Page 30 - Risk quantification and modification in older patients with colorectal cancer
P. 30
Chapter 2
Dekker et al.15 the studied intervention was left-sided colorectal surgery and in the studies of Hu et al.22, Battersby et al.31 and Hoshino et al.33 only rectal surgery was studied. The study populations for the 25 studies originated from single centres (13), multicentre studies (5) and registry data or administrative data (7). In the majority of the studies (19 out of the 25) data were collected prospectively. Two prediction model studies used a meta-analysis to select predictors for the final model instead of a primary database.15,27
Patients and outcomes
The number of patients that were included ranged from 119 to 23,5407. Average age ranged from 61 to 76. Two models were exclusively developed for patients of 65 and over.29,34
Of the 26 models, ten models studied mortality as an outcome and seven anastomotic leakages (Table 1). Two models with mortality as an outcome were also developed to predict major complications or major morbidity.13,14 Deep organ space infections, wound disruptions, stroke, renal failure and sepsis were considered major complications and anastomotic leakage, abscess, bleeding or postoperative bowel obstruction as major morbidity in these studies. No models focused on quality of life or postoperative functional dependency.
Predictors
For model development, predictors were mostly selected based on their statistical significance (with p<.10 or p<.05) with a corresponding weight (OR), before constructing the final model.10-14,16-26,28-30,32-34 For three models, the choice of predictors depended exclusively on the research of the literature or clinical experience.15,27,31 The median number of predictors included was 6 (range 4-22). In Table 2A-C, predictors in the different prediction models are depicted, categorised by outcome (mortality, anastomotic leakage and “other outcomes”) which include all other surgical complications including ileus, post-acute care discharge, cardiac events and readmission.
28
