Multi- or uni-dimensional focus
As outlined earlier the interventions fell into two broad categories - interventions which explicitly aimed to improve functioning in both outcome domains (physical and psychosocial) and interventions which focused on one domain only. The first category referred to as interventions with a multidimensional focus (MDF) and the second category is referred to as interventions with a uni-dimensional focus (UDF).
The seven MDF studies (Bai 2004; Carmack Taylor 2006; Daubenmier 2006; Fillion 2008; Giesler 2005; May 2008a; Rogers 2009) recruited 653 participants in total. However, the findings were inconsistent. Four studies used cognitive behaviour therapy (CBT) combined with physical exercise (Bai 2004; Carmack Taylor 2006; Fillion 2008; May 2008a). As highlighted above, Carmack Taylor 2006 found no statistically significant improvement in either physical or psychosocial outcomes. This study, entitled 'Active for Life' (Carmack Taylor 2006), evaluated the efficacy of a six-month group-based lifestyle physical activity programme to improve the quality of life (QoL) of prostate cancer patients compared to an educational programme and standard care. The lifestyle programme participants were taught CBT skills to help them undertake and maintain regular physical activity (30 minutes of moderate activity prescribed for most days of the week). The control group was provided with the same educational component, but did not receive CBT. The authors cite high baseline scores and an insufficient sample size as possible reasons for no statistically significant improvement among intervention participants compared to the control group. They concluded that a lifestyle programme which focused on cognitive-skills training and education was insufficient to improve the QoL of prostate cancer patients.
May 2008a also utilised CBT, but combined it with a supervised physical training (PT) programme (PT plus CBT) and compared it to an attention control arm which received the physical training programme alone and a waiting list control group. The study was reported in four papers; one paper detailed the physical outcomes, one presented QoL outcomes, one paper combined both outcomes and the fourth paper examined cancer-related fatigue (May 2008a). While this study showed a significant improvement in physical fitness (as measured by the 12-item Physical Activity Scale for the Elderly (PASE)) from baseline in both intervention groups (PT mean difference 2.1 (95% CI 1.2 to 3.0); PT plus CBT mean difference 2.0 (95% CI 1.1 to 2.9); P < 0.0001) there was no statistically significant difference between the groups (mean difference -0.1, 95% CI -1.5 to -1.3, P not provided). No psychosocial improvements were detected between the intervention groups during the three-month follow-up period. However, when compared to a waiting list control group, statistically and clinically relevant improvements were noted for both rehabilitation intervention arms in terms of physical role limitations (change 20.8 (8.9 to 32.7) P < 0.001), physical functioning (change 9.4 (5.1 to 13.6) P < 0.001), vitality (change 9.8 (5.3 to 14.3) P < 0.001) and health change (change 25.7 (16.8 to 34.5) P < 0.001). The authors concluded that adding a structured CBT intervention to a group-based self management physical training programme did not enhance the overall effect.
Similarly, Bai 2004 examined the effect of combined CBT and exercise therapy on the QoL of nasopharyngeal cancer patients compared to a standard care group. This study showed statistically significant improvements in five areas of functioning compared to the control group: physical (mean difference -4.61, P < 0.01), cognitive (mean difference 2.38, P < 0.05), emotional (mean difference -3.69, P < 0.01), fatigue (mean difference -5.84, P < 0.01) and general QoL (mean difference -4.26, P < 0.01). However, it is important to note that this was a small study (n = 45) and it was classified as having a high risk of bias because details of random sequence generation, group allocation or blinding of the outcomes were not provided. In addition, information about a sample size calculation was not included and it was not clear whether or not they met their target recruitment number to identify statistically significant findings.
Fillion 2008 also developed a group intervention that combined stress management using cognitive and behavioural strategies with physical activity to reduce fatigue and improve energy levels, QoL, fitness and emotional distress. This short four-week intervention for breast cancer survivors showed that participants in the intervention group had greater improvement in fatigue score (intervention mean (M) = 2.40 (standard deviation (SD) 0.84), control M = 2.75 (SD 0.93); P = 0.03), energy levels (intervention M = 2.63 (SD 0.72), control M = 2.24 (SD 0.88); P = 0.01) and emotional distress (intervention M = 11.15 (SD 3.85), control M = 13.13 (SD 5.44); P = 0.04) compared to the standard care control group. We assessed this study as having a moderate risk of bias because the authors did not state whether or not the outcomes assessment had been blinded.
The intervention in the Rogers 2009 study was based on a social cognitive model and comprised group discussions facilitated by a Clinical Psychologist, supervised exercise, home-based exercise and face-to-face counselling sessions with an exercise specialist. This 12-week intervention for breast cancer survivors receiving hormone therapy found improvements in objective measures of physical fitness and psychosocial measures including physical activity as measured by accelerometer (mean difference (MD) 72.103; 95% CI 25,383 to 119,000; effect size (d) = 1.02; P = 0.004), back/leg muscle strength (MD 12.3; 95% CI 0.4 to 15.9; (d) = 0.81; P = 0.017), waist-to-hip ratio (MD -0.05; 95% CI -0.01 to -0.08; (d) = -0.77; P = 0.018) and social well-being as measured by the Functional Assessment of Cancer Therapy - Breast (FACT-B) measure (MD 2.0; 95% CI 0.3 to 3.8; (d) 0.76; P = 0.03) compared to a standard care control group who received information only. However, it should be noted that the study was based on a small sample (n = 41), and it was not possible to analyse potential moderating factors such as demographic variables, age or readiness to change. In addition, we assessed the study as having a high risk of bias because exercise specialists who completed the objective physical measures were not blinded to participant group allocation.
The remaining two MDF studies (Daubenmier 2006; Giesler 2005) did not use a specific psychotherapeutic approach such as CBT. Instead, they provided general psycho-education (Giesler 2005) or psycho-education specific to stress management (Daubenmier 2006).
Daubenmier 2006 developed the most intensive intervention for men undergoing active surveillance for prostate cancer. Experimental group patients were prescribed an intensive lifestyle programme that included a vegan diet supplemented with soy (one daily serving of tofu plus 58 g of a fortified soy protein powdered beverage), fish oil (3 g daily), vitamin E (400 IU daily), selenium (200 µg daily) and vitamin C (2 g daily), moderate aerobic exercise (walking 30 minutes six days weekly), stress management techniques (gentle yoga-based stretching, breathing, meditation, imagery and progressive relaxation for a total of 60 minutes daily) and participation in a weekly one-hour support group to enhance adherence to the intervention. The intervention was maintained for one year and outcomes were compared with a standard care control group. This study was reported in two papers. No statistically significant group differences were observed for physical or psychosocial QoL scores. However, the authors highlighted that the baseline scores for participants were high and that this ceiling effect left little room for improvement, though individuals who improved their lifestyle enhanced their QoL. The intervention did affect cancer progression: men in the intervention group had a significant reduction in their PSA (prostate-specific antigen) values (intervention -4%, control +6%, P = 0.01) and in the growth of LNCaP prostate cancer cells, which is a measure of prostate cancer progression (intervention 70%, control 9%, P < 0.001).
The final MDF study (Giesler 2005) developed a unique nurse-led, computer-assisted intervention that randomised patient-spouse dyads to the intervention or a standard care arm. Dyads in the intervention arm met once per month for six months with a nurse. The nurse identified and tracked QoL problems using a computer-based assessment programme. An extensive range of physical and psychoeducational strategies were available for each problem identified. The intervention had statistically significant long-term (up to 12 months) beneficial effects on sexual outcomes (intervention mean (M) 12.35 (SD 17.28), control M = 3.11 (SD 19.61); P = 0.02) and on cancer worry (intervention M = 14.15 (SD 25.12), control M = 3.07 (SD 17.68); P = 0.03) compared to patient dyads receiving standard care. Some participants in the intervention group had high levels of baseline depression and they fared worse, relative to the control group. However, we classified the results from this study as having a high risk of bias because the authors did not specify their method of randomisation, allocation concealment or undertake intention-to-treat (ITT) analysis. In addition, the study had insufficient statistical power due to subject recruitment difficulties.
In summary, there is limited evidence to support the use of MDF programmes for cancer survivors. Only three studies (Bai 2004; Fillion 2008; Rogers 2009) showed both physical and psychosocial benefits for their programmes. However, as discussed, the studies had a moderate risk of bias (Fillion 2008) or a high risk of bias (Bai 2004; Rogers 2009). May 2008a also found positive outcomes for their rehabilitation programme, though the positive outcomes were attributed to the physical component per se and were not enhanced by including a psychosocial component.
The majority of UDF rehabilitation programmes (four out of five) found a significant effect for the stated aim or focus of the given programme. Berglund 2007 did not find improvement in the intervention or control group. Studies that had a stated aim of improving physical functioning showed significant improvements in at least one physical outcome. The interventions did not have any significant effect on psychosocial outcomes, though this was not a goal of the programmes.
Bennett 2007 used motivational interviewing (MI) to increase physical activity and improve aerobic fitness, health and fatigue in a small (n = 56) mixed group of participants. The intervention arm self reported a significantly higher level of physical activity with a mean increase in energy expenditure of 1556 kcal/wk compared to an increase of 397 kcal/wk in the control group (P < 0.05). However, it should be noted that the mean level of regular activity at baseline was significantly lower (P = 0.04) in the intervention group than in the control group. Motivational interviewing did not have a statistically significant effect on mental health status.
Demark-Wahnefried 2006 developed a home-based diet and exercise programme known as Project LEAD, to improve lifestyle behaviours and ultimately enhance physical functioning. Project LEAD showed a statistically significant improvement in diet quality over the six-month period (intervention +2.2, control -2.9; P = 0.003). Change in physical function did not reach significance (P = 0.23). The improvements in the intervention arm did diminish during the post-intervention period. Although the study was relatively large (n = 182) it did not achieve the expected or required recruitment rate. In addition, the authors did not provide clear information about random sequence generation, allocation concealment or blinding of outcome assessors and the study was classified as having a high risk of bias.
Similarly, the Demark-Wahnefried 2007 second study, entitled the FRESH START trial, aimed to improve diet and exercise practices of breast and prostate cancer survivors through a tailored printed intervention delivered via mail. This study achieved a high recruitment rate (n = 543) and observed a significant mean difference between arms in the practice of two or more lifestyle behaviours (intervention +34%, control +18%; P < 0.001). While the attention control arm showed improvement in outcomes, the intervention arm showed significantly greater improvements in mean scores for exercise minutes per week (intervention +59.3, control +39.2; P = 0.02), fruit and vegetable consumption (intervention +1.1 servings, control +0.6; P = 0.01), decreased fat intake (intervention -4.4%, control -2.1%; P < 0.001) and reduced body mass index (BMI) (intervention -0.3, control +0.1 kg/m2; P = 0.004). Significant changes in psychosocial outcomes such as depression, social support and quality of life were not noted for either arm over the duration of the study, though there may have been a ceiling effect at baseline measurement.
Only one UDF study, Lepore 1999, explicitly stated the aim of promoting psychosocial aspects of care though they incorporated and assessed physical aspects. Lepore 1999 included an educational session on ‘Cancer, diet and exercise’ and assessed physical and mental health outcomes using the SF-36; interpersonal conflict using the Lepore Social Conflict Scale; and a variety of measures of cognitive processing. This small study (n = 24) showed that intervention participants had greater improvements in mean mental health scores over time (intervention +14.33, control +0.67; P < 0.05), fewer interpersonal conflicts with their wife (intervention +0.3, control -0.05; P < 0.01), improved self efficacy (intervention +0.19, control +0.06; P < 0.05) and lower levels of distress associated with cancer-related thoughts (intervention -0.46, control +0.17; P < 0.05). Mean physical functioning scores were not significantly affected by the intervention. Similar to other studies, we classified Lepore 1999 as having a high risk of bias because it did not provide information about the method of random sequence generation, allocation concealment or blinding of outcome assessors.
Overall, current rehabilitation programmes with a uni-dimensional or focused aim rather than a multidimensional aim or focus appear to be more successful in terms of effecting positive change in the domain directly related to their focus.