Sustained Exercise & Obesity Cancer Prevention: US Study Results

Here’s a breakdown of the ⁣sensitivity analyses conducted, as described in the text:

1. Lagged Covariate Data: Covariate data was‍ shifted back by 2 years ⁣to ensure it ⁣preceded physical activity data,⁢ reducing ⁣the risk of reverse‍ causation. This changed the risk assessment period ⁤from 11 years to 9 years.
2. Different BMI Functional Forms: ⁣ The analysis was repeated using both linear ⁤and log-transformed⁢ BMI values.
3. Temporal Ordering of Covariates: The order in⁢ which time-varying covariates were considered when modeling their relationships‍ was altered.
4. Age Restriction & Adherence: Participants were restricted to under 90 years at baseline, and ⁣adherence requirements were relaxed for those over 90 due ⁤to‍ limited data.
5. COPD Exclusion & Adherence: Participants with baseline COPD were excluded,and those who developed COPD during follow-up were also excused from ⁣adherence ⁤requirements.
6. Cardiovascular Event History & Adherence: Participants with a history of cardiovascular events at baseline were included, and⁢ adherence wasn’t impacted by subsequent events.
7. Cancer Screening Adjustment: The analysis was adjusted for baseline ⁢and time-varying cancer screening history (breast, colorectal, prostate) to account for potential healthcare-seeking biases.
8. Smoker Exclusion: ⁤ Current and recent smokers (less than 20 years since quitting at baseline) were excluded.

Positive Control:

* CVD Mortality as a Positive Control: Cardiovascular disease (CVD)⁤ mortality (using ICD-10 codes I00-I99) was used as ⁣a positive control ⁢outcome to validate the modeling approach, as a strong effect of moderate-to-vigorous physical activity (MVPA) is ⁣expected.

These analyses were performed to assess how robust the study’s findings were to different assumptions and data handling choices.