Basen-Engquist 2013

·    Design: prospective longitudinal study

·    Funding/CoI: National Institutes of Health Grants; CoI not reported

·    Setting: 1 university and 1 private centre, US

·    Sample size: N=100

·    Duration: recruitment Jan 2007 – Sept 2010

·   Eligibility criteria: women who had been diagnosed with Stage I, II, or IIIa endometrial cancer and were at least 6 months posttreatment with no evidence of disease; exclusion if they met the public health recommendations for physical activity (moderate or greater intensity on at least 5 days per week for 30 min or more, or vigorous intensity activity for 20 min or more on at least 3 days per week) and had maintained that level of activity for 6 months or longer

·   A priori patient characteristics:

o Mean age: 57y

o Mean time since diagnosis: 26m

o Mean BMI: 34.2

Exercise recommendation tailored to fitness level provided by masters-level exercise physiologist

Variables included in analysis:

·      Social-Cognitive Theory variables: exercise self-efficacy, outcome expectations (positive and negative), barriers self-efficacy

·      BMI

·     Exercise self-efficacy was the only variable that significantly predicted exercise minutes at the next time point (p=0.0069 in multivariate model)

Level of evidence: high risk of bias

·    Strong selection (643 potentially eligible persons)

·    Linear mixed-effects models, which account for the correlation among repeated measurements within subjects over time

·    Self-efficacy: measured with self-developed questionnaire

·    Outcome expectations: idem

·    Physical activity expressed in exercise minutes

Bélanger 2012

·    Design: cross-sectional survey

·    Funding/CoI: Lisa Belanger is supported by the Alberta Innovates: Health Solutions studentship award; Kerry Courneya is supported by the Canada Research Chairs Program; Alexander Clark is supported by career awards from the Canadian Institutes for Health Research and Alberta Innovates: Health Solutions

·    Setting: Alberta, Canada

·    Sample size: N=588

·    Duration: patients diagnosed in 2008

·   Eligibility criteria: young adult cancer survivors being diagnosed with invasive cancer between the ages of 20-44 and currently still between the ages of 20-44

·   A priori patient characteristics:

o Mean age: 38.2y

o Females: 70%

o Mean time since diagnosis: 73.6m

o Mean BMI: 26.5

Dependent variable: physical activity (Leisure Score Index from the Leisure-Time Exercise Questionnaire) and %of participants meeting public health physical activity guidelines (2008 physical activity guidelines for Americans)

Independent variables:

·      Theory of Planned Behavior variables (affective attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, planning, intention)

·      Demographic, medical and behavioural variables

·     Path analysis explained 38% (p<0.001) of the variance in physical activity with significant contributions from intention, planning, affective attitude, education, and general health

·     56% (p<0.001) of the variance in intention was explained by perceived behavioral control, instrumental attitude, and affective attitude

Level of evidence: high risk of bias

·    Participants randomly selected from persons identified through the Alberta Cancer Registry

·    588/2000 respondents

Blaney 2013

·    Design: cross-sectional survey

·    Funding/CoI: funded by the Department for Employment and Learning, Northern Ireland; no CoI

·    Setting: service users of supportive care cancer charity in Northern Ireland

·    Sample size: N=456

·    Duration: carried out in 2008

·   Eligibility criteria: cancer survivors

·   A priori patient characteristics:

o Median age: 61y

o Females: 76%

o Mainly breast cancer (64.4%)

o Median BMI: 29.04

Exercise frequency and intensity were measured using the Leisure Score Index (LSI) of the Godin Leisure-Time Exercise Questionnaire

·     Top 10 barriers interfering with exercise participation: illness/other health problems (37.3%), joint stiffness (36.9%), fatigue (35.7%), pain (30.1%), lack of motivation (26.5%), weather extremes (26.2%), lack of facilities (25.5%), weakness (21.5%), lack of interest (20.7%) and fear of falling (19.5%)

·     Top 10 facilitators: fun (88.0%), included a variety of exercises (81.8%), gradually progressed (78.9%), flexible (75.5%), involved personal goal setting (73.9%), included good music (73.2%), tailored to the individual (73.1%), included feedback (66.2%) and approved by their oncologist (65.7%) or general practitioner (60.3%)

Level of evidence: high risk of bias

·    456/975 respondents

·    No multivariate analysis

Brunet 2011

·    Design: cross-sectional analysis of prospective longitudinal study

·    Funding/CoI: supported by a Canadian Institutes of Health Research Grant awarded to the second author; the first author is supported by a Joseph-Armand Bombardier Canada Graduate Scholarship from Social Sciences and Humanities Research Council of Canada and a Psychosocial Oncology Research Training doctoral award

·    Setting: Quebec, Canada

·    Sample size: N=169

·    Duration: 2009-2010

·   Eligibility criteria: women aged 18+, 0-20 weeks after primary treatment for stage I-III breast cancer; no health concerns that prevent them from engaging in physical activity

·   A priori patient characteristics:

o Mean age: 55.06y

o Mean BMI: 26.21

o Time since diagnosis: 10.59m

Variables:

·      Self-presentation processes : SPEQ

·      Social cognitive constructs: SPES (self-presentation efficacy scale)

·      Physical activity behaviour: LTEQ (Leisure-Time Exercise Questionnaire)

·     Impression motivation was a significant correlate of moderate-to-vigorous physical activity (β = 0.25)

·     SPEE (β = 0.21) and SPOV (β =0 .27) were moderators of this relationship

·     The final models accounted for 12–24% of the variance in moderate-to-vigorous physical activity

Level of evidence: high risk of bias

·    Participants were recruited through advertisements and oncologist referrals from various local medical clinics and hospitals

·    Each analytical model controlled for age and BMI

·    Disease-related variables did not change the pattern or significance of the results

Chipperfield 2013

·    Design: cross-sectional survey

·    Funding/CoI: supported by Abbott Pharmaceuticals grant #IIS MET-11-0029 and Cabrini Institute, Cabrini Health Scholarship #1068651; no other CoI

·    Setting: three centres, Melbourne, Australia

·    Sample size: N=356

·    Duration: 2010-2011 (data collection over 12-month period)

·   Eligibility criteria: men aged 40-80y at completion of radiotherapy for prostate cancer; radiotherapy between 9-30m ago

·   A priori patient characteristics:

o Mean age: 67.4y

o Mean time since diagnosis: 33.1m

Dependent variable: patients meeting National Physical Activity Guidelines of Australia (NPAGA); physical activity measured with IPAQ (International Physical Activity Questionnaire)

Independent variables:

·      Quality of life: prostate cancer subscale of the FACT-P

·      Depression and anxiety: HADS

·      Demographic and medical characteristics

·     The odds of meeting NPAGA were significantly higher with lower depression scores (OR 0.84 [95%CI 0.76-0.94], p<0.01)

·     Participants with a tertiary level education were significantly more likely to be meeting NPAGA than those with primary/secondary school (OR 0.61 [95%CI 0.38-0.97], p<0.05) or TAFE/apprenticeship qualifications (OR 0.25 [95%CI 0.09-0.68], p<0.01)

·     Treatment category, comorbid conditions, age, anxiety and QoL were not significantly associated with meeting NPAGA

Level of evidence: high risk of bias

·    356/638 respondents

Cox 2009

·    Design: cross-sectional survey as part of longitudinal cohort study

·    Funding/CoI: NIH, NINR RO3 NR009203, Robert Wood Johnson Foundation, NIH NCI U24 CA55727, American Lebanese Syrian Associated Charities

·    Setting: multicentre study, US & Canada

·    Sample size: N=838

·    Duration: unclear

·   Eligibility criteria: persons who had survived five or more years after treatment for malignant disease diagnosed (before age 21) between 1970 and 1986

·   A priori patient characteristics:

o Not reported

Dependent variable: physical activity participation (participants were asked: “During the past month, did you participate in any physical activities or exercises such as running, calisthenics, golf, bicycling, swimming, wheelchair basketball, or walking for exercise?”)

Independent variables:

·      Directly observed variables: primary-care physician's familiarity with cancer-related problems, current pain resulting from cancer or its treatment, frequency of fatigue, whether survivors had discussed the risk of recurrent cancer with their primary-care physician, baseline frequency of aerobic exercise, age at diagnosis, current anxiety as a result of cancer or its treatment, current highest school grade completed, whether the survivor had seen a primary care physician since cancer treatment ended, intrinsic motivation, extrinsic motivation

·      Latent variables: survivor-provider interaction, fear, affect, and stamina

·     40% of the variance in male survivors' recent physical activity participation was explained directly and/or indirectly by self-reported health fears (p=0.01), perceived primary-care physician expertise (p=0.01), baseline exercise frequency (p≤0.001), education level (p=0.01), self-reported stamina (p=0.01), cancer-related pain (p≤0.001), fatigue (p≤0.001), age at diagnosis (p=0.01), cancer-related anxiety (p≤0.001), motivation (p=0.01), affect (p=0.01), and discussion of subsequent cancer risk with the primary-care physician (p≤0.001)

·     31% of the variance in females' recent physical activity participation was explained directly and/or indirectly by self-reported stamina (p≤0.001), fatigue (p=0.01), baseline exercise frequency (p=0.01), cancer-related pain (p≤0.001), cancer-related anxiety (p=0.01), recency of visits with primary-care physician (<0.001), quality of interaction with the primary-care physician (p=0.01), and motivation (p≤0.001)

Level of evidence: high risk of bias

·    Original study contacted 20346 persons; of 12872 persons that remained alive, 1600 persons were randomly sampled; of these, 838 completed the 2 surveys

·    No multivariate analysis

Gjerset 2011

·    Design: cross-sectional survey

·    Funding/CoI: funded by the Norwegian Foundation for Health and Rehabilitation and the Norwegian Cancer Society

·    Setting: Norwegian Radium Hospital

·    Sample size: N=975

·    Duration: 2/2007 – 9/2007

·   Eligibility criteria: patients aged 18-75y that had received curatively intended treatment for malignant lymphoma, breast, testicular, cervical, ovarian or prostate cancer

·   A priori patient characteristics:

o Age 45-64y: 48%

o Females: 56%

o BMI < 25: 48%

o Time since diagnosis ≥2y: 89%

Dependent variable: level of physical activity participation (modified version of Godin Leisure-Time Exercise Questionnaire)

Independent variables:

·      Medical and demographic variables

·     Increasing age and weight, low education, comorbidity and smoking were associated with physical inactivity after treatment

·     Change in level of physical activity from active to inactive was associated with comorbidity, distant disease and smoking, while a change from inactive to active was associated with high education

Level of evidence: high risk of bias

·    975/2024 analysable patients

Harrison 2009

·    Design: longitudinal cohort study

·    Funding/CoI: National Breast Cancer Foundation, Australia

·    Setting: Queensland, Australia

·    Sample size: N=287

·    Duration: unclear

·   Eligibility criteria: women with primary, invasive, unilateral breast cancer (diagnosed in 2002), aged 20-74y

·   A priori patient characteristics:

o Mean age: 55y

Dependent variable: physical activity (Behavioral Risk Factor Surveillance System), converted to metabolic equivalent task (MET) hours/week, and categorized according to national physical activity guidelines

Independent variables:

·      Medical, behavioural and demographic variables

·     Nine variables showed associations with change in physical activity levels from 6 to 18 months following diagnosis, collectively explaining 35% of variance

·     The only statistically significant factor was treatment-related complications: mean adjusted change in MET = 17.7 (95%CI 3.0-32.4) if no complications (p=0.01)

Level of evidence: high risk of bias

·    287/511 randomly selected women were analysable

Hsu 2011

·    Design: prospective longitudinal study

·    Funding/CoI: Department of Defense of US Army (DAMD17 – 03 – 1 – 0521), excellence for cancer research center grant, No: DOH99-TD-C-111-002, Department of Health, Executive Yuan, Taiwan and grants from the Kaohsiung Medical University Hospital, Taiwan (KMUH95-5D10, KMUH96-6G17); no CoI

·    Setting: 3 teaching hospitals in metropolitan areas of north and south Taiwan

·    Sample size: N=196

·    Duration: 2003-2005

·   Eligibility criteria: women aged 18+ with confirmed first diagnosis of breast cancer and completed therapy; currently in remission; absence of recurrent disease after initial breast cancer treatment

·   A priori patient characteristics:

o Mean age: 47y

Dependent variable: exercise frequency (21-item exercise log)

Independent variables:

·      demographic variables, fatigue, perceived health status, social support for exercise, perceived barriers for exercise, exercise self-efficacy, exercise outcome expectancy

·     Baseline exercise frequency was the best significant predictor of exercise frequency

·     The effect of social support for exercise on exercise frequency was apparently larger in older subjects, especially those over 40 years old, than in younger subjects

·     Mental health, exercise barriers and exercise outcome expectancy significantly contributed to change in exercise frequency

Level of evidence: high risk of bias

Huy 2012

·    Design: retrospective cohort study

·    Funding/CoI: Deutsche Krebshilfe e. V. [Grant No. 70-2892-BR I and 108523/108419], the Hamburg Cancer Society, the German Cancer Research Centre, and the German Federal Ministry for Education and Research [Grant No. 01KH0402]

·    Setting: German region

·    Sample size: N=1067

·    Duration: 2002-2010

·   Eligibility criteria: women with primary invasive breast cancer or carcinoma in situ (that had undergone mastectomy or lumpectomy)

·   A priori patient characteristics:

o Mean age: 63.5y

o Mean BMI: 26.3

Dependent variable: physical activity measured with questionnaire and converted to MET-hours per week

Independent variables:

·      breast cancer-related variables, patient-related variables

·     Patients treated with chemotherapy, radiotherapy, or both had a stronger decline in physical activity during therapy and the first 3 months after surgery, respectively, compared to patients without therapy or those treated only with hormones (adjusted β = -9.73 [95%CI -18.55 to -0.91] to -13.54 [-21.93 to ‑5.15]; p<0.05)

·     Overall decline in physical activity was greater in patients treated with chemo- (β = ‑15.41 [‑30.28 to -0.55]; p=0.042) or radiotherapy (β = -12.56 [-24.97 to -0.15]; p=0.047)

·     Participation in rehabilitation was positively associated with an increase in physical activity after breast cancer therapy (β = 7.62 [2.63 to 12.61]; p=0.003)

·     There was a negative association for age considering overall change in physical activity after controlling for other covariates (β = -0.66 [-1.22 to -0.10] per year; p=0.020)

·     No significant associations with BMI, WHR, or other patient-related variables

·     Patients with medical risk factors had a stronger decline in physical activity during therapy compared to those without these conditions (β = -5.56 [-9.59 to -1.53]; p=0.007)

·     The presence of medical risk factors was also a negative predictor for overall change in total leisure-time physical activity (β = -8.25 [‑14.26 to -2.24]; p=0.007)

·     No further significant results for other clinical characteristics

·     Patients with a higher prediagnostic physical activity level had a greater decline in physical activity during therapy (β = -0.77 [-0.83 to ‑0.72) per MET-h/week; p < .001)

·     Significant associations for change after therapy and overall change in total leisure-time physical activity

·     Smoking and alcohol consumption were not significantly associated with change in physical activity in adjusted analyses

Level of evidence: high risk of bias

·    Of the 5969 invited persons (in 2 German regions), 3919 completed baseline assessment; out of these, 2542 completed the follow-up assessment; the data presented here are from 1 German region

Karvinen 2009

·    Design: retrospective cohort study

·    Funding/CoI: University of Alberta–EFF Support for the Advancement of Scholarship Small Faculties Research Grant and a Research Team Grant from the National Cancer Institute of Canada with funds from the Canadian Cancer Society and the NCIC/CCS Sociobehavioral Cancer Research Network

·    Setting: Alberta, Canada

·    Sample size: N=397

·    Duration: 10/2005 – 2/2006

·   Eligibility criteria: patients 18+ with diagnosis of bladder cancer within the last 15 years

·   A priori patient characteristics:

o Mean age: 70.2

o Females: 25.3%

o Mean time since diagnosis: 72.4m

Dependent variable: exercise behaviour (Leisure Score Index from the Godin Leisure Time Exercise Questionnaire)

Independent variables:

·      Theory of Planned Behavior variables (affective attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, planning, intention)

·      Demographic and medical variables

·     Intention (β=0.25, p<0.001), perceived behavioral control (β=0.18, p=0.001), and planning (β=0.12, p=0.018) explained 20.9% of the variance in exercise over a 3-month period

·     Perceived behavioral control (β=0.32, p<0.001), affective attitude (β=0.18, p=0.002), instrumental attitude (β=0.15, p=0.025) and descriptive norm (β=0.10, p=0.032) explained 39.1% of the variability in exercise intention

·     Constructs from the TPB mediated the associations between adjuvant therapy, cancer invasiveness, age, and exercise

·     Age and adjuvant therapy also moderated some of the associations within the TPB

Level of evidence: high risk of bias

·    1027 persons received questionnaire: no differences between respondents and non-respondents

McGowan 2013

Speed-Andrews 2012

·    Design: cross-sectional survey

·    Funding/CoI: not reported

·    Setting: Alberta, Canada

·    Sample size: N=600

·    Duration: May – Aug 2008

·   Eligibility criteria: patients diagnosed with colorectal cancer aged 18+ that completed adjuvant therapy

·   A priori patient characteristics:

o Mean age: 67.3y

o Females: 41.7%

o Mean time since diagnosis: 51m

Dependent variable: (1) physical activity (Leisure Score Index from the Godin Leisure Time Exercise Questionnaire) and percentage of participants meeting 2008 Physical Activity Guidelines for Americans; (2) sport participation rate, sport preferences

Independent variables:

·      Theory of Planned Behavior variables (affective attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, planning, intention)

·      Demographic and medical variables

·     The TPB explained 34% (p<0.001) of the variance in physical activity behaviour with direct associations for intention (β = 0.22; p=0.001) and planning (β = 0.18; p=0.015)

·     Intention had 62% (p<0.001) of its variance explained by perceived behavioural control (β = 0.43; p<0.001), affective attitude (β = 0.25; p<0.001) and instrumental attitude (β = 0.15; p<0.001)

·     33.0% (p=0.001) of the variance in sport participation was explained by being male (β=0.12; p=0.006), in better general health (β=0.12; p=0.006), and ≥5 years post-diagnosis (β=0.09; p=0.031)

·     The most common barriers to sport participation were time, age/agility, and no interest/dislike of sports

·     The most common anticipated benefits of sport participation were improved physical fitness, meeting people, and improved health

Level of evidence: high risk of bias

·    Completion rate: 30% (600/2000)

·    Significant but small differences were found and identified that responders were younger in age by about 4 years compared to non-responders (67.3 vs. 71.1 years; p<0.001) and nearly 2 months further away from the date of diagnosis (56.4 vs. 54.6 months; p=0.012)

Milne 2008

·    Design: cross-sectional survey

·    Funding/CoI: Courneya is supported by the Canada Research Chairs Program and a Research Team Grant from the National Cancer Institute of Canada with funds from the Canadian Cancer Society and the NCIC/CCS Sociobehavioral Cancer Research Network

·    Setting: Western Australia

·    Sample size: N=558

·    Duration: May – Dec 2004

·   Eligibility criteria: women diagnosed in 2002 with breast cancer aged 18+, no longer undergoing active treatment, no secondary cancers

·   A priori patient characteristics:

o Mean age: 59y

o Mean time since diagnosis: 25.2m

Dependent variable: physical exercise (Godin Leisure Time Exercise Questionnaire)

Independent variables:

·      Demographic and medical variables

·      Self-determination theory (SDT) motivation continuum: Behavioural Regulation for Exercise Questionnaire-2

·      Competence and autonomy support: Perceived Competence Scale (PCS) and modified Health Care Climate Questionnaire (mHCCQ)

·     SDT constructs explained 20.2% (p<0.01) of the physical activity variance

·     Significant independent SDT predictors included identified regulation (β = 0.14, p<0.05) and competence (β = 0.23, p<0.01), with autonomy support approaching significance (β = 0.9, p=0.057)

Level of evidence: high risk of bias

·    558/1045 completers

Ng 2008

·    Design: cross-sectional survey

·    Funding/CoI: funding not reported; no CoI

·    Setting: 4 Harvard-affiliated hospitals, US

·    Sample size: N=511

·    Duration: diagnosis made between 1969 and 1996

·   Eligibility criteria: patients with Hodgkin’s lymphoma aged 18+, 5 or more years from diagnosis

·   A priori patient characteristics:

o Median: 26y

o Females: 50%

Dependent variable: health practice (routine physical examination and dental visit in the past year; smoking; daily alcohol consumption; physical activity)

Independent variables: age at Hodgkin’s lymphoma diagnosis (≤50 vs. >50), gender, time since Hodgkin’s lymphoma treatment (<10 years, 10–15 years vs. >15 years), annual household income (<$60,000 vs. $$60,000), educational level (<college level vs. college level or higher), history of Hodgkin’s lymphoma relapse or second cancer, and reported level of concern regarding future health and cancer risks

·     Higher household income (OR=1.48, 95%CI 1.09-2.02; p=0.01) independently predicted for having had a physical examination in the past year

·     Lower educational level (OR=3.3, 95%CI 1.64-5.56; p=0.0004) and history of relapsed Hodgkin’s lymphoma (OR=2.1, 95%CI 1.07-3.91; p=0.03) were independent predictors for smoking, moderate/heavy alcohol use, and/or physical inactivity

Level of evidence: high risk of bias

·    Completion rate: 50% (511/1023)

·    Significant differences between responders and non-responders: responders were older at the time of diagnosis; more females

Peddle 2008

·    Design: cross-sectional survey

·    Funding/CoI: University of Alberta–Social Sciences Research Grant Program; CoI reported in detail

·    Setting: Alberta, Canada

·    Sample size: N=413

·    Duration: June – Sept 2004

·   Eligibility criteria: patients aged 20-80y diagnosed with colorectal cancer and completed adjuvant therapy for at least 1y; no evidence of recurrent disease

·   A priori patient characteristics:

o Mean age: 60y

o Females: 46%

o Mean BMI: 29.0

Dependent variable: exercise behaviour (Leisure Score Index from the Godin Leisure Time Exercise Questionnaire)

Independent variables:

·      Self-determination theory (SDT) variables (behavioural regulation, perceived autonomy support, psychological need satisfaction in exercise)

·      Demographic and medical variables

·     SDT and education explained 16% of the variance in exercise behavior: identified regulation (β=0.17, p=0.031), introjected regulation (β=0.15, p=0.006), and education (β=0.16, p<0.001) each making a significant independent contribution

Level of evidence: high risk of bias

·    Response rate: 51.1% (413/809)

Soerjomataram 2012

·    Design: cross-sectional survey

·    Funding/CoI: internal grant from the Public Health Department of Erasmus MC; data collection was supported by Comprehensive Cancer Centre South; no CoI

·    Setting: Eindhoven, the Netherlands

·    Sample size: N=1349

·    Duration: conducted in 2009

·   Eligibility criteria: patients diagnosed with colorectal cancer between 1998 and 2007

·   A priori patient characteristics:

o Age 65+: 57%

o More than 5y since diagnosis: 30%

Dependent variable: lifestyle behaviour (smoking, alcohol consumption, weight)

Independent variables:

·      Demographic and medical variables

·      Having received chemotherapy was significantly associated with being overweight (adjusted OR=1.5, 95%CI 1.05-2.3) and consuming alcohol (adjusted OR=1.7, 95%CI 1.1-2.7)

·      Female patients were less likely than males to currently smoke (OR=0.5, 95%CI 0.4-0.8), consume alcohol (OR=0.3, 95%CI 0.2-0.4) or be overweight (OR=0.6, 95%CI 0.5-0.8)

·      Survivors from the lowest socioeconomic group were more likely to be current smokers (OR=1.8, 95%CI 1.1-3.0) and overweight (OR=1.5, 95%CI 1.1-2.1)

Level of evidence: high risk of bias

·    Response rate: 74% (1349/1682)

·    Significant differences between responders and non-responders: non-responders were more likely to be older than 65 years, had cancer of the colon and had one or more comorbidity at time of diagnosis

Stevinson 2009

·    Design: cross-sectional survey

·    Funding/CoI: niversity of Alberta and a Research Team Grant from the National Cancer Institute of Canada, with funds from the Canadian Cancer Society and the NCIC/CCS Sociobehavioral Cancer Research Network

·    Setting: Alberta, Canada

·    Sample size: N=359

·    Duration: May – Oct 2006

·   Eligibility criteria: women aged 18+, diagnosed with ovarian cancer between 1985 and 2005

·   A priori patient characteristics:

o Age 60+: 22%

o Time since diagnosis <5y: 25%

Dependent variable: physical activity (Leisure Score Index from the Godin Leisure Time Exercise Questionnaire)

Independent variables:

·      Theory of Planned Behavior variables (affective attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, planning, intention)

·      Demographic and medical variables

·      36% of the variance in physical activity guidelines was explained by the Theory of Planned Behaviour variables, with intention being the sole independent correlate (β = 0.56; p<0.001)

·      Adding significant medical and demographic variables explained an additional significant 6% of the variance in physical activity behavior, with being disease-free (β = 0.09; p=0.03), having a healthy BMI (β = 0.12; p=0.005), and being better educated (β = 0.14; p=0.001) achieving independent associations with behavior, although intention remained the most important correlate (β = 0.51; p<0.001)

Level of evidence: high risk of bias

·    Response rate: 51.4%

Trinh 2012

·    Design: cross-sectional survey

·    Funding/CoI: grants described in article; no CoI

·    Setting: Alberta, Canada

·    Sample size: N=703

·    Duration: unclear

·   Eligibility criteria: patients aged 18+ diagnosed with kidney cancer between 1996 and 2010

·   A priori patient characteristics:

o Mean age: 64.4y

o Females: 37.6%

o Mean time since diagnosis: 68.6m

o Mean BMI: 28.6

Dependent variable: physical activity (Leisure Score Index from the Godin Leisure Time Exercise Questionnaire)

Independent variables:

·      Theory of Planned Behavior variables (affective attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, planning, intention)

·      Demographic and medical variables

·      42% of the variance in physical activity guidelines was explained by the Theory of Planned Behaviour variables

·      There were significant pathways from perceived behavioural control (β = 0.18, p=0.02), planning (β = 0.22, p<0.01) and intention (β = 0.31, p<0.01) to physical activity

·      There were strong significant total effects of perceived behavioural control (β = 0.43, p<0.01) and intention (β = 0.49, p<0.01) on physical activity

·      There were significant total effects of instrumental attitude (β = 0.14, p=0.02), descriptive norm (β = 0.04, p=0.01), and planning (β = 0.22, p<0.01) on physical activity

Level of evidence: high risk of bias

·    Completion rate: 703/1985

·    Responders were approximately one year closer to their date of diagnosis, had a slightly higher rate of treatment with systemic therapy, and less likely to have renal cell carcinoma and more likely to have clear cell carcinoma

Vallance 2012

·    Design: cross-sectional survey

·    Funding/CoI: Project Interface Grant from Alberta Health Services – Cancer Corridor

·    Setting: Alberta, Canada

·    Sample size: N=524

·    Duration: Sept – Oct 2009

·   Eligibility criteria: women aged 18+ with stage I-IIIa breast cancer who had completed adjuvant therapy (except hormonal therapy)

·   A priori patient characteristics:

o Mean age: 62.4y

o Mean time since diagnosis: 76.4m

Dependent variable: physical activity (Leisure Score Index from the Godin Leisure Time Exercise Questionnaire)

Independent variables:

·      Theory of Planned Behavior variables (affective attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, planning, intention)

·      Demographic and medical variables

·      Physical activity intention explained 12% of the variance in physical activity behaviour (p<0.01) while the Theory of Planned Behavior constructs together explained 43% of the variance in physical activity intention (p<0.01)

·      Intention had a significant direct effect on physical activity behaviour (β = 0.26, p<0.001)

Level of evidence: high risk of bias

·    Response rate: 30% (524/1735)

Yang 2013

·    Design: cross-sectional survey

·    Funding/CoI: National Cancer Center (grant no. 0910191 and 1210150); no CoI

·    Setting: 10 centres, South-Korea

·    Sample size: N=493

·    Duration: conducted in 2009

·   Eligibility criteria: patients aged 18+ diagnosed with cancer

·   A priori patient characteristics:

o Mean age: 59.1y

o Mean time since diagnosis: 2.4y

o Females: 8.1%

Dependent variable: continued smoking

Independent variables:

·      Perceived social support

·      Demographic and medical variables

·      Current alcohol consumption (OR = 3.29; 95%CI 1.91-5.65), early cancer stage (p for trend < 0.01), lung cancer diagnosis (OR = 0.41; 95%CI 0.19-0.88), and high perceived social support (OR = 0.59; 95%CI 0.37-0.96) showed significant associations with continued smoking

Level of evidence: high risk of bias

·    Completion rate: 25.2% (493/1956)