Digital health technologies in improving efficiency in reproductive medicine

Abstract

In this thesis, the potential for digital health technologies to improve efficiency in reproductive medicine is explored. An efficient clinical system is one achieving high levels of performance (outcomes) relative to the inputs (resources, time, money) consumed. The primary goals of reproductive medical care are, where possible, to achieve a healthy live birth and, if this is not possible, to sustain the psychological wellbeing of those with an unfulfilled child wish. More efficient care would achieve these goals with reduced practical and psychological burden on patients, lower costs, and optimal use of resources.
To provide an understanding of the landscape of digital tools in reproductive medicine, this thesis starts with a systematic review and meta-analysis of existing digital support tools. Although digital tools for use alongside fertility treatment have been developed, few are supported by research evidence. The work identified that the small number of digital support tools that have been evaluated in randomised trials overall have a small positive impact on pregnancy rates, but no significant impact on psychological outcomes.
The thesis proceeds to address two aspects of the way in which the burden of treatment can potentially be reduced for patients, firstly through emotional and psychological support by means of an emotional support app (MediEmo) and secondly, through streamlining clinical processes using machine learning and algorithms to develop digital health interventions.
Digital tools could help empower patients through reduction of IVF treatment burden, thereby reducing treatment dropout, the latter being a significant problem in the fertility treatment journey. One way of reducing treatment burden is by reducing the emotional strain of treatment and this thesis explores the use of a novel app MediEmo in doing just that. MediEmo is a smartphone app designed to provide practical and psychological support to patients during IVF treatment. A 3-year descriptive evaluation study demonstrated high engagement and usage of the app, with 80% of eligible patients entering app data. MediEmo was found to be a sensitive tool to examine the emotional experiences of patients during IVF treatment, indicating the emotional intensity of treatment cycles, particularly the two-week wait prior to pregnancy test. In addition, MediEmo data revealed that elevated levels of negative emotions are also experienced during both intrauterine insemination and frozen embryo transfer cycles and that more support may be required for patients undertaking these treatments. Furthermore, the observational study of rates of return for more IVF treatment within 1 year of a failed IVF cycle found that active app usage was associated with a higher rate of return for further treatment within one year of cycle failure compared to non-users of the app. A higher rate of return could, in theory, improve the chances of pregnancy via cumulative success. Further work planned to test this hypothesis is described.
Development of new algorithms has led to considerable research and clinical interest in using machine learning to develop smart digital health interventions. The latter part of this thesis explores applications of these techniques aiming to develop digital tools for use in clinical reproductive medicine. A retrospective study using machine learning aimed to optimise timing of the trigger injection during an IVF stimulation cycle. After this work proved unsuccessful, similar techniques were used to address the need for frequent transvaginal ultrasound monitoring of ovarian follicular growth. This monitoring is an onerous and invasive aspect of fertility treatments, that is traditionally performed, but the frequency of this intervention lacks evidence to support efficiency. Models were developed to predict key variables in an IVF cycle using data from scans on single days. The results identified that ultrasound scans are most useful between day 8 and 10 of an IVF cycle, when accurate predictions of oocyte maturation trigger day, and ovarian hyperstimulation syndrome (OHSS) risk, can be made. With further work, it may be possible to omit the less useful earlier scans, particularly in circumstances such as the Covid pandemic and the resulting need for social distancing. The development of algorithms to improve mid-cycle communication with patients is also summarised.
Digital health technologies will play an increasing role in reproductive medicine. This thesis demonstrates this emerging contribution via two specific aspects aiming to reduce patient and clinical treatment burden: 1) emotional support by way of a digital app and 2) utilising advanced analytics in streamlining clinical processes. However, any new digital tools in fertility care must be robustly evaluated to show they truly improve efficiency. The tools should then be tested in the ‘real-world’ and the thesis concludes by considering how this field of work could develop in the future.

More information

Identifiers

Catalogue record

Export record