“Optimal patient selection” is a very common argument to support the continued use of low-value surgeries such as arthscopic partial meniscectomy or re-insertion of degenerative cuff tears in elderly patients. While on average these surgeries offer no benefit to our patients, the proponents say that it is just about patient selection: you need to choose the right patient who will benefit. This is a myth which nobody has been able to prove it works.
Two recent studies had very interesting results which further give insight to this topic.
Teemu Karjalainen et al. asked “Can Surgeons or Patients Predict the Likelihood of Improvement With Nonoperative Treatment of Chronic Tennis Elbow?” They concluded that
Surgeons’ predictions about the recovery were not associated with any of the measured patient characteristics, indicating that the predictions were based on heuristics, that is, mental shortcuts or rules of thumb that clinicians commonly use in clinical decision-making.
…surgeons are unable to reliably predict who will or will not improve with nonoperative treatment
This study very efficiently debunks the claim that “optimally selected would benefit from operative treatment in chronic tennis elbow”.
Another very interesting study was published in early January this year in JAMA Network Open. It was a secondary analysis of a previous trial comparing lumbar decompression surgery with or without spinal fusion. In the original trial, patients were randomized to receive either decompression alone or decompression with fusion, and surgeons were also asked whether they preferred the treatment assigned to the patient. Results were as follows:
At 2-year follow-up, 87 of 116 patients (75%) who received surgery in agreement with the surgeons’ recommendations and 77 of 106 (73%) who received surgery in disagreement with the surgeons’ recommendations reached the primary outcome (difference, 2.4 percentage points; 95% CI, −9.1 to 13.9 percentage points).
In essence, this means that whether or not the surgeon would have chosen a different procedure than what randomization assigned had no effect on the outcome. For example, if a patient received decompression only, even though the surgeon would have preferred fusion, the results were still the same, and vice versa. This suggests that patient selection by the surgeon would not have improved outcomes in this patient group.
These studies offer more cumulative evidence against the myth of “optimal patient selection”. Sure, someone will argue that “yes, those surgeons were not competent enough, and experienced surgeons will succeed”. Still, no study has managed to show that there are surgeons who can pick the winner.
