INTRODUCTION

Smoking cessation treatment is considered crucial for tobacco users as it helps prevent chronic diseases and premature deaths causally linked to tobacco use¹. The health consequences of tobacco use include cancers, stroke, diabetes, and other chronic diseases, as well as more general adverse effects on the body, such as inflammation and impairment of immune function². The short-term clinical outcomes of tobacco cessation include a lower risk of death from respiratory diseases and cancer³, a lower risk of secondary coronary heart disease events and death^4,5, and a lower rate of diabetes⁶.

Article 14 of the World Health Organization’s (WHO) Framework Convention on Tobacco Control (FCTC) recommends the implementation of tobacco cessation interventions – brief advice in primary healthcare settings, national toll-free quitlines, and pharmacotherapy, including at least nicotine replacement therapy (NRT) – as an important component of comprehensive tobacco control programs^7,8. Many countries have offered a combination of these interventions as the most effective method of helping people quit smoking^9-11. Following the WHO recommendations, population-based approaches, such as quitlines, the internet, and self-help interventions, have been successfully implemented in countries where governments can invest in promoting those services⁷. Healthcare providers in the existing healthcare system have also been encouraged by the government to be on the frontline of tobacco cessation for more consistent and effective identification and treatment of tobacco users. However, as the existing healthcare systems are not sufficiently organized to address the issue of tobacco use, such attempts were not very successful even in high-income countries, and there is a marked lack of coverage of the tobacco control measure on ‘offering cessation’ in MPOWER (32% of the world population in 2018)⁸.

In Korea, the successful implementation of population-based cessation approaches, e.g. public health center-based cessation services (since 2005), a national toll-free quitline (since 2006), and internet-based cessation services (since 2013), combined with the enforcement of other tobacco control measures, significantly reduced the smoking rate among men from 51.6% in 2001 to 39.3% in 2015¹². However, the prevalence of smoking remains high, and the economic burden attributable to smoking is also significant, as illustrated by the direct medical costs of cancer attributable to smoking: US$531 million in 2014, or 1.2% of the total Korean national healthcare expenditure for that year¹³. With the tobacco tax increase in 2015, the National Health Insurance Service (NHIS) started to cover the cost of tobacco cessation consultation and drug fees in hospitals and clinics across the country. The service introduced systematic changes to the screening and documentation of tobacco use in patient health records; implemented training for healthcare providers on effective, evidence-based cessation support; defined rules for including cessation support in routine care; and developed a budget and resources for delivering effective cessation support. The cessation service currently covers approximately 300000 smokers who visit clinics and hospitals for medical care or health check-ups and decide to quit by the physicians’ brief advice, annually¹⁴.

In this study, to see the effect of cessation service use, we measured changes in the prevalence and related medical costs of selected diseases among cessation service users 1 year before and after the service enrollment, which have never been reported, and compared these amounts with those of never smokers and smokers who did not use the service.

METHODS

Study design and population

In the current study, we obtained two analytical samples from the NHIS database. One is for comparison between pre and post of smoking related diseases prevalence and their medical cost among users of the NHIS smoking cessation intervention program. Smokers who used the NHIS smoking cessation intervention program from 2015 to 2017 and participated in the NHIS general health screening were included in the study as the baseline population (n=783494). After excluding individuals with unknown age or gender (n=612), invalid date of death (n=65), or death within 1 year after enrollment of the service (n=3502), we included 779315 service users in the first analytical data to evaluate changes in disease prevalence and related medical costs, before and after the calendar year when NHIS smoking cessation intervention program was used (Figure 1). The other is for comparison of smoking-related diseases prevalence and their medical cost between cases (smokers who used the cessation program) and controls (smokers who did not use the cessation program and lifelong never smokers). From the 30 million individuals in the NHIS general health screening participants between 2013 and 2018, a subgroup of NHIS smoking cessation intervention program users in 2016 (n=320667) was extracted. After excluding 89473 individuals who were never screened, those with unknown matching variables (i.e. age, sex, income level, and residential area), those aged <19 years, and those with an invalid death date, we finally included 231204 users as cases in the second analytical data. The same number of current smokers (n=231204) and never smokers (n=231204) were selected after frequency matching for age, sex, income level, and residential area (Figure 2). Under internal regulation of NHIS to minimize the amount of data released regarding privacy protection and analytical eligibility, the NHIS cessation service users in 2016 were only included as cases in the second analytical data to reduce the huge amount of control data added from NHIS general health screening participants in 2013–2018.

Figure 1

Flow chart for selection of study population for comparing the diseases prevalence and the medical costs before and after enrollment in the National Health Insurance Service (NHIS) smoking cessation program

Figure 2

Selection of population in cases (National Health Insurance Service (NHIS) smoking cessation service users) and controls (smokers who did not use the service and never-smokers) for comparing the diseases prevalence and the medical costs between them

RESULTS

Table 1 shows the general characteristics of the study population. The NHIS smoking cessation service users were predominantly men (86.5%) and aged ≥40 years (74.3%). Half of the users (49.1%) were in the upper middle or high-income quintiles. Seventy-three percent of smokers enrolled in the service once, and 27% enrolled multiple times. Of all the service users, 35.2% completed the full cessation program by attending six hospital/clinic visits or receiving a prescription for 8–12 weeks. About one-third (34.8%) were obese (i.e. BMI ≥25 kg/m²), 9.7% had high blood sugar (FBS ≥126 mg/dL), 10.7% had high TC (≥240 mg/dL), and 68.7% drank alcohol at least once per week.

Table 2 shows the medical costs of selected smoking-related diseases during the 1 year before and after the first enrollment in the NHIS smoking cessation service. Among the 779315 service users, 97.2% (n=757510) used medical services for any kind of disease at least once in the year before enrolling in the smoking cessation program. The most prevalent diseases were acute upper respiratory infections (37.0%), followed by acute lower respiratory infections (31.4%), other upper respiratory diseases (20.3%), chronic lower respiratory diseases (12.7%), and diabetes mellitus (10.8%). After cessation service enrollment, the number of patients who accessed care for any kind of disease increased by 1% (n=7487), the number of days spent in the hospital increased by 13%, and medical costs increased by 23%. The number of patients, days spent in the hospital, and medical costs for 1 year consistently increased for all 15 selected diseases after service enrollment. The biggest changes in the number of patients were observed in influenza (64%), cancer (26%), and diseases of the arteries, arterioles, and capillaries (21%). We observed decreased prevalence and medical costs of pneumonia among women and participants aged < 50 years, whereas an increasing pattern was consistently observed for the remaining diseases in both sexes and older age groups (Supplementary file Tables 1 and 2).

Figure 3 and Supplementary file Table 3 show the disease prevalence among cessation service users, smokers who never used the cessation service, and never smokers. Generally, disease prevalence was highest for cessation service users and lowest for smokers who never used the cessation service. For example, the prevalence of all diseases (ICD-10 codes A00–Z99) in 2017 was 97.7%, 88.8%, and 91.1% for cessation service users, smokers who did not use the service, and never smokers, respectively. The prevalence of acute upper respiratory infections in 2017 was highest among cessation service users (36.8%), followed by never smokers (32.9%) and smokers who never used the cessation service (26.8%). Disease prevalence increased more among the cessation service users than among the other groups. For example, the prevalence of ischemic heart diseases increased from 2.9% in 2015 to 3.9% in 2017 (35.6% change) among cessation service users, whereas 20% increases were noted in the other groups. The same pattern was observed for other smoking-related diseases (Supplementary file Table 3-1). However, the number of days spent in the hospital and medical costs for chronic diseases were lowest in never smokers (Supplementary file Tables 3-1 and 3-2).

Figure 3

The prevalence of the five most common diseases in 2015 and 2017 among the National Health Insurance Service (NHIS) smoking cessation service users (service users), smokers who did not use the NHIS cessation service (smokers), and never smokers in 2016

DISCUSSION

In the current study, the number of patients, days spent in the hospital, and medical costs for 1 year before and after the use of the NHIS smoking cessation program were measured and compared with those of control groups (smokers who do not use the service and never smokers) to estimate the short-term effect of the cessation service use. From the study results, it was found that the disease-specific prevalence and their medical cost were highest and increased more in the cessation service users compared with the control groups. Particularly, decreased prevalence and medical costs for pneumonia were observed among cessation service users aged <50 years.

Although specific smoking cessation interventions have been suggested to improve the involvement of healthcare providers in cessation support, the interventions are not frequently successful in healthcare settings, and their effectiveness has been evaluated by the cessation success rate or process improvements¹⁹. Therefore, more studies are needed to measure other outcomes, such as tobacco-related disease prevalence and medical costs, to assess the ultimate goal of cessation interventions: to minimize the health risks and economic burden caused by tobacco use. Furthermore, research on the use of smoking cessation services and healthcare costs using real-world data is necessary to provide a strong context in terms of generalizability, as they are far less.

In these perspectives, a few studies suggested a decrease in some healthcare costs in the year post-smoking cessation treatment, but these findings resulted from the limited study design and target population, such as the comparison of smokers with or without inpatient tobacco dependence treatment during the hospitalization or the comparison of varenicline versus NRT users without comparison to the non-treated group^20,21. While some previous studies, including two recently published studies from Ontario, Canada, based on the longitudinal follow-up design using real-world measurement of the healthcare cost linked with administrative healthcare data, showed an increase in or relatively higher medical utilization and cost among the smoking cessation service users^22-25. One study reported higher total healthcare costs in the treatment cohort pre- and post-treatment periods²⁴, and another study identified a significantly greater rate of outpatient visits, emergency department visits and hospitalizations among patients who enrolled in smoking cessation treatment offered through primary care clinics over a 5-year follow-up period²⁵. The above suggested that the increase, which contradicts other prior studies’ results, may be explained by more access to medical care for smoking cessation service users for their various health concerns or differences in treatment settings, patient populations, and healthcare service outcome measures among the studies.

Our findings correspond to the two recent publications and have some strong points in the data used, study design applied, and evidence introduced. The current study is based on the NHIS claims data, which include comprehensive information on disease code, medical cost for each inpatient and outpatient service or care, and other information on socioeconomic and medical status with individual base linkage to the data from NHIS general health screening participants. Moreover, NHIS claims data are highly representative of the Korean population, as they cover the entire population in Korea. The change in disease prevalence and medical cost were compared with smokers who do not use the smoking cessation service and never smokers, as well as within smoking cessation service users in the 1-year period before and after service enrollment. Known smoking-related diseases are classified and introduced for each comparison on the change in medical cost, which could give information on the different short-term effects of smoking cessation service use by disease-specific characteristics.

The long-term health benefits of smoking cessation are well known. Cessation reduces mortality, carrying respective gains of approximately 3, 6, 9, and 10 years of life expectancy for those aged 30–60 years²⁶. The short-term economic and health benefits of smoking cessation have also been demonstrated for acute myocardial infarction (AMI) and stroke, in modelling studies using US²⁷ and Australian data²⁸. In the Australian simulation, if smoking prevalence had decreased by 1% in the first year, approximately 1000 hospitalizations for AMI and 350 for stroke would have been avoided over 7 years, saving approximately $20.4 million in healthcare costs²⁸. However, our observational study based on real data showed increased medical costs and the prevalence of diseases post-cessation service use. This was true even for acute respiratory infections, which can be considered a target disease for measuring short-term health benefits. Our results are comparable with those of other observational studies. Previous research on the healthcare costs associated with former smokers has suggested that those who quit smoking may incur greater healthcare costs than continuing smokers. For example, in a cross-sectional study in Germany, the probability of hospitalization was highest among those who quit 1–3 years ago and decreased thereafter²⁹. In a retrospective cohort study of enrollees in a health maintenance organization in the US, former smokers’ costs were significantly greater in the year immediately following cessation relative to those incurred by continuing smokers²². More specifically, the healthcare costs among former smokers began to rise in the quarter prior to cessation and were significantly greater than those of continuing smokers in the quarter immediately following cessation²³. However, former smokers’ costs fell in year two, and this decrease was maintained throughout the 6-year follow-up period²². Increased hospital utilization in the first year after quitting smoking is more likely to be a cause of smoking cessation than a consequence. Specifically, short-term excess healthcare utilization among former smokers may result from smoking-related conditions that led to smoking cessation²⁹. After quitting, former smokers may seek medical care that they delayed while they were smokers, resulting in increased healthcare utilization³⁰. On the other hand, in a randomized controlled trial of pharmacotherapy and counselling for smoking cessation, sustained quitters amassed lower healthcare costs by the sixth quarter post-quitting than trial participants who continued smoking³¹.

The results of the present study help us to understand the characteristics of smokers who use the cessation program. Among the cessation service users, the prevalence of smoking-related diseases and their medical costs increased in the year following service enrollment. Notably, the disease prevalence and medical costs for pneumonia were lowest among individuals aged <50 years. This finding may provide evidence that the short-term benefits of the cessation program are greater for younger smokers than for older adult smokers. On the other hand, compared with matched control groups by age, sex, income, and residential area, the prevalence of most of the selected diseases was highest among the cessation service users. For example, the prevalence of acute upper respiratory infections in 2017 was 36.8%, 32.9%, and 26.8% in the groups of cessation service users, never smokers, and smokers who never used the cessation service, respectively. As another example of chronic disease, the prevalence of diabetes mellitus in 2017 was 12.8%, 9.1%, and 7.7% in the groups of cessation service users, smokers who never used the cessation service, and never smokers, respectively. We also observed increased prevalence and medical costs between 2015 and 2017 for most diseases in all three groups, though the increase was highest among the cessation service users. This finding may be explained by cessation service users being more vulnerable to diseases and the cessation program simply targeting the right people. Interestingly, for most of the selected diseases, the smokers who never used the NHIS cessation service showed a lower disease prevalence than both service users and never smokers. Moreover, in our results, healthcare utilization for any kind of disease was highest among cessation service users and lowest among smokers who did not use the service. One plausible explanation is that smokers who need medical services may be motivated to use the cessation service, whereas the remaining smokers are less likely to need and seek healthcare. Thus, smokers who never use the cessation service are healthy enough to continue indulging in their smoking habit. Quitters may seek medical care that they delayed while they were smokers, resulting in increased healthcare utilization upon service enrollment^29,30.

CONCLUSIONS

Cessation service users were more likely to seek healthcare. Increased healthcare utilization in the first year after cessation service use may have resulted from smoking-related conditions that led individuals to attempt smoking cessation.