br Our study has some limitations
Our study has some limitations. NORMO data for physical activity are self-reported, and it is well known that people tend to overestimate their level of physical activity . In the collection of data, the participants were told to round the level of both moderate and vigorous activities to the nearest half hour, which could potentially lead to overestimation of the physical activ-ity level. To counteract this potential bias, we used conservative estimates of MET-h. Second, we used RR estimates for breast cancer to estimate the number of preventable cancers of endometrial cancer, a similar approach as Parkin . Third, we based our calcula-tions on a single domain of physical activity (leisure time including transportation), and it might not reflect the
total level of physical activity of the population. Addi-tionally, we did not differentiate the type of recreational physical activity, i.e. running, gardening etc., because the data do not allow this distinction. This does not allow for a nuanced estimation for each of the Nordic countries, where e.g. cycling as transportation (4.0e6.8 MET-h) is dominant in Denmark, in Diphenylterazine (DTZ) to cross country skiing (6.8e9.0 MET-h) which is more common in Norway, Finland and Sweden .
One limitation of the Prevent model is that it does not provide any uncertainty measure, such as confidence in-tervals. It is therefore important to do sensitivity analyses, which in our study indicates that the percentage of avoidable cancers is fairly robust to changes in LAT and LAG times, as well as to incorporating a trend in the cancer incidence. Even so, the results should be inter-preted with caution, as the main purpose of the model is not to produce valid estimates of the future cancer burden but rather show the difference in the number of cases under different levels of exposure prevalence.
We did not take any other changes in modifiable risk behaviour into consideration, which could have an un-defined impact on the results. Increased levels of phys-ical activity could potentially result in reduced sedentary behaviour and healthier dietary habits. Evidence sug-gests that physical activity and sedentary behaviour may be inversely correlated , but we did not include data for sedentary behaviour. However, it would be relevant to perform similar calculations on preventable cancer cases and the potential of reducing sedentary behaviour such as TV-viewing and computer screen time in the Nordic countries, since sedentary behaviour is consid-ered an independent risk factor of several types of can-cer [32,36].
Furthermore, we did not adjust for the interaction with overweight. An increase in physical activity could also lead to a lower prevalence of overweight and obesity, which would add to the number of avoidable cancer cases. Our previous study showed that an elimi-nation of overweight and obesity in the Nordic countries would avoid 40,000 postmenopausal breast cancer cases in a 30-year period, 45,000 colon cancer cases and 33,000 endometrial cancer cases .
In addition, we have chosen to include the cancer sites for which there is strong evidence of a protective effect of physical activity. Studies have shown that more cancer sites than those included here could be associated with insufficient physical activity [5,6], and the number of avoidable cancers could therefore be larger. For instance, Moore et al. found an inverse association be-tween physical activity and the risk of oesophageal adenocarcinoma, cancers of the liver, lung, kidney, bladder, head and neck, rectum, gastric cardia as well as myeloid leukaemia and myeloma, in addition to the three sites included in our study, namely colon, endo-metrial and postmenopausal breast cancer . Still, the literature is conflicting, and many studies only compare
groups with highest and lowest physical activity levels which is not enough for our calculations.
Our results show a potential of increasing physical activity for cancer prevention in the Nordic countries. From a public health perspective, it is also important to increase the level of physical activity since regular physical activity prevents several other non-communicable diseases, e.g. cardiovascular diseases, type 2 diabetes mellitus and overweight and obesity [37e40]. Hence, interventions aiming to raise the level of physical activity in the Nordic countries should be pri-oritised. However, interventions as well as physical ac-tivity recommendations to the public should be held at a realistic level to encourage the population to adopt a more physically active lifestyle. Engaging in 5 h physical activity with moderate intensity per week corresponding to at least 15 MET-h might be considered realistic for the Nordic populations to achieve (scenario A), but it would probably take years of structural changes and interventions to achieve. In addition, our estimations revealed an effect of about 4600 preventable cases, if all groups with insufficient levels of physical activity were reduced by 50%, or if the group with least physical ac-tivity was eliminated. These are more realistic scenarios. Either way, it requires political endorsement, infra-structural policy development and effective health pol-icies as well as the engagement of health organisations, public health authorities and other relevant stakeholders working in close cooperation and over a long time to plan and implement effective structural changes and long-term interventions aiming to increase the levels of physical activity in the Nordic countries.