Table 2.
Summary of findings
| Post-diagnosis physical activity and all-cause mortality for patients with breast cancer, COPD, type 2 diabetes and IHD | ||||
|---|---|---|---|---|
| Outcome | Relative effect per 10 MET-h/week (95% CI) | № of participants (studies) | Certainty of the evidence (GRADE) | Informative statements |
| All-casue mortality for breast cancer | HR 0.78 (0.71 to 0.86) | 27,248 (12 observational studies) | ⨁⨁◯◯ LOW a,b,c | The evidence suggests post-diagnosis physical activity results in a slight reduction in all-cause mortality for individuals with breast cancer. |
| All-cause mortality for COPD | HR 0.70 (0.45 to 1.09) | 4784 (2 observational studies) | ⨁◯◯◯ VERY LOW b,c,d,e,f | The evidence is very uncertain about the effect of post-diagnosis physical activity on all-cause mortality for individuals with COPD. |
| All-cause mortality for IHD | HR 0.88 (0.83 to 0.93) | 42,027 (8 observational studies) | ⨁⨁◯◯ LOW b,c,g | The evidence suggests post-diagnosis physical activity results in a slight reduction in all-cause mortality for individuals with IHD. |
| All-cause mortality for type 2 diabetes | HR 0.96 (0.93 to 0.99) | 32,221 (6 observational studies) | ⨁⨁◯◯ LOW a,b,c | The evidence suggests post-diagnosis physical activity results in a slight reduction in all-cause mortality for individuals with type 2 diabetes. |
Explanations:
aDowngraded by two levels since five studies were judged as serious risk of bias regarding confounding or selection bias based on ROBINS-I
bDowngraded by one level because although exposure was assessed in all studies using validated questionnaires, there were differences in the assessment and calculation of physical activity levels
cUpgraded by one level due to the dose-response gradient
dDowngraded by two levels since two studies were judged as serious risk of bias regarding confounding or selection bias based on ROBINS-I
eDowngraded by one level because the 95% CI includes the null value (HR = 1.0) and includes important benefits HR < 0.75
fDowngraded by one level because publication bias could not be assessed due to limited number of studies (< 5 studies)
gDowngraded by two levels since three studies were judged as serious risk of bias regarding confounding or selection bias based on ROBINS-I