The Intersection of Oncology Care and Worker Well-Being

Labor supply and well-being outcomes will be evaluated at 3 timepoints (baseline, start of treatment, 12 months). The likelihood of remaining employed will be estimated with linear probability models as is the convention in labor economics. The study team also estimate employment using analogous logistic models, reporting marginal effects. Weekly hours worked will be estimated using ordinary least squares regression. In addition to measuring raw hours, the team estimate change and percent change in hours worked from baseline. These models can be expanded to accommodate control variables for worksite and job characteristics as well as accommodations, satisfaction, and other covariates predictive employment and hours worked.

Well-being will be assessed as specified by the Worker Well-Being Questionnaire (WellBQ) Manual23 and Patient Reported Outcomes Measurement Information Systems (PROMIS) domains. For PROMIS, subjects respond to a 4-item questionnaire for each domain. Responses are summed and standardized to t-scores, with higher t-scores representing higher levels of the domain measured. For example, an individual with a PROMIS depression t-score of 50 is the average observed in the general U.S. population, while a t-score of 60 indicates one standard deviation above the average score in the general population and represents moderate depressive symptoms. Guided by previous health literature, the team consider a half standard deviation (5-point) increase from initial scores a meaningful improvement for all PROMIS domains. The literature also suggests that meaningful change ranges from 2 to 6 points. Secondary to Aim 1, the team will assess the correlation between PROMIS measures and WellBQ. The WellBQ is at an earlier stage of adoption, warranting side-by-side comparison with established instruments like PROMIS. The InCharge instrument will be evaluated as suggested by the developers.