Abstract

Background

It is well known that patients’ health risk behavior will affect the survival outcome of diseases. Smoking alone is a significant health risk behavior that affects the survival outcome of lung cancers; other health risk behaviors remained unknown. Therefore, this study discusses the effects of health risk behaviors on the survival outcome in lung cancer patients.

Methods

The study sample consists of 1410 newly diagnosed, histologically confirmed lung cancer patients from a medical center hospital in central Taiwan. The patient medical records were collected from 1 January 1998 to May 2004. Besides descriptive statistical analyses, t-test and analysis of variance were used to analyze the relationship between patient characteristics, patient health risk behavior and survival. Chi-square tests were used to analyze the relationship between patient characteristics, patient health risk behavior and stage of disease. Cox's proportional hazards regression was computed for the risk of survival by patient characteristics and patient health risk behavior.

Results

The results showed that there is a significant difference between smoking (P < 0.001), alcohol consumption (P = 0.027), routine physical check-ups (out-of-pocket) (P < 0.001) and survival time. Patients with betel-nut consumption did not have a significant effect upon survival than non-betel-nut consumption patients. When holding constants all the variables in Cox's proportional hazards model, smoking (P = 0.02), routine physical check-ups (out-of-pocket) (P = 0.017) and stage at diagnosis of lung cancer (P < 0.001) will affect lung cancer patients’ survival.

Conclusions

This study presented evidence showing that smoking, alcohol consumption behavior, routine physical check-ups and stage at diagnosis play an important role in determining the survival of lung cancer patients.

INTRODUCTION

Cancer is a complex disease. The survival of patients with cancer depends on the localization of the tumor, histology/pathology, tumor stage, host–tumor interaction and subjective assessment of the other diseases, illnesses or conditions. Several studies indicate that smoking and alcohol consumption behaviors are related to the survival of lung cancer patients. However, very few of them included betel-nut consumption or routine physical check-ups into their study. Therefore, this study investigates the relationship between patient's health behaviors and survival time.

Several studies pointed out that health behaviors affect lung cancer patients’ survival. Johnston-Early et al. (1) found that continuing smoking during the treatment of lung cancer was associated with a poor prognosis, while discontinuing smoking may have beneficial effects on survival. Hinds et al. (2) concluded that ever-smokers demonstrated a significantly greater probability of dying during the 12 months after diagnosis than never-smokers. Tammemagi et al. (3) also found that current smoking at diagnosis is an important independent predictor of shortened lung cancer survival. Another study diagnosed the relationship between smoking and lung cancer patients’ survival in Japan and concluded that smoking is an important clinical prognostic factor in evaluating overall long-term survival (Fujisawa et al., 1999).

Alcohol consumption was first proposed as a possible risk factor for the incidence of lung cancer by Potter and McMichael (4). However, the relationship between alcohol consumption and lung cancer patients’ survival has been regarded with skepticism for decades. Paull et al. (5) examined the survival difference for alcohol-abusing patients (more than five drinks or cans of beer a day) and non-abusing patients. The outcome indicated that alcohol-abusing patients have fewer survival days than non-abusing patients. Freudenheim et al. (6) also concluded that the risk of lung cancer was slightly greater in subjects who consumed at least 30 g of alcohol (approximately two drinks) per day than in those who abstained.

There is clear evidence that smoking and alcohol consumption cause lung cancer, but few studies focus on the relationship between betel-nut consumption and lung cancer. The National Health Research Institutes in Taiwan (2005) found that patients with betel-nut consumption have 120 times the risk of having cancer (including lung cancer, oral cancer and esophageal cancer, etc.). Otherwise, according to the Taiwanese culture, in rural areas, people with smoking behavior usually come with alcohol and betel-nut consumptions. Therefore, this study discusses the relationship between betel-nut consumption and lung cancer.

Costs to routine physical check-ups in Taiwan are considerably more affordable when compared with those in many Western countries. The out-of-pocket amounts were about 120 to 1500 US dollars depending on the services. People with self-awareness to personal health tend to be willing to pay for routine cancer screening and preventive interventions such as chest computed tomography (CT).

The effect of smoking on lung cancer risks has been well documented, but the effects of alcohol and betel-nut remain controversial. Therefore, this study will discuss the effects of health behaviors (such as smoking, alcohol consumption, betel-nut consumption and routine physical check-ups) on survival outcome in lung cancer patients .

METHODOLOGY

Sampling

This study consisted of 1410 histologically confirmed lung cancer patients from a medical center in central Taiwan who were newly diagnosed between January 1998 and May 2004. Patients’ information and health behavior information were extracted from patients’ medical records by three trained medical record nurses for the study. Patients’ medical records included patients’ administration records, patients’ progress notes and nursing assessment notes. Samples were drawn from the records of two clinical services. The sample included 650 lung cancer patients who were at early stage diagnosis (stages I, II and IIIa) consecutively admitted for surgery, and 760 patients randomly drawn from among 1850 patients who were at late stage diagnosis (stages IIIb and IV). After excluding the missing data, the analyzed sample was composed of 1139 patients with lung cancer. To test the representativeness of the randomly selected group, chi-square tests were used to analyze the relationship between a sample of 760 patients randomly selected from the stages IIIb and IV and the entire 1850 sample from stages IIIb and IV. No significant differences existed between the two groups regarding age (P = 0.3094) and sex (P = 0.8049).

Data Exclusion

Data excluded from this study include patients who were not evaluated due to a lack of follow-up; lung cancer as a metastatic cancer diagnosis; and surgery/operative mortality from operation or complication of surgery. The final data sample included 1139 cases.

Staging

Stage at diagnosis was defined in accordance with the classification outlined for clinical staging of lung cancer in the American Joint Committee on Cancer's (AJCC) Cancer Staging Manual in 1997. Lung cancer is classified as stage I, stage II, stage III (both IIIa and IIIb) and stage IV. In stage I, the cancer is found only in the lung, with normal tissue being present around the tumor. In stage II, cancer spreads to nearby lymph nodes or to the chest wall, the diaphragm, the mediastinalpleura or the parietal pericardium. In stage III, cancer has either spread to the lymph nodes in the mediastinum, or spread to the lymph nodes on the opposite of the chest or in the lower neck. In stage IV, cancer spreads to other parts of the body or to another lobe of the lungs. Staging was determined by a senior physician through a review of sampling pathology and cytology.

Data Analysis

The SAS Statistics System was used to analyze the data. Only individual-level analysis was involved in the study.

Descriptive statistical analyses were first conducted to characterize the study population of lung cancer cases and study variables. T-test and ANOVA were used to analyze the relationship between patient characteristics, patient health behavior and survival. Chi-square tests were used to analyze the relationship between patient characteristics, patient health behavior and stage of disease. Cox's proportional hazards regression was computed for the risk of survival by patient characteristics and patient health behavior.

RESULTS

Basic Characteristics

Patient characteristics

The patients’ mean age was 65.5 years. The age range of the study population was 23–90. For chi-square tests, ANOVA and Cox regression analysis, the age at diagnosis was divided into six categories: <30, 31–40, 41–50, 51–60, 61–70 and >71 years.

Of the 1139 patients, 73.9% (N = 842) were men; 26.1% (N = 297) were women. Regarding marital status, 88.6% (N = 990) patients were currently married; 11.4% (N = 128) were single (including widowed and separated/divorced). In terms of education, 63.5% (N = 672) of them were lower than elementary school; 13.3% (N = 141) had 7–9 years of education; 16.2% (N = 171) had 10–12 years of education; 7% (N = 74) had more than 13 years of education. About seven of every 10 patients sampled (71.9%; N = 773) reported having a religion. With regard to work status, 30.6% (N = 337) of patients were working when they were diagnosed with lung cancer; 27.4% (N = 302) of them were retired and 42.0% (N = 463) were not employed (Table 1).

Table 1.

The relationship between patient characteristics and survival time (n = 1139)

Variables N (%) Mean SD P 
Age    0.24 
 <30 8 (0.7) 1023.3 930.5  
 31–40 42 (3.7) 635.0 559.9  
 41–50 88 (7.7) 724.4 666.1  
 51–60 163 (14.3) 772.1 631.8  
 61–70 345 (30.3) 838.1 714.3  
 >70 448 (39.3) 756.0 661.3  
Gender    0.02 
 Male 842 (73.9) 738.8 651.0  
 Female 297 (26.1) 843.3 697.7  
Religion    0.996 
 No 302 (28.1) 764.6 667.9  
 Yes 773 (71.9) 764.9 663.6  
Marital status   0.09  
 Currently married 990 (88.6) 775.2 666.2  
 Single (included devoice) 128 (11.4) 669.5 641.9  
Education (Year)   0.314  
 ≤6 672 (63.5) 750.4 680.7  
 7–9 141 (13.3) 754.1 614.7  
 10–12 171 (16.2) 841.9 680.8  
 ≥13 74 (7.0) 688.9 618.2  
Occupation   0.175  
 Currently working 337 (30.6) 730.8 653.0  
 Retired 302 (27.4) 731.3 670.9  
 Not working 463 (42.0) 806.4 661.7  
Health behaviors     
 Smoking    <0.0001 
  Smokera 624 (55.9) 694.2 662.6  
  Non-smoker 493 (44.1) 858.8 657.3  
 Alcohol    0.027 
  Alcohol consumption 112 (10.2) 635.8 614.2  
  Non-alcohol consumption 986 (89.8) 783.0 670.7  
 Betel-nut consumption    0.463 
  Betel-nut consumption 37 (4.0) 698.2 593.2  
  Non-betel-nut consumption 892 (96.0) 779.4 662.4  
 Physical check-up (out-of-pocket)    <0.0001 
  Yes 140 (12.3) 1036.3 731.1  
  No 989 (87.6) 726.9 645.0  
Stage    <0.0001 
 I 215 (19.4) 1226.3 656.8  
 II 106 (9.5) 1045.1 673.9  
 III 406 (36.6) 727.4 621.8  
 IV 383 (34.5) 488.7 536.1  
Variables N (%) Mean SD P 
Age    0.24 
 <30 8 (0.7) 1023.3 930.5  
 31–40 42 (3.7) 635.0 559.9  
 41–50 88 (7.7) 724.4 666.1  
 51–60 163 (14.3) 772.1 631.8  
 61–70 345 (30.3) 838.1 714.3  
 >70 448 (39.3) 756.0 661.3  
Gender    0.02 
 Male 842 (73.9) 738.8 651.0  
 Female 297 (26.1) 843.3 697.7  
Religion    0.996 
 No 302 (28.1) 764.6 667.9  
 Yes 773 (71.9) 764.9 663.6  
Marital status   0.09  
 Currently married 990 (88.6) 775.2 666.2  
 Single (included devoice) 128 (11.4) 669.5 641.9  
Education (Year)   0.314  
 ≤6 672 (63.5) 750.4 680.7  
 7–9 141 (13.3) 754.1 614.7  
 10–12 171 (16.2) 841.9 680.8  
 ≥13 74 (7.0) 688.9 618.2  
Occupation   0.175  
 Currently working 337 (30.6) 730.8 653.0  
 Retired 302 (27.4) 731.3 670.9  
 Not working 463 (42.0) 806.4 661.7  
Health behaviors     
 Smoking    <0.0001 
  Smokera 624 (55.9) 694.2 662.6  
  Non-smoker 493 (44.1) 858.8 657.3  
 Alcohol    0.027 
  Alcohol consumption 112 (10.2) 635.8 614.2  
  Non-alcohol consumption 986 (89.8) 783.0 670.7  
 Betel-nut consumption    0.463 
  Betel-nut consumption 37 (4.0) 698.2 593.2  
  Non-betel-nut consumption 892 (96.0) 779.4 662.4  
 Physical check-up (out-of-pocket)    <0.0001 
  Yes 140 (12.3) 1036.3 731.1  
  No 989 (87.6) 726.9 645.0  
Stage    <0.0001 
 I 215 (19.4) 1226.3 656.8  
 II 106 (9.5) 1045.1 673.9  
 III 406 (36.6) 727.4 621.8  
 IV 383 (34.5) 488.7 536.1  

aSmoker included current and former smokers.

Patient health behavior

In this study, smokers are defined as  individuals who are currently smoking or were former smokers who smoked for more than 1 year before. Regarding patients’ smoking history, 55.9% (N = 624) of the patients smoked for more than 1 year; 89.8% (N = 986) did not drink alcohol regularly and 4.0% (N = 37) of the patients consumed betel-nut. Only 12.3% (N = 140) patients had routine physical check-ups (out-of-pocket) (Table 1).

Staging

According to the AJCC (1997), 215 of the 1139 lung cancer patients (or 19.4%) belonged to stage I at the time of diagnosis; 106 (9.5%) were in stage II; 406 (36.6%) were listed under stage III and 383 (34.5%) were grouped into stage IV (Table 1).

Survival

All subjects were followed until the end of December 2004. At the end of 2004, 385 patients were alive (33.8%); 753 (66.2%) patients of this dynamic cohort died during follow-up and were, thus, uncensored.

Health Behaviors vs. Survival

Smoking vs. survival

The t-test was used to compare the differences between smoking and survival time (Table 1). Survival time showed a significant difference for smokers (P < 0.05). The relationship between patient smoking history and survival showed that non-smoking patients were more likely to have significantly better survival rates than smoking patients (the mean of survival of non-smoking history patient is 858.8 days compared with the survival of smoking patient, which is 694.2 days).

Alcohol consumption vs. survival

Table 1 also indicated that the mean survival time showed there was a significant difference between alcohol consumption and survival time; abstainers from alcoholic beverages showed longer survival time than those patients who consumed alcohol (783.0 days for abstainers compared with 635.8 days for alcohol users).

Betel-nut vs. survival

The t-test was used to compare the difference between betel-nut consumption and survival time. There was no significant difference between betel-nut consumption status and survival time (P = 0.463) (Table 1). However, patients with no betel-nut consumption behavior had a tendency of longer survival time (779.4 days) than patients with betel-nut consumption behavior (698.2 days).

Physical check-ups (out-of-pocket) vs. survival

In Table 1, the mean survival time shows that there was a significant difference between patients with routine physical check-ups (out-of-pocket) and survival time (P < 0.0001); patients with routine physical check-ups (out-of-pocket) had longer survival time than those patients who did not have routine physical check-ups (out-of-pocket) (1036.3 days for routine physical check-ups (out-of-pocket) compared with 726.9 days for non-routine physical check-ups (out-of-pocket)).

Health Behaviors vs. Patients’ Characteristics

Smoking vs. patients’ characteristics

Of the 1139 patients, after excluding missing data, 55.9% (N = 624) smoked for more than 1 year and 44.1% (N = 493) were non-smoking. The chi-square test was used to detect the relationship between smoking, patient characteristics (age, sex, marital status, religion, education and occupation), lung cancer stage at diagnosis and routine physical check-ups (out-of-pocket). Table 2 shows a significant relationship between smoking and gender (P < 0.0001), occupation (P < 0.0001) and routine physical check-ups (out-of-pocket) (P < 0.001). Women were less likely to be smokers than men. Non-smoking lung cancer patients were more likely to go for routine physical check-ups (out-of-pocket) to diagnose lung cancer. Half of the smoking lung cancer patients did not have work at the time we analyzed these data.

Table 2.

The relationship between patient characteristics and health behavior (N = 1139)

Variables Smoking, N (%) Non-Smoking, N (%) P Alcohol, N (%) Non-alcohol, N (%) P Betel-nut, N (%) Non-betel-nut, N (%) P 
Age (years)   P = 0.771   0.541   0.816 
 <30 4 (00.7) 4 (0.8)  0 (0.0) 8 (0.8)  0 (0.0) 7 (0.8)  
 31–40 24 (04.0) 18 (3.8)  5 (5.0) 35 (3.7)  0 (0.0) 36 (4.2)  
 41–50 49 (08.2) 38 (8.0)  7 (6.9) 77 (8.1)  3 (8.6) 66 (7.7)  
 51–60 98 (16.4) 63 (13.3)  16 (15.8) 143 (15.0)  5 (14.3) 135 (15.8)  
 61–70 189 (31.6) 150 (31.6)  38 (37.6) 291 (30.5)  13 (37.1) 266 (31.1)  
 >70 234 (39.1) 201 (42.4)  35 (34.7) 399 (41.9)  14 (40.0) 346 (40.4)  
Gender   P < 0.0001   <0.0001   <0.0001 
 Male 603 (96.6) 225 (45.6)  111 (99.1) 698 (70.8)  36 (97.3) 664 (74.4)  
 Female 21 (03.4) 268 (54.4)  1 (0.9) 288 (29.2)  1 (2.7) 228 (25.6)  
Education (year)   P = 0.621   0.332   0.199 
 ≤6 374 (64.0) 291 (63.0)  72 (68.6) 583 (63.0)  27 (73.0) 513 (61.8)  
 7–9 82 (14.0) 56 (12.1)  14 (13.3) 121 (13.1)  8 (21.6) 113 (13.6)  
 10–12 89 (15.2) 81 (17.5)  11 (10.5) 158 (17.1)  1 (2.7) 142 (17.1)  
 >13 39 (06.7) 34 (7.4)  8 (7.6) 64 (6.9)  1 (2.7) 62 (7.5)  
Religion   P = 0.088   0.819   0.007 
 Yes 408 (69.7) 356 (74.5)  75 (70.8) 675 (71.8)  32 (91.4) 600 (70.7)  
 No 177 (30.3) 122 (25.5)  31 (29.2) 265 (28.2)  3 (8.6) 249 (29.3)  
Marital status   P = 0.441   0.331   0.016 
 Currently married 544 (88.0) 435 (89.5)  96 (85.7) 865 (88.8)  36 (100.0) 777 (87.9)  
 Single 74 (12.0) 51 (10.5)  16 (14.3) 109 (11.2)  0 (0.0) 107 (12.1)  
Occupation   P < 0.0001   0.013   0.012 
 Currently working 206 (34.0) 129 (26.7)  43 (39.1) 285 (29.7)  19 (51.4) 264 (30.4)  
 Retired 214 (35.3) 83 (17.2)  35 (31.8) 257 (26.7)  4 (10.8) 239 (27.5)  
 Not working 186 (30.7) 271 (56.1)  32 (29.1) 419 (43.6)  14 (37.8) 366 (42.1)  
Physical check-ups (out-of-pocket)   P < 0.0001   0.033   0.464 
 Yes 58 (09.3) 82 (16.7)  7 (6.3) 131 (13.3)  4 (10.8) 116 (13.0)  
 No 566 (90.7) 408 (83.3)  105 (93.8) 852 (86.7)  33 (89.2) 773 (87.0)  
Stage   P = 0.377   0.584   0.309 
 I 113 (18.6) 101 (20.9)  16 (15.7) 197 (20.4)  7 (20.0) 178 (20.4)  
 II 61 (10.1) 43 (8.9)  10 (9.8) 90 (9.3)  6 (17.1) 81 (9.3)  
 III 231 (38.1) 164 (34.0)  42 (41.2) 343 (35.4)  14 (40.0) 314 (36.1)  
 IV 201 (33.2) 175 (36.2)  34 (33.3) 338 (34.9)  8 (22.9) 298 (34.2)  
Variables Smoking, N (%) Non-Smoking, N (%) P Alcohol, N (%) Non-alcohol, N (%) P Betel-nut, N (%) Non-betel-nut, N (%) P 
Age (years)   P = 0.771   0.541   0.816 
 <30 4 (00.7) 4 (0.8)  0 (0.0) 8 (0.8)  0 (0.0) 7 (0.8)  
 31–40 24 (04.0) 18 (3.8)  5 (5.0) 35 (3.7)  0 (0.0) 36 (4.2)  
 41–50 49 (08.2) 38 (8.0)  7 (6.9) 77 (8.1)  3 (8.6) 66 (7.7)  
 51–60 98 (16.4) 63 (13.3)  16 (15.8) 143 (15.0)  5 (14.3) 135 (15.8)  
 61–70 189 (31.6) 150 (31.6)  38 (37.6) 291 (30.5)  13 (37.1) 266 (31.1)  
 >70 234 (39.1) 201 (42.4)  35 (34.7) 399 (41.9)  14 (40.0) 346 (40.4)  
Gender   P < 0.0001   <0.0001   <0.0001 
 Male 603 (96.6) 225 (45.6)  111 (99.1) 698 (70.8)  36 (97.3) 664 (74.4)  
 Female 21 (03.4) 268 (54.4)  1 (0.9) 288 (29.2)  1 (2.7) 228 (25.6)  
Education (year)   P = 0.621   0.332   0.199 
 ≤6 374 (64.0) 291 (63.0)  72 (68.6) 583 (63.0)  27 (73.0) 513 (61.8)  
 7–9 82 (14.0) 56 (12.1)  14 (13.3) 121 (13.1)  8 (21.6) 113 (13.6)  
 10–12 89 (15.2) 81 (17.5)  11 (10.5) 158 (17.1)  1 (2.7) 142 (17.1)  
 >13 39 (06.7) 34 (7.4)  8 (7.6) 64 (6.9)  1 (2.7) 62 (7.5)  
Religion   P = 0.088   0.819   0.007 
 Yes 408 (69.7) 356 (74.5)  75 (70.8) 675 (71.8)  32 (91.4) 600 (70.7)  
 No 177 (30.3) 122 (25.5)  31 (29.2) 265 (28.2)  3 (8.6) 249 (29.3)  
Marital status   P = 0.441   0.331   0.016 
 Currently married 544 (88.0) 435 (89.5)  96 (85.7) 865 (88.8)  36 (100.0) 777 (87.9)  
 Single 74 (12.0) 51 (10.5)  16 (14.3) 109 (11.2)  0 (0.0) 107 (12.1)  
Occupation   P < 0.0001   0.013   0.012 
 Currently working 206 (34.0) 129 (26.7)  43 (39.1) 285 (29.7)  19 (51.4) 264 (30.4)  
 Retired 214 (35.3) 83 (17.2)  35 (31.8) 257 (26.7)  4 (10.8) 239 (27.5)  
 Not working 186 (30.7) 271 (56.1)  32 (29.1) 419 (43.6)  14 (37.8) 366 (42.1)  
Physical check-ups (out-of-pocket)   P < 0.0001   0.033   0.464 
 Yes 58 (09.3) 82 (16.7)  7 (6.3) 131 (13.3)  4 (10.8) 116 (13.0)  
 No 566 (90.7) 408 (83.3)  105 (93.8) 852 (86.7)  33 (89.2) 773 (87.0)  
Stage   P = 0.377   0.584   0.309 
 I 113 (18.6) 101 (20.9)  16 (15.7) 197 (20.4)  7 (20.0) 178 (20.4)  
 II 61 (10.1) 43 (8.9)  10 (9.8) 90 (9.3)  6 (17.1) 81 (9.3)  
 III 231 (38.1) 164 (34.0)  42 (41.2) 343 (35.4)  14 (40.0) 314 (36.1)  
 IV 201 (33.2) 175 (36.2)  34 (33.3) 338 (34.9)  8 (22.9) 298 (34.2)  

Alcohol consumption vs. patients’ characteristics

Table 2 indicates the relationship between patient characteristics (age, sex, marital status, religion, education and occupation), lung cancer stage at diagnosis, routine physical check-ups (out-of-pocket) and alcohol consumption. Table 2 shows a significant relationship between alcohol consumption and gender (P < 0.0001), occupation (P = 0.013) and routine physical check-ups (out-of-pocket) (P =0.033). Women were less likely to consume alcohol than men. Among patients with alcohol consumption behavior when diagnosed with lung cancer, nearly 40% of patients were working, whereas 32% of patients were retired. Non-alcohol lung cancer patients were more likely to have routine physical check-ups than alcohol patients (Table 2).

Betel-nut vs. patients’ characteristics

Table 2 indicates the relationship between betel-nut consumption and patient characteristics (age, sex, marital status, religion, education and occupation), lung cancer stage at diagnosis and routine physical check-ups (out-of-pocket). Table 2 shows a significant relationship between betel-nut consumption and gender (P < 0.001), religion (P = 0.007), marital status (P = 0.016) and occupation (P = 0.012). Among the total number of lung cancer patients, women were less likely to be betel-nut consumers than men; single people were less likely to be betel-nut consumers than those who are married. Non-religious patients were less likely to be betel-nut consumers than religious patient cases (Table 2).

Cox's Proportional Hazards Model

In multivariate analysis, when holding other variables constant, smoking behavior, routine physical check-ups (out-of-pocket), stage and age at diagnosis (31–40 years old) will affect lung cancer patients’ survival (P < 0.05) (Table 3). In reviewing the relationship between patients behavior and survival, this study found that non-smoking patients were shown to have a lower risk of death than smoking patients (smoker as the reference group vs. non-smoker, HR = 0.781, P = 0.02). Non-alcohol-consuming patients were not found to be significantly associated with lower risk of death than alcohol-consuming patients (alcohol consumption as the reference group vs. non-alcohol consumption, HR = 0.905, P = 0.481). Non-betel-nut consumption among patients did not have significant effect upon death than betel-nut consumption (betel-nut consumption as the reference group vs. non-betel-nut consumption, HR = 0.998, P = 0.992). Patients with routine physical check-up (out-of-pocket) experience were shown to be at decreased hazard of death than patients with no routine physical check-up (out-of-pocket) experience (with non-routine physical check-ups as the reference group vs. with routine physical check-ups, HR = 0.714, P = 0.017).

Table 3.

Survival analysis: patient characteristics, health behaviors and survival (N = 1139)

Variables Number Hazard ratio P value 95% CI 
Health behaviors 
 Smoking 
  Smoker (reference) 490    
  Non-smoker 411 0.781 0.02 (0.634–0.962) 
 Alcohol 
  Alcohol consumption (reference) 78    
  Non-alcohol consumption 823 0.905 0.481 (0.686–1.195) 
 Betel-nut consumption 
  Betel-nut consumption (reference) 35    
  Non-betel-nut consumption 784 0.998 0.992 (0.672–1.482) 
 Physical check-ups (out-of-pocket) 
  No (reference) 784    
  Yes 117 0.714 0.017 (0.543–0.940) 
Age 
 <30 0.691 0.53 (0.219–2.187) 
 31–40 21 1.773 0.017 (1.107–2.842) 
 41–50 50 0.941 0.745 (0.652–1.358) 
 51–60 119 0.883 0.351 (0.679–1.147) 
 61–70 244 0.98 0.841 (0.801–1.199) 
 >70 (reference) 460    
Gender 
 Male (reference) 682    
 Female 219 0.824 0.111 (0.647–1.046) 
Marital status 
 Single (reference) 113    
 Married 788 1.115 0.113 (0.957–1.611) 
Stage 
 I 186 0.185 <.001 (0.139–0.248) 
 II 86 0.331 <.001 (0.242–0.454) 
 III 324 0.584 <.001 (0.487–0.700) 
 IV (reference) 305    
Variables Number Hazard ratio P value 95% CI 
Health behaviors 
 Smoking 
  Smoker (reference) 490    
  Non-smoker 411 0.781 0.02 (0.634–0.962) 
 Alcohol 
  Alcohol consumption (reference) 78    
  Non-alcohol consumption 823 0.905 0.481 (0.686–1.195) 
 Betel-nut consumption 
  Betel-nut consumption (reference) 35    
  Non-betel-nut consumption 784 0.998 0.992 (0.672–1.482) 
 Physical check-ups (out-of-pocket) 
  No (reference) 784    
  Yes 117 0.714 0.017 (0.543–0.940) 
Age 
 <30 0.691 0.53 (0.219–2.187) 
 31–40 21 1.773 0.017 (1.107–2.842) 
 41–50 50 0.941 0.745 (0.652–1.358) 
 51–60 119 0.883 0.351 (0.679–1.147) 
 61–70 244 0.98 0.841 (0.801–1.199) 
 >70 (reference) 460    
Gender 
 Male (reference) 682    
 Female 219 0.824 0.111 (0.647–1.046) 
Marital status 
 Single (reference) 113    
 Married 788 1.115 0.113 (0.957–1.611) 
Stage 
 I 186 0.185 <.001 (0.139–0.248) 
 II 86 0.331 <.001 (0.242–0.454) 
 III 324 0.584 <.001 (0.487–0.700) 
 IV (reference) 305    

Status: death (n = 687); time: survival days.

DISCUSSION

From Table 1, it can be inferred that women have longer survival days than men. Many studies concluded that patient's gender and marital status will affect lung cancer patient's survival (7–10). Women usually have longer survival rate than men and patients who are married have longer survival rate than patients who remain single. This study also confirmed the conclusions. Although the study did not show significant difference regarding marital status, the married patients showed the tendency of longer survival days than single patients.

The goal of this study was to examine the impact of different patient health behaviors (smoking, alcohol, betel-nut consumption and routine physical check-ups (out-of-pocket)) on survival in lung cancer patients. These differences are all statistically significant except that the patient with a betel-nut consumption history is not correlated with survival (Table 1). When holding constant all of the variables, smoking and routine physical check-up (out-of-pocket) experience showed differences in survival (Table 3).

The results of this study showed that smoking did have a significant effect upon survival, which was the same as with previous studies (3,11,12). The smoking status was defined based on the information at the time when patients’ medical records were taken. Here, smokers are defined as current smokers (those who were still smoking at the time when his/her medical record was taken) and former smokers (those who already quit smoking but had smoked for more than 1 year before the medical record was written). From this study, lung cancer patients with smoking experience had less survival days than patients with no smoking behavior. Tobacco plays an important role in the survival outcome of lung cancer patients (Tables 1 and 3).

Alcohol consumption has also been associated with lung cancer survival time in our study (Table 1). However, after adjusting for all of the variables, alcohol consumption experience did not affect the survival of lung cancer patients' (Table 3). A previous study indicated that alcohol-abusing patients have fewer survival days than non-abusing patients (5). But many researchers reported that alcohol had a direct role in the development of lung cancer (13,14), but Zang and Wynder (7) reported that the association between alcohol and lung cancer risk can be fully explained by the confounding effect of cigarette use (15). Their study showed that alcohol may not itself be a carcinogen; however, in combination with a chemical carcinogen, it is known to enhance carcinogenesis.

This study indicated that patients with routine physical check-up (out-of-pocket) experience had significantly longer survival than patients without routine physical check-up (out-of-pocket) behavior (Table 1). This study further indicated that patients with routine physical check-ups (out-of-pocket) had a significant relationship with patients’ smoking and alcohol consumption behaviors (Table 2). Lung cancer screening behaviors may be related to patient perceptions, beliefs and diverse cultural, social economics status (SES) and political environments. In addition, other potential variables, such as family history of lung cancer, health-promoting behaviors (non-alcohol use and non-tobacco use), geographic proximity to the healthcare institution, transportation availability, and other realized access variables might have influenced lung cancer outcomes. Healthcare interventions should consider all aspects of access. Presently, low-dose CT (LDCT) screening is capable of detecting very small, early-stage cancers (16–20). Based on low-dose CT screening trials, the prevalence rate for lung cancer ranged from 0.1 to 2.9% (18). The increase in lung cancer detection is associated with a higher proportion of stage I disease (71–100%) and patients whose lung cancers were detected by CT screening and whose cancers were surgically resected had a 10-year survival in excess of 80% (17,18). This shows that CT screening might increase detection of lung cancer and survival rate. In Taiwan, we had a early lung cancer LDCT screening program for high-risk family members conducted under National Health Research Institutes GELAC demonstration project. The prevalence rate for lung cancer was 1.7% among 1560 examinations, and the cost was US$9000 per finding case (21). It seems that early lung cancer screening with LDCT maybe applicable in high-risk lung cancer family members or some groups of physical check-ups (out-of-pocket). So for improving early diagnosis of lung cancer, prevention approaches should be considered for all sectors (including the healthcare system and public population) in order to improve lung cancer outcomes. However, the related lung cancer screening trials (such as the NLST and NELSON trials) are still ongoing and did not have specific outcomes (18,22).

The important factors to be considered for early diagnosis are increasing public awareness of lung cancer and getting media attention to the importance of early diagnosis. A few studies found stage I and II lung cancer to be often asymptomatic. In the absence of symptoms in high-risk group patients, consideration should be given to recommending early detection and screening programs that may contribute to improved diagnosis of stage  and shorten patient-dependent delay of diagnosis. The only possibility to detect asymptomatic lung cancers is to screen high-risk groups and increase patient awareness. A general education of the population concerning early symptoms of lung cancer (i.e. long-standing coughing, chest pain, body weight loss and hemoptysis) may increase patients’ awareness of the symptoms and the urgency with regard to medical attention, thereby leading to earlier definitive treatment. Screening of high-risk groups, such as smokers, alcoholics or with lung cancer family history, should be considered in Taiwan during the routine physical check-ups that is part of NHI (National Health Insurance) coverage. Therefore, people in Taiwan must be encouraged to participate in lung cancer prevention and healthcare intervention programs.

One limitation of this study is the lack of distinction in the cell type (i.e. small cell and non-small cell cancer) from the retrieved data. However, we realize that it is essential to consider the different cell types in the data collection and analyses of future studies.

The information provided allowed only very crude measures for consumption, because this study did not include quality, quantity, duration and intensity of smoking/alcohol/betel-nut consumption. A few limitations that could have affected the results should be recognized and discussed. The following considerations should be taken into account when interpreting this study's findings. The study was based on the data collected from the medical records of patients at hospital as the primary source of data. Retrospective studies are disadvantageous by nature. In the future, medical professionals should propose to carry out a prospective study that is based on a structured patient interview targeting patient health behavior. A further limitation was that this study used data from medical records, which presented the problem of missing information. Another limitation of this study is the inability to collect information on SES, including demographic variables that may explain the difference of patients' behavior in survival and stage. In the future study, medical professionals should suggest obtaining SES data from a prospective study.

CONCLUSIONS

The goal of this study was to examine the impact of health behaviors on the survival of lung cancer patients. Many studies reported that tobacco, alcohol and betel nut play an important role in increasing the chance of acquiring lung cancer, but a few studies that examine lung cancer survival associated with patients’ health behavior. Patients with no smoking and no alcohol consumption behavior had significantly longer survival time than patients with smoking and alcohol consumption behavior. Patients without betel-nut consumption behavior had a tendency of longer survival time than patients with betel-nut consumption behavior, although it is not significantly different. Patients with routine physical check-up (out-of-pocket) behavior had significantly longer survival time than patients without routine physical check-ups (out-of-pocket) behavior. When holding constant all of the variables, smoking behavior, routine physical check-ups, gender and the lung cancer stage at diagnosis will affect lung cancer patients' survival.

It is imperative that we broaden our understanding of the important factor that affects the survival of lung cancer patients. The results of this study presented evidence showing that smoking, alcohol consumption behavior and routine physical check-ups play an important role in determining the survival of lung cancer patients.

Funding

This study was supported by Taiwan Department of Health, China Medical University Hospital Cancer Research Center of Excellence (DOH99-TD-C-111-005).

Conflict of interest statement

None declared.

References

1
Johnston-Early
A
Cohen
MH
Minna
JD
Paxton
LM
Fossieck
BE
Ihde
DC
, et al.  . 
Smoking abstinence and small cell lung cancer survival
JAMA
 , 
1980
, vol. 
244
 (pg. 
2175
-
9
)
2
Hinds
MW
Yang
HY
Stemmermann
G
Lee
J
Kolonel
LN
Smoking history and lung cancer survival in women
J Natl Cancer Inst
 , 
1982
, vol. 
68
 (pg. 
395
-
9
)
3
Tammemagi
CM
Neslund-Dudas
C
Simoff
M
Kvale
P
Smoking and lung cancer survival, the role of comorbidity and treatment
Chest
 , 
2004
, vol. 
125
 (pg. 
27
-
37
)
4
Potter
JD
McMichael
AJ
Alcohol, beer and lung cancer—a meaningful relationship?
Int J Epidemiol
 , 
1984
, vol. 
13
 (pg. 
240
-
2
)
5
Paull
DE
Updyke
GM
Baumann
MA
Chin
HW
Little
AG
Adebonojo
SA
Alcohol abuse predicts progression of disease and death in patients with lung cancer
Ann Thorac Surg
 , 
2005
, vol. 
80
 (pg. 
1033
-
9
)
6
Freudenheim
JL
Ritz
J
Smith-Warner
SA
Albanes
D
Bandera
EV
Brandt
PA
, et al.  . 
Alcohol consumption and risk of lung cancer: a pooled analysis of cohort studies
Am J Clin Nutr
 , 
2005
, vol. 
82
 (pg. 
657
-
67
)
7
Zang
AE
Wynder
LE
Differences in lung cancer risk between men and women: examination of the evidence
J Natl Cancer Inst
 , 
1996
, vol. 
88
 (pg. 
183
-
92
)
8
Kreuzer
M
Boffetta
P
Whitley
E
Ahrens
W
Gaborieau
V
Heinrich
J
, et al.  . 
Gender differences in lung cancer risk by smoking: a multicentre case–control study in Germany and Italy
Br J Cancer
 , 
2000
, vol. 
82
 (pg. 
227
-
33
)
9
Ferguson
KM
Wang
J
Hoffman
CP
Haraf
JD
Olak
J
Masters
AG
Sex-associated differences in survival of patients undergoing resection for lung cancer
Ann Thorac Surg
 , 
2000
, vol. 
69
 (pg. 
245
-
9
)
10
Øystein
K
The impact of marital status on cancer survival
Soc Sci Med
 , 
2000
, vol. 
52
 (pg. 
357
-
68
)
11
Hackett
TP
Cassem
NJ
Baker
JW
Patient delay in cancer
N Engl J Med
 , 
1973
, vol. 
289
 (pg. 
14
-
20
)
12
Marcus
PM
Lung cancer screening: an update
J Clin Oncol
 , 
2001
, vol. 
19
 (pg. 
83s
-
6s
)
13
Potter
JD
Sellers
TA
Folsom
AR
McGovern
PG
Alcohol, beer, and lung cancer in postmenopausal women. The Iowa women's health study
Am J Epidemiol
 , 
1992
, vol. 
2
 (pg. 
587
-
95
)
14
Prescott
E
Gronbaek
M
Becker
U
Sorensen
TI
Alcohol intake and the risk of lung cancer: influence of type of alcoholic beverage
Am J Epidemiol
 , 
1999
, vol. 
149
 (pg. 
463
-
70
)
15
Zang
EA
Wynder
EL
Reevaluation of the confounding effect of cigarette smoking on the relationship between alcohol use and lung cancer risk, with larynx cancer used as a positive control
Prev Med
 , 
2001
, vol. 
32
 (pg. 
359
-
70
)
16
Henschke
CK
McCauley
DI
Yankelevitz
DF
Naidich
DP
McGuinness
G
Miettinen
OS
, et al.  . 
Early lung cancer action project: overall design and findings from baseline screening
Lancet
 , 
1999
, vol. 
354
 (pg. 
99
-
105
)
17
International Early Lung Cancer Action Program Investigator
Survival of patient with stage I lung cancer detected on CT screening
N Engl J Med
 , 
2006
, vol. 
355
 (pg. 
1763
-
71
)
18
Harvey
IP
David
PC
David
HJ
John
DM
Giorgio
VS
Andrew
TT
Principles and Practice of Lung Cancer: The Official Reference Text of IASLC
 , 
2010
4th edn
Lippincott Williams & Wilkins
19
Sone
S
Nakayama
T
Honda
T
Tsushima
K
Li
F
Haniuda
M
, et al.  . 
Long-term follow-up study of a population-based 1996–1998 mass screening programme for lung cancer using mobile low-dose spiral et
Lung Cancer
 , 
2007
, vol. 
58
 (pg. 
329
-
34
)
20
Sobue
T
Moriyama
N
Kaneko
M
Kusumoto
M
Kobayashi
T
Tsuchiya
R
, et al.  . 
Screening for lung cancer with low dose helical CT: anti-lung cancer association Project
J Clin Oncol
 , 
2002
, vol. 
20
 (pg. 
911
-
20
)
21
Tsai
YH
Lung cancer early detection in Taiwan. Presented at the meeting of Early lung cancer symposium in Taichung
 
Taiwan 2010
22
Aberle
DR
Brown
K
Lung cancer screening with CT
Clin Chest Med
 , 
2008
, vol. 
29
 (pg. 
1
-
14
)