Although the World Health Organization advocates that countries provide comprehensive smoking cessation services, few countries do so.1 While the United States,2 the United Kingdom,3 France4 and Australia5 have developed tobacco treatment guidelines, few physicians adhere to them.6 7 In fact, according to the US Preventive Services Task Force, tobacco cessation counselling, which is a high priority service, has the lowest rate of delivery.8 Physician participation is low in such programmes because of time constraints, perceived low effectiveness,9^–11 perceived lack of training and materials,12 perceived lack of receptiveness by smokers to counselling,13 14 the fear that negative messages might lose customers, lack of financial incentives and physicians being smokers themselves.6 15 16

Taiwan implemented the Smoking Cessation Outpatient Services programme in 2002, the same year that it enacted a tobacco health and welfare tax. Taiwan’s smoking cessation counselling guidelines recommend a six-minute counselling session to present a five-point agenda termed the 5As: Ask, Advise, Assess, Assist and Arrange. During that session, subjects are to receive a brief consultation about smoking cessation and a prescription of medication that might help them quit smoking (for example, nicotine replacement therapy (NRT) and bupropion). Physicians are required to attend a 6-hour smoking cessation consultation training workshop before signing the contract with government. Between 2002 and 2005, about 3500 physicians (10 of all physicians in Taiwan) completed the training and were certified to provide smoking cessation services. Evaluation studies showed that the training programme enhanced physicians’ knowledge of smoking cessation practices.17

Telephone surveys of adult smoking behaviour showed the smoking rates to be 42.8 and 39.9 for men and 4.5 and 4.8 for women in 2004 and 2005, respectively.18 Taiwan is estimated to have four million smokers, over half of whom have attempted to quit at least once. However, only 20 000 smokers (0.5) had received treatment for their addiction in 2003. To promote these services to smokers, the Taiwanese government increased reimbursement rates and medication subsidies in 2005. Under the new funding policy started in January 2005, participating physicians started receiving 350 New Taiwan dollars ($NT, $US11) per visit instead of $NT250 ($8) they were receiving in 2004. Also, the government offered a subsidy of cessation medications to eligible subjects 19 years old or older who were smoking at least 10 cigarettes per day. In 2005, subjects could receive medication subsidies of up to $NT400 ($13) per week instead of $NT250 ($8) per week before the increase.17 Under the new programme, in addition to the subsidies, every patient could receive two courses (8 weeks per course) on smoking cessation each year. Previously, only family practitioners, psychiatrists and internal medicine physicians could offer cessation counselling. After the change, physicians in any specialty of medicine were allowed to participate and to offer government sponsor cessation services.

Although some studies have shown a positive impact of financial incentives and smoking cessation insurance coverage on the utilisation of smoking cessation services and quitting outcomes,19^–24 few studies have evaluated the effectiveness of this financing of nationwide smoking cessation services. This study aims to evaluate the effect of the 2005 increase in reimbursement rates and medication subsidies on physician participation in the programme, patient utilisation of services and cessation outcomes at six-month follow-up. In addition, this study examines whether these policy changes expanded the breadth and number of physicians and subjects offering and receiving these services.

METHODS

We linked five existing provider and patient datasets. These datasets have been continuously collected by the Office of Taiwan Smoking Cessation Services (OTSCS) since the launch of smoking cessation programme in September 2002. We used two provider datasets: the Contracted Physician Dataset and the Health Organization Registration Dataset. All participating physicians must complete training requirements and contract with the government to receive reimbursement. The Contracted Physician Dataset provides information about physicians’ demographic characteristics and medical specialty. The Contracted Health Organization Dataset provides information about the type of organisation, accreditation level and geographic area of the provider.

We used three patient datasets including (1) Patient Claims Data, (2) Patient Intake Survey and (3) Patient Six-Month Follow-Up Survey. After each patient completes a cessation course, his or her physician is required to submit patient claims data and patient intake survey data to the OTSCS for reimbursement. The patient intake survey collects data on the patient’s gender, age, daily cigarette consumption, frequency of quit attempts the previous year, nicotine dependence and chronic health problems. Each month a random sample of 1300 participants who had enrolled six months earlier is selected to complete the follow-up survey. On average, around 650 of the 1300 completed the survey, which includes questions about the patient’s educational level, occupation, current smoking status and reasons that he or she succeeded or failed to quit smoking.

As for our dependent variables, the number of monthly consultations provided per physician and the annual number of cessation consultations received per patient were collected by linking 2004–2005 Patient Claims Data, Contract Physician Data and Contracted Health Organization Data. The quitting outcomes were collected by asking participants whether they were currently smoking at the six-month follow-up survey. Our independent variables were increased funding, consumer characteristics (gender, age, education, employment, nicotine dependence status, previous attempts to quit and chronic disease status) and provider characteristics (physician age, gender, specialty, organisation type, accreditation level and location). In this study, smoking cessation service or consultation refers to any routine visit in which a physician advises the patient to quit smoking by providing brief counselling and prescribing cessation medications.

The provider and patient demographic variables found in the Contracted Physician Dataset, Contracted Organisation Dataset and the Patient Claims Dataset have high validity, because these data are based on reimbursements that have already been collected. Expert evaluation and pretests were used to improve content validity and reliability in the Patient Intake Survey and Six-month Follow-up Survey. Some instruments with proved validity, such as the Fagerstrom Test for Nicotine Dependence,25 26 were adapted and included in the Patient Intake Survey. About 300 subjects of those who reported quitting smoking at the six-month follow-up survey were asked to undergo a carbon monoxide (CO) biochemical validation test to confirm they had quit. Only half of the subjects participated in CO measurement. Eighty-five per cent of the CO test results were lower than 8 parts per million (ppm), confirming the self-reported outcomes.

SAS software was used to link and analyse provider and patient datasets. The generalised estimating equations (GEE) method27 28 was conducted to examine the effects of increasing reimbursement rates and medication subsidies on monthly smoking cessation services provided per physician and the annual consultations received per patient. Since the data of these outcome variables are over-dispersed count data, the GENMOD program with REPEATED statement and negative binomial distribution was selected. Logistic regression was used to examine the effect on smoking cessation, which is a dichotomous outcome.

RESULTS

Contracted physicians

After the increase in financial support and the broadening of medical specialisations that can offer cessation services in 2005, the number of contracted physicians increased from 1841 (5.5 of all physicians) in 2004 to 3466 (10.4) in 2005. Most increases were seen in private organisations (1503 vs 3092 physicians) and non-hospital clinical settings (970 vs 2335 physicians) (table 1). Ninety per cent of the participating physicians were men and their mean age was 46 years old. In both years, a larger percentage of those participating in the programme were practitioners of family medicine (2004 vs 2005: 51.2 vs 39.2) and internal medicine (2004 vs 2005: 32.6 vs 26.0) (table 2).

Annual number of cessation services delivered increased five times, from 22 167 subjects (44 554 visits) in 2004 to 109 508 subjects (273 754 visits) in 2005. The average monthly number of smoking cessation consultations per contracted physician increased from two in 2004 to eight in 2005. The number of subjects they saw increased from one to five. After excluding half of contracted physicians who did not provide cessation services in each month, the average monthly number of smoking cessation services provided per physician who delivered services increased from five visits in 2004 to 15 visits in 2005. The number of subjects they saw increased from three to nine.

After adjusting for other variables, our multivariate analyses showed that the increases in reimbursement rates and medication subsidies almost doubled the average number of smoking cessation services a contracted physician delivered (table 3). Younger physicians, male physicians, surgeons and otolaryngologists provided more consultations than older physicians, female physicians and family practitioners. Clinics, privately owned facilities or facilities located in Taipei provided more smoking cessation services than medical centres, regional and district hospitals, publicly owned facilities or facilities located in central Taiwan. For example, the average contracted physician in a clinic provided five times the monthly number of consultations than one in a medical centre. Similarly, contracted physicians working in privately owned facilities provided twice the monthly consultations that were being provided by physicians working for publicly owned ones.

Patients receiving smoking cessation services

As can be seen in table 4, the number of subjects receiving cessation services increased five times, from 22 167 subjects in 2004 to 109 508 in 2005, representing from 0.5 to 2.75 of the total population of smokers in Taiwan. There was a greater increase (0.4 vs 2.1) in the proportion of younger smokers seeking cessation services than in older smokers (0.9 vs 3.6) (table 4). The mean age of those receiving these services was slightly lower in 2005 than in 2004 (44.1 vs 45.4 years old). A slightly smaller percentage of women used these services in 2005 than in 2004 (14.6 vs 15.5). Annual number of cessation visits per patient increased from 2.0 visits in 2004 to 2.5 visits in 2005.

Based on our multivariate analysis, the increases in government reimbursement for consultation and subsidies for medication for tobacco treatment significantly increased the annual number of cessation visits per patient. In addition, subjects who were older, those seen by physicians providing more monthly consultations, those seen by psychiatrists and those receiving smoking cessation services in the private sector or in clinics were more likely to be provided with more smoking cessation consultations per year.

Determinants of successful cessation

Of the yearly random sample of 15 600 smokers who had received smoking cessation services, 7708 and 7053 smokers participated in the six-month follow-up interview in 2004 and 2005, respectively. The mean age of 2004 sample was 46 years old and that of the 2005 sample 45 years old. Eighty-five per cent of respondents were male. The percentage of subjects with college level educations or higher was lower in 2005 than in 2004 (27.9 vs 32.1). There was a lower percentage of subjects with high nicotine dependence in 2005 than in 2004 (32.5 vs 41.0). Forty-one per cent of the smokers receiving services in 2004 reported that they had attempted to quit smoking the previous year, while 31 reported in 2005 that they had attempted to quit the previous year. Thirty-six per cent of respondents in 2004 and 31 in 2005 reported some type of chronic disease such as hypertension, diabetes or heart disease.

At the six-month follow-up, 1942 subjects (25.2) in 2004 and 1500 subjects (21.3) in 2005 reported they were not smoking. Basing our estimates of the quit rates at the six-month follow-up in 2004 (25.2) and 2005 (21.3) and the number of subjects received smoking cessation services in 2004 (22 167) and 2005 (109 508), 5586 smokers in 2004 and 23 325 smokers in 2005 successfully quit smoking at the six-month follow-up interviews. In addition, total government expenditures to reimburse the cost of consultation and subsidise costs of cessation medicines increased from $0.8 million in 2004 to $7 million in 2005. Based on these figures, the average cost for cessation consultations and medication per quitter was $135 and $300 in 2004 and 2005, respectively.

After adjusting for consumer and provider factors, our multivariate analysis revealed that the 2005 increase in funding for smoking cessation services could not be significantly associated with an increase in successful cessation outcomes (table 5). We found more successful treatment outcomes for those who had attempted to quit the previous year (odds ratio (OR)  1.26). In addition, after adjusting for other factors, we found a strong association between age and successful cessation. For instance, compared with those between the ages of 18 and 34 years old, subjects over 65 years old were more likely to successfully quit smoking (OR  2.50), followed by subjects between 50 and 64 years old (OR  1.65), and those between 35 and 49 years old (OR  1.15). We found a negative relation between severity of nicotine dependence and quitting smoking after adjusting for other factors. The odds ratios of quitting smoking for subjects with high nicotine dependence were 0.57 times the odds ratios for subjects with low nicotine dependence. In addition, subjects who received smoking cessation services through public providers (OR  1.21) compared to private sector, or district hospitals (OR  1.22) instead of clinic settings were more likely to quit after adjusting for other factors. On the other hand, subjects seen by physicians who delivered more than 36 smoking cessation consultations per month were less likely (OR  0.57) to quit smoking than subjects seen by physicians who delivered less than five consultations per month.

DISCUSSION

This study found the increase in funding for smoking cessation services encouraged providers to promote more cessation consultations and encouraged subjects to make more cessation visits. These findings are consistent with other studies24 29 30 that have found that subjects with full coverage for smoking cessation treatment were more likely to use the cessation services than subjects with partial coverage, though smoking cessation outcomes were not found to be significantly different. We did find that the increases in utilisation came with a fivefold increase in number of quitters at about twice the cost per quitter. Using a high extrinsic motivation strategy, the 2005 increase encouraged those with low intrinsic motivation to join the cessation programme. Therefore, as suggested by the US National Cancer Institute, interventions that impact larger number of smokers can have cumulative effects on disease rates and create substantive public health benefits.31

In addition, this study and another study32 found that the types of providers played an important part in number of cessation services offered and cessation outcomes. This study, like one previous review study,33 found that increasing reimbursement rates attracted physicians working in private clinics more than those working in pubic hospitals. However, we found that clinics and hospitals in the public sector had more successful smoking cessation outcomes than those in the private sector. Different organisations, depending on their resources and constraints, may adopt different level of elements recommended in the “5As” clinical practice guidelines for tobacco treatment and perform them differently. Some hospitals may involve more nurses in counselling and follow-up contacts, while some clinic physicians may provide only brief advice. As a result, the differences in services may lead to different outcomes. To control the quality of cessation services, we should establish quality measures of smoking cessation outcomes and a claims audit system. One study has indicated that financial incentives and smoking cessation quality measures could promote smoking cessation services.34

This study found that consumer characteristics affected the use of smoking cessation services. These findings are consistent with those of other studies35^–38 that reported smokers who had chronic diseases and who were older to be more likely to receive cessation consultation from providers. Still other studies39^–42 have found older subjects and those with more quit attempts to be more likely to quit, while subjects with higher nicotine dependence were less likely to succeed at quitting smoking.41 43^–45 In our study, over half of the smokers that continued to smoke reported that they failed to quit because they were unable to overcome the craving they felt for a cigarette. Thirty per cent gave other reasons, including as social, peer and work pressures, bad moods and stress. These results are similar to other studies46^–48 that found social influences, such as peer smoking, situational factors and negative effect were important factors triggering relapse to smoke. However, this study and another study49 found that over half of relapsed smokers reported a higher interest in attending cessation services again. This finding suggests that the promotion of repeated tobacco dependence treatment is important in assisting relapsed smokers attempt to quit again. In addition, other intensive interventions, such as individual behaviour therapy and group support, may be needed for some highly nicotine dependent subjects.41

This study has several limitations. First, since Taiwan’s Smoking Cessation Outpatient Services programme was designed to provide cessation services at a national level, so this study did not have a control group. Validity could also be influenced by other situations not included in this study. For example, some smoking cessation media campaigns and other projects, such as the establishment of smoker identification systems in public hospitals, may also promote the utilisation of outpatient cessation services. Second, this study used existing datasets which did not collect information regarding subjects’ psychological characteristics, which might also influence smoking cessation. Third, a major challenge of longitudinal studies is attrition. To understand the representativeness of the sample in the six-month follow-up, demographic characteristics of follow-up sample subjects were found to be similar to the characteristics of subjects obtained from claims data. However, at the six-month follow-up survey, 46 of respondents in 2004 and 55 of respondents in 2005 did not complete the interview—no one answered the phone or the subjects were not at home. This study and England’s smoking treatment study45 found that younger smokers were more likely to be lost to follow-up. Also, the successful quit rate in this study may be overestimated because younger smokers were more likely to relapse. Finally, another problem in estimating a successful quit rate is to consider who should be included in the denominator for computing this outcome. By considering dropouts in the denominator, the low and high estimates of six-month follow-up smoking cessation rates in this study were 13–25 in 2004, while the low-high quit rates were 10–21 in 2005.

Although the World Health Organization advocates the establishment of smoking cessation services at a national level, most countries face financial challenges to achieve this aim.1 Similarly, Taiwan has faced a financial shortage that may reverse the expansion of cessation outpatient services. In January 2006, a new tobacco tax law passed in Taiwan, which increases the tobacco health tax from $NT5 ($0.16) per pack in 2005 to $NT10 ($0.32) in 2006. However, the new tobacco tax law decreases the proportion of heath tax funding allocated for tobacco control programmes from 10 (2002–2005) to 3 in 2006. Facing this funding shortage for tobacco control programmes, in April 2006 the Taiwanese government decided to bring the reimbursement rates and medication subsidies back to the pre-January 2005 levels. The experience in England suggests that integrating smoking cessation services as a mainstream national health service and providing stable financial resources is an important means of speeding up implementation of smoking cessation services.50 However, England44 and Taiwan have found that providing dedicated funding for nationwide smoking cessation outpatient services over a longer period to be a major challenge.

In conclusion, the policy of increasing funding for smoking cessation services appears to have been successful in promoting these services. However, as the number of smoking cessation services increased, total cost also increased dramatically. The efficient allocation of resources that ensure that smoking cessation services remain in agreement with the requirements of policies that finance them is important if their effectiveness is to be improved and maintained. Several studies31 51 52 report that population-wide strategies, such as increasing the tobacco tax, smoke-free environments, media advocacy and accessible treatment, play an important part in changing societal norms and promoting smoking cessation. Finally, it is critical to continue to examine the long-term effects of changing smoking cessation financing on provider delivery of smoking cessation services, consumer utilisation, smoking cessation outcomes and population-based smoking prevalence and cessation rates.