Page 36 - Prevention and Treatment of Incisional Hernia- New Techniques and Materials

Chapter 2
- length of hospital stay
- length of stay in ICU
- reinterventions
Data on extramural care, work absence and other patient costs will be gathered via questionnaires at each follow-up (1 and 12 months).
For a description of the calculation of the e ect measures see paragraph ‘outcome parameters’. Discounting of future costs and e ects is not relevant because of the limited time horizon of 1 year. When costs of a treatment are similar across subgroups, the absolute bene t determines the cost- e ectiveness of a treatment for a speci c subgroup. Randomized controlled trials are designed to evaluate the e ects of treatment at the group level, and cost-e ectiveness is usually calculated for this group as a whole. There could however be substantial and relevant between subgroup variability. It is therefore common to consider subgroup speci c e ects of interventions. The subgroup speci c cost-e ectiveness will be estimated by rst deriving a prognostic index, based on the prede ned predictors of incisional hernia: abdominal aneurysm aorta (AAA), obesity, diabetes, COPD, corticosteroid usage, radiotherapy, cardiovascular disease, smoking, age, cancer, other abdominal wall hernias and collagen disorders.
Sample size calculation
Millbourn et al. found a decrease in the incidence of incisional hernia from 18% to 5,6% in a randomized controlled trial. [34] In this trial, follow-up consisted of clinical instead of radiological examination for incisional hernia occurrence. In this trial, ultrasound examination will be used in order to be able to diagnose incisional hernia with higher sensitivity. It is expected that a relative decrease of the incidence incisional hernia after one year of 50% is reasonable. The mean reported one year incidence of incisional hernia in literature is 15%(1- 5). In order to reduce the mean incidence of incisional hernia from 15 to 7.5%, power calculations showed that two groups of 259 evaluable patients each are needed (power=0.80, alfa=0.05). Loss to follow-up is estimated at 10% of included patients. A total of 576 patients (2 x 288) will be included in the study to correct for loss to follow-up. Overall e ects will be calculated adjusted for predictive baseline characteristics, which will lead to a higher statistical power.
34