Cost-effectiveness of time-lapse monitoring with or without the use of embryo selection software compared to routine incubation and selection

Abstract

STUDY QUESTION: Is time-lapse monitoring (TLM) with or without a machine learning selection algorithm for embryo selection cost-effective when compared with conventional culture and selection? SUMMARY ANSWER: Neither TLM with or without the use of machine learning for embryo selection is likely to be cost-effective when compared with the conventional approach. TLM adds cost to IVF treatments without improving clinical results. WHAT IS KNOWN ALREADY: Recent randomized controlled trials (RCTs) provide evidence that TLM is unlikely to increase pregnancy or live birth rates (LBRs) relative to conventional embryo culture and selection. To date no cost-effectiveness analysis of TLM has been published. STUDY DESIGN, SIZE, DURATION: A cost-effectiveness analysis (CEA) from a healthcare perspective was performed based on data from our multicentre RCT on TLM, which was conducted from 2017 to 2020. Women scheduled for Day 3 single embryo transfer during their first, second, or third oocyte retrieval were included. The trial included three strategies: (i) TLE: embryo selection based on a machine learning algorithm and uninterrupted culture; (ii) TLR: routine morphological embryo selection and uninterrupted culture; and (iii) CON (Control): routine morphological embryo selection and interrupted culture. PARTICIPANTS/MATERIALS, SETTING, METHODS: In total, 1731 couples were included. The difference in mean costs between treatment groups was calculated with the following cost items: laboratory procedures, laboratory facilities (incubator, disposables, service), embryo transfers, pregnancies, miscarriages, and deliveries. We did not consider costs of stimulation, oocyte retrieval, and IVF/ICSI procedures as these were identical in each group. The effectiveness measure for the CEA was the cumulative LBR. Incremental cost-effectiveness ratios were calculated for direct comparisons between strategies. MAIN RESULTS AND THE ROLE OF CHANCE: The 12-month cumulative LBR rate was similar in the three treatment groups: TLE 48.7%, TLR 48.4%, CON 48.2%. The mean difference in costs was €237 (95% CI: −35 to 508) between TLE and CON, and €55 (95% CI: −203 to 313) between TLR and CON. The incremental cost-effectiveness ratio between TLE and CON for one additional live born was €45 950; for TLR versus CON, it was €26 905. The probability that TLE or TLR are cost-effective was low for all willingness-to-pay level, with several scenarios resulting in comparable findings. LIMITATIONS, REASONS FOR CAUTION: This study reports the cost-effectiveness for one time-lapse incubator; however, more systems are currently available. WIDER IMPLICATIONS OF THE FINDINGS: While our cost-effectiveness analysis provides evidence that TLM is unlikely to be cost-effective, its use is not inferior to conventional methods in terms of clinical results and can offer certain advantages for IVF laboratories. STUDY FUNDING/COMPETING INTEREST(S): The authors received a grant from the Netherlands Organization for Health Research and Development (ZonMw) for the execution of the SelecTIMO study (Health Care Efficiency Research programme grant 843001602). Merck (Germany and the Netherlands) supplied the six time-lapse incubators, funded the laboratory adjustments, and provided technical support and training to laboratory personnel before and during the study. The following declarations of competing interests are outside the submitted work: J.M.J.S. reports research grants from Ferring BV and Merck BV, speakers’ fees from Merck BV and travel support for attending meetings from Merck BV and Goodlife BV. C.B.L. reports speaker’s honoraria from Merck and Organon (The Netherlands), travel support from Merck and Organon, and was Editor-in-Chief of Human Reproduction. M.v.W. is Editor-in-Chief of Human Reproduction Update. The remaining authors have no conflicts of interest to declare. TRIAL REGISTRATION NUMBER: NTR5423: ICTRP Search Portal (who.int).