From health outcomes to value assessments : preference-based evaluation of interventions & valuation of productivity costs among working adults

Abstract: Background: This thesis applies value assessment frameworks to non-specific low back pain (LBP) in a working population to evaluate health outcomes and to estimate productivity costs. LBP is a common diagnosis in primary health care. It is characterized by recurrent pain episodes and is a major factor in the increasing sickness absence among workers. LBP contributes substantially to the burden of disease and to the economic burden in terms of productivity loss to employer and society at large. Interventions are thus needed to prevent recurring LBP and its associated burden. Such interventions need to be properly evaluated from a variety of perspectives, including that of the worker, the employer and society at large, in order to help us better understand how to support worker-health policies and a sustainable working life. Aims: In the first part of this thesis, the frameworks used aim to investigate the cost-effectiveness of and preferences for secondary prevention interventions for LBP. The second part of the thesis focuses on the production loss measure validity test and on deriving wage multipliers to estimate productivity costs. The thesis investigates productivity costs associated with sickness absence, sickness presenteeism and work environment-related problems from the employer’s perspective. Methods: In Study 1, a cost-effectiveness analysis was conducted among working adults (n =159) to compare the effect and costs of yoga for the prevention of LBP with strength exercises and evidence-based advice. This analysis was conducted from two perspectives, namely that of the employer and that of society at large. The outcome quality adjusted-life years (QALYs) were examined in a 12-month follow-up randomized controlled trial. In Study 2, a discrete choice experiment was conducted among working adults with LBP (n =112). This experiment used the conditional logit model to examine the influence of exercise attributes and individual characteristics on preferences for exercise to prevent LBP. In Study 3, using the Pearson Correlation and the Bland and Altman’s test of agreement, the convergent validity of the Swedish health-related and work environment-related production loss measures (HRPL and WRPL respectively) were tested against the Health and Work Performance Questionnaire (HPQ) among working adults (n = 88). The HPQ is an extensively psychometrically tested and widely-used instrument. The Intraclass Correlation Coefficient (ICC) and Bland and Altman’s tests of repeatability were used as tests of stability (n = 44). In Study 4, wage multipliers for managers (n = 758) were derived using an ordinal probit model to predict the costs of productivity loss from sickness absence, sickness presenteeism and work environment-related problems. Results: The cost-effectiveness analysis in Study 1 demonstrates that yoga is less costly and improves quality of life (QALY) compared with strength exercises and evidence-based advice. For an additional QALY worth EUR 11,500 for society, yoga yielded a positive incremental net benefit of EUR 1,542 and EUR 2,860 compared with strength exercise and evidence-based advice respectively. Yoga could also be cost-effective, compared with evidence-based advice, if an employer considers that the improvements in QALY justify the additional cost of the intervention (EUR 150 per worker with LBP). These results only hold for those who adhered to the recommendations of exercising twice a week. The discrete choice experiment in Study 2 demonstrates that the most preferred exercise option was medium to high-intensity cardiovascular training performed in a group with trainer supervision at least once to twice per week. The most preferred types of incentive were exercise during work hours and wellness allowances (Friskvårdbidrag). The individual characteristic that most consistently influenced preferences for exercise was age. The convergent validity test of the production loss measures conducted in Study 3 showed moderately strong correlations (i.e. r = 0.46 for the HRPL and r = 0.31 for WRPL), as expected. The ICC for HRPL assessments was 0.90 and 0.91 for WRPL between the different testing occasions. This suggests that the test–retest reliability was good. Study 4 demonstrates that sickness absence, sickness presenteeism and work environment problems significantly impact team workers’ productivity when job characteristics (i.e. teamwork, ease of substitution of workers and time sensitivity of output) are taken into account. To determine the economic implications for the employer in terms of the cost, it was estimated that the median wage multipliers from the sampled occupations in the study were 1.92 for sickness absence, 1.65 for health-related presenteeism due to acute illness, 1.58 for health-related presenteeism due to chronic illness, and 1.70 for work environment problems. Conclusions: Studies 1– 4 gave rise to the following conclusions: a) Yoga may be considered a cost-effective early intervention for the prevention of LBP, but further investigations are warranted. b) Where preferences for exercise to prevent LBP are concerned, the most important factors were the age and exercise attributes such as type of exercise, frequency, level of supervision and incentives. This implies that providers and employers could improve participation in exercise for working adults with non-specific LBP by focusing on the exercise characteristics which are most attractive. c) The validity test of the HRPL and WRPL suggests that the measures have convergent validity and good stability. This finding may have implications for advancing methods of assessing production loss as an outcome, which represents a major cost for employers. d) After job characteristics had been taken into account, the economic impact of sickness absence, sickness presenteeism and work environment-related problems on team productivity exceeded the cost of wages in a number of occupations. This implies that there could be productivity gains for employers if the cost of health and work environment-related productivity losses can be reduced.