A decision analysis to identify the ideal treatment for adult spinal deformity: is surgery better than non-surgical treatment in improving health-related quality of life and decreasing the disease burden?

Adult spinal deformity (ASD) is a major public health problem. There are pros and cons of the available treatment alternatives (surgical or non-surgical) and it had been difficult to identify the best treatment modality. To construct a statistical DA model to identify the optimum overall treatment in ASD. From an international multicentre database of ASD patients (968 pts), 535 who had completed 1 year follow-up (371 non-surgical-NS, 164 surgical-S), constitute the population of this study. DA was structured in two main steps of: (1) baseline analysis (assessing the probabilities of outcomes, assessing the values of preference-utilities-, combining information on probability and utility and assigning the quality adjusted life expectancy (QALE) for each treatment) and (2) sensitivity analysis. Four hundred and thirty-two patients (309 NS, 123 S) had baseline and 1 year follow-up ODI measurements. Overall, 104 (24.1 %) were found to be improved (a decrease in ODI > 8 points), 225 (52.1 %) unchanged (-8 > ODI > 8) and 65 deteriorated. Surgery presented with a higher chance of improvement (54.2 %) versus NS (9.7 %). The overall QALE ranged from 56 to 69 (of 100 years) and demonstrated better final QALE in the NS group (60 vs. 65, P = 0.0038), this group having started with higher QALE as well (56 vs. 65 years, P < 0.0001). There were improvements in overall QALE in both groups but this was significant only in the surgical group (S from 56 to 60 years, P < 0.0001; NS from 65 to 65 years, P = 0.27). In addition, in the subgroup of patients with significant baseline disability (ODI > 25) surgery appeared to yield marginally better final QALE (58 vs. 56 years, P = 0.1) despite very a similar baseline (54 vs. 54 years, P = 0.93). This study demonstrated that a single best treatment modality for ASD may not exist. Conservative treatment appears to yield higher (up to 6 %) QALE compared to surgery, most probably secondary to a higher baseline QALE. On the other hand, surgery provides a significantly higher increase in QALE. Especially in patients with significant disability at baseline, the final QALE tended higher in the S group (although not significant). Finally, chances of a relevant improvement at first year turned out to be significantly lower with NS treatment.