled Bao and Bawendi to develop their CQD spectrometer, and the simplicity of their design overcomes the constraints usually seen in conventional spectrometers. Future developments in nanotechnology and their potential for commercialization may provide the full complement of CQD materi als needed for spectral measurements beyond the visible range. One promising research area concerns chalcogenide CQDs, which have optical emission and absorption properties that extend to longer, infrared wavelengths7. Further technical challenges must be overcome to improve CQD materials and reduce optical losses. If practical automated quantum-dot
patterning on image sensors can be realized, then the costs that limit widespread integration of this technology into consumer electronics will be reduced. In the future, we may see tiny, high-resolution CQD spectrometers used on space missions or as ubiquitous sensing ele ments in household devices connected to the Internet. ■ Norm C. Anheier is at the Pacific Northwest National Laboratory, Richland, Washington 99352, USA. e-mail: [email protected] 1. Faraday, M. Phil. Trans. R. Soc. Lond. 147, 145–181 (1857). 2. Bao, J. & Bawendi, M. G. Nature 523, 67–70 (2015). 3. Ekimov, A. I. & Onushchenko, A. A. JETP Lett. 34, 345–349 (1981). 4. Newton, I. Opticks (Smith & Walford, 1704). 5. Kaltenegger, L. et al. Astrobiology 10, 89–102 (2010). 6. Wiens, R. C., Maurice, S. & the ChemCam Team. Geochem. News 145, 41–48 (2011). 7. Kershaw, S. V. et al. Chem. Soc. Rev. 42, 3033–3087 (2013).
PU BL I C H EA LT H
The case for pay to quit A randomized controlled trial of four financial-incentive programmes for smoking cessation finds that reward-based schemes lead to sustained abstinence, but low public acceptability of such schemes threatens their adoption. THERESA M. MARTEAU & ELENI MANTZARI
T
obacco remains the most lethal legal product1, killing up to half of all users2 and accounting for more than half of the difference in life expectancy between rich and poor members of society3. Higher pricing, which equates to a financial penalty, is thought to be the most successful interven tion for reducing smoking. A 50% increase in inflation-adjusted prices is estimated to cut consumption by 20% in high-, middle- and low-income countries, with the largest impacts on the young and on poor people4. For those wanting to quit, behavioural interventions and pharmacotherapy can help, although few attempts result in sustained quitting5. Financial incentives are a relatively new addition to the repertoire of behavioural interventions. Writ ing in the New England Journal of Medicine, Halpern et al.6 present a trial of four incentive schemes that adds to the growing evidence base (Fig. 1). The trial is the first to compare four incen tive schemes for smoking cessation against usual care. It involved 2,538 employees of a US company, their relatives and friends. Two of the schemes targeted individuals and two
4 0 | N AT U R E | V O L 5 2 3 | 2 J U LY 2 0 1 5
© 2015 Macmillan Publishers Limited. All rights reserved
targeted groups of six participants who were incentivized through the collective perfor mance of the group. One individual and one group-based scheme provided rewards of US$800 for smoking cessation, and the other two schemes required refundable $150 depos its, together with a $650 reward, for successful participants. Acceptance of the reward-based schemes — assessed as the proportion of participants enrolling in the scheme to which they were assigned — was 90%, much higher than for the deposit schemes, at 13.7%. There were no differences in the acceptance of individual and group-based schemes. When comparing the entire cohort of participants, regardless of whether or not they enrolled on the scheme they were offered (with those not enrolling in the scheme assumed to have remained smokers), the quit rates in the intervention groups at 6 months ranged from 9.4% to 16% — all higher than the 6% quit rate in the usual-care group. Quitting was higher for the reward- than for the deposit-based schemes (15.7% versus 10.2%) and similar for individ ual and group-based schemes. Among the small proportion of participants who accepted the deposit schemes, 52% had quit at 6 months, compared with 17% of those
PETER DAZELEY/GETTY
NEWS & VIEWS RESEARCH
Figure 1 | Carrots work for smokers. Halpern et al.6 show that reward-based financial-incentive schemes are more commonly accepted by smokers than are deposit-based schemes, but that both approaches lead to higher rates of quitting than usual care.
on reward schemes. This outcome led Halpern et al. to conclude that use of deposits is better. However, the lack of interest in these schemes limits their usefulness, as does the absence of sustained effects. At 12 months, 6 months after the incentives were stopped, about 50% of quit ters were smoking again. Only those who had been enrolled in reward-based schemes, who retained quit rates of 7.5% for the individual and 8.7% for the group schemes, showed an advantage over the 3.4% quit rate at 12 months achieved through usual care. The problem of ‘gaming’ — faking being a smoker to qualify for enrolment on a scheme or being a non-smoker to remain on a scheme — is of particular concern when using financial incentives for smoking cessation. Although reports of quitting by participants in Halpern and colleagues’ trial were validated by checking cotinine levels (a component of tobacco) in their saliva, smoking at enrolment was confirmed in only a small minority of par ticipants. An estimated 20% of those recruited were non-smokers, although this did not affect the study’s findings. By contrast, a study of pregnant smokers who were offered participa tion in a financial-incentive scheme revealed that none of the 239 enrolled, all of whom were biochemically tested at baseline, were non-smokers7. This raises questions about the contextual features of incentive schemes that can be used to minimize gaming. Halpern and colleagues’ well-designed study, the largest known of its kind so far, makes a considerable contribution to the growing evidence about financial incentives
for smoking cessation. Two recent systematic reviews8,9, one of which8 includes the results of this trial, suggest that financial-incentive schemes can be effective in achieving sustained quitting, particularly when substantial incen tives are used and when offered to pregnant smokers. The effect of incentives is also dou bled in more-deprived populations9, suggest ing that this approach motivates greater change in people on low incomes, among whom rates of smoking are not only markedly higher but also resistant to reduction. This finding high lights the potential contribution of financial incentives for reducing the health inequalities related to smoking, including the significantly higher death rates from tobacco among poor people. Were such schemes a pill or a behavioural intervention not involving money, they would doubtless be included in the range of interventions on offer to smokers want ing to quit. We are, however, “funny about money” and in particular about “money out of place”10. Given their equal effectiveness to other interventions, financial incentives are less acceptable to the public, and in turn potentially to health professionals and policymakers, than other interventions for changing behaviours11. Why is there not greater acceptability of a seemingly cost-effective intervention for reducing the leading cause of preventable premature death worldwide? The low accept ability of paying people to stop smoking (or to lose weight or take medication) reflects sev eral concerns about ideas of fairness, including
‘coercing the vulnerable’, ‘rewarding the feckless’ and ‘not rewarding the responsible’. The first of these often arises in the context of incentivizing the disadvantaged, because it is assumed that such individuals are less free to resist monetary temptations. This com promises the use of incentives for addressing health inequalities. An example of the effect of the latter two concerns is a situation in the United States12, in which objections from non-smoking employees led to the rejection by management of a $750 reward scheme for smoking cessation (for which there was strong evidence of effectiveness) and the adoption of a $625 penalty deducted from smokers’ salaries — an intervention for which at the time there was no evidence of effectiveness. The acceptability of financial incentives can change: it increases with information about effectiveness and varies with incentive type (acceptability is lower for cash incentives and higher for voucher schemes)13. Nonetheless, financial interventions to change behav iour — whether they involve price increases through taxation or reward schemes — are much less acceptable than other, often less effective, interventions14. Studies are needed that explore the scale and duration of finan cial-incentive schemes in different popula tions of smokers, and that aim to identify the characteristics of schemes producing the greatest changes. In parallel with this, research is needed to develop ways of increas ing the acceptability of interventions that could significantly help to reduce the death toll from tobacco, particularly among poor individuals. ■ Theresa M. Marteau and Eleni Mantzari are in the Behaviour and Health Research Unit, University of Cambridge, Cambridge CB2 0SR, UK. e-mail: [email protected] 1. www.vox.com/2014/5/19/5727712/the-threedeadliest-drugs-in-america-are-all-totally-legal 2. Peto, R. et al. Br. Med. Bull. 52, 12-21 (1996). 3. Jarvis, M. J. & Wardle, J. in Social Determinants of Health (eds Marmot, M. & Wilkinson, R.) 240–255 (Oxford Univ. Press, 1999). 4. Jha, P. & Peto, R. N. Engl. J. Med. 370, 60–68 (2014). 5. US Department of Health & Human Services. The Health Consequences of Smoking — 50 Years of Progress: A Report of the Surgeon General, 2014 (HHS, 2014). 6. Halpern, S. D. et al. N. Engl. J. Med. 372, 2108–2117 (2015). 7. Ierfino, D. et al. Addiction 110, 680–688 (2015). 8. Cahill, K., Hartmann-Boyce, J. & Perera, R. Incentives for Smoking Cessation (Cochrane Library, 2015). 9. Mantzari, E. et al. Prevent. Med. 75, 75–85 (2015). 10. Sandel, M. J. What Money Can’t Buy: The Moral Limits of Markets (Farrar, Straus & Giroux, 2012). 11. Promberger, M., Brown, R. C. H., Ashcroft, R. E. & Marteau, T. M. J. Med. Ethics 37, 682–687 (2011). 12. Volpp, K. G. & Galvin, R. J. Am. Med. Assoc. 311, 909–910 (2014). 13. Promberger, M., Dolan, P. & Marteau, T. M. Soc. Sci. Med. 75, 2509–2514 (2012). 14. Diepeveen, S., Ling, T., Suhrcke, M., Roland, M. & Marteau, T. M. BMC Public Health 13, 756 (2013). 2 J U LY 2 0 1 5 | V O L 5 2 3 | N AT U R E | 4 1
© 2015 Macmillan Publishers Limited. All rights reserved
