Long-term telephone follow-up of older adults attending a smoking cessation clinic in Türkiye: a retrospective cohort study Authors: Fatma Gökşin Cihan, Hatice Küçükceran, Nur Demirbas, Aynur Yalçıntaş Published: 2026 ABSTRACT AbstractBackgroundSmoking cessation remains beneficial in later life, yet long-term information on smoking status and cessation experiences among older adults attending smoking cessation services is limited. This study descriptively evaluated long-term smoking status and cessation experiences in older adults previously attending a smoking cessation clinic.MethodsThis study used a retrospectively identified cohort with a one-time long-term telephone outcome assessment. Medical records of adults aged ≥ 60 years who attended a smoking cessation outpatient clinic between January 2015 and March 2020 were reviewed, and telephone follow-up was conducted in October 2024, approximately 4.6–9.8 years after clinic attendance. The primary outcome was self-reported smoking status at interview. Recorded treatment modality reflected the treatment plan documented at baseline; medication initiation, duration, and adherence were not available. Multivariable analyses were exploratory.ResultsOf 297 eligible participants, 26 (8.8%) were recorded as deceased. Among the remaining 271 participants, 97 (35.8%) were interviewed. Of these 97 respondents, 35 (36.1%) reported that they were not currently smoking at interview, representing a respondent-only estimate. Under extreme-case assumptions for the 174 non-contacted participants, the proportion not currently smoking ranged from 12.9% to 77.1%. Forty-three respondents (44.3%) reported a longest smoking-free period of ≥ 12 months, although its timing relative to follow-up was unknown. No robust association between recorded treatment modality and smoking status was identified: the overall respondent-only model was not statistically significant, and the treatment-modality association was not retained in the extreme-case sensitivity analysis. Among the 62 respondents currently smoking, 50 (80.6%) expressed willingness to receive support for another quit attempt.ConclusionsLong-term smoking-status estimates were substantially limited by loss to follow-up and should be interpreted as respondent-only findings rather than as an abstinence rate for the original cohort. The treatment-modality analyses were exploratory and do not support comparative treatment-effectiveness conclusions. The high willingness to attempt quitting again among respondents who were still smoking highlights the potential for continued re-engagement with cessation care in later life.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12875-026-03550-9.Keywords: Smoking, Older adults, Nicotine dependence, Tobacco use cessation, Follow-up, Primary care Background Population ageing is accelerating worldwide, with the number of adults aged 60 years or older expected to increase substantially over the coming decades [1]. Smoking remains an important preventable contributor to morbidity and mortality in later life [2]. In Türkiye, daily tobacco-product use among adults aged ≥ 65 years was 11.3% in 2023, with a substantially higher prevalence among men than women [3]. Importantly, smoking cessation continues to provide meaningful health benefits even at advanced ages, including reductions in premature mortality and gains in life expectancy [4, 5]. Despite these benefits, tobacco use may be insufficiently assessed and treated in older adults. Multimorbidity, polypharmacy, concerns about adverse effects and drug interactions, perceived treatment contraindications, and the misconception that quitting offers limited benefit in later life may reduce treatment engagement [6]. Smoking cessation is also a dynamic process characterized by repeated quit attempts and relapse; therefore, continued assessment, renewed behavioral support, and re-engagement with treatment remain important after an unsuccessful quit attempt [7]. Long-term evidence on smoking status after attendance at smoking cessation services remains limited in older populations. Previous studies involving older adults have primarily evaluated outcomes within the first few years after treatment [8, 9]. A study from Türkiye reported that 36% of adults aged ≥ 65 years achieved at least one year of abstinence [9], while a three-year follow-up study from Taiwan reported seven-day point-prevalence abstinence of 37.3% among adults aged ≥ 65 years [10]. However, differences in outcome definitions, follow-up duration, and handling of loss to follow-up limit direct comparisons, and relatively little is known about smoking status, repeated quit attempts, relapse experiences, and willingness to re-engage with cessation care several years after initial clinic attendance. Accordingly, the primary purpose of this study was descriptive: to characterize self-reported current smoking status and subsequent cessation experiences among adults aged ≥ 60 years who had attended a smoking cessation outpatient clinic approximately 4.6–9.8 years earlier. Secondary objectives were to describe subsequent quit attempts, the longest self-reported period of abstinence, willingness to make another quit attempt, and medical conditions reported after initial clinic attendance. Analyses examining associations between selected baseline characteristics (including age, nicotine dependence, cumulative smoking exposure, and recorded treatment modality) and smoking status at follow-up were considered exploratory and hypothesis-generating. No causal or comparative treatment-effectiveness hypothesis was prespecified. Methods Study design, setting, and participants This study used a retrospectively identified cohort with a one-time long-term telephone outcome assessment. The study was conducted at the Smoking Cessation Outpatient Clinic (SCOC) of Necmettin Erbakan University Faculty of Medicine in Konya, Türkiye. Medical records of all adults aged ≥ 60 years who attended the SCOC between 1 January 2015 and 1 March 2020 were reviewed. All patients meeting this age criterion who attended the clinic at least once during the study period were included in the original cohort, irrespective of whether they set a target quit date, attended a minimum number of counseling sessions, or completed the three-month program. No additional exclusion criteria were applied. During the study period, 297 adults aged ≥ 60 years attended the clinic, representing 14.8% of the 2,006 total clinic attendees. The one-time long-term telephone outcome assessment was conducted between 5 and 15 October 2024. Before telephone contact was attempted, mortality status was checked individually for all members of the original clinic cohort through the hospital electronic information system. This system is linked to the Central Civil Registration System (MERNIS), Türkiye’s national civil-registration database; therefore, mortality ascertainment was not restricted to deaths occurring within the study hospital. Participants recorded as deceased were not contacted. No additional independent mortality-registry linkage was undertaken specifically for this study. For the remaining participants, telephone contact was attempted up to three times at different times of the day (before noon, in the afternoon, and in the evening). A total of 97 participants were successfully contacted, provided verbal informed consent, and completed the telephone interview. Based on the clinic-attendance period and the date of the telephone assessment, the possible interval between initial clinic attendance and follow-up ranged from approximately 4.6 to 9.8 years. Participant-specific initial attendance dates were not retained in the analytical dataset and could not be reliably reconstructed for the present analysis. Consequently, individual follow-up duration could not be calculated or included as a covariate in the statistical models. Participants with longer follow-up periods had more opportunity both to relapse and to make additional quit attempts; therefore, time-related residual bias cannot be excluded. Data sources and baseline measurements Baseline demographic information, including age, sex, educational level, employment status, and marital status, was extracted from medical records. Marital status was originally recorded as married, single, or widowed; where a dichotomous variable was required for analysis, single and widowed participants were combined as “not married.” Information on medical conditions diagnosed after the initial clinic attendance was obtained retrospectively from participants during the telephone interview. Smoking-related baseline data included age at smoking initiation, daily cigarette consumption, smoking duration, cumulative tobacco exposure, and exhaled carbon monoxide (CO) level. Pack-years were calculated by multiplying the number of cigarette packs smoked per day by the number of years of smoking. Baseline exhaled CO was measured using a piCO Smokerlyzer device (Bedfont Scientific Ltd., Maidstone, England) and recorded in parts per million (ppm). Baseline nicotine dependence was assessed using the six-item Fagerström Test for Nicotine Dependence (FTND), with total scores ranging from 0 to 10 [11, 12]. Higher scores indicate greater nicotine dependence. For descriptive purposes, scores < 6 were categorized as low-to-moderate nicotine dependence and scores ≥ 6 as high nicotine dependence, consistent with previous studies using an FTND score of ≥ 6 to identify high nicotine dependence [13]. Smoking cessation treatment and recorded treatment modality The SCOC provided a structured three-month smoking cessation program. Behavioral counseling was part of the standard clinic approach and included assessment of smoking behavior, motivational support, identification of smoking triggers, development of coping strategies, and relapse-prevention advice. At the initial clinic assessment, the treatment modality documented in the medical record was categorized as behavioral counseling alone or pharmacotherapy plus behavioral counseling. Pharmacological agents recorded in the clinic files included nicotine replacement therapy, bupropion, and varenicline. Selection of the treatment plan was based on clinical assessment, nicotine dependence, health status, contraindications, patient preference, medication availability, and the clinical guidelines in use at the time [14]. Importantly, recorded treatment modality reflected the treatment plan documented at the initial clinic visit and should not be interpreted as confirmation that medication was initiated, dispensed, taken for the recommended duration, or adhered to. Standardized information on medication initiation, treatment duration, adherence, discontinuation, the number of counseling sessions attended, and formal program completion was not consistently available in the analytical dataset. Information on subsequent smoking-cessation treatments or the method used in later quit attempts was also not collected. Smoking status at the end of the three-month program was not systematically recorded. Consequently, the study could not evaluate actual treatment exposure, adherence, program completion, immediate treatment effectiveness, or comparative treatment effectiveness. Telephone follow-up and outcome assessment A structured five-item telephone interview guide was developed by the investigators specifically for this study. Participants were asked: (1) “What is your current smoking status?”; (2) “How many times have you tried to quit smoking?”; (3) “What was the longest period you were able to remain without smoking?”; (4) “Have you received any new diagnosis after your initial attendance at the smoking cessation clinic?”; and, for participants who were currently smoking, (5) “Would you like to receive support to try quitting smoking again?” Before the main follow-up, the interview guide was pretested with three older adults to assess clarity, comprehensibility, sequence, and feasibility of telephone administration. These individuals were not included in the final analysis, and no substantive changes were required following the pretest. The guide did not undergo formal cognitive testing, psychometric validation, or external validation. An English-language version of the interview guide is provided as supplementary material. All telephone interviews were conducted by a single investigator who was one of the study authors and had training and clinical experience in smoking cessation care. The interviewer was unaware of the participants’ recorded baseline treatment modality at the time of the telephone interview. Interviews were conducted directly with participants, and proxy responses from relatives or caregivers were not accepted. Responses were entered directly into a standardized Microsoft Excel data-entry form during the interview. When a response was unclear or internally inconsistent, the interviewer asked additional questions during the same call to clarify the response before recording the final response. The primary outcome was self-reported smoking status at the time of the telephone interview. Participants who reported that they were not smoking at the time of the interview were classified as “self-reported not currently smoking at interview,” whereas those who reported smoking were classified as “self-reported currently smoking at interview.” No fixed recall window, such as the preceding 7 or 30 days, was specified, and smoking status was not biochemically verified. Non-daily smoking, use of other tobacco products, and nicotine-containing products were not assessed separately. Accordingly, this outcome should not be interpreted as conventional point-prevalence, continuous, or prolonged abstinence. The number of quit attempts and the longest self-reported period without smoking were obtained from the telephone interview. The timing and sequence of quit and relapse episodes were not recorded; therefore, individual cessation–relapse trajectories could not be reconstructed. The longest self-reported smoking-free period was additionally categorized descriptively as < 1 month, 1 to < 6 months, 6 to < 12 months, and ≥ 12 months. These categories describe the longest smoking-free period reported by participants and do not indicate continuous abstinence immediately preceding the telephone assessment. Exact cessation and relapse dates were unavailable, and time-to-relapse or survival analyses could not be performed. Statistical analysis Continuous variables were evaluated according to their distribution. Normally distributed variables are presented as mean ± standard deviation, whereas non-normally distributed variables are presented as median and first and third quartiles (Q1–Q3). Minimum and maximum values are also reported where informative. Categorical variables are presented as numbers and percentages. No a priori sample-size calculation was performed because the study included all eligible adults aged ≥ 60 years who attended the clinic during the predefined study period. The original cohort comprised 297 eligible clinic attendees. Long-term smoking-status information was obtained from 97 interviewed participants, of whom 35 reported that they were not currently smoking at interview. Given the limited number of outcome events, multivariable analyses were intentionally parsimonious and were considered exploratory and hypothesis-generating. Baseline characteristics were compared between contacted and non-contacted participants after excluding participants recorded as deceased. Baseline characteristics were also compared according to recorded treatment modality among interviewed participants. Student’s t-test was used for normally distributed continuous variables and the Mann–Whitney U test for non-normally distributed continuous variables. Categorical variables were compared using Pearson’s chi-square test or two-sided Fisher’s exact test, as appropriate. The categories of the longest self-reported smoking-free period were summarized descriptively according to smoking status at interview. The multivariable analysis was not designed to estimate a causal treatment effect. A clinically informed framework was used to consider factors that could plausibly influence both recorded treatment modality and long-term smoking status, including sociodemographic characteristics, baseline smoking severity, comorbidity, treatment contraindications, motivation and previous quit history, medication availability and adherence, and follow-up duration. However, several of these factors were unavailable or incompletely measured, and only 35 respondents reported that they were not currently smoking. To reduce the risk of overfitting and unstable estimates, the exploratory multivariable model was restricted to age at initial clinic attendance, baseline FTND score, pack-years, and recorded treatment modality. The model should therefore not be interpreted as a causal adjustment model or as an estimate of comparative treatment effectiveness. An exploratory multivariable binary logistic regression analysis was performed among all 97 interviewed participants. Self-reported smoking status at interview was coded as 1 for not currently smoking and 0 for currently smoking. Recorded treatment modality was entered as a categorical variable, with pharmacotherapy plus behavioral counseling as the reference category. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were reported. Model performance was evaluated using the omnibus likelihood-ratio test, Nagelkerke’s R², and the Hosmer–Lemeshow goodness-of-fit test. Participant-specific follow-up duration could not be incorporated into the model because exact individual clinic-attendance dates were unavailable. To illustrate the uncertainty associated with missing follow-up smoking-status data, two extreme-case estimates were calculated after excluding the 26 participants recorded as deceased. In one extreme scenario, all 174 non-contacted participants were classified as currently smoking; in the opposite extreme scenario, all non-contacted participants were classified as not currently smoking. These scenarios were intended to illustrate the range of possible overall smoking-status estimates and were not considered formal corrections for missing follow-up data. An additional exploratory multivariable analysis was performed among all 271 participants not recorded as deceased under the extreme assumption that all 174 non-contacted participants were currently smoking. The same four baseline variables were included as in the respondent-only model. Because the probability of successful follow-up may have differed according to recorded treatment modality or other baseline characteristics, this analysis was not considered inherently conservative for the treatment-modality association and should not be interpreted as correcting for differential loss to follow-up. All statistical tests were two-sided, and P < 0.05 was considered statistically significant. Results Between January 2015 and March 2020, 297 adults aged ≥ 60 years attended the smoking cessation outpatient clinic, representing 14.8% of the 2,006 total clinic attendees during this period (Fig. 1). Their mean age at initial clinic attendance was 64.80 ± 4.34 years (range, 60–83 years). Baseline sociodemographic characteristics are presented in Table 1. Flowchart of patient selection for the study. The flowchart shows the stepwise selection of patients from all smokers attending the smoking cessation outpatient clinic (SCOC) between January 2015 and March 2020. Of 2006 total attendees, 297 were aged ≥ 60 years and were eligible for inclusion. Of these, 26 (8.8%) were recorded as deceased by the time of telephone follow-up in October 2024 and were not contacted. Of the remaining 271 participants, 97 were successfully reached and interviewed, whereas 174 could not be contacted. The n values shown in each box represent participant counts at each stage of the study. SCOC: smoking cessation outpatient clinic Table 1Demographic characteristics of older adult patients at the smoking cessation clinic (N = 297)Characteristic n %Sex Female268.8 Male27191.2Education Illiterate3612.1 Primary school19064.0 High school3913.1 University3210.8Working status Working8127.2 Not working21672.8Marital status Married23378.4 Single3511.8 Widowed299.8 Total297100 Demographic characteristics of older adult patients at the smoking cessation clinic (N = 297) At the time of the long-term follow-up, 26 of the 297 participants (8.8%) were recorded as deceased and were not contacted. Of the remaining 271 participants, 97 (35.8%) were successfully contacted and completed the telephone interview, whereas 174 could not be reached because their registered telephone numbers were inactive or repeated telephone calls were unanswered. Baseline characteristics of contacted and non-contacted participants are compared in Supplementary Table S1. Contacted participants were less frequently married than non-contacted participants (70.1% vs. 82.8%; P = 0.016) and had higher baseline exhaled carbon monoxide levels [median 9.0 (Q1–Q3: 6.0–11.5) vs. 7.5 (Q1–Q3: 5.0–11.0) ppm; P = 0.019]. No statistically significant differences were observed in age, sex, educational level, employment status, baseline FTND score, pack-years, or recorded treatment modality (all P > 0.05). In the overall cohort, the median cumulative smoking exposure was 50 pack-years, the mean baseline FTND score was 5.69 ± 2.50, and the median baseline exhaled carbon monoxide level was 8 ppm (Q1–Q3: 5.5–11.0; range, 1–26 ppm). Baseline nicotine-dependence characteristics and recorded treatment modalities are presented in Table 2. Table 2Baseline nicotine-dependence characteristics and recorded smoking-cessation treatment plan (N = 297)Variable n %Cigarettes per day (FTND item) ≤10279.1 11–2012742.8 21–306722.5 ≥317625.6Time to first cigarette after waking (FTND item) Within 5 min10435.0 6–30 min7224.3 31–60 min6923.2 After 60 min5217.5Difficulty refraining in no-smoking areas (FTND item) Yes16956.9 No12843.1Most hated cigarette to give up (FTND item) First one in the morning17358.2 Any other12441.8Smoke more in first hour after waking (FTND item) Yes13645.8 No16154.2Smoke when ill in bed (FTND item) Yes17257.9 No12542.1Recorded treatment plan BC + Varenicline12843.1 BC + Bupropion6923.2 BC + NRT (patches)4013.4 BC only6020.3Recorded treatment plan reflected the treatment modality documented at the initial clinic visit and did not confirm medication initiation, dispensing, duration, or adherence. All six items from the Fagerström Test for Nicotine Dependence (FTND) are listed. FTND total scores range from 0 to 10. For descriptive purposes, scores < 6 were categorized as low-to-moderate nicotine dependence and scores ≥ 6 as high nicotine dependenceBC Behavioral counseling, NRT Nicotine replacement therapy (transdermal patches) Baseline nicotine-dependence characteristics and recorded smoking-cessation treatment plan (N = 297) Recorded treatment plan reflected the treatment modality documented at the initial clinic visit and did not confirm medication initiation, dispensing, duration, or adherence. All six items from the Fagerström Test for Nicotine Dependence (FTND) are listed. FTND total scores range from 0 to 10. For descriptive purposes, scores < 6 were categorized as low-to-moderate nicotine dependence and scores ≥ 6 as high nicotine dependence BC Behavioral counseling, NRT Nicotine replacement therapy (transdermal patches) Among the 97 interviewed respondents, 35 (36.1%) reported that they were not currently smoking at the time of the interview, whereas 62 (63.9%) reported that they were currently smoking. To illustrate the uncertainty associated with missing follow-up smoking-status data, two extreme-case estimates were calculated after excluding participants recorded as deceased. When all 174 non-contacted participants were classified as currently smoking, the proportion not currently smoking was 12.9% (35/271). Under the opposite extreme assumption that all non-contacted participants were classified as not currently smoking, the corresponding proportion was 77.1% (209/271). Respondents reported a median of one quit attempt (Q1–Q3: 1–3; range, 1–8). The median longest self-reported period without smoking was 4 months (Q1–Q3: 0.19–36 months; range, 0–120 months). When categorized descriptively, the longest reported smoking-free period was < 1 month for 27 respondents (27.8%), 1 to < 6 months for 23 (23.7%), 6 to < 12 months for 4 (4.1%), and ≥ 12 months for 43 (44.3%). Among the 35 respondents who reported not currently smoking at interview, 33 (94.3%) reported a longest smoking-free period of ≥ 12 months. Of the 43 respondents whose longest reported smoking-free period was ≥ 12 months, 10 (23.3%) were currently smoking at interview. The timing of the longest smoking-free period relative to the follow-up interview was not recorded; therefore, these data should not be interpreted as continuous abstinence for 1, 6, or 12 months immediately preceding the interview. In the bivariate comparison of recorded treatment modality, 9 of 14 respondents (64.3%) with behavioral counseling alone documented as their baseline treatment plan and 26 of 83 (31.3%) with pharmacotherapy plus behavioral counseling documented as their baseline treatment plan reported that they were not currently smoking at interview (P = 0.018). Baseline characteristics according to recorded treatment modality are presented in Supplementary Table S2. No statistically significant differences were observed between the recorded treatment-modality groups in the measured baseline characteristics (all P > 0.05). Comparisons of sociodemographic and smoking-related characteristics according to smoking status at interview are presented in Table 3. Table 3Characteristics of interviewed participants according to self-reported smoking status at interview (N = 97)VariableAll interviewed,n (%)Not currently smoking,n (%)Currently smoking,n (%)P valueRecorded treatment modality0.018ᵃ Behavioral counseling alone14 (14.4)9 (64.3)5 (35.7) Pharmacotherapy plus behavioral counseling83 (85.6)26 (31.3)57 (68.7)Recorded pharmacological agent (n = 83)0.084ᵃ Varenicline + BC46 (55.4)19 (41.3)27 (58.7) Bupropion + BC19 (22.9)3 (15.8)16 (84.2) NRT + BC18 (21.7)4 (22.2)14 (77.8)Age at initial clinic attendance, years0.734ᵃ 60–6456 (57.7)21 (37.5)35 (62.5) ≥ 6541 (42.3)14 (34.1)27 (65.9)Sex1.000ᵇ Female9 (9.3)3 (33.3)6 (66.7) Male88 (90.7)32 (36.4)56 (63.6)Marital status0.255ᵃ Married68 (70.1)27 (39.7)41 (60.3) Not married29 (29.9)8 (27.6)21 (72.4)Education0.511ᵃ Primary school or lower73 (75.3)25 (34.2)48 (65.8) High school or higher24 (24.7)10 (41.7)14 (58.3)Number of quit attempts< 0.001ᵃ 156 (57.7)30 (53.6)26 (46.4) 213 (13.4)1 (7.7)12 (92.3) ≥ 328 (28.9)4 (14.3)24 (85.7)Nicotine dependence (FTND)0.710ᵃ Low-to-moderate (FTND < 6)44 (45.4)15 (34.1)29 (65.9) High (FTND ≥ 6)53 (54.6)20 (37.7)33 (62.3)New diagnosis after SCOC visit0.077ᵃ Yes64 (66.0)19 (29.7)45 (70.3) No31 (32.0)15 (48.4)16 (51.6) Missing2 (2.1)1 (50.0)1 (50.0)Data are presented as n (%). Percentages in the “All interviewed” column were calculated using N = 97, except for recorded pharmacological-agent categories, for which percentages were calculated among the 83 participants with pharmacotherapy plus behavioral counseling documented as their baseline treatment modality. Percentages in the “Not currently smoking” and “Currently smoking” columns are row percentages. Two participants had missing information on new diagnoses, and the corresponding P value was calculated using available dataAge refers to age at initial clinic attendance. “Not married” included participants who were single or widowed. “Primary school or lower” included participants who were illiterate or primary-school graduates; “high school or higher” included high-school and university graduates. Recorded treatment modality reflected the treatment plan documented at the initial clinic visit and did not confirm medication initiation, duration, or adherenceBC behavioral counseling FTND Fagerström Test for Nicotine Dependence, NRT Nicotine replacement therapy, SCOC Smoking cessation outpatient clinicᵃ Pearson’s chi-square test. ᵇ Two-sided Fisher’s exact test Characteristics of interviewed participants according to self-reported smoking status at interview (N = 97) Data are presented as n (%). Percentages in the “All interviewed” column were calculated using N = 97, except for recorded pharmacological-agent categories, for which percentages were calculated among the 83 participants with pharmacotherapy plus behavioral counseling documented as their baseline treatment modality. Percentages in the “Not currently smoking” and “Currently smoking” columns are row percentages. Two participants had missing information on new diagnoses, and the corresponding P value was calculated using available data Age refers to age at initial clinic attendance. “Not married” included participants who were single or widowed. “Primary school or lower” included participants who were illiterate or primary-school graduates; “high school or higher” included high-school and university graduates. Recorded treatment modality reflected the treatment plan documented at the initial clinic visit and did not confirm medication initiation, duration, or adherence BC behavioral counseling FTND Fagerström Test for Nicotine Dependence, NRT Nicotine replacement therapy, SCOC Smoking cessation outpatient clinic ᵃ Pearson’s chi-square test. ᵇ Two-sided Fisher’s exact test Among the 97 interviewed respondents, 64 (66.0%) reported receiving at least one new medical diagnosis after their initial clinic attendance. The most frequently reported diagnoses were chronic obstructive pulmonary disease (n = 23) and malignancy (n = 6). Among the 62 respondents who reported currently smoking at interview, 50 (80.6%) stated that they would like to receive support for another quit attempt. An exploratory multivariable binary logistic regression analysis was conducted among all 97 interviewed respondents, with self-reported smoking status at interview as the dependent variable (1 = not currently smoking; 0 = currently smoking). Age at initial clinic attendance, baseline FTND score, pack-years, and recorded treatment modality were included in the model. The overall model was not statistically significant, χ²(4) = 7.68, P = 0.104, with a Nagelkerke R² of 0.104. The Hosmer–Lemeshow test showed no evidence of poor model fit, χ²(8) = 9.08, P = 0.336. Age at initial clinic attendance (adjusted OR = 0.93, 95% CI: 0.83–1.05; P = 0.250), baseline FTND score (adjusted OR = 0.95, 95% CI: 0.76–1.17; P = 0.605), and pack-years (adjusted OR = 1.01, 95% CI: 0.99–1.04; P = 0.246) were not clearly associated with smoking status at interview. Respondents with behavioral counseling alone recorded as their baseline treatment modality had higher odds of reporting that they were not currently smoking than those with pharmacotherapy plus behavioral counseling recorded at baseline (adjusted OR = 4.82, 95% CI: 1.38–16.86; P = 0.014). However, given the non-significant overall model, small behavioral counseling-only group, wide confidence interval, non-random treatment allocation, and inability to account for several important potential confounders and individual follow-up duration, this association was considered exploratory and should not be interpreted as evidence of comparative treatment effectiveness (Table 4). Table 4Exploratory multivariable logistic regression analysis of selected baseline characteristics associated with self-reported not currently smoking at interview (N = 97)PredictorAdjusted OR95% CIP valueAge at initial clinic attendance, per one-year increase0.930.83–1.050.250Baseline FTND score, per one-point increase0.950.76–1.170.605Pack-years, per one-unit increase1.010.99–1.040.246Recorded treatment modality: behavioral counseling alone¹4.821.38–16.860.014The dependent variable was self-reported smoking status at the telephone interview (1 = not currently smoking; 0 = currently smoking). All predictors were entered simultaneously using the Enter method. The analysis included all 97 interviewed respondents, with no cases excluded because of missing dataOR odds ratio, CI confidence interval, FTND Fagerström Test for Nicotine Dependence¹The reference category was recorded pharmacotherapy plus behavioral counseling. Recorded treatment modality reflected the treatment plan documented at the initial clinic visit and did not confirm medication initiation, duration, or adherenceOverall model: χ²(4) = 7.678, P = 0.104; Nagelkerke R²=0.104; Hosmer–Lemeshow χ²(8) = 9.080, P = 0.336 Exploratory multivariable logistic regression analysis of selected baseline characteristics associated with self-reported not currently smoking at interview (N = 97) The dependent variable was self-reported smoking status at the telephone interview (1 = not currently smoking; 0 = currently smoking). All predictors were entered simultaneously using the Enter method. The analysis included all 97 interviewed respondents, with no cases excluded because of missing data OR odds ratio, CI confidence interval, FTND Fagerström Test for Nicotine Dependence ¹The reference category was recorded pharmacotherapy plus behavioral counseling. Recorded treatment modality reflected the treatment plan documented at the initial clinic visit and did not confirm medication initiation, duration, or adherence Overall model: χ²(4) = 7.678, P = 0.104; Nagelkerke R²=0.104; Hosmer–Lemeshow χ²(8) = 9.080, P = 0.336 An additional exploratory extreme-case sensitivity analysis was conducted among all 271 participants not recorded as deceased, with all 174 non-contacted participants classified as currently smoking. The overall model was not statistically significant, χ²(4) = 1.95, P = 0.746, with a Nagelkerke R² of 0.013; the Hosmer–Lemeshow test was P = 0.202. None of the selected baseline variables was clearly associated with smoking status under this extreme-case assumption. In particular, the association involving recorded behavioral counseling alone was attenuated and was no longer statistically significant (adjusted OR = 1.42, 95% CI: 0.62–3.24; P = 0.409). This analysis was intended to illustrate sensitivity to missing follow-up smoking-status data and was not considered a formal correction for differential loss to follow-up or a conservative estimate of the treatment-modality association (Supplementary Table S3). Discussion This study involved a retrospectively identified cohort with a one-time long-term telephone outcome assessment among adults aged ≥ 60 years who had attended a smoking cessation outpatient clinic approximately 4.6–9.8 years earlier. Among the 97 participants who were successfully contacted, 35 (36.1%) reported that they were not currently smoking. However, this proportion represents the percentage of respondents who reported not currently smoking at interview and should not be interpreted as the long-term smoking-status estimate of the entire original clinic cohort. After excluding 26 participants recorded as deceased, smoking status remained unknown for 174 of 271 participants. The best- and worst-case sensitivity estimates ranged from 12.9% to 77.1%, demonstrating the substantial uncertainty introduced by missing follow-up data. The proportion reporting not currently smoking at interview was numerically similar to findings from some previous studies of older adults attending smoking cessation services. Ozturk et al. reported that 36% of adults aged ≥ 65 years achieved at least one year of abstinence [9]. In a three-year follow-up study from Taiwan, prolonged abstinence and seven-day point-prevalence abstinence among participants aged ≥ 65 years were 20.1% and 37.3%, respectively [10]. However, direct comparisons should be made cautiously because studies differ in follow-up duration, treatment setting, handling of non-responders, biochemical verification, and abstinence definitions. The outcome in the present study reflected self-reported smoking status at a single follow-up interview and did not represent continuous, prolonged, or conventionally defined point-prevalence abstinence. Because the present respondent-only proportion spans a multi-year interval, its clinical meaning differs from single-year quit rates. If the 36.1% respondent-only estimate is annualized over the possible 4.6–9.8-year follow-up range, it corresponds to an approximate annual rate of 3.7–7.8%; applied instead to the extreme worst-case 12.9% estimate for the full eligible cohort, the annualized rate is lower still. For context, the 2016 Turkey Global Adult Tobacco Survey reported that 24.6% of Turkish tobacco users had attempted to quit in the preceding year and that 13.6% of former daily tobacco users nationally had successfully quit [15]. These figures are not directly comparable — they describe a general adult population rather than treatment-seeking older clinic attendees, use different reference periods, and are not annualized in the same way — but they suggest that the present cohort’s long-term outcomes are not obviously more favorable than national experience, reinforcing the need for caution in interpreting the respondent-only estimate as evidence of durable treatment benefit. The additional follow-up findings illustrate the recurrent nature of smoking cessation and relapse. Participants reported a median of one subsequent quit attempt, and the median longest self-reported abstinence period was four months. Overall, 43 participants (44.3%) reported having achieved at least one abstinence period lasting 12 months or longer. Of these, 10 (23.3%) were currently smoking at follow-up, indicating a return to smoking after a prolonged abstinence period. Because exact cessation and relapse dates were unavailable, these findings should not be interpreted as standardized 12-month continuous abstinence or time-to-relapse outcomes. Nevertheless, they demonstrate that a return to smoking may occur even after relatively long periods without smoking. Importantly, 80.6% of participants who were currently smoking expressed willingness to make another quit attempt. This finding supports viewing tobacco dependence as a chronic, relapsing condition in which renewed treatment remains appropriate after relapse. A randomized primary care trial found that immediate encouragement to re-enter treatment after relapse increased treatment re-engagement and subsequent abstinence compared with referral alone [16]. Thus, relapse should prompt renewed assessment and treatment rather than withdrawal of cessation support. In the respondent-only multivariable analysis, age at initial clinic attendance, baseline FTND score, and pack-years were not independently associated with self-reported smoking status at interview. These null findings should not be interpreted as evidence that these factors have no role in long-term smoking cessation. Only 35 participants reported that they were not currently smoking at interview, and the regression model was deliberately restricted to four clinically selected variables to reduce the risk of overfitting. Furthermore, the overall model was not statistically significant and explained only a modest proportion of the variability in current smoking status. The adjusted findings should therefore be regarded as exploratory rather than as definitive evidence of factors associated with long-term smoking status. Among contacted participants, having behavioral counseling alone recorded as the baseline treatment plan was associated with higher odds of reporting not currently smoking at interview than having pharmacotherapy plus behavioral counseling recorded. However, several findings argue against interpreting this association as evidence that behavioral counseling alone is more effective. First, only 14 contacted participants had behavioral counseling alone recorded as their baseline treatment plan, resulting in a wide confidence interval and an imprecise estimate. Second, although the measured baseline characteristics (including age, sex, education, marital status, FTND score, pack-years, and exhaled carbon monoxide level) did not differ significantly between treatment groups, the small behavioral counseling-only group limited the power of these comparisons. The absence of statistically significant differences does not establish clinical comparability. Third, treatment allocation was not randomized and was influenced by clinical assessment, patient preference, health status, contraindications, and medication availability. Unmeasured differences in comorbidity burden, motivation, previous cessation experiences, adverse effects, and clinician judgment may therefore have influenced treatment selection. Information on program completion, counseling-session attendance, medication initiation, duration of use, adherence, and discontinuation was also unavailable. Residual confounding and confounding by indication cannot be excluded. Most importantly, the treatment-modality association observed in the respondent-only analysis was not retained in the extreme-case sensitivity analysis in which all non-contacted participants were classified as currently smoking. When all 174 non-contacted participants were classified as currently smoking, having behavioral counseling alone recorded as the baseline treatment plan was no longer associated with reporting not currently smoking. This indicates that the original association was sensitive to assumptions regarding loss to follow-up. Previous research among older clinic attendees has found combined pharmacological and behavioral treatment and greater follow-up attendance to be associated with one-year abstinence [9]. The present finding should therefore be interpreted only as an exploratory observational association and not as evidence of comparative treatment effectiveness. Two-thirds of the contacted participants reported receiving at least one new medical diagnosis after their initial clinic attendance, most commonly chronic obstructive pulmonary disease or malignancy. However, the study did not provide precise evidence of an association between a reported new diagnosis and self-reported smoking status at interview. Interpretation of this finding is further limited by the retrospective and self-reported nature of the diagnosis data. Nevertheless, clinical encounters related to newly diagnosed conditions may provide opportunities to reassess tobacco use and offer evidence-based smoking-cessation support. More than half of the original cohort reported smoking even when ill in bed, indicating substantial nicotine dependence and the difficulty of stopping during illness. Hospitalization and acute illness may nevertheless provide important opportunities to initiate treatment. A 2024 Cochrane review found that smoking cessation counseling initiated during hospitalization and continued for more than one month after discharge increased cessation rates; continuation of counseling and pharmacotherapy following discharge also improved outcomes [17]. Linking hospital-based cessation interventions with continued primary care follow-up may therefore be particularly valuable. Previous research from France found that smoking status was not routinely assessed or documented in many older patients and that cessation support was not consistently offered [6]. Together with the high willingness to attempt quitting again observed in the present study, these findings suggest that primary care encounters may contain missed opportunities for repeated assessment, brief advice, individualized pharmacological treatment, and re-referral to structured cessation services. Older age should not be regarded as a reason to withhold evidence-based smoking cessation support. Strengths and limitations This study provides relatively uncommon long-term follow-up information on older adults who attended a smoking cessation outpatient clinic. Additional strengths include the use of baseline clinical records, comparison of measured baseline characteristics between contacted and non-contacted participants, extreme-case estimates for missing outcomes, comparison of treatment groups at baseline, and multivariable analyses incorporating an extreme-case sensitivity scenario. The additional analyses allowed the robustness of the respondent-only findings to be examined and demonstrated the substantial influence of loss to follow-up. Several important limitations should be considered. First, only 97 of the original 297 eligible participants were successfully contacted. Although non-contact was primarily attributable to inactive telephone numbers or unanswered calls rather than explicit refusal, smoking status remained unknown for 174 participants after known deaths were excluded. Contacted participants were less frequently married and had slightly higher baseline exhaled carbon monoxide levels than non-contacted participants. Although the absolute difference in carbon monoxide level was small, these differences indicate that respondents may not have been fully representative of the original cohort. The respondent-only proportion reporting not currently smoking at interview (36.1%) therefore cannot be generalized to the entire eligible cohort. In addition, because participants who died before the long-term assessment could not be interviewed, the 97 respondents necessarily represent identifiable survivors of the original cohort; competing mortality occurring over the 4.6–9.8-year interval may have altered the composition of the population available for follow-up, and the respondent-only findings should be interpreted as describing outcomes among survivors rather than the original clinic cohort as a whole. Second, the primary outcome was self-reported smoking status at the time of the telephone interview, categorized as not currently smoking or currently smoking. Participants who had previously relapsed and subsequently quit again were classified according to their smoking status at follow-up. The outcome did not establish continuous or prolonged abstinence because no fixed recall period, such as seven or 30 days, was used. Smoking status was not biochemically verified; consequently, misclassification and social-desirability bias cannot be excluded. Third, the long interval between initial clinic attendance and telephone follow-up introduced substantial potential for recall bias. Participants may not have accurately recalled the number and timing of subsequent quit attempts, relapse episodes, their longest abstinence period, or newly diagnosed medical conditions. These variables may therefore have been under- or overestimated, particularly among participants who experienced several quit–relapse cycles. Current smoking status required less retrospective recall because it referred to smoking at the time of the interview, but it remained based entirely on self-report. Fourth, exact individual clinic-attendance and program-completion dates were unavailable in the analytical dataset. Follow-up duration could therefore be described only using the overall possible range of approximately 4.6–9.8 years; its individual distribution, mean, and median could not be calculated. Participant-specific follow-up duration could not be included in the multivariable analyses. Variation in follow-up time may have influenced both the probability of relapse and the observed associations. Fifth, several important limitations concern the recorded treatment-modality comparison. Treatment was not randomly assigned, and residual confounding by indication cannot be excluded, since treatment selection reflected clinical assessment, patient preference, health status, contraindications, and medication availability rather than random allocation. Recorded treatment modality reflects the plan documented at the initial clinic visit rather than confirmed exposure: standardized information on program completion, the number of counseling sessions attended, medication initiation, adherence, discontinuation, and adverse effects was unavailable, so actual treatment exposure could not be confirmed and exposure misclassification cannot be excluded. The behavioral counseling-only group was small, the overall respondent-only regression model was not statistically significant, and its treatment estimate had a wide confidence interval. Individual follow-up duration could not be calculated and therefore could not be evaluated for potential differential follow-up by treatment group. In addition, the treatment-modality association observed in the respondent-only analysis was not retained in the extreme-case sensitivity analysis in which all non-contacted participants were classified as currently smoking. Consequently, the study cannot evaluate immediate program effectiveness, standardized 12-month treatment outcomes, or differences between program completers and non-completers, and the treatment-modality finding should not be interpreted causally or as evidence of comparative treatment effectiveness. Sixth, although the author-developed telephone interview guide was pretested with three older adults for clarity and feasibility, these individuals were not included in the main study, and the guide did not undergo formal cognitive testing, psychometric validation, or external validation. Variation in the interpretation of questions, particularly those concerning historical cessation and relapse experiences, cannot be excluded. Finally, this was a single-center study with a predominantly male cohort, which limits generalizability to other settings and to older women who smoke. The COVID-19 pandemic may also have contributed to mortality, changes in residence, inactive telephone numbers, and loss of contact during the follow-up period. Conclusions Among 97 interviewed respondents, 35 reported that they were not currently smoking at the time of the long-term telephone assessment. However, substantial loss to follow-up and uncertainty related to missing smoking-status data prevent this respondent-only proportion from being interpreted as a long-term smoking-status estimate for the original clinic cohort. Most respondents who were currently smoking expressed willingness to receive support for another quit attempt, highlighting the continued potential for re-engagement with cessation care in later life. The exploratory association between recorded behavioral counseling alone and smoking status at interview should not be interpreted as evidence of comparative treatment effectiveness. Treatment allocation was not randomized, recorded treatment modality did not confirm actual medication exposure or adherence, the behavioral counseling-only group was small, individual follow-up duration was unavailable, and the association was not retained in the extreme-case sensitivity analysis. Consistent with established smoking-cessation guidance, older adults should continue to be assessed for tobacco use and offered appropriate evidence-based cessation support, including renewed treatment opportunities after relapse. This clinical recommendation reflects established evidence rather than a comparative treatment effect demonstrated by the present study. Prospective studies using standardized abstinence definitions, biochemical verification, individual follow-up durations, detailed treatment-exposure and adherence data, and strategies to minimize loss to follow-up are needed to better characterize long-term smoking cessation outcomes in older adults. Supplementary Information Supplementary Material 1. Supplementary Material 2. Acknowledgements The authors acknowledge all study respondents for their participation. Authors’ information FGC is a Professor of Family Medicine at Duzce University School of Medicine, Duzce, Türkiye, with research interests in tobacco cessation, AI in healthcare, and primary care. HK is an Assistant Professor of Family Medicine at Necmettin Erbakan University. ND is an Associate Professor of Family Medicine at Necmettin Erbakan University. AY is a General Practitioner at Karaaslan Family Healthcare Center, Konya, Türkiye. Abbreviations Behavioral counseling Carbon monoxide Chronic obstructive pulmonary disease Fagerström Test for Nicotine Dependence Nicotine replacement therapy Parts per million Smoking cessation outpatient clinic Authors’ contributions FGC: conceptualization, methodology, investigation, writing – review & editing. HK: methodology, investigation, data curation, formal analysis, writing – original draft. ND: conceptualization, methodology, investigation, data curation, formal analysis, writing – original draft, writing – review & editing. AY: conceptualization, data curation, writing – original draft, writing – review & editing. All authors read and approved the final manuscript. Funding No funding was received for the preparation or submission of this manuscript. Data availability The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate Ethical approval was obtained from the Necmettin Erbakan University Ethics Committee on 4 October 2024 (decision no. 2024/5241). The approval covered retrospective review of clinic records, telephone follow-up of contactable participants, and inclusion of participants recorded as deceased or not contacted in descriptive and sensitivity analyses. The study was conducted in accordance with the Declaration of Helsinki. Before the telephone interview, the purpose of the study, voluntary participation, and confidentiality were explained verbally, and verbal informed consent was obtained and recorded by the interviewer. Consent for publication Not applicable. This manuscript does not contain data from any individual person in any identifiable form. Competing interests FGC acts as the Guest Editor for the “Managing substance use disorders collection”. The other authors declare that they have no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.