The use of patient-reported preoperative activity levels as a stratification tool for short-term and long-term outcomes in patients with adult spinal deformity.

OBJECTIVE Given the recent shift in health care toward quality reporting requirements and a greater emphasis on a cost-quality approach, patient stratification with respect to long-term outcomes and the use of health care resources is of increasing value. Stratification tools may be effective if they are simple and evidence based. The authors hypothesize that preoperative patient-reported activity levels might independently predict postoperative outcomes in patients with adult spinal deformity. METHODS This is a retrospective cohort. A total of 575 patients in a prospective adult spinal deformity surgical database were identified with complete data regarding the preoperative level of activity. Answers to question 5 of the Scoliosis Research Society-22r Patient Questionnaire (SRS-22r) were used to stratify patients into active and inactive groups. Outcomes were length of hospital stay (LOS), level of activity, and reaching the minimum clinically important difference (MCID) for SRS-22r domains and the Physical Component Summary (PCS) of the SF-36 at 2 years postoperatively. The 2 groups were compared with respect to several potential confounders. Covariates with p < 0.1 were controlled for. The impact of activity on LOS was assessed using multivariate negative binomial regression analysis. Multivariate logistic regression models additionally controlling for the respective baseline health-related quality of life (HRQOL) scores were used to assess the association between preoperative activity levels and reaching the MCID at 2 years postoperatively. RESULTS A total of 420 (73%) of the 575 patients who met the inclusion criteria had complete data at 2 years postoperatively. The inactive group was more likely to be significantly older, have a higher Charlson Comorbidity Index, worse baseline radiographic deformity, and greater correction of most radiographic parameters. After controlling for possible confounders, the active group had a significantly shorter LOS (incidence risk ratio 0.91, p = 0.043). After adding respective baseline HRQOL scores to the models, active patients were significantly more likely to reach the MCID for the SRS-22r pain domain (OR 1.72, p = 0.026) and PCS (OR 1.94, p = 0.013). Active patients were also significantly more likely to be active at 2 years postoperatively on multivariate analysis (OR 8.94, p < 0.001). CONCLUSIONS The authors' results show that patients who belong to the inactive group are likely to have a longer LOS and lower odds of reaching the MCID in HRQOL or being active at 2 years postoperatively. Inquiring about patients' preoperative activity levels might be a reliable and simple stratification tool in terms of long- and short-term outcomes in ASD patients.