A new model-based study published by the Aquarius team in BMJ Open suggests that AI-assisted digital cytology could substantially reduce the laboratory staff time needed to review and report cervical screening slides.
Cervical screening plays an important role in preventing cervical cancer by identifying cell changes before they progress. In the NHS Cervical Screening Programmes, samples are first tested for high-risk HPV. Samples with a positive result are then examined through cytology to identify cellular abnormalities that may require further investigation.
This part of the screening pathway depends on a highly skilled laboratory workforce. However, cytology services are experiencing continuing staffing pressures. In 2024, 14% of relevant cytology posts across England, Scotland and Wales were reported as vacant, and workforce shortages are expected to increase as experienced staff retire.
What the study aimed to assess
Researchers from Aquarius Population Health and experts from NHS laboratories examined whether AI-assisted digital cytology could help laboratories use their available workforce more efficiently.
The study evaluated the GeniusTM Digital Diagnostics System, which creates digital images of cervical cytology slides and uses Artificial Intelligence (AI) to identify areas of interest for laboratory professionals to examine. The system does not make the final diagnosis or replace trained staff. Interpretation and reporting remain the responsibility of qualified laboratory professionals.
What we did
The research team developed a decision-tree model based on the NHS cervical screening laboratory workflow in England. The model compared the staff time required to process 479,125 cytology slides (number seen in England over one year) using either manual microscopy or AI-assisted digital cytology. The cytology workflow included primary screening, quality review stages and consultant reporting.
Published evidence, laboratory data and input from co-authors who are experts working in laboratories across England and Scotland informed the analysis. The team also tested different assumptions to explore how changes in screening times and laboratory workflows might affect the results.
What we learnt
The model estimated that screening and reporting the annual volume of slides in England would require 31,842 staff hours with AI-assisted digital cytology, compared with 103,151 hours using manual microscopy. This equated to an estimated saving of more than 71,000 staff hours and a potential 69% improvement in overall productivity.
On average, reviewing and reporting a slide took an estimated 4.0 minutes with AI-assisted digital cytology, compared with 12.9 minutes using manual microscopy. The greatest efficiency gains were identified during primary screening, although substantial savings were also estimated for consultant review.
A separate analysis using a Scottish laboratory workflow also indicated considerable time savings.
Why the findings matter
Cytology is a key part of the cervical screening pathway and can become a bottleneck when staffing capacity is limited. Reducing the time required to review slides could help laboratories manage demand, maintain turnaround times and strengthen the resilience of screening services. It may also mean that people could also get their results faster, which can reduce anxiety and also mean people who need further treatment may get it faster.
The findings may be particularly relevant as screening programmes respond to changing patterns of demand. While HPV vaccination is expected to reduce HPV positivity in younger groups over time and potentially reduce the number of people who require cytology, the introduction of HPV self-sampling may help more under-screened people take part in screening and could increase the number of HPV-positive samples requiring follow-up cytology.
The study provides early evidence to inform decisions about AI-assisted digital cytology, but it does not capture every aspect of implementation. Laboratories would also need to consider digital infrastructure, training, validation, quality management and local working practices. Real-world evaluations are needed to confirm the time savings and understand the wider effects on costs, backlogs and service capacity.
Co-author note
Dr Elisabeth Adams, the CEO of the Aquarius Population Health and one of the co-authors remarks, “This was a great piece of work to show how a new technology could support healthcare services that are currently under workforce pressure in the UK. It builds on initial work we published on the impact on AI-assisted digital cytology when it was first being trialled. In this case, we wanted to create a simple model with real data from experts across the UK, to demonstrate to decision-makers how a switch to a new technology could alleviate workforce challenges. We’ve seen that AI products in healthcare can have an important role, but we must understand and estimate the impact that introducing them can have. We hope that this work can inform key decisions and guidelines in the UK and more widely.”
Aquarius are experts in developing qualitative and quantitative evidence to support decisions about screening programmes, health technologies and service delivery. To learn more, visit our website or contact info@aquariusph.com.
Citation
Wilson A, Cropper A, Ma Y, et al. Improving laboratory workforce efficiency using AI-assisted digital cytology within an HPV-based cervical screening programme: A model-based evaluation for the NHS Cervical Screening Programmes. BMJ Open 2026;16:e113298. doi: 10.1136/bmjopen-2025-113298
https://bmjopen.bmj.com/content/16/7/e113298