Table 1 Data-merging strategy in PODESA

Accelerometer data

Identical accelerometer hardware and software was used to measure physical work demands in both the NOMAD and in the DPhacto cohorts, making the two cohorts highly comparable; thus, the accelerometer data were added in a simple merge.

Questionnaire data

The majority of survey items from the questionnaires in the NOMAD and DPhacto cohorts are identical or comparable (77 items) enabling adding them in a simple merge.

However, for the minority of survey items which were not identical, and merely similar, we assessed the comparability of the items and possible modification. Specifically, four items had accordance of question wording but dissimilar response scale size (e.g. nine versus ten categories), and five items that differed in wording (e.g. used non-identical time frames). A total of 14 items were non-comparable and therefore not merged.

For similar, but not identical, survey items, the following procedure was used: firstly, based on findings from the literature within the field, we evaluated whether dissimilarities in wording or response scales could influence the answers, and secondly, using descriptive statistics we assessed the answer distribution in both cohorts. Similar items were merged if the literature indicated no difference in answers due to wording or response scale of the items and if the answer distribution on the items was similar in the two cohorts.

The questionnaire data contain, e.g., background information to be used as covariates (such as age, sex, smoking status, alcohol intake).

(An overview of the merging of the questionnaire data is available upon request).

Health check data

The health check data from the two studies derive from a health check and a physical testing session at baseline conducted by trained research professionals. Because identical health check procedures were followed in the NOMAD and DPhacto cohorts that are the basis of the PODESA cohort, we added the health check from each study in a simple merge.

The health check data includes data on, e.g., height, weight, hip and waist circumference, percentage of body fat, blood pressure, maximal oxygen uptake, maximal hand grip strength, back extension endurance and back flexibility.

Register data on long-term sickness absence

In addition to the PODESA cohort data, we add register data on primarily long-tern sickness absence from the DREAM register dataset which includes weeks of subsidized sickness absence spells (typically granted after 30 days of sickness absence).

Register data on short-term sickness absence

We also add register data including short-term sickness absence from the ‘Danish Register of Work Absences’ which includes daily employer-reported sickness absence.