Costs and quality of life of small-incision open cholecystectomy and laparoscopic cholecystectomy - an expertise-based randomised controlled trial

Eligible patients were asked to enrol in the trial at the time when cholecystectomy was decided. Written informed consent was obtained before randomisation. The trial was not blinded.

Patients aged 18 years or older with biliary disease where cholecystectomy was considered the best treatment according to published recommendations [15, 16] were asked to participate in the trial. Both elective and emergency patients were enrolled. Patients with acute cholecystitis, choledocholithiasis, jaundice, pancreatitis, obesity and co-morbidity were included if cholecystectomy was considered the best treatment. The exclusion criteria were: inability to understand given information, surgeons from either the LC team or the SIOC team unavailable for emergency cholecystectomy, cholecystectomy performed for malignant disease or suspicion of malignancy, and cholecystectomy performed as part of another operation.

Randomisation was made using an internet-based system with computer-generated random numbers. Patients were automatically stratified at three levels: hospital, age (< or > = 70 years), and degree of priority (elective, emergent, or surgery for life-threatening disease). The patients were randomised when the decision for surgery was made.

The economic analysis was conducted from a societal perspective, calculating both direct public healthcare costs and indirect costs generated by the loss of productivity (sick leave). Cost generating posts in health care were calculated from the time the patient was admitted for surgery. Preoperative investigations (preoperative radiology, preoperative visits to outpatient clinic, etc.) were not included. Costs for pre-, intra-, and postoperative endoscopic retrograde cholangiogram (ERC), sphincterotomy and endoscopic stone extraction were regarded as a part of the operative procedure and were thus included. Detailed calculations of cost items were performed in one hospital (Umeå University Hospital). All costs are given as USD. Where reliable item costs could be found, they were retrieved from the Hospital Department of Economics (costs for perioperative and secondary cholangiogram, in hospital stay, postoperative recovery unit). Other costs were calculated manually, (disposables, reusable instruments, including laparoscopic equipment). Costs for LC were calculated with the presumption of reusable trocars. Personnel costs per minute were calculated using mean wages for each category of personnel obtained from the Department of Economics, Umeå University Hospital. Costs for social benefits and employer fees, at a rate of 43.71% (mean for hospital staff), were added to the mean wages. Operational costs were calculated with one resident and one senior surgeon participating from incision to last stich for each operation. A standard cost for counselling and administration (hospital record, sick leave certificate and prescription of drugs) were added. The cost for surgical theatre staff was calculated from the moment patients arrived at the theatre until they left for the recovery unit, individually for each patient. All patients were estimated to spend 4 h in the postoperative recovery unit. Patients undergoing ambulatory surgery were estimated to spend 8 h in the surgical ward. An average cost for ERC, sphincterotomy and endoscopic stone extraction of 1216 USD, was estimated by the Hospital Department of Economics. Indirect costs due to loss of production were calculated based on mean wages obtained from Statistics Sweden (SCB) for men and women, respectively, including social benefits and employer fees at a rate of 41.15% (mean for all employed). Indirect costs were calculated only for patients that reported taking a sick leave from work due to the operation. Cost-generating posts are listed in Table 1.

Health-related quality of life was assessed using the EQ 5-D-3L (EuroQol Group, Rotterdam, The Netherlands [11]) which consists of five questions concerning patient mobility, self-care, usual activities, pain/discomfort and anxiety/depression. In this study, a value set derived from the Swedish population was used [17]. The EQ-5D-3L index values were further calculated to AUC and quality adjusted life years (QALYs), please see Sample size, analysis and statistics.

The operating surgeon registered the operative data (except for operative time) in the Swedish registry for gallstone surgery, GallRiks [13]. An independent assessor, as defined by GallRiks, registered operative time, length of hospital stay and complications within 30 days. Complications were graded according to Dindo-Clavien [18]. Health-related QoL was recorded by the patient before surgery and on postoperative days 3, 7, 11, and 30, as well as 1 year after surgery. The patient reported length of sick leave at 30 days. Data from GallRiks were controlled against patient hospital records for all patients. Where registration errors were detected, they were corrected.

The power calculation to determine the sample size for the trial was based on earlier publications on cholecystectomies where the main differentiating factor between the two interventions was duration of the operation. It was assumed that SIOC would take 16 min less compared to LC, based on three previous studies [19‐21]. The calculation was made for 90% power, 5% significance level and an anticipated data loss of 25%. With these assumptions, it was calculated that the study should consist of at least 350 patients. The sample size was estimated to be sufficient for detecting significant differences of routine costs, but not for comparing relatively rare complications between groups. All analyses were made with the intention-to-treat principle. Quantitative results were presented as median values, 25 and 75^th percentiles whenever the distribution of the data was skewed. A non-parametric Mann-Whitney U test was used for statistical tests of significance. AUC (Area Under Curve) values up to 1 year were calculated using the EQ-5D-3L scores at five different postoperative time points and the assumption that the change between time points was linear. Missing EQ-5D-3L values were imputated if the case had at least two registered genuine EQ-5D-3L values at different time points, otherwise the case was excluded from AUC calculation. We used the principle to add (or subtract) the mean change between specific time points for the group (LC or SIOC) to the last genuine value to generate the value for the next time point if it was missing. In all, 128 imputations for missing values were added to the 1372 genuine values. Thirty-three cases were excluded because of more than three missing EQ-5D-3L scores. The EQ-5D-3L scores and AUC values were calculated from raw data using Microsoft® Excel for Mac version 14.2.3 (Microsoft Corporation, Redmond, Washington, USA), and Stata® software release 13.1 (StataCorp LP, College Station, Texas, USA) was used for statistical calculations.

Age, sex, preoperative EQ-5D-3L score, proportion of smokers, proportion employed and education level were similar between the groups (Table 2). The SIOC-group had a longer time between decision for surgery and operation [10], 156 versus 103 days, respectively (P = 0.03), slightly more emergency admissions, and more cases with acute cholecystitis than the LC group, 25.6 and 14.7%, respectively (p = 0.012).

The median value for direct cost was (p25-p75) 4210 USD (3851–5073) for the LC group and 3963 USD (3512–4848) for the SIOC group, P = 0.002. Operation costs and total costs, including indirect costs, were significantly higher for the LC group, p = <0.001 and p = 0.003 respectively (Table 3).

EQ-5D-3L-AUC within 1 year was, median, (p25-p75) 349 (337–351) for the LC group and 349 (338–350) for the SIOC group, P = 0.8096. QALYs had a median of 0.9639 (0.9313–0.9683) and 0.9636 (0.9343–0.9679) in the LC group and SIOC group, respectively. EQ-5D-3L-AUC at 30 days was, median (p25-p75) 25.17 (23.70–25.85) for the LC group and 25.10 (23.67–25.76) for the SIOC group. Index values for EQ-5D-3L at all time points are shown in Fig. 2. Subgroup analyses for costs and QoL of patients that had emergency cholecystectomy, elective cholecystectomy, complications and no complications are presented in Additional file 1: Table S1.