Casper Tax1, Paulien H M Govaert2, Martijn W J Stommel3, Marc G H Besselink4, Hein G Gooszen1,5, Maroeska M Rovers1,5. 1. Department of Operating Rooms, Radboud University Medical Centre, Radboudumc Institute for Health Sciences, Nijmegen, The Netherlands. 2. Radboud University Student Biomedical Sciences, Nijmegen, The Netherlands. 3. Department of Surgery, Radboud University Medical Centre, Radboudumc Institute for Health Sciences, Nijmegen, The Netherlands. 4. Department of Surgery, Academic Medical Center, Amsterdam, The Netherlands. 5. Department of Health Evidence, Radboud University Medical Centre, Radboudumc Institute for Health Sciences, Nijmegen, The Netherlands.
Abstract
OBJECTIVE: To illustrate how decision modeling may identify relevant uncertainty and can preclude or identify areas of future research in surgery. SUMMARY BACKGROUND DATA: To optimize use of research resources, a tool is needed that assists in identifying relevant uncertainties and the added value of reducing these uncertainties. METHODS: The clinical pathway for laparoscopic distal pancreatectomy (LDP) versus open (ODP) for nonmalignant lesions was modeled in a decision tree. Cost-effectiveness based on complications, hospital stay, costs, quality of life, and survival was analyzed. The effect of existing uncertainty on the cost-effectiveness was addressed, as well as the expected value of eliminating uncertainties. RESULTS: Based on 29 nonrandomized studies (3.701 patients) the model shows that LDP is more cost-effective compared with ODP. Scenarios in which LDP does not outperform ODP for cost-effectiveness seem unrealistic, e.g., a 30-day mortality rate of 1.79 times higher after LDP as compared with ODP, conversion in 62.2%, surgically repair of incisional hernias in 21% after LDP, or an average 2.3 days longer hospital stay after LDP than after ODP. Taking all uncertainty into account, LDP remained more cost-effective. Minimizing these uncertainties did not change the outcome. CONCLUSIONS: The results show how decision analytical modeling can help to identify relevant uncertainty and guide decisions for future research in surgery. Based on the current available evidence, a randomized clinical trial on complications, hospital stay, costs, quality of life, and survival is highly unlikely to change the conclusion that LDP is more cost-effective than ODP.
OBJECTIVE: To illustrate how decision modeling may identify relevant uncertainty and can preclude or identify areas of future research in surgery. SUMMARY BACKGROUND DATA: To optimize use of research resources, a tool is needed that assists in identifying relevant uncertainties and the added value of reducing these uncertainties. METHODS: The clinical pathway for laparoscopic distal pancreatectomy (LDP) versus open (ODP) for nonmalignant lesions was modeled in a decision tree. Cost-effectiveness based on complications, hospital stay, costs, quality of life, and survival was analyzed. The effect of existing uncertainty on the cost-effectiveness was addressed, as well as the expected value of eliminating uncertainties. RESULTS: Based on 29 nonrandomized studies (3.701 patients) the model shows that LDP is more cost-effective compared with ODP. Scenarios in which LDP does not outperform ODP for cost-effectiveness seem unrealistic, e.g., a 30-day mortality rate of 1.79 times higher after LDP as compared with ODP, conversion in 62.2%, surgically repair of incisional hernias in 21% after LDP, or an average 2.3 days longer hospital stay after LDP than after ODP. Taking all uncertainty into account, LDP remained more cost-effective. Minimizing these uncertainties did not change the outcome. CONCLUSIONS: The results show how decision analytical modeling can help to identify relevant uncertainty and guide decisions for future research in surgery. Based on the current available evidence, a randomized clinical trial on complications, hospital stay, costs, quality of life, and survival is highly unlikely to change the conclusion that LDP is more cost-effective than ODP.
Authors: J van Hilst; E A Strating; T de Rooij; F Daams; S Festen; B Groot Koerkamp; J M Klaase; M Luyer; M G Dijkgraaf; M G Besselink Journal: Br J Surg Date: 2019-04-23 Impact factor: 6.939