In this study, we aimed to evaluate paid and unpaid productivity losses in patients with ischemic stroke at various time points up to 2 years after stroke according to HCA and FCM. Higher mRS scores increase both paid and unpaid productivity loss costs. mRS was significantly associated with paid productivity loss at all time points.
Several studies have previously examined lost paid productivity in stroke patients. Overall, our findings on absenteeism and overtime are consistent with those of Barral et al. (19), which takes into account changes in hourly wages over the years and time-frame differences in reported costs. However, an advantage of our study, compared to Barral et al (19) who described the costs of mRS 0–2 versus 3–5, is that we were able to present separate estimates for each mRS category from 0 to 2 (e.g., 1 year, €604 (mRS 0), €7215 (mRS 1), €10014 (mRS 2)). As we found significant differences between cost estimates for mRS 0, 1, and 2, our results suggest that using dichotomized mRS scores may lead to less accurate results when using these values in economic evaluations. Due to the limited sample size in these categories, we grouped mRS scores into 3–5, but this may be less problematic than grouping mRS 0–2. Because of the severity of functional impairment, substantial and similar productivity losses are expected across all categories in this group. Barral et al. A friction period of 50 days was applied to FCM, and 80% of productivity loss was considered as actual cost. Using this approach, they estimated the overall average cost of lost productivity due to FCM in the first year after stroke to be €3685 (19), which is similar to our study. Other studies did not report productivity loss as a function of functional outcome or focused only on patients who returned to work, precluding direct comparison with our results.16,17,18).
Considering uncompensated work costs, previous studies have mainly focused on indirect costs or uncompensated work costs of retired patient caregivers. In contrast, our study provides a more comprehensive view of productivity loss by specifically addressing uncompensated productivity loss in patients engaged in professional activities before stroke (18, 19). Our estimates, along with previous findings on unpaid and indirect costs for caregivers and retired patients, highlight that unpaid costs can be significant and their inclusion can have a significant impact on the results of cost-effectiveness analyzes (27).
Although no previous studies have examined the association between EVT and return to work, a recent systematic review and meta-analysis including 39 studies identified male gender, working a desk job, stroke severity, and independence in activities of daily living as positive prognostic factors for return to work, and aphasia as a negative prognostic factor (12). Our study did not identify any significant predictors of return to work. However, these analyzes are exploratory and no firm conclusions can be drawn from the results as the data have not been utilized for this purpose.
Economic evaluations are playing an increasingly important role in guiding the allocation of limited health care resources. Paid and unpaid lost productivity costs account for a significant portion of the total societal cost of stroke, but are often underestimated in economic assessments. These omissions pose a risk of underestimating the economic impact of interventions, especially as the incidence of stroke increases in the working-age population (3, 5, 6). Previous studies have highlighted the economic burden of stroke by assessing medical and societal costs per mRS score (20, 28), but few studies have focused specifically on working-age stroke survivors. Including productivity losses in cost-effectiveness analyzes can provide a more comprehensive understanding of the economic value of treatment and better inform healthcare resource allocation (27). However, this is currently hampered by significant international differences in how these costs are integrated, identified, measured and evaluated (27). Dutch guidelines recommend the use of FCM due to its social perspective (10) rather than HCA, which tends to overestimate productivity losses by adopting an employee-centric perspective (19, 29). Exploring the impact of both methods in studies and using standardized questionnaires such as the validated iPCQ (questionnaire-based) may improve the incorporation of productivity losses in economic assessments across countries.
Strengths and Limitations
An important strength of our study is that we were able to assess the cost of lost productivity by functional outcome, making the results easily applicable and useful for future research. Moreover, using both HCA and FCM methods allows the results to be more generalizable to international differences in productivity loss assessments. Additionally, a questionnaire based on the validated iPCQ was used to assess productivity loss. Including patients from 18 stroke centers in the Netherlands increased the generalizability of the results. Finally, our study included patients at multiple time points, up to 2 years after stroke, allowing us to observe changes in productivity beyond the initial 90 days, when recovery may continue to impact patients’ ability to return to work or resume other productive activities.
Our study also has limitations. First, despite including 280 records out of 249 patients, the distribution across mRS groups was uneven, resulting in small sample sizes in some categories. Average work time and cost estimates based on these small sample sizes may be unreliable and should be interpreted with caution. Additionally, 31 of the 280 respondents participated at two different time points, reducing the variability in the results. Additionally, the use of self-report questionnaires may lead to recall bias. Because we used both clinical trial data and cross-sectional study data, our population may differ somewhat from the general population of stroke patients under 65 years of age. However, using trial data allowed us to include patients across all intervention centers in the Netherlands, which improved the generalizability of our findings. Additionally, the MR CLEAN-LATE trial had broad and pragmatic inclusion criteria that limited inclusion bias. Incorporating cross-sectional data improved generalizability to non-test populations. Because data collection for both studies took place in MUMC+, the questionnaires and procedures used were consistent between the two datasets. Considering the representativeness of the cohort, it is pertinent to note that the available patient characteristics presented in this study (with regard to medical history and stroke characteristics) are consistent with the previously reported characteristics of a young stroke population in the Netherlands (30). Because our study was only conducted in the Netherlands, the results may not be generalizable to other countries. However, if populations and working conditions are considered comparable, our results can be applied to estimate productivity losses due to different cost structures or currencies.
