Assessment of spine surgery quality is difficult as the most important outcomes are subjective and typically measured with patient reported outcomes (PROs). Specialties like cardiothoracic surgery and oncology can easily measure mortality, a meaningful outcome in those disciplines. Outcomes such as infection and readmission rates can be measured in spine surgery, but those outcomes are generally less important than PROs and tend to affect only short-term results. Measuring PROs has become standard for spine investigations, though very few studies have compared PROs across hospitals and surgeons. In order to better determine the effect of hospital system and individual surgeon on PROs, Sara Khor and colleagues from Washington analyzed a statewide database that evaluated baseline and 1-year Oswestry Disability Index (ODI) scores following lumbar fusion from 2012-2016. Of over 1,000 lumbar fusion patients with baseline ODI scores, 737 had one-year ODI scores and formed the cohort in this study. Seventeen hospitals and 58 surgeons contributed patients to the database, and 13 hospitals and 16 surgeons had at least 10 cases and were included in the analysis. They performed two analyses in which success was defined as a 15-point improvement on the ODI (minimal important difference or MID) or a one year ODI score of less than 22 (minimal disability). Overall, 59% of patients reached MID and 43% had minimal disability at 1 year. The proportion reaching MID ranged from 44-79% across hospitals and 33-84% across surgeons. A similarly broad distribution was observed for those reaching minimal disability, ranging from 29-53% across hospitals and 8.3-70% across surgeons. Interestingly, controlling for patient factors and then performing reliability adjustment eliminated any significant differences across hospitals and surgeons, indicating that patient factors, rather than hospital and surgeon factors, were driving outcomes. A secondary analysis using a previously developed outcomes prediction algorithm demonstrated that, at baseline, 64% of patients had a greater than 50% chance of reaching MID, while 36% had a less than 50% chance of reaching MID. The observed to expected ratio of those reaching MID tended to be around 1 for hospitals and surgeons in the high likelihood of success group, while the results were much more variable in the low likelihood of success group.
This is an interesting study that included a large number of patients treated across 13 hospitals and by 16 surgeons and demonstrated that patient factors tended to drive outcomes more than hospital or surgeon factors. Some of the differences among hospitals and surgeons may have been obscured by the relatively low number of patients treated at some of the hospitals and by some lower volume surgeons. While there were substantial differences in unadjusted outcomes, once patient factors were controlled for, most of these differences vanished. This suggests that patient selection rather than hospital quality or surgeon skill was driving outcomes. Given prior studies on outcome predictors, this comes as no surprise given the major role patient characteristics and comorbidities play in predicting PROs. One interesting finding is that there was much more inter-hospital and inter-surgeon variation in outcomes among the low likelihood of success patients, indicating that hospital or surgeon factors may be more important in the more challenging patients. The authors indicated that limiting surgery to those in whom success is likely (i.e. > 50%) would markedly improve outcomes, however, this would result in 36% of patients being denied surgery. The overall proportion reaching MID, 64%, seems relatively low, especially since the majority of patients were diagnosed with stenosis, spondylolisthesis, or radiculopathy, all of which are associated with fairly good outcomes. The relatively low level of "success" may in part be attributed to a high bar to reach MID, a 15-point improvement on the ODI. Other studies have suggested that a 10-point improvement on the ODI is a clinically important difference, and using that standard would have markedly changed the results. While the authors do not go so far as to suggest limiting care to only those with a high likelihood of success, some readers could make that inappropriate logical leap. The United States' healthcare system does not currently ration care, and applying any such standard would result in the denial of care to some patients who would benefit from surgery.
Please read Ms. Khor's article in the April 1 issue. Does this change your view on the role hospitals and surgeons play in affecting PROs following lumbar fusion?
Adam Pearson, MD, MS
Associate Web Editor