INTRODUCTION: Achieving a “successful outcome” may be the most understandable and clinically relevant outcome for a clinical trial but this approach is limited by a lack of consensus on the “right” definition of success.
METHODS: We reanalyzed the one‐year results of the Spine Patient Outcomes Research Trial randomized cohort comparing surgery and non‐operative treatment for lumbar disc herniation using different criteria for composite definitions of success: 1) > 15 points improvement on ODI and > 10 points improvement on SF‐36 BP; 2) global self‐rated “major improvement” and “satisfied” with symptoms; 3) final leg pain bothersomeness score of 0 (no leg pain); 4) #1 and #2 and #3 and no opioids; 5) final leg pain bothersomeness of 0, final SF‐36 BP score > 80, and final ODI score < 10. The surgery group included all patients having surgery within 3 months of enrollment. In all analyses, repeat spine surgery was considered a failure in the surgery group and cross‐over to surgery after 3 months was considered a failure in the non‐operative group.
The Table summarizes success rates and the number needed to treat (NNT) to obtain one additional “success” for each definition.
CONCLUSIONS: The proportion of patients achieving a successful outcome was highly dependent on the definition of success and even greater variation might be expected if additional criteria such as return to work were incorporated into the definitions. The NNT to achieve one additional success between treatment groups was less sensitive to the definition of success.