Govert Veldhuijzen

119 CBE before colonoscopy is reducing operational costs, with lower patient and societal expense group, 21.5% required additional information, 18.5% could be counselled via telephone alone. But the remaining 3.0% of patients reported an extra outpatient clinic visit. From the endoscopy unit cost perspective, we defined three patient’s education routes: nurse counselling, CBE alone and CBE with additional action (telephone or visit). The additional action group is combined as the key additional cost driver (nurse wage) is comparable. To compare both groups (217 in nurse counselling versus 280 patients in the CBE group) we had to resolve the missing data in the CBE group. We decided to impute missing data on the patient route after CBE, as we assumed the missing values occurred at random without selective dropout, precluding selection bias. The proportion of patients in each route was used to randomly assign the missing 145 patients to each of the routes. For more detail, we performed this imputation stratified on trial site level. This resulted in complete data on 280 patients in the CBE group. The second study described a cost model which was developed to evaluate the costs effected by CBE implementation. 21 We included all relevant components of the work process to afford better insight in costs and savings for hospital management. Development of standardized cost model For the development of the cost model, we use a mixed-methods approach. We reviewed the business case of CBE provided by the supplier of the CBE software. We mapped the patient flow in several endoscopy units where the CBE was already implemented. We identified and validated key components of the cost model through literature research and qualitative interviews with financial experts (e.g. hospital department managers) in the field. The following phase consisted of the construction the final model. An overview of this process is depicted in figure 1., reprinted with permission. 21 6