Wei Zhang

The objectives of this thesis were to: (1) examine the sex/gender-specific impacts that employment status change (employment transitions, ETs) have on body weight and waist circumference (WC) changes in middle- and older-age adults and (2) assess the contribution of changes in health behaviours (HBs: sleep duration and quality, smoking status, alcohol consumption, and physical activity) in the association between ETs and anthropometric changes.Chapter 1: Eight bibliometric databases were systematically searched, with citations followed up. Twelve studies were included. All studies examined retirement but reported mixed results. Retirement either led to weight gain or did not alter weight compared to non-retirement. Occupation type modified the association: weight gain was more commonly reported among retirees from physically demanding occupations. Two included studies also examined job-loss and results were also mixed. Key confounders and commonly studied HBs were identified.Chapter 2 & 3: Two waves of data from the Canadian Longitudinal Study on Aging were used to classify 10,117 working women and men into three ET groups: stayed working, entered retirement, and stopped working. The outcomes were change in weight and WC on continuous scales, and change as categories (≥5% gain, ≥5% loss, no change). Multivariable linear and multinomial logistic regressions were adjusted for confounders and HBs.Multivariable linear regression models showed that weight changes did not differ across ETs in women, although changes in WC showed different directions across ET. By contrast, men who entered retirement lost more weight and had greater reductions in WC relative to men who stayed working (-0.59 kg, [95% CI: -1.11, -0.08] and -0.83 cm, [95% CI: -1.46, -0.20]). Final multinomial logistic regression models did not show significant associations; however, the direction of the effects remained. In models that included HB change variables, estimates were not attenuated. Retirement may result in small amounts of weight loss in middle-aged and older Canadian men. Work stoppage appears to lead to more WC increases in women, but the evidence is uncertain. Contrary to hypotheses that place HBs along the pathway between ETs and anthropometric change, the findings from the empirical study suggest that they are independent risk factors instead.

View record

This study aimed to: (a) adapt the previously validated Valuation of Lost Productivity (VOLP) questionnaire for people with health problems, to a caregiver version to measure work productivity loss associated with caregiving responsibilities, (b) evaluate measurement feasibility and validity of an online version of the caregiver VOLP questionnaire, and (c) compare two caregiver populations based on whether they worked from home.A mixed methods design was utilized. Qualitative methods were used for VOLP adaptation and online conversion. Quantitative methods were used to evaluate feasibility and validity of the online VOLP and compare the two caregiver groups. The Work Productivity and Activity Impairment (WPAI) questionnaire for caregivers was included to compare its outcomes and their correlations with VOLP outcomes.When adapting the VOLP for caregivers, our qualitative analysis showed the importance of adding three major components: caregiving time, work productivity loss related to volunteer activities and caregivers’ lost job opportunities. A total of 382 caregivers who completed online survey were included in our final quantitative analysis. We found small Spearman rank correlations between VOLP and WPAI, observing a larger correlation between absenteeism [r = 0.49 (95% confidence interval: 0.37–0.60)] than presenteeism [r = 0.36 (0.24–0.47)]. Correlations between VOLP outcomes and total caregiving hours were larger for absenteeism [r = 0.38 (0.27–0.47)] than presenteeism [r = 0.22 (0.10–0.34)]. Correlations between WPAI outcomes and total caregiving hours were smaller for absenteeism [r = 0.27 (0.15–0.38)] than presenteeism [r = 0.35 (0.23–0.46)].When comparing caregivers who are working from home and those not working from home, differences in education, occupations and caregiver health status were observed. The VOLP absenteeism and presenteeism outcomes were significantly higher in caregivers working from home than caregivers not working from home, with small effect sizes from 0.16 to 0.19. This was not the case with the WPAI outcomes.The study provides evidence of the feasibility and preliminary validity evidence of the adapted VOLP caregiver questionnaire in measuring work productivity loss due to caregiving responsibilities, when compared with the results for WPAI and results from the previous patient-VOLP validation study.

View record