Artificial Intelligence for Planning Assisted Conception Protocols

  • Research type

    Research Study

  • Full title

    The Application and Validation of Artificial Intelligence to Automate the Planning of Assisted Conception Protocols

  • IRAS ID

    360571

  • Contact name

    Rima Dhillon-Smith

  • Contact email

    Rima.dhillon@doctors.org.uk

  • Sponsor organisation

    Care Fertility, Manchester

  • Clinicaltrials.gov Identifier

    N/A, N/A

  • Duration of Study in the UK

    0 years, 7 months, 24 days

  • Research summary

    This study looks at whether computer-based methods, known as machine learning, can accurately predict how many eggs are collected during IVF treatment. The number of eggs retrieved is an important part of IVF and can influence how treatment is planned and discussed with patients.

    The research uses information from over 40,000 IVF treatment cycles carried out at CARE Fertility centres in the UK between 2012 and the present. This data already exist in clinic records and include routine details such as age, hormone test results, treatment medication, and the number of eggs collected. All data used in the study will be fully anonymised, meaning no individual patient can be identified.

    Previous studies have explored similar prediction tools, but many were based on small numbers of patients or used methods that may produce unreliable results. This study aims to address these limitations by using a large dataset and well-established research standards.

    Three different machine learning approaches will be tested to see how accurately they can predict the number of eggs retrieved: a statistical regression model, a random forest model, and an artificial neural network. The accuracy of these models will be assessed by comparing predicted egg numbers with the actual numbers collected, using clear measures of prediction error.

    This is a retrospective study, meaning it does not involve any new treatments, tests, or contact with patients. No consent is required because only anonymised historical data are used. The findings will help improve understanding of how reliable these prediction tools are and may inform future research into personalised IVF treatment planning.

  • REC name

    South Central - Oxford B Research Ethics Committee

  • REC reference

    26/SC/0149

  • Date of REC Opinion

    5 May 2026

  • REC opinion

    Favourable Opinion