We thank Dr. Dexter for his comments on our article [1] about interpreting odds ratios, and we agree with him about the value of relative risk information. As pointed by Dr. Dexter, odds ratios can be an overestimator of relative risk, even though levels of statistical significance remain the same [2]. Dr. Dexter is correct in that our paper provided odds ratios, not relative risk, and therefore our paper’s discussion section contained overestimates of relative risk. Using the conversion formula provided by Zhang and Yu [2] and as suggested in Dr. Dexter’s letter, we calculated the adjusted relative risk. When compared cases with consistent surgical teams, the relative risk of cases by inconsistent teams for prolonged operative time, prolonged hospital stay, and 30-day hospital readmission was 1.37 (95 % confidence interval [CI] 1.15–1.61), 1.30 (95 % CI 1.15–1.46), and 1.36 (95 % CI 1.06–1.74), respectively.

Consequently, we wish to correct the errors in the discussion section in our paper and restate relevant sentences as below:

  • When compared with consistent teams, inconsistent teams were associated with a 36 % higher risk of hospital readmission within 30 days of discharge after adjustment for key patient characteristics, surgeon, and time of day.

  • Inconsistent teams were also associated with a 30 % increase in occurrences of prolonged hospital stay after controlling patient characteristics, surgeon, and time of day.

  • There was a 37 % increase in occurrences of prolonged operative time, suggesting higher likelihood of variant events and other inefficiencies when cases were performed by inconsistent teams.

The percentage of ad hoc teams was 18.1 %, which may be calculated from the information in Table 2. The relative risk of prolonged operative times for cases with ad hoc teams compared to consistent teams was 1.78 (95 % CI 1.17–2.48).