Supportive Care in Cancer

, Volume 22, Issue 12, pp 3125–3134

Smoking status, service use and associated factors among Japanese cancer survivors—a web-based survey

Authors

    • Psycho-Oncology DivisionNational Cancer Center East
    • Center for Psychiatric Oncology and Behavioral SciencesMassachusetts General Hospital
    • Department of NeuropsychiatryKeio University School of Medicine
  • Shino Umezawa
    • Psycho-Oncology DivisionNational Cancer Center East
    • Section of Liaison Psychiatry and Palliative Medicine Division of Comprehensive Patient CareTokyo Medical and Dental University
  • Ado Basaki-Tange
    • Psycho-Oncology DivisionNational Cancer Center East
  • Maiko Fujimori
    • Psycho-Oncology DivisionNational Cancer Center East
    • National Center for Neurology and Psychiatry
  • Mitsunori Miyashita
    • Division of Palliative Nursing, Health SciencesTohoku University Graduate School of Medicine
Original Article

DOI: 10.1007/s00520-014-2284-2

Cite this article as:
Fujisawa, D., Umezawa, S., Basaki-Tange, A. et al. Support Care Cancer (2014) 22: 3125. doi:10.1007/s00520-014-2284-2

Abstract

Purpose

This study aims to investigate smoking status and its associated factors among Japanese cancer survivors. We stretched our focus on association with health-related behaviors other than smoking (alcohol intake, physical exercise, and social activity) and the smoking cessation strategies used by cancer survivors.

Methods

An anonymous cross-sectional web-based survey was conducted, enrolling survivors of various types of cancer up to 10 years after diagnosis. Smoking status, socioeconomic status, health-related behaviors other than smoking, and smoking cessation resource that the participants used were evaluated. Factors associated with continuous smoking after cancer diagnosis were explored using multivariate analysis.

Results

Among 168 participants who were smoking at the time of cancer diagnosis, 96 participants (57.1 %) continued smoking. Sixty-seven survivors (69.8 %) were willing to reduce or quit smoking, however, only 39 survivors (40.6 %) were provided with counseling or intervention on smoking cessation. Male gender, shorter time after cancer diagnosis, and lack of regular physical exercise associated with continuous smoking. Higher level of fear of cancer recurrence had trend-level significance of association with smoking cessation.

Conclusions

Substantial proportions of Japanese cancer survivors continue smoking after diagnosis of cancer. The majority of them are not provided with relevant information or support, despite their willingness of reducing or quitting smoking. Smoking cessation is associated with other health behaviors (i.e., physical exercise). This suggests considerable missed opportunities for health-care providers to provide cancer survivors with counseling and evidence-based interventions. Promotion of professional support on smoking cessation and education to encourage healthy behaviors are needed.

Keywords

Cancer survivorSmoking cessationRisk factorHealth behaviorExercise

Introduction

With advances in cancer detection and treatment, increasingly more attention has been paid towards care for cancer survivors. Lifestyle factors are potential means to improve cancer outcomes and quality of life among cancer survivors [1]. Smoking status especially has critical significance [2]. Continued smoking after cancer diagnosis is associated with poor response to cancer treatment [3, 4], increased morbidity and mortality [5, 4] (both cancer related [6] and noncancer related [7]), increased risk of cancer recurrence and secondary cancer [8], and decreased quality of life [9, 10]. There is also evidence that smoking cessation leads to risk reduction of cancer relapse and increase of quality of life (QOL) [11].

In Japan, the smoking rate among general population is as high as 23.4 % (male, 38.2 %; female, 10.9 %); despite the rate is on gradual decrease. Smoking accounts for 19.5 % of cancer incidence and causes and 23.2 % of cancer death in Japanese population [12] and increases the rate of secondary cancer [13]. Studies on smoking among cancer survivors has been scarce in Japan, and they are limited to the studies that addressed small numbers of lung cancer patients who underwent cancer treatments [14, 15].

Although awareness of the risks associated with continued smoking is growing, the smoking rate of cancer survivors seems to be still high. According to the National Health Interview Survey 1998–2001 in the USA, the smoking rate of cancer survivors was estimated to be about 20 % and was not different from that of the general adult population. This trend is even more remarkable among younger population [16, 17]. The situation seems to be similar in Australia [18], UK [19], and other European countries [20, 21].

Despite there is good evidence that shows effectiveness of smoking cessation interventions among cancer survivors [22], relatively small proportion of survivors has been reported to be informed of useful resource (e.g., 42 % in US cancer survivors). Even smaller proportion of smoker survivors use those evidence-based treatments (33.5 % for pharmacotherapy and 3.8 % for behavioral treatment [23, 24]). Whether appropriate information and support on smoking cessation has been provided to cancer survivors is an important issue that needs to be explored.

For successful smoking cessation, behavioral and psychological issues should be considered. Affective and behavioral issues correlate with smoking cessation in the context that (1) several negative health behaviors, namely, inactivity, poor diet, and tobacco use share common risk factors [17], (2) substance use problems (nicotine and alcohol) frequently coincide, and (3) smoking can be a coping strategy to alleviate psychological distress [25]. However, these issues have not been well explored in cancer survivor population.

Therefore, the current study aimed to investigate smoking status and its associated factors among Japanese cancer survivors. We stretched our focus on association with health-related behaviors other than smoking (alcohol intake, physical exercise, and social activity), and the smoking cessation strategies used by cancer survivors.

Method

Participants

A cross-sectional, web-based survey was conducted as a part of the study to measure quality of life of community-dwelling cancer survivors. We aimed to recruit 600 cancer survivors using a stratified sampling method by gender (male/female = 1:1) and time since cancer diagnosis (200 survivors each from the following three categories—within 2 years from cancer diagnosis, 2 to 5 years, and 5 to 10 years). The sample size was set as such because we considered 200 participants each was enough to illustrate difference in characteristics of each group, since a meta-analysis of supportive care needs surveys showed majority of past studies involved up to 200 participants [26]. Among these participants, we selected the participants who were smoking at their cancer diagnosis as our study sample.

Procedure

The survey was conducted during the period between December 7 and 13, 2012. The study was approved by the institutional review board of Tohoku University. We used a website-monitoring system called INTAGE research monitor, Inc. (http://intage.co.jp). In this system, approximately 1,300,000 people from any part of Japan were spontaneously registered. Among them, 2,059 participants were registered as age of 20 years or older and as being diagnosed of cancer within 10 years. Based on the past statistics of this monitoring system, we estimated the response rate as 66 %, therefore, we asked 300 candidate participants from each category (900 in total) to participate in the study, by sending an e-mail illustrating the details of the survey. The questionnaire was completely anonymous. Six hundred twenty-eight participants responded and completed the questionnaire (response rate, 69.7 %). Those who had never smoked or who had stopped smoking before the diagnosis of their cancer were excluded, and finally 168 survivors were identified as being a smoker at the time of their cancer diagnosis and were subjected to analysis (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00520-014-2284-2/MediaObjects/520_2014_2284_Fig1_HTML.gif
Fig. 1

Recruitment flow

Measures

Smoking status

To evaluate smoking status, we adopted questions that have been used in the Japan Comprehensive Survey of Living Conditions (CSLC) 2007 [27]. The CSLC is a nationally representative cross-sectional survey of Japanese household members in Japan, which has been conducted by the Ministry of Health, Labor and Welfare every 3 years. This questionnaire inquires current and past smoking habit, and willingness to quit smoking (for current smokers). The smoking status was classified into two categories: (1) cessation group at the time of survey (the participants who was smoking at the diagnosis of cancer but quitted afterward) and (2) continued smoking group (the participants who were smoking at the time of their cancer diagnosis, and had been smoking regularly until the time of survey).

Smoking cessation resources

The participants were inquired of whether and from whom they were provided with support for smoking cessation. The participants were asked to choose from the following five sources of support (multiple responses were allowed): (1) advice and/or support from their oncology staff, (2) advice and/or support from non-oncology medical staff, (3) smoking cessation clinic, (4) internet-based smoking cessation program, (5) other self-help resource, and (6) no resource used at all.

Other health behaviors

Alcohol consumption, physical activity, and participation in social activities were assessed using CSLC survey questionnaire [27], with minor modification of adding description of “before cancer diagnosis” and “after cancer diagnosis” to highlight change after cancer diagnosis. The participants were inquired whether they drink alcohol regularly, and how many units of alcohol they consume (if they drink regularly). We dichotomized the data according to the definition of hazardous drinking [28]. The participants were inquired of their physical activity level with the question “In your everyday life, do you intentionally exercise in order to promote/maintain your health?” They were asked to choose one response that best describes their condition from among the following five possible responses: (1) regularly, (2) sometimes, (3) stopped exercising after cancer diagnosis, (4) stopped exercising before cancer diagnosis, and (5) almost no habit of exercise since before cancer diagnosis. The participants were defined as “With no habit of regular exercise” when they endorsed 3, 4, or 5. The participants were provided with a list of social activities (e.g., hobby, physical exercise, job, training and education, art and creative works, community gathering, etc.) and were asked whether they had participated in any of them within one year before the survey. The participants were defined as “Not participating in social activities” when they denied participation in any of those activities.

Psychiatric morbidity

Psychiatric morbidity was assessed using the K6 scale, a self-rated six-item questionnaire that inquires frequency of psychological distress symptoms during the past 30 days. Each question has a five-point scale ranging from “none of the time” (0) to “all of the time” (4). The total K6 score ranges from 0 to 24, with higher scores indicating more severe psychological distress. A validation study of K6 demonstrated that the total score of 15 or higher indicates psychiatric morbidity with stratum-specific likelihood ratios (SSLRs) of 16 (6.1–34) [29].

Fear of cancer recurrence

Fear of cancer recurrence was assessed using a single scale adopted from Quality of Life-Cancer Survivors Instrument (QOL-CS) [30, 31]. This scale measures fear of cancer progression in a 10-point Likert scale. The higher score indicates more intense fear of cancer progression. In the original validation study, Cronbach’s alpha co-efficient of the whole scale was 0.93. The overall QOL-CS correlation coefficient with the Functional Assessment of Cancer Therapy scale (FACT-G) [32] was 0.78.

Perceived social support

Perceived social support was assessed with the short-version Multidimensional Scale of Perceived Social Support (MSPSS) [33, 34], a well-validated seven-item questionnaire that measures perceived social support on seven-point scales. The mean higher score indicates higher perceived social support. The scale has shown satisfactory internal consistency (Cronbach’s alpha = 0.85) and has criterion validity with General Health Questionnaire in a Japanese community-dwelling sample.

Demographic, socioeconomic, and medical characteristics

Background characteristics were collected, including gender [35, 36], age [37, 36], annual income (above or below four million Japanese Yen (national average annual income, approximately US$40,000 or 30,000 €) [38], marital status [36], cancer type, and numbers of co-morbid conditions [39].

Statistical analysis

First, we conducted univariate analyses to screen differences between the continued smoking group and the cessation group. Chi-square tests were used for categorical variables and Mann-Whitney’s tests were used for continuous variables, with a cutoff score of p < 0.2. Then, we conducted multivariate analysis using logistic regression analysis to explore for associated factors. Variables that have consistently been reported to associate with smoking behaviors, namely, gender, age, marital status, income, and time after cancer diagnosis, were entered regardless of statistical significance in univariate analysis. The level of statistical significance was set at p < 0.05 (two-tailed). All analyses were conducted using SPSS version 21.0 IBM.

Results

Characteristics of the participants

Characteristics of the participants are shown in Table 1. The participants’ mean age was 58.9 (SD = 11.9). The majority of the participants were male (66.1 %). Breast, gastric, and prostate cancer were the top three cancer groups. Most of the participants were in good performance status (≥1).
Table 1

Demographic background

Characteristics

 

N

Percentage

Demographic and social

 Age

 

Mean = 58.9 (SD = 11.9)

Range: [28–84]

 Gender

Male

111

66.1

Female

57

33.9

 Marital status

Married

126

75.0

Single or divorced/ widowed

42

25.0

 Having child(ren)

No

41

24.4

Yes

127

75.6

 Single or cohabiting

Single

24

14.3

Cohabiting

144

85.7

 Employment

Unemployed

88

52.4

Employed

80

47.6

 Change in working status

Changed

91

54.2

Not changed

77

45.8

 Income

<4 million yen

67

39.9

≥4 million yen

101

60.1

Clinical

 Primary cancer site

Head and neck

9

5.4

Thyroid

5

3.0

Lung

7

4.2

Gastric

22

13.1

Esophageal

15

8.9

Hepatobiliary

4

2.4

Colorectal

19

11.3

Breast

25

14.9

Prostate

21

12.5

Other urological

11

6.6

Gynecological

14

8.3

Hematological

6

3.6

Others

10

6.0

 Time since cancer diagnosis

<2 years

51

30.4

2–5 years

54

32.1

6–10 years

63

37.5

 Performance status

0

101

60.1

1

62

36.9

2

5

3.0

Smoking status

Of the 168 participants who were smoking at the time of cancer diagnosis, 96 survivors (57.1 %) were still smoking at the time of the survey. Of these 96 continued smokers, 67 survivors (69.8 %) were willing to cut down (n = 30; 31.2 %) or to quit smoking (n = 37; 38.5 %).

Smoking cessation resource

Among all the participants, only 39 survivors (40.6 %) had been provided by medical providers with advice or support for smoking cessation. Seventeen survivors (17.7 %) used self-help. Forty-five survivors (46.9 %) did not receive any resource for smoking cessation (Fig. 2). Cessation group were more likely to be provided with advice and/or support by oncology medical staff, than continuous smokers (36.1 vs. 25.0 %; p = 0.12).
https://static-content.springer.com/image/art%3A10.1007%2Fs00520-014-2284-2/MediaObjects/520_2014_2284_Fig2_HTML.gif
Fig. 2

Resources used for smoking cessation

Factors associated with smoking cessation after cancer diagnosis

In univariate analyses, marital status (single, widowed or divorced), shorter time after cancer diagnosis, lack of habit of regular physical exercise, lack of participation in social activities, hazardous drinking, lower level of fear of recurrence, and lack of advice on smoking cessation by oncology staff were extracted as possible associated variables for continued smoking (p < 0.2; shown in Table 2). The logistic regression analysis demonstrated that male gender, shorter time after cancer diagnosis, and lack of regular physical exercise were significantly associated with continuous smoking after cancer diagnosis. In addition, lower level of fear for cancer recurrence has a trend-level association with continuous smoking. The model explained 27.1 % of the variance (Table 3).
Table 2

Proportion of smoking cessation and continues smoking group

  

Smoking cessation

Continued smoking

p

n

%

n

%

Demographic and social variables

 Gender

Male

47

42

64

58

0.85

Female

25

44

32

56

 Age

<60

37

41

54

59

0.53

≥60

35

45

42

55

 

 Marital status

Married

59

47

67

53

0.07

Single/divorced/widowed

13

31

29

69

 

 Having child(ren)

No

15

37

26

63

0.35

Yes

57

45

70

55

 

 Single or cohabiting

Single

10

42

14

58

0.90

Cohabiting

62

43

82

57

 

 Current working status

Currently not working

40

45

48

55

0.48

Currently working

32

40

48

60

 

 Change in work status

Changed

40

44

51

56

0.75

Not changed

32

42

45

58

 

 Income

<4 million yen

31

46

36

54

0.47

≥4 million yen

41

41

60

59

 

Clinical variables

 Primary cancer site

Gastric-esophageal

19

51

18

49

0.26

Colorectal

10

53

9

47

Breast

13

52

12

48

Prostate

8

38

13

62

Others

22

33

44

67

 Time after diagnosis

<2 years

14

27

37

73

p<0.01

2–5 years

22

41

32

59

5–10 years

36

57

27

43

 Performance status

0

42

42

59

58

0.75

1 or 2

30

45

37

55

 Comorbidities

None

29

40

44

60

0.53

Present

43

45

52

55

Affective and behavioral variables

 Psychiatric morbidity (K6 ≥ 15)

None

52

45

63

55

0.36

Present

20

38

33

62

 Fear of cancer recurrence

Low

29

34

56

66

0.02

High

43

52

40

48

 Perceived social support

Low

40

41

58

59

0.53

High

32

46

38

54

 Regular physical exercise

None

21

33

43

67

0.04

Yes

51

49

53

51

 Social activities

None

17

35

32

65

0.17

Yes

55

46

64

54

 Regular alcohol intake

None

29

48

32

52

0.35

Yes

43

40

64

60

 Hazardous drinking

No

35

44

45

56

0.20

Yes

8

30

19

70

 Advice/support by oncology staff for smoking cessation

No

46

39

72

61

0.12

Yes

26

52

24

48

 Participation in cancer support group

Never

67

42

92

58

0.5

Yes (current/ past)

5

56

4

44

Table 3

Associated factors for continuous smoking after cancer diagnosis

 

B

S.D.

Wald

p

OR

95 % CI

Lower

Upper

Age <60

0.69

0.54

1.63

0.20

1.99

0.69

5.74

Male gender

1.47

0.62

5.69

0.02

4.36

1.30

14.6

Married

−0.69

0.62

1.25

0.26

0.50

0.15

1.69

Income (<4 million yen)

−0.02

0.49

0.001

0.98

0.99

0.38

2.55

Time after cancer diagnosis (reference, 5–10 years)

  

5.21

0.07

   

 <2 years

1.23

0.60

4.19

0.04

3.42

1.05

11.11

 2–5 years

0.91

0.54

2.86

0.09

2.49

0.87

7.16

Regular physical exercise

−1.21

0.57

4.53

0.03

0.30

0.10

0.91

Social activities

0.66

0.58

1.31

0.25

1.94

0.62

6.05

Hazardous drinking

0.66

0.59

1.27

0.26

1.94

0.61

6.13

Advice/support by oncology staff

−0.24

0.51

0.22

0.64

0.79

0.29

2.13

Fear of cancer recurrence

−0.14

0.08

3.061

0.08

0.87

0.74

1.02

Constant

−0.24

1.05

0.05

0.82

0.79

  

Discussion

This web-based study demonstrated that substantial proportion (57.1 %) of Japanese cancer survivors continue smoking even after diagnosis of cancer, and only small proportion of them are provided with appropriate support for smoking cessation. Male gender, shorter time after cancer diagnosis, and lack of habit of regular physical exercise are predictors for continuation of smoking after cancer diagnosis.

High rate of continuous smoking is consistent with past studies in other countries and in different cancer situations (23–42.9 %)[16, 18, 40] . A notable gap was found between large proportion of smoker survivors who are willing to cut down smoking (approximately 70 %) and small proportion of survivors who received support for smoking cessation (approximately 40 %). Provision of support was considerably lower than those reported in past studies (63 to 74 %) [41, 42, 24, 43], warranting further intervention in Japan.

As we hypothesized, cancer survivors’ health behavior, more specifically, habit of regular physical exercise, was associated with higher probability of smoking cessation. Habit of regular physical exercise may be a representation of cancer survivors’ awareness toward healthier behaviors in general, and/or may be working as an alternative stress-coping strategy that replaces smoking. Considering that higher level of fear of cancer recurrence had trend-level association with smoking cessation, we assume that high level of fear of cancer recurrence contributed to high awareness toward good health behaviors, which lead to smoking cessation and regular exercise. Psycho-behavioral smoking cessation programs usually start with raising the motivation of participants. Providing personalized information on the risk of cancer progression/relapse and how smoking and smoking cessation can influence it, can be the vital first step for smoking cancer survivors. Also, our results suggest research implication for combining smoking cessation program with promotion of other healthy behaviors such as exercise. Health counseling, including information provision on cancer prognosis and promotion of healthy behaviors, should be considered as an important aspect of care and should be remunerated.

As opposed to past studies [19, 44], male gender associated with continued smoking. One attributable reason is the age difference between the male (mean age = 60s) and female (mean age = 40s) survivors in our sample. The male participants are likely to have longer history of smoking compared with female participants, although we have not obtained detailed smoking history. Another possible reason is due to the difference in study settings between our study and past studies. Our study is a cross-sectional observational study, while the findings that women are less likely to quit smoking mostly came from interventional studies. Success rate in a smoking cessation program is different from a naturalistic study where patients may or may not participate in smoking cessation programs are different. In fact, in our study, female survivors were more likely to have used smoking cessation clinic (15.8 vs 6.3 %, p = 0.056).

The smoking cessation rate increased as time passed after cancer diagnosis, which is consistent with past studies [45]. Medical providers should not hold negative attitudes toward cancer survivors who are smoking at one time point. Rather, those survivors should be viewed as struggling their way to quit smoking. They should be provided with continuous advice and support, since repetitive advice has been proven to help smokers succeed in cessation [46].

“Teachable moments” of smoking cessation for cancer patients include instances such as on cancer screening [47], on cancer diagnosis, before surgery [48], during hospitalization and on discharge [49]. Considering that only a small proportion of our participants have ever received advice or support from oncology staff, any opportunity listed above can be a good opportunity. Since it is not realistic for Japanese oncologists to provide smoking counseling due to their heavy case load, multidisciplinary approach is essential.

Strengths and limitations

The strengths of this study include enrollment of participants with well-balanced distribution of time since their cancer diagnoses, use of web-based questionnaire, and consideration of health behaviors other than smoking. Use of web-based survey is advantageous in that genuine anonymity is guaranteed. This minimizes the risk of reporting bias, because, if this kind of study were conducted in a medical facility, participants might not report their unhealthy behaviors honestly to the researchers who have possible connection with medical providers.

Use of web-based questionnaire can also be a limitation due to insufficient representativeness of the sample. The participants only consist of cancer survivors with internet literacy and with relatively good performance status. Other limitations include skewed distribution of type of cancer (higher proportion of breast and prostate cancer and lower proportion of lung cancer, even if compared with Japanese cancer survivor population [50]), clinical information being based upon patient report, lack of detailed smoking history, and lack of information on participants’ readiness to change their behaviors (motivation and self-efficacy). The cross-sectional design cannot tell causal relationships.

Despite these limitations, the current study is worthy of reporting because research on this topic has been scarce in Japan and even in other Asian countries [51, 52]. The study is also unique in that it assessed cancer survivors’ health behaviors as well as relevant service use.

Conclusion

This web-based survey demonstrated that substantial proportions of Japanese cancer survivors continue smoking after diagnosis of cancer. Majority of them are not provided with relevant information or support, despite their willingness to reduce or quit smoking. This suggests considerable missed opportunities for health-care providers to provide cancer survivors with counseling and evidence-based interventions. The current study also illuminated populations who are at higher risk of continuing smoking after cancer diagnosis, especially in regard with other health behaviors. Further promotion of education and support to encourage healthy behaviors is needed.

Acknowledgments

This study was fully supported by a Grant-in-Aid from Japanese Ministry of Health Labor and Welfare. The authors express gratitude to Professors Yosuke Uchitomi and Masatoshi Inagaki (Okayama University) and Ms. Junko Ueda-Nouno (National Cancer Center East) for their valuable inputs.

Copyright information

© Springer-Verlag Berlin Heidelberg 2014