The 20-year impact of tobacco price and tobacco control expenditure increases in Minnesota, 1998-2017.

INTRODUCTION: Tobacco control programs and policies reduce tobacco use and prevent health and economic harms. The majority of tobacco control programs and policies in the United States are implemented at local and state levels. Yet the literature on state-level initiatives reports a limited set of outcomes. To facilitate decision-making that is increasingly focused on costs, we provide estimates of a broader set of measures of the impact of tobacco control policy, including smoking prevalence, disease events, deaths, medical costs, productivity and tobacco tax revenues, using the experience of Minnesota as an example.
METHODS: Using the HealthPartners Institute's ModelHealth™: Tobacco MN microsimulation, we assessed the impact of the stream of tobacco control expenditures and cigarette price increases from 1998 to 2017. We simulated 1.3 million individuals representative of the Minnesota population.
RESULTS: The simulation estimated that increased expenditures on tobacco control above 1997 levels prevented 38,400 cancer, cardiovascular, diabetes and respiratory disease events and 4,100 deaths over 20 years. Increased prices prevented 14,600 additional events and 1,700 additional deaths. Both the net increase in tax revenues and the reduction in medical costs were greater than the additional investments in tobacco control.
CONCLUSION: Combined, the policies address both short-term and long-term goals to reduce the harms of tobacco by helping adults who wish to quit smoking and deterring youth from starting to smoke. States can pay for initial investments in tobacco control through tax increases and recoup those investments through reduced expenditures on medical care.

Introduction

Increasing tobacco price and investments in comprehensive tobacco control are high-impact evidence-based strategies for reducing smoking prevalence.[1, 2] To inform policy decisions, simulation models of United States (U.S.) populations have predicted policy effects on smoking prevalence, reductions in mortality, and medical costs.[2-9] For example, the U.S. Congressional Budget Office (CBO) simulated the 10-year impact of a federal cigarette tax increase, predicting a variety of outcomes that included mortality, medical costs, earnings, federal government revenue and federal spending.[9]

Most tobacco policy changes in the United States occur at state and local levels. State lawmakers may be motivated to support increases in tobacco taxes and/or tobacco control expenditures for various reasons, including improved health, reduced medial costs, improved productivity, and increased general revenue funds. Analyses of state-level policies have evaluated their impact on a variety of smoking-related behaviors,[1] and simulations have estimated the current and future impacts on prevalence, mortality, and medical spending,[10-14] but have not addressed the broader array of outcomes.

To expand the evidence base for formulating state-level tobacco control policy, we simulated the 20-year impact of tobacco price increases and increased investments in tobacco control in Minnesota from 1998 through 2017. From 1997 to 2016, adult smoking prevalence in Minnesota fell from 21.8% to 15.2%,[15] a 30% relative decrease. Part of this trend arose from earlier tobacco control policies, including the nation’s first statewide clean indoor air law in 1975, the nation’s first state-funded tobacco control program in 1985, and four state tax increases from 1985 to 1992. Significant policy changes after 1997 added to the trend, including strengthening the 1975 clean indoor air law to include all indoor public places, strengthening youth access laws, further increases in cigarette taxes and substantial increases in tobacco control expenditures. These changes were driven in part by ClearWay MinnesotaSM which was established through Minnesota’s tobacco settlement to implement tobacco control programs and research using 3% of the state’s tobacco settlement funds.[16] Since ClearWay Minnesota’s inception, state tobacco control expenditures increased by as much as 10-fold, varying by year, while cigarette prices increased nearly 4-fold in response to settlement costs and increases in federal and state tobacco taxes. Large tobacco price increases are an effective known strategy to reduce tobacco use. Tobacco control expenditures support the state’s free quit line, anti-tobacco media campaigns, tobacco education, communication, and community collaborations. Increased expenditures also support initiatives to reduce the harms of commercial tobacco to American Indian communities, and education and communication activities likely contributed to passing local clean air and Tobacco 21 laws and local restrictions on the sale of menthol cigarettes.

To provide a more comprehensive assessment of policy impacts, we used a detailed microsimulation model to simulate a broad range of outcomes, including smoking prevalence, disease events, mortality, smoking-attributable medical costs, productivity gains and state cigarette tax revenues within a 20-year timeframe. The results demonstrate the potential for health and economic gains in states that have not acted aggressively against tobacco.

Methods

We simulated changes in policies using the HealthPartners Institute’s ModelHealthTM: Tobacco MN–a microsimulation model.[17] The model simulates annual changes in cigarette smoking behavior over the lifetimes of individuals and estimates the health and economic consequences of cigarette smoking. The simulation model, data inputs, and policy parameters are described in S1 and S2 Supplements, including details on the use of databases and literature to inform the model. All planned model outcomes are reported in results. The model is capable of producing a wide range of intermediate and more granular outcomes such as number of quits and smoking initiations, lung cancer cases, and results by demographic factors that are not presented for brevity. The simulation does not include other tobacco products and the complex array of multiple product use that has developed in recent years.[18]

Demographics and smoking status

We chose 1997 as our base year to assess 20 years of change, 1998–2017. We simulated 1.3 million individuals with age, sex, race/ethnicity, and educational attainment distributions representative of the Minnesota population of all ages in 1997.[19] Each simulated individual’s insurance status and probabilities of changing insurance status over time varies with disability, employment and poverty status, as estimated for the U.S. from the U.S. Current Population Survey[19] and the Survey of Income and Program Participation.[20]

Youth smoking status is determined by its association with demographics as estimated from the first Minnesota Youth Tobacco Survey (MYTS) in 2000.[21] Year 2000 prevalence rates may produce a conservative baseline estimate of youth smoking in 1997, because national rates for youth were trending slightly downward at that time.[22] From MYTS we also estimated “net initiation” to represent the combined probability that one more youth becomes a smoker after accounting for quits during the year.

We estimated adult cigarette smoking status in 1997 from Minnesotans who responded to the 1996 or 1997 Behavioral Risk Factor Surveillance Surveys (BRFSS).[23] We combined two years of data to improve the precision of age- and sex-specific smoking behaviors, and in particular with cessation rates for which the sample is limited to current smokers. For ages 65 and older, we calibrated the initial smoking probabilities derived from BRFSS to be consistent with prevalence estimates from the first Minnesota Adult Tobacco Survey (MATS) in 1999.[24] We estimated cessation rates from combined 1996 and 1997 BRFSS data. Too few Minnesotans ages 18–24 were represented in BRFSS to yield reliable estimates. Therefore, we assumed that cigarette smoking status from ages 18 to 24 was the same as the model’s predicted rates for 25-year-olds. This assumption could understate peak adult prevalence during the lifetime but provides a reliable estimate of prevalence prior to the ages of high harms from smoking-attributable chronic disease. To assign relapse probabilities, we constructed a curve based on literature that reveals a declining probability of relapse with greater time since quit.[25-29]

The number of cigarettes smoked per day was estimated by age, sex, educational status and race/ethnicity for the U.S. population and calibrated so the model reproduces cigarette packs smoked per capita in Minnesota in 1997.[30]

Consequences of cigarette smoking

The model includes smoking-attributable cancers, and cardiometabolic and respiratory diseases identified in Smoking-Attributable Mortality, Morbidity, and Economic Costs (SAMMEC) estimates (S1 pages 33–46).[31] We obtained Minnesota death rates by age and sex for smoking-attributable conditions in 1996–1998 from Detailed Mortality Data.[32] We disaggregated these rates into never, current and former smokers using Minnesota adult smoking prevalence and relative risks of disease from the 2014 Surgeon General’s Report.[31] We approximated Minnesota smoking-attributable non-fatal disease rates using the inverse of case-fatality rates for the United States in an update of the previously described U.S. version of ModelHealthTM: Tobacco.[33]

Smoking-attributable medical costs measure the additional total cost of medical care of current and former smokers in excess of those of never smokers. To approximate smoking-attributable medical costs of Minnesotans (S1 page 17), we scaled estimates for current smokers derived for the United States from 2000–2010 linked Medical Expenditure Panel Survey and National Health Interview Survey (MEPS-NHIS) data[34] using the ratio of Minnesota-to-U.S. per-capita healthcare expenditures.[35] We inflation-adjusted these costs to 2017 U.S. dollars using the Medical Care Consumer Price Index.[36] MEPS and other claims data include former smokers who quit smoking only after diagnoses that lead to increased healthcare utilization after quitting.[37-39] Therefore, for former smokers, we fit an exponential function to the relationship of current and former expenditures based on time since quit, using the relationship between current and former smoker mortality risks reported by the CBO (S1 page 16).[9]

In the model, productivity losses reflect absence from work,[40] lower productivity at work,[40] and lost years of work, including unpaid household productivity (S1 page 19).[41] We scaled U.S. measures of productivity by the ratio of Minnesota to U.S. per-capita earnings.[42]

Simulation scenarios

We separately assessed the gains made by increasing investments in tobacco control expenditures in Minnesota and by increasing cigarette prices. We compared policy scenarios to the baseline scenario in which initiation and cessation rates are held constant at 1997 levels.

The baseline scenario predicts what the burden of cigarette smoking in Minnesota would have been if no policies, programs or trends had changed initiation and cessation rates from 1998 through 2017. This includes holding constant both changes aimed at reducing tobacco use and those aimed at increasing or maintaining tobacco use, such as tobacco industry marketing. Smoking prevalence falls over time in the baseline scenario, even with initiation and cessation rates held constant.

We had planned to analyze the effect of state clean indoor air legislation, but it was excluded because we were unable to simulate secondhand smoke exposure and its harms within the study timeframe. The analysis plan was not pre-registered.

Increased Tobacco Control Investments (ITCI) scenario

In the 1997 baseline year, $0.81 per capita was appropriated for tobacco control in Minnesota from state, federal and foundation sources (S2 Supplement of Table S2.4),[43] expressed in 2017 dollars.[44] That amount grew to $9.85 per capita in 2001 before falling back to $4.07 by 2014[43] after state-controlled tobacco control endowments were used to balance the state budget in 2004. Nevertheless, an additional $376 million ($US 2017) was invested in tobacco control from 1998 through 2014 than would have if per-capita appropriations had remained at 1997 levels.

Comprehensive tobacco control programs coordinate multiple strategies, such as large- and small-scale media, school-based education, quit lines, and programs to decrease tobacco’s accessibility to minors.[45] The effect of increasing investments in tobacco control in the simulation model is based on statistical studies that compare the association between tobacco control expenditures and smoking behaviors across US states. As described in detail in S2 Supplement, we reviewed studies that provide estimates of expenditure elasticities (the percent change in prevalence for each percent change in expenditure) that we could apply in the simulation model. The model estimates how annual percent changes to investments in tobacco control above 1997 levels[43] would modify the baseline initiation and cessation rates that define the baseline scenario. For the years 2015 to 2017, when appropriations data were not available at the time of analysis, we assumed that appropriations were maintained at 2014 per-capita levels.[43] We adjusted per-capita expenditures to 2017 dollars to remove inflation from the annual percent changes.

ITCI + Price scenario

Minnesota average retail cigarette prices increased $0.76 per pack in 1999, the year after the state tobacco lawsuit settlement. In subsequent years, federal taxes increased $0.77, and state taxes increased $3.10 unadjusted for inflation.[30] We created an ITCI + Price Scenario by adding to the ITCI Scenario changes in per-pack cigarette prices associated with tax increases and the conceptually similar increase in per-pack price following the 1998 Minnesota tobacco settlement.[16, 30] The difference in health and economic outcomes between this ITCI + Price Scenario and the ITCI scenario provides estimates of the impact of tax and 1998 settlement-associated price increases alone. Therefore, we assume that one policy does not impact the relative effectiveness on smoking behaviors of the other policy, either positively through synergistic effects or negatively through reducing the portion of remaining smokers or would-be smokers whose decisions are most amenable by policy.

Taxes in the model operate through price-prevalence elasticities (the percent change in smoking prevalence per one percent change in cigarette price), and price-intensity elasticities (the percent change in quantity smoked among continuing smokers per one percent change in price). These elasticities were drawn from literature with careful consideration of which studies provide the most appropriate price elasticity estimates as inputs to the simulation model, and on how the results of multiple studies should be combined to estimate price elasticity (see S2 Supplement). We assumed that tax increases have a one-time impact on adult cessation probabilities in the first year and a permanent impact on youth initiation. We inflation-adjusted each year’s average tobacco price and each increase in price to 2017 U.S. dollars.

Sensitivity analysis

We explored the influence of model parameters on adult prevalence, deaths, and medical care costs in sensitivity analysis. We explored the potential impact of any systematic bias in measuring cessation probabilities by 25% in either direction. Similarly, we examined the effects of systematic bias of up to 25% in the relative risks of disease of former and current smokers compared with never-smokers. Such bias may arise from the methods or data sources used to estimate the relative risk of disease for the U.S. population, or in applying those relative risks to the Minnesota population.

We explored systematic bias of up to 35% in estimating medical costs. Costs are based on MEPS, which produces lower estimates of health care costs per capita than National Health Expenditure Accounts (NHEA).[46] Therefore, the predicted cost-savings are more likely to be understated than overstated. We also explored the impact of alternative estimates of smoking costs derived to support the 2014 Surgeon General’s Report[47] because they align more closely with NHEA costs.

We increased and decreased by 50% the estimates of expenditure and price elasticities. To compare tobacco control investments to resulting savings, we discounted the time sequence of investments in tobacco control and medical care costs at 3% per year.

Finally, we constructed the Price Increase Scenario for sensitivity analysis, in which the price increases associated with increased taxes and the 1998 tobacco settlement are added to the baseline scenario. We used this scenario to explore the impact of estimating increased investments in tobacco control marginal to price increases and to explore the impact of not estimating the impact of price increases marginal to increased tobacco control investments.

Results

The model estimates that, due to increased tobacco control expenditures, 7,400 fewer youth and 139,000 fewer adults smoked cigarettes in 2017. Although investments fell after 2001, the impact on youth smoking stays at about 1 percentage point through 2017 (Fig 1; shown by the distance between the Baseline and ITCI scenarios for youth). In contrast the simulation estimates that adult prevalence continued to decline as a result of ITCI, even as investments declined from their peak, resulting in a 3.4 percentage point decline by 2017. This reflects additional quits among adults each successive year and additional non-smoking youth aging into the adult population.

Fig 1

Predicted trends in youth and adult prevalence in Minnesota under baseline and policy scenarios.

The impact of the simulated price changes is shown by the difference between the ITCI and the ITCI + Price Increase scenarios (Fig 1). For youth, the estimated impact of price increases was greater than for investments in tobacco control. For adults, the opposite result was found.

Table 1 shows the 20-year cumulative outcomes for the main scenarios and the policy impacts. Annual results for each scenario are shown in S3 Supplement. The model estimates that $1.6 billion in smoking-attributable medical care costs was saved through 2017 as a result of increased investments in tobacco control. In addition, there were $1.2 billion gains in paid and unpaid productivity. The model estimates that these economic gains were accompanied by reductions in smoking-attributable cancer cases (4,100 fewer), combined cardiovascular and diabetes hospitalizations (24,600 fewer), respiratory disease hospitalizations (9,800 fewer) and deaths (4,100 fewer). If tax rates had stayed at 1997 levels, state tobacco tax revenues would have been an estimated $800 million lower over the 20 years in 2017 dollars as a result of reduced smoking prevalence that followed increased investments in tobacco control. However, the reduction in medical cost savings would have been twice as large as the lost tax revenues (see also S3 Supplement of Table S3.7).

Table 1

Cumulative impact of tobacco control policies, Minnesota 1998–2017.

	Scenarios			Policy impact
Outcome	Baseline	ITCI	ITCI + Price	ITCI1	Price2	Combined3
Youth smoking prevalence, ages 9–174	13.7%	12.7%	7.5%	-1.0%	-5.1%	-6.1%
Adult smoking prevalence, ages 18+4	17.8%	14.4%	11.6%	-3.4%	-2.8%	-6.2%
Person-years of cigarette smoking, all ages	16,803,500	14,754,100	13,084,000	-2,049,400	-1,670,100	-3,719,500
SA cancer cases	174,700	170,600	169,000	-4,100	-1,600	-5,700
SA CVD and diabetes hospitalizations	1,489,300	1,464,700	1,455,500	-24,600	-9,300	-33,800
SA respiratory disease hospitalizations	442,900	433,200	429,400	-9,800	-3,700	-13,500
SA deaths	186,000	182,000	180,300	-4,100	-1,700	-5,700
SA medical costs (millions of 2017 $US)	27,900	26,300	25,600	-1,600	-700	-2,300
Productivity (millions of 2017 $US)	4,808,300	4,809,500	4,810,400	1,200	900	2,100
Cigarette tax revenues5 (millions of 2017 $US)	5,400	4,600	7,800	-800	3,200	2,400

1ITCI Scenario compared to Baseline.

2ITCI+Price scenario compared to ITC.

3ITCI + Price Scenario compared to Baseline.

4Prevalence rates shown are for 2017; all other figures show cumulative estimates from 1998 to 2017.

5Tax revenues are shown using 1997 rates in the Baseline and ITCI scenarios (to show the effect of ITCI alone) and with actual tax rates in the ITCI + Price scenario. Cumulative revenues in the ITCI scenario using actual tax rates would be $10,400 million. SA = Smoking-attributable. CVD = Cardiovascular disease. ITCI = Increased Tobacco Control Investments.

The increases in tobacco prices reduced smoking-attributable medical care costs by an estimated $700 million and increased productivity by $900 million, while preventing 1,600 cancer cases, 9,300 cardiovascular and diabetes hospitalizations, 3,700 respiratory disease hospitalizations and 1,700 smoking-attributable deaths. Although tobacco price increases reduced cigarette sales, state tobacco tax revenues increased an estimated $3.2 billion in 2017 dollars over those 20 years, due to the increase in state tax rates behind some of the price increases.

The price changes following the tobacco settlement and tax increases had a larger impact on youth prevalence from 1998 to 2017 than did the increases in tobacco control expenditures (Fig 2), while the impact on adult prevalence was approximately the same in the two scenarios, indicating the potential for price changes to have greater impact in the future.

Fig 2

Relative decline in tobacco use and harms by policy compared to baseline scenarios from 1998 to 2017 in Minnesota.

SA = smoking attributable. CVD—cardiovascular disease. hosp = hospitalization.

Results of sensitivity analysis and secondary economic analyses

Changing baseline cessation rates by 25% changed the impact on adult prevalence in 2017 by 10% in relative terms compared to the base case (0.3 and 0.4 percentage points for decreased and increased baseline cessation rates, respectively), and had a similar relative impact on deaths and medical care costs prevented (Table 2). Changing the effectiveness of ITCI by 50% had a nearly proportional effect on adult prevalence in 2017, but less impact on deaths prevented and medical care costs saved. Using alternative cost estimates that align with those in 2014 Surgeon General’s report increased savings by about 35%. Discounting the stream of medical costs to present value in 1997 reduced the measured costs prevented through 2017 by 34%, to $1.1 billion. The discounted additional investments in tobacco control over 1997 per capita levels were $270 million over 1998 to 2017 (S3 Supplement of Table S3.7). Therefore, discounted savings realized through 2017 were 3 times greater than additional investments made through 2017. In addition, discounted medical costs saved remain greater than discounted investments even when the effectiveness of additional investments is reduced by 50%.

Table 2

Sensitivity analysis.

Scenario	Increased tobacco control investments; ITCI compared to Baseline Scenario			Increased cigarette prices; ITCI + Price compared to ITCI Scenario
	Percentage point change in adult prevalence	Change in SA deaths	Change in SA costs (millions of 2017 $US)	Percentage point change in adult prevalence	Change in SA deaths	Change in SA costs (millions of 2017 $US)
Base case	-3.4%	-4,100	-1,600	-2.8%	-1,700	-700
Baseline cessation rates +25%	-3.1%	-3,500	-1,500	-2.8%	-1,800	-800
Baseline cessation rates -25%	-3.8%	-4,500	-1,700	-2.8%	-1,200	-600
Relative risk of SA disease +25%	-3.4%	-5,000	-1,600	-2.8%	-1,900	-700
Relative risk of SA disease -25%	-3.4%	-3,200	-1,600	-2.8%	-1,200	-700
SA medical costs +35%	-3.4%	-4,100	-2,200	-2.8%	-1,700	-900
SA medical costs -35%	-3.4%	-4,100	-1,100	-2.8%	-1,700	-400
Alternative SA medical costs	-3.4%	-4,100	-2,200	-2.8%	-1,600	-900
Effect of increased ITCI +50%	-4.4%	-5,100	-2,100	na	na	na
Effect of increased ITCI -50%	-2.3%	-2,800	-1,200	na	na	na
ITCI effect marginal to price effect	-2.9%	-4,000	-1,600	na	na	na
SA medical costs discounted 3%	-3.4%	-4,100	-1,100	na	na	na
Effect of increased prices +50%	na	na	na	-3.8%	-2,300	-900
Effect of increased prices -50%	na	na	na	-1.8%	-1,100	-500
Price effect not marginal to ITCI effect	na	na	na	-3.3%	-1,700	-700

SA = Smoking-attributable. ITCI = Increased Tobacco Control Investments. na = not applicable

Discussion

The simulation results suggest that from 1998 to 2017, the level of investments in tobacco control reduced tobacco harms in Minnesota more than did the increases in the price of cigarettes that occurred in the state. Minnesota’s additional investments in tobacco control reduced smoking-attributable deaths by an estimated 4,100 and smoking-attributable medical costs by $1.6 billion. Medical cost savings were twice the size of the reduction in tax revenue that were attributable to declining cigarette sales with increased tobacco control investments. Over the same period, Minnesota’s price increases reduced deaths by an additional 1,700 and reduced medical costs by $700 million. With increasing tax rates, cigarette tax revenues increased more than the expenditures on tobacco control despite declining sales attributable to both increased tobacco control expenditures and cigarette prices.

The health impacts from increased investments in tobacco control seem disproportionate to those of price increases given their relative impact on adult prevalence in 2017. However, the simulation estimates that the increased investments in tobacco control had a larger impact on adult prevalence in early years (Fig 1). Therefore, within the 20-year period, increased investments in tobacco control accumulated more years of people living as former smokers and a longer average time since quit. Minnesotans who were younger than 18 anytime between 1998 and 2017 remain decades away from their peak risk of smoking-attributable disease, so the large impact of price increases on youth prevalence did not translate into larger reductions in the harms of tobacco by 2017.

Our time horizon does not capture all of the benefits of changes between 1998 and 2017. If the analyses were projected into the future, we would see greater benefits as youth who avoided smoking initiation become adults, and as adult former smokers place more years between their last cigarette and their current health risk. After 2017, the impact of price changes could exceed those of increased investments in tobacco control.

The range of outcomes reported here highlight the impact of tobacco control policy and its relevance to the varied interests of individual legislators and their constituents. Most prior studies of state tobacco policy report effects on prevalence and mortality, and less frequently, medical costs. This study adds effects on disease events, productivity, and state cigarette tax revenues to the literature. For Minnesota, Levy et al. estimated that price increases during 1994 to 2011 reduced smoking prevalence by 13.5% and 13.6%, in relative terms, for men and women ages 15 and above, respectively.[10] In our analysis, smoking prevalence among adults decreased 15.7% in relative terms in response to price changes from 1997 to 2017 (Fig 2). The difference in estimates may be driven largely by the $1.60 tax increase in 2013 and subsequent inflation indexing of state tobacco taxes at the end of the time period. Levy et al. also estimated the impact of tobacco control expenditures inclusive of media campaigns. The relative declines in prevalence they estimate for 1994 to 2011, 6.2% for men and 5.6% for women, were lower than our estimate for adults from 1997 to 2017, 19.1% (Fig 2). Levy et al. estimated expenditures in categories of intensity rather than as a continuous variable and used effectiveness estimates corresponding to those categories rather than expenditure elasticities. Differences in model structure also are likely to contribute to differences in results.

We previously estimated the health and economic gains that Minnesotans realized between 1998 and 2017 from reductions in tobacco prevalence during that period, without regard to the source of prevalence decline.[17] The ITCI + Price Scenario in the current study produced a lower adult prevalence in 2017 (11.6%) than was used to estimate the impact of all reductions in prevalence regardless of cause in the prior report (13.5%). This seemingly incongruent result may have several sources. Tobacco industry marketing efforts to counter the impact of investments in tobacco control and tax increases may have prevented initiation rates from declining and cessation rates from rising as much as they otherwise would have.[48, 49] The tobacco industry spent an average of $165 million on tobacco marketing in Minnesota each year from 1998 to 2017 as estimated from national expenditures and state tobacco sales.[50] The price and/or expenditure elasticities we derived from literature could be biased upward. While there is extensive literature indicating that both investments in tobacco control and increasing prices reduce tobacco use, the size of the reductions obtained from literature may be imprecise. In particular, model results by age group (not shown), indicate that prevalence is lower in the ITCI scenario than actual prevalence rates between ages 25 and 64, but not in younger or older adults. The effect sizes from the literature may not generalize to the size of additional tobacco control investments or tobacco price increases experienced in Minnesota between 1998 and 2017, or the Minnesota population may have responded differently. In addition, our effect sizes are mathematically adjusted to avoid double counting of relapse as described in S2 Supplement and the proper degree of adjustment is uncertain. Finally, the model assumes relapse rates are equal for all quits due to a lack of data and technical limitations to assigning a reason for each quit. Policy-induced quits could have a higher relapse rate, in which case the simulation would somewhat overstate long-term policy effects.

Other limitations should also be noted. The simulation does not include all harms of smoking, including direct costs outside the medical care sector, and therefore may understate the benefits of reduced tobacco use. For example, despite a requirement that only ‘fire safe’ cigarettes be sold in the state, in 2018 careless smoking was faulted in 10 of 30 fire fatalities with known cause in Minnesota.[51] In the U.S., smoking-related fires were responsible for $1.36 billion in direct costs such as property damage and firefighting expenses, and $1.16 billion in productivity losses in 1995 (the most recent comprehensive estimate we found).[52] Higher cigarette prices are associated with fewer residential fires and fire deaths.[53] Our estimates also exclude benefits of reducing secondhand smoke which contributes to cancer, respiratory disease, and cardiovascular disease, low birth weight, and their associated medical costs.[31] Max et al. estimated that the portion of attention deficit hyperactivity disorders associated with secondhand smoke costs the U.S. health care system $2 billion and the U.S. education system $9 billion when they used serum cotinine levels as a biomarker for secondhand smoke exposure.[54]

The simulation also excludes the impact of policies on use of tobacco products other than combustible cigarettes. Expenditures on tobacco control are likely to reduce other forms of tobacco use, which would cause us to understate the total benefits of ITCI. While Minnesota taxes all forms of tobacco, including e-cigarettes, higher taxes on cigarettes (other taxes held constant) might lead tobacco users to substitute other forms of tobacco. While other tobacco products may produce less harm, they are not without risk, and excluding them may cause us to slightly overstate the impact of cigarette price increases.

S1 and S2 Supplements detail model inputs, including the literature reviewed, decisions on study inclusion and how study results are used in the model and database analyses. Like all simulations, ours is limited by the precision of model inputs, such as using self-reported tobacco use and making adjustments to US data on costs and productivity to approximate Minnesota-specific values. Our assumption on tobacco use for adults less than 25 years of age would cause us to simulate too few former smokers if peak adult prevalence occurs before age 25. The most uncertain inputs to the model are the elasticity estimates for increased investments in tobacco control and increased cigarette prices. We used a broad range of values in sensitivity analysis to reflect the uncertainties noted above. Although estimating return on investment for ITCI was not a study goal, sensitivity analysis results indicate that even at lower estimates of effectiveness, investments in tobacco control were more than offset by reductions in smoking-attributable medical expenditures, indicating a positive return on investment during the limited time horizon, even when the value of improved health and productivity gains were excluded from the calculation. The additional state tobacco tax revenues collected during this period were greater than additional investments in tobacco control, even though the price increases reduced cigarette sales.

Conclusion

This study enhances our understanding of the impact of two key policy levers identified by the CDC as best practices for U.S. states to reduce the harms of tobacco–increasing tobacco prices and increasing investments in tobacco control.[55] Investments in tobacco control can more than pay for themselves in the mid to long-term. Although Minnesota did not fund investments in tobacco control through a tax increase, the state’s experience indicates that revenues from a tax increase can fund the upfront investments even as these policies reduce cigarette packs sold. Strategically, the combination of the policies targets both short-term and long term harms of tobacco by supporting established smokers in their quit attempts while deterring youth from starting smoking through price increases and youth-focused strategies such as media campaigns and reducing youth access. Therefore, policy makers who seek to improve health and reduce costs would be wise to consider aggressively implementing both of these policies. Comprehensive state tobacco control programs that combine price increases, investments in tobacco control and policies that prevention initiation, and help smokers who want to quit have been found to be the most effective strategy for reducing the harms of tobacco.[1, 55]

HealthPartners Institute ModelHealthTM: Tobacco MN model documentation model version 1.2.

(DOCX)

Click here for additional data file.

Background and model implementation of policies for Minnesota tobacco policy analyses.

(DOCX)

Click here for additional data file.

Annual results tables.

(DOCX)

Click here for additional data file.

3 Jan 2020

PONE-D-19-28016

The 20-year impact of tobacco price and tobacco control expenditure increases in Minnesota, 1998-2017

PLOS ONE

Dear Dr. Maciosek,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Feb 17 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Stanton A. Glantz

Academic Editor

PLOS ONE

Journal Requirements:

1. When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2.  Thank you for stating the following in the Financial Disclosure section:

[AL, MM and ZX received funding through a research contract between their employer and Clearway Minnesota (http://clearwaymn.org). Coauthors from Clearway Minnesota participated in study design and manuscript preparation.]

We note that one or more of the authors are employed by a commercial company: 'Clearway Minnesota'.

Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form.

Please also include the following statement within your amended Funding Statement.

“The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

2. Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

Within your Competing Interests Statement, please confirm that this commercial affiliation does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests) . If this adherence statement is not accurate and  there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

* Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

3. Please include a new copy of Table 1 in your manuscript; the current table is difficult to read. Please follow the link for more information: http://blogs.PLOS.org/everyone/2011/05/10/how-to-check-your-manuscript-image-quality-in-editorial-manager/

Additional Editor Comments (if provided):

As you revise the paper, please pay particular attention to the suggestions that would make the results and implications clearer to policy makers. At the same time, including more of the analytical details in the body of the paper can also be accommodated, so long as there are clear statements of the policy implications in the abstract, introduction and discussion highlighted with appropriate subheads.

Like most work in this area, the paper concentrates on health cost savings. Reductions in tobacco use also have important cost implications for other sectors of government and society in general. For example, Wendy Max's group has shown that the cost savings to the educational system by reducing AHDH related to secondhand smoke exposure among youth are substantially larger than the direct medical costs (Childhood secondhand smoke exposure and ADHD-attributable costs to the health and education system. Max W, Sung HY, Shi Y. J Sch Health. 2014 Oct;84(10):683-6. doi: 10.1111/josh.12191). There are other issues, such as fires that are also impacted. I am not suggesting that the authors expand the scope of their paper, but they should at least mention these other economic impacts because they are important to policy makers and show that the estimates in their paper are something of a lower bound.

The authors need to recognize that their numbers are estimates and report their results rounded to 2 or 3 significant digits, based on the precision of the input variables in the model.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This study assessed the 20-year impact of tobacco price increases and increased tobacco control expenditure during 1998-2017 in Minnesota using a microsimulation model developed by the authors. Different from most of similar studies in the literature, this study include a broader array of outcome: smoking prevalence, disease events, mortality, smoking-attributable medical costs, productivity gains, and state cigarette tax revenues. This a well-written article with details of their models and simulation methodologies presented in the appendices. Although this study made many assumptions about their model parameters, it provided sensitivity analyses to explore the influence of a wide range of changes in model parameters. This study demonstrated that increasing tobacco prices and investments in tobacco control simultaneously is effective to reduce the harms of tobacco use and is cost-saving in the mid- to long-term. I only have several minor comments.

1. Page 7, the methodology of estimating the consequences of cigarette smoking was not explained sufficiently. More description is needed regarding whether and how an econometric model was developed to estimate the smoking-attributable medical costs for current smokers. According to Table S3.1 under the baseline scenario, smoking-attributable medical costs was $1.324 billion in 2017. It would be great if the authors can add some discussion to compare their annual SA medical cost estimates with similar studies in the literature.

2. Page 9, lines 151-153: It is not clear whether the price increases mentioned here ($0.76, $0.67, and $3.10) reflect constant dollar or nominal term. For example, Line 152 indicates that in subsequent years after 1999, the federal excise tax on cigarettes increased by $0.67 per pack. However, during that period, the federal cigarette taxes increased 3 times: by $0.10 per pack in 2000, by $0.05 per pack in 2004, and by $0.61 per pack in 2009. The sum of these increases is $0.76 which does not match to $0.67.

3. Page 9, lines 155-157: “The difference in health … between this ITC + Price Scenario and the ITC scenario provides estimates of the impact of tax increases alone”. This sentence is not correct because this difference also reflects the impact of the price increase due to the Minnesota tobacco lawsuit settlement.

4. Page 11, lines 188-189. The reduced numbers of smokers in 2017 (7,434 for youth and 138,123 for adults) due to ITC do not add up to the number shown in Table S3.4 (146,402).

5. Page 11, lines 201-205; page 12, line 206 and lines 211-203: All the numbers presented in these sentences do no match to the numbers shown in Table 1 and Tables S3.4.

6. Page 12, line 207 and line 215: Add “(data not shown”) after “$784 million”, and similarly after “$3.20 billion”.

7. Page 13, the last row “Base case”: the number shown in the last column does not match to the corresponding cell in Table 1.

Reviewer #2: This manuscript presents the findings of a microsimulation to assess the impact of tobacco control policies, including tax increases and increased expenditures on tobacco control, in Minnesota during the twenty year period spanning 1998-2017. The study found that that increased expenditures on tobacco control prevented 37,421 cancer, cardiovascular, diabetes and respiratory disease events and 3,839 deaths over 20 years. Additionally, increased prices prevented an additional 14,702 additional events and 1,662 deaths. The authors conclude that both the net increase in tax revenues and the reduction in medical costs were greater than the additional investments in tobacco control, and that states can pay for initial investments in tobacco control through tax increases and recoup those investments through reduced healthcare expenditures.

This manuscript is generally well written, the analytic approach is free from any major methodological pitfalls, and would make a meaningful contribution to the scientific literature. Nonetheless, the manuscript could be improved, including more prominent recognition of various limitations of the analysis. Specific recommendations for revision are as follows:

1. Abstract. Page 2. Introduction. The first sentence of the abstract understates the considerable body of scientific evidence on the efficacy of tobacco control programs; specifically, the language needs to more clearly state the direction of the effect. They don’t just “influence” tobacco use and related health and economic outcomes – they reduce these outcomes.

2. Introduction. Page 4. Third paragraph. The third sentence references “prior initiatives” and “significant policy change during this period.” However, it’s not clear to a lay reader what this means, particularly those without tobacco control expertise. A sentence should be added that clearly articulates what policies were actually implemented.

3. Introduction. Page 4. Since the manuscript only focuses on tobacco control program expenditures and price increases (i.e. and not other specific proven strategies such as smoke-free policies), it would be helpful to better set the stage for inclusion of these variables in the Introduction, including the focus on increased price. One way to accomplish this would be to reinforce that tobacco control expenditures support a variety of efforts, including state and community programs that influence policy adoption, and that increasing price has previously been noted as the single most effective intervention for reducing consumption. That way you better set the stage for the reader around the measures focused upon in the manuscript.

4. Introduction. Page 5. First paragraph. The last two sentences of the Introduction section are implication and discussion points, not Introductory content. They should be deleted here, and reinforced in the Discussion section.

5. Methods. Page 5. Demographics and smoking status. It’s not clear what specific age groups from the “Minnesota population in 1997” the data were adjusted to – was it all individuals, those 18+, or something else?

6. Methods. Page 6. The paragraph beginning “we estimated adult cigarette smoking” lacks necessary precision for the reader to understand the rationale for the employed analytic approach. For example, it’s not clear why 1996 and 1997 data were combined (sample size issue?). Additionally, it’s not clear what is meant by “calibrating the initial smoking probabilities”. Furthermore, it’s not apparent how the 18-24 year old age group, for which there was insufficient sample, could be predicted using estimates for 25-year olds, which reflects only a single age year; did you mean some range that began with 25 as the lower bound? If sample size were an issue, it would seem that you could consider adding multiple years of data to afford sufficient sample for 18-24 year olds, which is a critical smoking initiation demographic. If you can’t this needs to be duly noted as a limitation of the analysis.

7. Methods. Page 7. Consequences of Cigarette Smoking. More clarity needs to be provided in the first paragraph of this section to articulate what smoking-attributable diseases were included. The current framing suggests it could be some, but not all. To clarify, it might be more prudent to place the onus of inclusion on outcomes found to be significantly associated with smoking in the 2014 Surgeon General’s Report, rather than focusing on SAMMEC, which is no longer in existence.

8. Methods. Page 8. It’s unfortunate that the authors could not include smoke-free policies in the analysis, which was undoubtedly a contributor to both declining secondhand smoke exposure, as well as reduced smoking, in Minnesota during the assessed period. This needs to be stated as limitation of the manuscript (along with other proven strategies that couldn’t be included). Additionally, the text in this paragraph of the Methods should use the more commonly recognized term of “secondhand smoke exposure” instead of “secondary smoke exposure”. It’s also not clear what is meant by the statement “the analysis plan was not pre-registered.”

9. Methods. Page 8. It’s strongly recommended that the authors identify another abbreviation than “ITC” for “Increased Investments in Tobacco Control.” The abbreviation “ITC” has long been associated with the International Tobacco Control Survey in the tobacco control field, and thus, using this abbreviation may cause confusion when findings are reported elsewhere or taken out of the direct context of this study. Instead, authors could consider something like Increased Tobacco Control Investments, or “ITCI”.

10. Methods. Page 9. It’s questionable whether it is safe to assume that one policy does not impact the relative effectiveness on smoking behaviors of the other policy. Increased expenditures for tobacco control programs include state and community interventions, as outlined in CDC’s Best Practices for Comprehensive Tobacco Control Programs. Price increases are considered a product of those programs. As such, increased expenditures for programs would be associated with increased momentum around pricing strategies like excise taxes. Since there’s not much else the authors can do post hoc, at the least this should be duly noted as a limitation of the analysis.

11. Discussion. Page 14. It is strongly recommended that the authors not pit individual interventions against one another. The framing around investments being more impactful than price would surely be used by opponents of cigarette excise taxes to state that they aren’t needed because they aren’t as effective. The major take home here is really that both are effective and important as part of a comprehensive approach to tobacco prevention and control at the state level.

12. Discussion. Page 15. The authors appropriately note that tobacco industry marketing efforts could counter the impact of tobacco control investments. However, this point could be made clearer for lay readers if the most recent data from the Federal Trade Commission were cited. In 2017, $9 billion was spent on advertising and promotion of cigarettes— more than $1 million every hour.

13. Discussion. Page 16. The manuscript needs a paragraph that clearly articulates the limitations of the employed analytic approach, including several of the items noted previously in this review (e.g. lack of smoking estimates for 18-24 year olds, omission of full scope of policies such as smoke-free, assumption about independent effect of price and tobacco control program expenditures, etc).

14. Discussion. Another key point to reinforce in the Discussion is that this analysis only accounts for cigarette smoking, not all tobacco products, including other combustible products, smokeless tobacco products, and e-cigarettes. During the later part of the assessed period, particularly from 2011-2017, massive increases in the use of e-cigarettes were observed among youth – both nationally and in Minnesota. Therefore, this model doesn’t account for the full scope of tobacco products being used, as well as associated health risks/costs. Cigarette smoking is certainly responsible for the overwhelming burden of death and disease caused by tobacco use, but these other products are not without risk. Accordingly, this needs to be explicitly stated as a limitation of the analysis, and ideally, a reference to the diversification of the tobacco product landscape should also be made in the narrative elsewhere in the manuscript.

15. Discussion. Another essential point that is briefly referenced, but should be more prominently focused upon, in the manuscript is that a comprehensive approach is warranted. There is no single panacea, and for several decades, tobacco control has most effectively functioned through a comprehensive lens that addresses the various drivers influencing initiation and cessation. As currently framed, the manuscript could be misinterpreted as suggesting that either expenditures and price are effective alone. It’s also important to reinforce that it’s not just making the expenditures, but what those dollars are actually used for; large amounts of dollars allocated to ineffective initiatives or strategies will do little for preventing initiation and reducing consumption at the population level. The expenditures needs to be focused of proven strategies for which evidence of efficacy and cost-effectiveness have been established.

Reviewer #3: This is an important paper with findings relevant to continuing policy decisions about the investment of millions of tobacco settlement dollars. The economic analyses are detailed and comprehensive, and provide the basis for important policy discussions and conclusions. As such, the strength and details of these analyses are needed. However, this leads to a key question: what is the primary audience of this paper? The decision appears to be health policy makers, and not health economists. The details in the Supplemental Files are dense but important, and if more were included in the paper it would become more of a health economist paper. However, as written, many of the critical technical details of the model, parameter definitions and selection, sensitivity analysis, and technical aspects of the results are left to the Supplemental Files. Even so, the language of the paper and topics discussed still focus too much on economic issues rather than presenting key findings most relevant to the policy makers. For example, while noted in the discussion that "estimating return on investment for increased ITC was not a study goal..." (lines 290-291), the findings of a positive return on investment even during the limited time horizon is very important, and should not be buried deep in the Discussion. Similarly, the sensitivity analysis is important to these potential return on investment discussions and conclusions, but as presented, these analyses are written more for economists, and as such, will be too confusing to most policy makers. The critical link between the sensitivity analyses and return on investment conclusions needs to be clearly defined when introducing them (line 226). Also, while the Figures (1 and 2) are good at communicating the impact of ITC and Price, the specific data in Table 1 and Supplemental Files S3 are not presented well in the text for policy makers, and especially not summarized well in Abstract. The focus is on deaths and SA incidents averted; however, the bottom line for many policy makers is dollars saved. The medical cost savings and productivity gain estimates are unique to these analyses and should be given primary emphasis, along with the resulting return on investment calculations based upon them.

Also, the wealth of literature review, technical discussions, and econometrics in Supplemental Files S1 and S2 are very impressive. As noted above, if the primary audience of the paper is health policy makers, then these details need to stay there. However, the descriptions of the technical approaches in the main paper should be reviewed and considered: are they communicating the technical quality of the work presented in the Supplemental Files to the policy makers as effectively as possible? In many places, the reader should be specifically referred to the excellent presentation of the issues and parameter selection criteria in S1 and S2.

More detailed comments:

Lines 43-48: I believe this is the only microsimulation model -- and the others, particularly the Levy model is a relatively crude macrosimulation model. This distinction should be emphasized.

Lines 59-65: the potential of these findings to provide data needed for cost-effectiveness and return on investment calculations should be introduced and emphasized.

Line 70: the details in literature review supporting selection of data inputs, policy parameters, and details in model should get more emphasis here.

Lines 128-130: The discussion of clean indoor air legislation should be revised. The specific impact on CIA on health effects of reduced exposures to secondhand smoke are not included (as noted); however, the policy impact of CIA on other outcomes by default are included in the ITC. While these are not specifically included in the model parameters, their impact is captured in the overall results. This should be noted.

Lines 131-137: The details of annual tobacco control investments are provided in Supplemental Files. Guide readers to these details. This also could apply to other data included in the model and only shown in detail in Supplemental Files; i.e., annual excise tax levels and average annual retail cigarette prices.

Lines 167+: as noted above, the important relevancy of these simulations for explaining return on investment after adjusting for possible biases and reduced parameter estimates for impact of ITC or price, i.e., expenditure and price elasticities, etc.

Line 192: "...due to lagged effect of prior investments." will not be understood my almost all readers as being primarily due to two model decisions: first, only one time impact of price increases applied at time of tax increase (standard econometrics) for adults, but impact on youth extended; and second, the use of cumulative funding with 25% discount (a la Farrelly et al models). Maybe this is too technical a point to discuss in results, but might merit being raised in discussion.

Line 209: the detailed impacts are shown in S3, but more of them could be shown in an extended Table 1. Additionally, details on annual ITC should be linked with results in S3 to highlight return on investments. Also, S3 could show more details on annual state tax revenues under Baseline Scenario as well as under actual tax increases. Finally, it would be outstanding to see a combined return on investment table combining tax revenues, ITC investments, and SA medical costs. Such a table could have very important policy implications.

Table 1: something does not look right. I did not search out all the details in Supplemental Files, but the deaths do not look right. SA cancers (incidents) and SA deaths are almost identical; however, SAMMEC estimates are that only about one-third of SA deaths are due to cancer. This needs to be checked, since SA deaths should be much higher than SA cancer incidents and deaths.

Figure 2 (lines 217-224): emphasis on percent changes, but policy impact in $s only shown in Table 1. Figure could be 2a (current) and 2b with numbers (number of youth prevented, increase number of adult quitters, SA incident cases prevented, SA deaths averted, and most importantly, SA costs saved (in $s), productivity gains (in $s) and (to be computed) total savings (SA medical averted and productivity gains).

Lines 225-237: as noted above, the impact of sensitivity analyses on return on investment calculations needs more emphasis. Lines 233-236 briefly discuss these findings; but, these calculations need much more emphasis and should be provided in greater detail; maybe even added as additional columns in Table 2.

Lines 239-255: this discussion is relevant, but may not strike the right balance. Make it clear from the first sentence that both ITC and price increases are needed: they are complimentary in timing and level of impacts produced.

Lines 256-267: discussion of Levy et al model results relevant, but this is a microsimulation model that differs in many ways from their model beyond parameter estimates.

Lines 287-296: discussion of uncertainties and imprecisions in model parameters and inputs is always relevant. The details in S1 and S2 provide very good discussion of how parameters and model specifications were done. Point readers to these discussions and reviews. Yes, uncertainties about elasticities, especially about ITC, limit model predictions. Thus, the need for good sensitivity analyses (as was done). Nevertheless, the positive return on investment calculations were found, as well as the positive revenue gains for tax increases (which need more emphasis with more specific numbers provided, even thru this has been documented in many prior econometric analysis).

FINALLY, it is suggested that future analyses estimate the even greater impacts (on all outcome measures, but particularly SA cost savings) if ITC levels would have been higher (like full BP levels or even 50% of BP level) or would not have been reduced in 2004.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

18 Feb 2020

We appreciate the reviewers and editor volunteering their time to help us improve the manuscript. We discuss each suggestion below.

Additional Editor Comments (if provided):

As you revise the paper, please pay particular attention to the suggestions that would make the results and implications clearer to policy makers. At the same time, including more of the analytical details in the body of the paper can also be accommodated, so long as there are clear statements of the policy implications in the abstract, introduction and discussion highlighted with appropriate subheads.

We appreciate that the editor recognized the conflict between suggestions to focus the paper and requests to add both technical details and further policy implications. As suggested, we have added both detail and policy statements rather than having a more focused paper. To keep a bound on the technical detail within the manuscript, we now refer readers to places in the supplements where they can find additional detail.

Like most work in this area, the paper concentrates on health cost savings. Reductions in tobacco use also have important cost implications for other sectors of government and society in general. For example, Wendy Max's group has shown that the cost savings to the educational system by reducing AHDH related to secondhand smoke exposure among youth are substantially larger than the direct medical costs (Childhood secondhand smoke exposure and ADHD-attributable costs to the health and education system. Max W, Sung HY, Shi Y. J Sch Health. 2014 Oct;84(10):683-6. doi: 10.1111/josh.12191). There are other issues, such as fires that are also impacted. I am not suggesting that the authors expand the scope of their paper, but they should at least mention these other economic impacts because they are important to policy makers and show that the estimates in their paper are something of a lower bound.

We have added the following paragraph to the discussion section that speaks to the excluded economic and other benefits that are not included in the model:

“Other limitations should also be noted. The simulation does not include all harms of smoking, including direct costs outside the medical care sector, and therefore may understate the benefits of reduced tobacco use. For example, despite a requirement that only ‘fire safe’ cigarettes be sold in the state, in 2018 careless smoking was faulted in 10 of 30 fire fatalities with known cause in Minnesota.(51) In the U.S., smoking-related fires were responsible for $1.36 billion in direct costs such as property damage and firefighting expenses, and $1.16 billion in productivity losses in 1995 (the most recent comprehensive estimate we found).(52) Higher cigarette prices are associated with fewer residential fires and fire deaths.(53) Our estimates also exclude benefits of reducing secondhand smoke which contributes to cancer, respiratory disease, and cardiovascular disease, low birth weight, and their associated medical costs.(31) Max et al. estimated that the portion of attention deficit hyperactivity disorders associated with secondhand smoke costs the U.S. health care system $2 billion and the U.S. education system $9 billion when they used serum cotinine levels as a biomarker for secondhand smoke exposure.(54)”

The authors need to recognize that their numbers are estimates and report their results rounded to 2 or 3 significant digits, based on the precision of the input variables in the model.

We have rounded results to 2 digits because differences between the results of scenarios in sensitivity analysis that test substantial changes in model parameters are not visible when rounded to 3 digits. This is an indication that model precision may be better than 3 digits, and rounding to 2 significant digits preserves the utility of the sensitivity analysis. Please note we reviewed all numbers in the text given that Reviewer 1 noted several inconsistencies between the numbers in the text and the tables. In all cases, the numbers in the tables were found to be correct. Therefore, changes to some numbers in track changes are a combination of rounding to 2 significant and corrections. These track changes may give the appearance of incorrect rounding as any changes to numbers in the text were small.

Review Comments to the Author

Reviewer #1: This study assessed the 20-year impact of tobacco price increases and increased tobacco control expenditure during 1998-2017 in Minnesota using a microsimulation model developed by the authors. Different from most of similar studies in the literature, this study include a broader array of outcome: smoking prevalence, disease events, mortality, smoking-attributable medical costs, productivity gains, and state cigarette tax revenues. This a well-written article with details of their models and simulation methodologies presented in the appendices. Although this study made many assumptions about their model parameters, it provided sensitivity analyses to explore the influence of a wide range of changes in model parameters. This study demonstrated that increasing tobacco prices and investments in tobacco control simultaneously is effective to reduce the harms of tobacco use and is cost-saving in the mid- to long-term. I only have several minor comments.

1. Page 7, the methodology of estimating the consequences of cigarette smoking was not explained sufficiently. More description is needed regarding whether and how an econometric model was developed to estimate the smoking-attributable medical costs for current smokers. According to Table S3.1 under the baseline scenario, smoking-attributable medical costs was $1.324 billion in 2017. It would be great if the authors can add some discussion to compare their annual SA medical cost estimates with similar studies in the literature.

Rather than adding additional technical detail to the text, we now refer readers to the places in these supplements where details can be found. We are aware of no other published estimates of smoking-attributable medical expenditures for the state of Minnesota. Due to the importance of medical cost inputs on the estimates of medical costs averted, we had already included results using alternate medical cost estimates that align better (on a per-person basis) with the national estimates provided in the 2014 Surgeon General’s Report. We now draw out these results in the Results of Sensitivity Analyses by adding the following sentence: “Using alternative cost estimates that align with those in 2014 Surgeon General’s report increased savings by about 35%.”

2. Page 9, lines 151-153: It is not clear whether the price increases mentioned here ($0.76, $0.67, and $3.10) reflect constant dollar or nominal term. For example, Line 152 indicates that in subsequent years after 1999, the federal excise tax on cigarettes increased by $0.67 per pack. However, during that period, the federal cigarette taxes increased 3 times: by $0.10 per pack in 2000, by $0.05 per pack in 2004, and by $0.61 per pack in 2009. The sum of these increases is $0.76 which does not match to $0.67.

Thank you for the careful read. We now clarify that these are not adjusted for inflation. There was a typo that listed the federal excise tax increases as $0.67 rather than $0.77. Please note that we calculate the total as $0.77 rather than $0.76 because our source shows the federal increase in 2009 to be $0.62 rather than the $0.61 stated in the reviewer’s comment.

3. Page 9, lines 155-157: “The difference in health … between this ITC + Price Scenario and the ITC scenario provides estimates of the impact of tax increases alone”. This sentence is not correct because this difference also reflects the impact of the price increase due to the Minnesota tobacco lawsuit settlement.

Thank you. We have revised the sentence. It was meant to convey that the estimates exclude ITC (now ITCI) effects, but that was not clear.

4. Page 11, lines 188-189. The reduced numbers of smokers in 2017 (7,434 for youth and 138,123 for adults) due to ITC do not add up to the number shown in Table S3.4 (146,402).

Thank you. The numbers in the tables were correct. The number in the text for adults was misread – it was another similar number in the underlying spreadsheet that was used when completing the text.

5. Page 11, lines 201-205; page 12, line 206 and lines 211-203: All the numbers presented in these sentences do no match to the numbers shown in Table 1 and Tables S3.4.

As above, numbers in tables were correct. We apologies for the careless entry of numbers in the text.

6. Page 12, line 207 and line 215: Add “(data not shown”) after “$784 million”, and similarly after “$3.20 billion”.

Another row has been added to provide these numbers in Table 1.

7. Page 13, the last row “Base case”: the number shown in the last column does not match to the corresponding cell in Table 1.

Thank you. This appears to be rounding error causing the numbers to differ by 1. This level of precision is no longer visible after following guidance from the editor to round all results to a justifiable number of significant digits.

Reviewer #2: This manuscript presents the findings of a microsimulation to assess the impact of tobacco control policies, including tax increases and increased expenditures on tobacco control, in Minnesota during the twenty year period spanning 1998-2017. The study found that that increased expenditures on tobacco control prevented 37,421 cancer, cardiovascular, diabetes and respiratory disease events and 3,839 deaths over 20 years. Additionally, increased prices prevented an additional 14,702 additional events and 1,662 deaths. The authors conclude that both the net increase in tax revenues and the reduction in medical costs were greater than the additional investments in tobacco control, and that states can pay for initial investments in tobacco control through tax increases and recoup those investments through reduced healthcare expenditures.

This manuscript is generally well written, the analytic approach is free from any major methodological pitfalls, and would make a meaningful contribution to the scientific literature. Nonetheless, the manuscript could be improved, including more prominent recognition of various limitations of the analysis. Specific recommendations for revision are as follows:

1. Abstract. Page 2. Introduction. The first sentence of the abstract understates the considerable body of scientific evidence on the efficacy of tobacco control programs; specifically, the language needs to more clearly state the direction of the effect. They don’t just “influence” tobacco use and related health and economic outcomes – they reduce these outcomes.

Thank you. The sentence now reads “Tobacco control programs and policies reduce tobacco use and prevent health and economic harms.”

2. Introduction. Page 4. Third paragraph. The third sentence references “prior initiatives” and “significant policy change during this period.” However, it’s not clear to a lay reader what this means, particularly those without tobacco control expertise. A sentence should be added that clearly articulates what policies were actually implemented.

This suggestion helps to make the text more concrete. We have divided the sentence into two so we can provide sufficient detail. It now reads: “Part of this trend arose from earlier tobacco control policies including the nation’s first state-wide clean indoor air law in 1975, the nation’s first state-funded tobacco control program in 1985, and four state tax increases from 1985 to 1992. Significant policy changes after 1997 added to the trend, including strengthening the 1975 clean indoor air law to include all indoor public places, strengthening youth access laws, further increases in cigarette taxes and substantial increases in tobacco control expenditures.”

3. Introduction. Page 4. Since the manuscript only focuses on tobacco control program expenditures and price increases (i.e. and not other specific proven strategies such as smoke-free policies), it would be helpful to better set the stage for inclusion of these variables in the Introduction, including the focus on increased price. One way to accomplish this would be to reinforce that tobacco control expenditures support a variety of efforts, including state and community programs that influence policy adoption, and that increasing price has previously been noted as the single most effective intervention for reducing consumption. That way you better set the stage for the reader around the measures focused upon in the manuscript.

Thank you for the suggestion. We believe this helps to focus the introduction as it leads into the methods. The penultimate paragraph of the introduction now ends with: “Large tobacco price increases are the most effective known strategy to reduce tobacco use. Tobacco control expenditures support the state’s free quit line, anti-tobacco media campaigns, tobacco education, communication, and community collaborations. Increased expenditures also support initiatives to reduce the harms of commercial tobacco to American Indian communities, and education and communication activities likely contributed to passing local Tobacco 21 laws and local restrictions on the sale of menthol cigarettes.”

4. Introduction. Page 5. First paragraph. The last two sentences of the Introduction section are implication and discussion points, not Introductory content. They should be deleted here, and reinforced in the Discussion section.

We agree. The last sentence was already well-covered in the Conclusions and therefore we deleted it entirely. The second to last section has been moved to the Discussion section.

5. Methods. Page 5. Demographics and smoking status. It’s not clear what specific age groups from the “Minnesota population in 1997” the data were adjusted to – was it all individuals, those 18+, or something else?

We have revised the sentence to clarify that this includes all ages.

6. Methods. Page 6. The paragraph beginning “we estimated adult cigarette smoking” lacks necessary precision for the reader to understand the rationale for the employed analytic approach. For example, it’s not clear why 1996 and 1997 data were combined (sample size issue?). Additionally, it’s not clear what is meant by “calibrating the initial smoking probabilities”. Furthermore, it’s not apparent how the 18-24 year old age group, for which there was insufficient sample, could be predicted using estimates for 25-year olds, which reflects only a single age year; did you mean some range that began with 25 as the lower bound? If sample size were an issue, it would seem that you could consider adding multiple years of data to afford sufficient sample for 18-24 year olds, which is a critical smoking initiation demographic. If you can’t this needs to be duly noted as a limitation of the analysis.

We have clarified that the two years were combined to add precision, particularly in estimating cessation rates. We would need to combine many years to obtain an adequate sample of surveyed Minnesotan’s to obtain statistically more precise estimates by sex for ages 18-24. However, with changing smoking behaviors during this time, using multiple years would not necessarily provide a more accurate estimate. We have added to the methods section the following sentence: “This assumption could understate peak adult prevalence during the lifetime but provides a reliable estimate of prevalence prior to the ages of high harms from smoking-attributable chronic disease.” We have added to the limitations part of the Discussion section the following sentence: “Our assumption on tobacco use for adults less than 25 years of age would cause us to simulate too few former smokers if peak adult prevalence occurs before age 25.”

7. Methods. Page 7. Consequences of Cigarette Smoking. More clarity needs to be provided in the first paragraph of this section to articulate what smoking-attributable diseases were included. The current framing suggests it could be some, but not all. To clarify, it might be more prudent to place the onus of inclusion on outcomes found to be significantly associated with smoking in the 2014 Surgeon General’s Report, rather than focusing on SAMMEC, which is no longer in existence.

We modified the sentence to note that cancers, cardiometabolic and respiratory diseases are included and we now refer readers to places in supplement S1 where each disease is listed. Please note that while SAMMEC no longer exists as an electronic warehouse on CDC’s webpage, it does exist in the 2014 Surgeon General’s Report as Chapter 12 Smoking-Attributable Morbidity, Mortality, and Economic Costs . The CDC’s Office of Smoking and Health refer to the 2014 SGR as the updated SAMMEC.

8. Methods. Page 8. It’s unfortunate that the authors could not include smoke-free policies in the analysis, which was undoubtedly a contributor to both declining secondhand smoke exposure, as well as reduced smoking, in Minnesota during the assessed period. This needs to be stated as limitation of the manuscript (along with other proven strategies that couldn’t be included). Additionally, the text in this paragraph of the Methods should use the more commonly recognized term of “secondhand smoke exposure” instead of “secondary smoke exposure”. It’s also not clear what is meant by the statement “the analysis plan was not pre-registered.”

We have changed the text language to secondhand smoke and the limitations portion of the Discussion section now notes the exclusion of the benefits of reducing secondhand smoke from the estimates.

9. Methods. Page 8. It’s strongly recommended that the authors identify another abbreviation than “ITC” for “Increased Investments in Tobacco Control.” The abbreviation “ITC” has long been associated with the International Tobacco Control Survey in the tobacco control field, and thus, using this abbreviation may cause confusion when findings are reported elsewhere or taken out of the direct context of this study. Instead, authors could consider something like Increased Tobacco Control Investments, or “ITCI”.

Thank you for the suggestion. We have adopted ITCI throughout to prevent confusion and prevent readers from being distracted by a familiar acronym being used for a different concept.

10. Methods. Page 9. It’s questionable whether it is safe to assume that one policy does not impact the relative effectiveness on smoking behaviors of the other policy. Increased expenditures for tobacco control programs include state and community interventions, as outlined in CDC’s Best Practices for Comprehensive Tobacco Control Programs. Price increases are considered a product of those programs. As such, increased expenditures for programs would be associated with increased momentum around pricing strategies like excise taxes. Since there’s not much else the authors can do post hoc, at the least this should be duly noted as a limitation of the analysis.

We had this concern in mind when reviewing literature to obtain policy effectiveness estimates (the elasticities). The literature estimates which provide the effect sizes for each policy in the model are estimated while controlling for the effects of other policies on tobacco behavior. Thus, the simulation results are independent of each other to extent that the statistical control for contemporaneous policies in estimating policy effects in the literature was successful.

11. Discussion. Page 14. It is strongly recommended that the authors not pit individual interventions against one another. The framing around investments being more impactful than price would surely be used by opponents of cigarette excise taxes to state that they aren’t needed because they aren’t as effective. The major take home here is really that both are effective and important as part of a comprehensive approach to tobacco prevention and control at the state level.

We appreciate the concern and have clarified the text on page 14 to make it clear that the results pertain to the size of price increase and size of expenditure increases implemented in Minnesota. Clearly, the impact of one or the other will depend on the magnitude of each increase. We can’t prevent a disingenuous tax opponent from misusing the information that is provided in the results section. Funding tobacco control, has proven more politically difficult than raising tobacco taxes (even when funding was endowed by tobacco settlements, states raided the endowments). The results show that funding tobacco control is an important component. The Conclusions make the case for the price increases and additional investments in tobacco control being complimentary in a comprehensive tobacco control strategy.

12. Discussion. Page 15. The authors appropriately note that tobacco industry marketing efforts could counter the impact of tobacco control investments. However, this point could be made clearer for lay readers if the most recent data from the Federal Trade Commission were cited. In 2017, $9 billion was spent on advertising and promotion of cigarettes— more than $1 million every hour.

Thank you for the suggestion. We have added the best-available estimates of tobacco industry marketing expenditures in Minnesota to the text: “The tobacco industry spent an average of $165 million on tobacco marketing in Minnesota each year from 1998 to 2017 as estimated from national expenditures and state tobacco sales.”

13. Discussion. Page 16. The manuscript needs a paragraph that clearly articulates the limitations of the employed analytic approach, including several of the items noted previously in this review (e.g. lack of smoking estimates for 18-24 year olds, omission of full scope of policies such as smoke-free, assumption about independent effect of price and tobacco control program expenditures, etc).

Substantial new text has been added that covers most of these topics. However, we do not agree that omitting tobacco policies that were outside our scope is a limitation of your results.

14. Discussion. Another key point to reinforce in the Discussion is that this analysis only accounts for cigarette smoking, not all tobacco products, including other combustible products, smokeless tobacco products, and e-cigarettes. During the later part of the assessed period, particularly from 2011-2017, massive increases in the use of e-cigarettes were observed among youth – both nationally and in Minnesota. Therefore, this model doesn’t account for the full scope of tobacco products being used, as well as associated health risks/costs. Cigarette smoking is certainly responsible for the overwhelming burden of death and disease caused by tobacco use, but these other products are not without risk. Accordingly, this needs to be explicitly stated as a limitation of the analysis, and ideally, a reference to the diversification of the tobacco product landscape should also be made in the narrative elsewhere in the manuscript.

Thank you for the suggestion. This has been added to the expanded limitations portion of the discussion section. Also, at the beginning of the methods section, we now note that the model does not cover the complex arena of multiple tobacco product use.

15. Discussion. Another essential point that is briefly referenced, but should be more prominently focused upon, in the manuscript is that a comprehensive approach is warranted. There is no single panacea, and for several decades, tobacco control has most effectively functioned through a comprehensive lens that addresses the various drivers influencing initiation and cessation. As currently framed, the manuscript could be misinterpreted as suggesting that either expenditures and price are effective alone. It’s also important to reinforce that it’s not just making the expenditures, but what those dollars are actually used for; large amounts of dollars allocated to ineffective initiatives or strategies will do little for preventing initiation and reducing consumption at the population level. The expenditures needs to be focused of proven strategies for which evidence of efficacy and cost-effectiveness have been established.

We have added text that creates a better picture of a comprehensive approach in response to item #3 above. In addition, the conclusion now states that “Comprehensive state tobacco control programs that combine price increases, investments in tobacco control and policies that prevention initiation, and help smokers who want to quit have been found to be the most effective strategy for reducing the harms of tobacco.”

Reviewer #3: This is an important paper with findings relevant to continuing policy decisions about the investment of millions of tobacco settlement dollars. The economic analyses are detailed and comprehensive, and provide the basis for important policy discussions and conclusions. As such, the strength and details of these analyses are needed. However, this leads to a key question: what is the primary audience of this paper? The decision appears to be health policy makers, and not health economists. The details in the Supplemental Files are dense but important, and if more were included in the paper it would become more of a health economist paper. However, as written, many of the critical technical details of the model, parameter definitions and selection, sensitivity analysis, and technical aspects of the results are left to the Supplemental Files. Even so, the language of the paper and topics discussed still focus too much on economic issues rather than presenting key findings most relevant to the policy makers. For example, while noted in the discussion that "estimating return on investment for increased ITC was not a study goal..." (lines 290-291), the findings of a positive return on investment even during the limited time horizon is very important, and should not be buried deep in the Discussion. Similarly, the sensitivity analysis is important to these potential return on investment discussions and conclusions, but as presented, these analyses are written more for economists, and as such, will be too confusing to most policy makers. The critical link between the sensitivity analyses and return on investment conclusions needs to be clearly defined when introducing them (line 226). Also, while the Figures (1 and 2) are good at communicating the impact of ITC and Price, the specific data in Table 1 and Supplemental Files S3 are not presented well in the text for policy makers, and especially not summarized well in Abstract. The focus is on deaths and SA incidents averted; however, the bottom line for many policy makers is dollars saved. The medical cost savings and productivity gain estimates are unique to these analyses and should be given primary emphasis, along with the resulting return on investment calculations based upon them.

We have struggled with balancing implications for policy makers of the secondary economic findings with the original intent of the analysis. We believe that scientific integrity requires that we present the results in a manner that is consistent with the research plan and the manner in which the results were revealed to us. The intent of the analysis was to delineate what the contributions of cigarette price increases and tobacco control expenditures to Minnesotan’s health and economic well-being have been. We did not set out to determine if investments in tobacco control were cost-effective or produced a positive return on investment in a 20-year timeframe. While we agree that the secondary results that compare the investments in ITCI to their economic impact are likely to be the most compelling to policy makers, we feel it would be incorrect to focus the article on secondary results that happened to be very compelling when we would not have done so had we found that the ITCI did not yield a positive ROI within 20 years. We feel that this journal, with its focus on transparency and methodological rigor, is the proper venue for a paper presented in this manner.

Along the same lines, while we appreciate that a paper focused on ROI could make a crisp, succinct policy paper, our experience is that decision makers, especially those who champion a policy, want to be able to say that the policy they advocate has a sound basis, and that may require a detailed read and opinion from a trusted technical advisor. At the same time, few health economists would be interested in the technical merits of a study that had no practical implications. For these reasons, the delineation between a paper written for decision makers and one written for economists is not clear.

Therefore, we agree with editor’s advice to provide the additional detail on methods being sought by all reviewers while reinforcing the policy implications in the paper. Where possible, we now refer readers who may be looking for methods detail to specific places in the supplementary materials and we have added statements in the introduction, results, and discussion to better highlight the policy implications.

(Reviewer:) Also, the wealth of literature review, technical discussions, and econometrics in Supplemental Files S1 and S2 are very impressive. As noted above, if the primary audience of the paper is health policy makers, then these details need to stay there. However, the descriptions of the technical approaches in the main paper should be reviewed and considered: are they communicating the technical quality of the work presented in the Supplemental Files to the policy makers as effectively as possible? In many places, the reader should be specifically referred to the excellent presentation of the issues and parameter selection criteria in S1 and S2.

Thank you for recognizing the substantial background work presented in the supplements that made the analyses possible. We now refer readers to the supplements in multiple places and better draw readers’ attention to the type of detail that can be found in the supplements. For example, the discussion of the ICTI + price scenario, we now state: “These elasticities were drawn from literature with careful consideration of which studies provide the most appropriate price elasticity estimates as inputs to the simulation model, and on how the results of multiple studies should be combined (see supplement S2).”

More detailed comments:

Lines 43-48: I believe this is the only microsimulation model -- and the others, particularly the Levy model is a relatively crude macrosimulation model. This distinction should be emphasized.

We now note in the discussion section that differences in the structures of the models may contribute to differences in their results. Tobacco microsimulations are becoming more common, particularly in the literature on the impact of e-cigarettes on the population harms of tobacco. To our knowledge none have the detail in outcomes of the model used for this analysis, and therefore we focus on that distinction in this paper. The Levy model is well-validated and has provided import insights into tobacco policy for many years.

Lines 59-65: the potential of these findings to provide data needed for cost-effectiveness and return on investment calculations should be introduced and emphasized.

We have added to the introduction the sentence: “The results demonstrate the potential for health and economic gains in states that have not acted aggressively against tobacco.” We hesitate state that the results from the size and timing of price increases and tobacco control expenditure increases in Minnesota for a specific time period can be directly used in cost-effectiveness analyses.

Line 70: the details in literature review supporting selection of data inputs, policy parameters, and details in model should get more emphasis here.

This sentence now reads: “The simulation model, data inputs, and policy parameters are described in supplements S1 and S2, including details on the use of databases and literature to inform the model.” We now call attention to the supplements in additional places throughout the text.

Lines 128-130: The discussion of clean indoor air legislation should be revised. The specific impact on CIA on health effects of reduced exposures to secondhand smoke are not included (as noted); however, the policy impact of CIA on other outcomes by default are included in the ITC. While these are not specifically included in the model parameters, their impact is captured in the overall results. This should be noted.

We agree that the impact of clean air laws can be attributed to ITC to the extent that communications funded by ITC contributed to passing clean-air laws. However the studies that inform the impact of ITC on tobacco used behaviors in the model control for the existence of clean air laws and therefore provide an estimate of the effects of ITC on tobacco use that is independent of the impact clean air laws.

Lines 131-137: The details of annual tobacco control investments are provided in Supplemental Files. Guide readers to these details. This also could apply to other data included in the model and only shown in detail in Supplemental Files; i.e., annual excise tax levels and average annual retail cigarette prices.

Thank you. This is done.

Lines 167+: as noted above, the important relevancy of these simulations for explaining return on investment after adjusting for possible biases and reduced parameter estimates for impact of ITC or price, i.e., expenditure and price elasticities, etc.

These policy implications of the simulations reported in this section are prominent in the Conclusions.

Line 192: "...due to lagged effect of prior investments." will not be understood my almost all readers as being primarily due to two model decisions: first, only one time impact of price increases applied at time of tax increase (standard econometrics) for adults, but impact on youth extended; and second, the use of cumulative funding with 25% discount (a la Farrelly et al models). Maybe this is too technical a point to discuss in results, but might merit being raised in discussion.

We agree and have deleted this phrase. In an early draft we had additional explanation for this phrase and decided it was too detailed for most readers and was distracting.

Line 209: the detailed impacts are shown in S3, but more of them could be shown in an extended Table 1. Additionally, details on annual ITC should be linked with results in S3 to highlight return on investments. Also, S3 could show more details on annual state tax revenues under Baseline Scenario as well as under actual tax increases. Finally, it would be outstanding to see a combined return on investment table combining tax revenues, ITC investments, and SA medical costs. Such a table could have very important policy implications.

We have added, and refer readers to, two new tables in S3 the show changes in tax revenues, ITC expenditures, medical costs and productivity effects side-by-side. We feel that adding a time dimension to Table 1 to show the detail provided in S3 would be very challenging and would distract from the bottom line results.

Table 1: something does not look right. I did not search out all the details in Supplemental Files, but the deaths do not look right. SA cancers (incidents) and SA deaths are almost identical; however, SAMMEC estimates are that only about one-third of SA deaths are due to cancer. This needs to be checked, since SA deaths should be much higher than SA cancer incidents and deaths.

These estimates are correct. Only a fraction of cancer cases are fatal. We have modified the label in Table 1 to make it clear that these are cancer cases.

Figure 2 (lines 217-224): emphasis on percent changes, but policy impact in $s only shown in Table 1. Figure could be 2a (current) and 2b with numbers (number of youth prevented, increase number of adult quitters, SA incident cases prevented, SA deaths averted, and most importantly, SA costs saved (in $s), productivity gains (in $s) and (to be computed) total savings (SA medical averted and productivity gains).

Figure 2 provides a different view on the results than provided in Table 1, with an emphasis on relative change that provides a glimpse of what may occur in future years. It is not clear to us what would be gained by repeating the numbers of Table 1 in a figure. We chose to display them in a figure so readers could easily read the numerical estimates and because the different outcomes are on much different scales and thus do not lend themselves being displayed a single chart.

Lines 225-237: as noted above, the impact of sensitivity analyses on return on investment calculations needs more emphasis. Lines 233-236 briefly discuss these findings; but, these calculations need much more emphasis and should be provided in greater detail; maybe even added as additional columns in Table 2.

Please see discussion above.

Lines 239-255: this discussion is relevant, but may not strike the right balance. Make it clear from the first sentence that both ITC and price increases are needed: they are complimentary in timing and level of impacts produced.

The first paragraph of the Discussion has now been split into two paragraphs, with the impacts on tax revenues added to the first paragraph. The Conclusion makes the case that the two policies are complimentary.

Lines 256-267: discussion of Levy et al model results relevant, but this is a microsimulation model that differs in many ways from their model beyond parameter estimates.

We now note in the discussion section that differences in the structures of the models may contribute to differences in their results.

Lines 287-296: discussion of uncertainties and imprecisions in model parameters and inputs is always relevant. The details in S1 and S2 provide very good discussion of how parameters and model specifications were done. Point readers to these discussions and reviews. Yes, uncertainties about elasticities, especially about ITC, limit model predictions. Thus, the need for good sensitivity analyses (as was done). Nevertheless, the positive return on investment calculations were found, as well as the positive revenue gains for tax increases (which need more emphasis with more specific numbers provided, even thru this has been documented in many prior econometric analysis).

Additional referrals to S1 and S2 have been added in the text as noted above. The tax revenue results have been added to the first paragraph of the discussion. Tables comparing revenues and the other economic measures have been added to S3.

FINALLY, it is suggested that future analyses estimate the even greater impacts (on all outcome measures, but particularly SA cost savings) if ITC levels would have been higher (like full BP levels or even 50% of BP level) or would not have been reduced in 2004.

Thank you for the suggestions. We agree; estimates of full potential of these policies are needed.

24 Feb 2020

PONE-D-19-28016R1

The 20-year impact of tobacco price and tobacco control expenditure increases in Minnesota, 1998-2017

PLOS ONE

Dear Dr. Maciosek,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. You have done a nice job of responding to the reviewers.

There is only one small change left.  In the Introduction, you say that "Large tobacco price increases are the most effective known strategy to reduce tobacco use.”  While this is a commonly made statement, as you note elsewhere in the paper, how effective a tax increase is depends on the size of the tax increase.  For example, a smokefree law could have a bigger effect than a small tax increase.  Please replace "the most effective" with "an effective" and make any other needed wording adjustments to the paper to avoid overstating the importance of tax increases.

We would appreciate receiving your revised manuscript by Apr 09 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Stanton A. Glantz

Academic Editor

PLOS ONE

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

26 Feb 2020

Editor Comments:

There is only one small change left. In the Introduction, you say that "Large tobacco price increases are the most effective known strategy to reduce tobacco use.” While this is a commonly made statement, as you note elsewhere in the paper, how effective a tax increase is depends on the size of the tax increase. For example, a smokefree law could have a bigger effect than a small tax increase. Please replace "the most effective" with "an effective" and make any other needed wording adjustments to the paper to avoid overstating the importance of tax increases.

Response:

We agree that this is a contradiction. We have made the suggested change in the introduction and removed characterizations of the effect size of tobacco taxes (e.g. ‘large’, ‘substantial’) from the results section.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

28 Feb 2020

The 20-year impact of tobacco price and tobacco control expenditure increases in Minnesota, 1998-2017

PONE-D-19-28016R2

Dear Dr. Maciosek,

We are pleased to inform you that your manuscript has in been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Stanton A. Glantz

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

5 Mar 2020

PONE-D-19-28016R2

The 20-year impact of tobacco price and tobacco control expenditure increases in Minnesota, 1998-2017

Dear Dr. Maciosek:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Stanton A. Glantz

Academic Editor

PLOS ONE