Table 2. Rates, Unadjusted Risk Differences, and Multilevel Logistic Regression Model of Odds of Conversion to Open Laparotomya.
| Effectb | No. With Conversion to Open Laparotomy/Total No. (%) | Unadjusted Risk Difference (95% CI), %c | Adjusted Odds Ratio (95% CI)d | P Value |
|---|---|---|---|---|
| Treatment | ||||
| Conventional laparoscopic surgery | 28/230 (12.2) | 4.1 (−1.4 to 9.6) | 1 [Reference] | .16 |
| Robotic-assisted laparoscopic surgery | 19/236 (8.1) | 0.61 (0.31-1.21) | ||
| Sex | ||||
| Male | 39/317 (12.3) | 6.9 (1.8 to 12.1) | 1 [Reference] | .04 |
| Female | 8/149 (5.4) | 2.44 (1.05-5.71) | ||
| BMI class, overweight vs underweight or normale | ||||
| Underweight or normal | 13/179 (7.3) | 2.3 (−2.7 to 7.2) | 1 [Reference] | .19 |
| Overweight | 9/180 (5.0) | 0.54 (0.21-1.37) | ||
| BMI class, obese vs underweight or normale | ||||
| Underweight or normal | 13/179 (7.3) | −16.1 (−25.0 to −7.2) | 1 [Reference] | <.001 |
| Obese | 25/107 (23.4) | 4.69 (2.08-10.58) | ||
| Previous radiotherapy or chemoradiotherapy | ||||
| No | 27/262 (10.3) | 0.5 (−5.0 to 6.0) | 1 [Reference] | .86 |
| Yes | 20/204 (9.8) | 1.07 (0.50-2.26) | ||
| Intended procedure, high vs low anterior resectionf | ||||
| Low anterior resection | 37/312 (11.9) | 3.0 (−4.6 to 10.7) | 1 [Reference] | .26 |
| High anterior resection | 6/68 (8.8) | 0.55 (0.19-1.56) | ||
| Intended procedure, abdominoperineal resection vs low anterior resectionf | ||||
| Low anterior resection | 37/312 (11.9) | 7.2 (1.5 to 12.9) | 1 [Reference] | .007 |
| Abdominoperineal resection | 4/86 (4.7) | 0.18 (0.05-0.63) |
Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared).
The intracluster correlation coefficient (ICC) for operating surgeon as a random effect was 0.05 (95% CI, 0.01-0.06). The ICC is a measure of the proportion of variance in the outcome that is explained by the operating surgeon (eg, an ICC of 0 would indicate that the odds of conversion for a given patient would not be affected at all if he or she underwent surgery by a different operating surgeon). The ICCs for a range of outcomes across a number of surgical trials are reported in the ICC database.
The variables in this column were included in the model as fixed effects, and operating surgeon was included in the model as a random effect; these are all of the variables included in the model.
Risk differences are unadjusted estimates.
Odds ratios are adjusted estimates yielded by the model.
Derived from World Health Organization classification of obesity based on BMI, with a BMI of 25.0 to 29.9 indicating overweight and a BMI of 30.0 or greater indicating obese.
Intended procedure, rather than actual procedure, is included in the model. The intended procedure was collected at randomization and used for stratification. A sensitivity analysis adjusting for actual procedure instead of intended procedure showed no notable changes to the effect estimates, with the exception of the odds ratio comparing abdominoperineal resection vs low anterior resection, which was less pronounced (adjusted odds ratio = 0.43 [95% CI, 0.16-1.13]; P = .09).