Healthcare Costs of New Functional Impairment in the Year after Noncardiac Surgery

Postoperative functional impairment is a patient-centered outcome associated with important sequelae such as poor quality of life yet its impact on healthcare expenditure is poorly understood. Such data could shed light on the potential cost savings of interventions that support patients either before surgery (e.g., prehabilitation) or during their recovery period (e.g., home care, rehabilitation) to enhance functional recovery. Therefore, we sought to characterize the healthcare costs associated with postoperative functional impairment in a contemporary cohort of surgical patients.

We conducted a cost-of-illness analysis of data from the Measurement of Exercise Tolerance before Surgery (METS) study—a prospective multicenter cohort study that evaluated the prognostic validity of preoperative measures of functional capacity.  For this present analysis, we used data from participants in Ontario, Canada (32% of the entire METS cohort) that were linked to administrative healthcare databases (Supplemental Digital Content 1, https://links.lww.com/ALN/D203). In these databases, patient-level healthcare costs are calculated according to validated algorithms.  Thus, we provide a comprehensive summary of healthcare expenditure given that in Ontario nearly 100% of hospital care, community and hospital physician services, and prescription drugs for most adults are publicly funded. The Unity Health Toronto Research Ethics Board approved this secondary analysis (#12-319).

Eligible participants were adults (40 yr or older) who had one or more cardiovascular risk factors and underwent elective inpatient noncardiac surgery (orthopedic, abdomino-pelvic, head-neck, vascular, intrathoracic).  The EQ-5D-3L  was administered on enrollment and at 30 days and 1 yr after surgery. We defined postoperative functional impairment as a one or more level decline at 1 yr (from baseline) in the performance of either mobility, self-care, or usual activities as measured by the EQ-5D-3L.

We used mean values and SDs to summarize 1-yr healthcare costs. We fit quasi-Poisson regression with an identity link to model mean cost differences. We also used the log link to assess mean cost ratios, or the multiplicative cost difference between groups. Point estimates are presented with 95% CIs. To further characterize which areas of healthcare generated these costs, five cost subcategories were predefined: acute care, ambulatory, provider, continuing care, and emergency services (Supplementary Table 1, https://links.lww.com/ALN/D203). For the subcategories of continuing care and emergency costs, we fit two-part models due to excess of zero-cost observations.  Missing functional status was handled using listwise deletion, with an additional sensitivity analysis using multiple imputation by chained equations (10 imputations).  All monetary values were inflation-adjusted to 2016 Canadian dollars. 

The analytic cohort included 443 participants, of whom 100 (22.6%) had new functional impairment at 1 yr. Most healthcare costs were incurred from acute care and ambulatory settings as well as provider billings. Continuing care and emergency costs made up a smaller proportion of the total costs (table 1).

The total 1-yr healthcare costs per participant were $32,901.09 ± $27,116.38 for those with functional impairment and $24,863.98 ± $17,819.45 for those without, corresponding to a difference in mean costs of $8,037.11 (CI $3,380.99 to $12,986.17; table 1). Except for the first 29 days, costs were higher among participants with functional impairment compared to those without functional impairment at all timepoints after surgery (fig. 1). Notably, there were no significant cost differences before surgery (P = 0.65). Participants with functional impairment had 58% higher ambulatory costs (CI, 16 to 113%) and 28% higher provider costs (CI, 11 to 47%) compared to those without functional impairment. Functional impairment was associated with a 73% increase (CI, 18 to 150%) in continuing care costs, compared to those without functional impairment, among participants who required at least one of these services. Participants with functional impairment also had 2.32 higher odds (CI, 1.48 to 3.67) of emergency service utilization.

Postoperative functional status was missing for 39 (8.1%) participants. A sensitivity analysis using multiple imputation yielded qualitatively similar results (Supplementary Table 2, https://links.lww.com/ALN/D203).

In summary, one in five participants experienced new functional impairment in the year after surgery, which was associated with a 32% increase in healthcare costs corresponding to an increase in expenditure of $8,040 per patient. These disparities were not evident before surgery. Notably, cost differences were largely attributed to provider billings, ambulatory costs, and continuing care services. Although emergency-related costs made up a small proportion of total healthcare cost, participants with functional impairment more frequently utilized these services. Overall, our findings provide supporting evidence that interventions for preventing postoperative functional impairment may plausibly be associated with significant cost savings. Considering that more than 900,000 inpatient surgeries happen in Canada each year this may represent incremental healthcare costs of more than $1.45 billion annually across Canada. Future studies should also evaluate patient-level expenditure, for example, income loss, which is not captured in our current analysis.