Quality assurance in lung cancer screening programmes

Our proposal for a real-time dashboard in NHS England's Targeted Lung Health Checks

color blob

There is a bigger health crisis at the end of covid-19: the cancer backlog. In the UK, an estimated 50,000 diagnoses have been ‘missed’ during the pandemic, with appointments for suspected lung cancer suffering the most substantial decline.

Recent reports acknowledge the immense value of lung cancer screening in post-covid recovery. The UK Lung Cancer Coalition’s (UKLCC) first recommendation is to establish a UK-wide lung cancer screening programme as soon as possible. The All-Party Parliamentary Group (APPG) for Respiratory Health report concludes that:

“The roll-out of a full lung cancer screening programme across all four nations will do more to improve lung cancer survival than any other intervention”.

So, it’s no wonder that the NHS England Targeted Lung Health Check (TLHCs) programme, launched as a pilot in 2019, was recently expanded and is most likely becoming a long-term national initiative. The programme is also a leading implementer of technology and AI, and we are proud to be the preferred AI partner.

At Aidence, we are very invested in lung cancer screening, beyond providing tailored AI lung nodule management solutions. For one, we are also building a new application for patient follow-up. Furthermore, we aim to facilitate data aggregation and analysis, meeting this further recommendation from the UKLCC report:

“Work should be undertaken to determine how we can improve access to more near real-time data in lung cancer so that services can be monitored in a way that means action can be taken more quickly if data shows that key performance metrics are being missed”.

This article explains why and how we can create a quality dashboard for lung cancer screening, building on the lung health checks example.

Room for improvement

End-to-end lung cancer screening involves input from many healthcare professionals. Beyond image analysis, there are many opportunities for technology to add further support for effective programmes. Talking to clinicians involved in the TLHCs, we mainly recognised two areas for improvement:

Data aggregation

Public health programmes must prove their quality and cost-effectiveness against set targets to sustain funding. For example, a target in the NHS long term plan is the 28-day turnaround time, from referral to diagnosis or rule-out. But before assessing performance, the data must be gathered.

Data is being generated across all stages of the lung cancer screening pathway, supporting patient management and reporting. Currently, data collection is a manual process. Seeing the scale and complexity of the screening programme, manual data aggregation is a time-consuming, error-prone, and uninspiring task.

Manual processing is often the cause of reporting delays. The latest National Lung Cancer Audit annual report, released in 2020, covered lung cancer services and patient outcomes for patients diagnosed during 2018 – thus with a two-year delay!

We can do a better job bringing data together across screening sites.

Recall rates analysis

The success of any screening programme depends on the benefit to patients outweighing the potential harms. A big concern is bringing back patients for follow-up scans which turn out unnecessary. These additional scans increase radiation exposure and create anxiety for the patient.

Thus, recall rates are an essential performance indicator in lung cancer screening. Analysing them per hospital, region, and against a national average provides valuable information for programme managers.

There’s room for improving the current recall rates analysis in the TLHCs.

Where we stand

We already have a solid starting point for improving the collection and analysis of lung cancer screening data. Our lung nodule reporting tool, Veye Reporting, is currently processing data from NHS lung health check sites. The software produces standardised reports of lung nodules following the protocols and requirements of the TLHC programme. It also allows exporting a list with all verified reports as a .csv file, simplifying reporting to the national team.

As the data that goes through Veye Reporting is standardised and delivered according to specific protocols, it can be aggregated and exported to provide management information and insights for quality assurance of the screening programme. Some of the data points that can inform these insights, per site and region, are:

  • Amount of patient scans
  • Number of lung nodules
  • Follow-up decisions (lung and non-lung-related)
  • Turnaround time
  • Recall rate

By aggregating data from multiple sites to a regional or national level, we can further provide a comparison with the national average – in real-time!

Quality dashboard benefits

A live quality dashboard for the lung health checks would allow sites and programme leads to have their finger on the pulse of the programme and take appropriate action. This translates into (at least) the following benefits:

Efficiency gain

Going from spreadsheets stitched together to an automated dashboard is simply a more efficient approach, better suited for a large-scale, life-saving population health programme.

Lower recall rates

Patients should (ideally) receive the same high-quality care regardless of where they are helped. However, there are variations in early diagnoses and outcomes across the trusts. A quality assurance tool could close the gaps between trusts.

As mentioned above, the recall rate is a key indicator of performance in lung cancer screening; this applies at a radiologist level, as well as site and region. A benchmark for recall rates would make it possible to highlight areas where this rate is too high and focus on improving it.

The recall rates presented by our reporting software can be compared against other radiologists, other sites, regions, the national target and the national average. Hospitals can thus assess the performance of their own radiologists and also learn from each other.

Monitoring goes hand in hand with adequate training. A good example from breast cancer screening is PERFORMS, a self-assessment and training platform for mammogram readers. Participants get a set of real-world, anonymised, complex cases to read and receive feedback on their interpretation. A similar platform would serve lung cancer screening as well.

Better targeting

Another approach to limiting lung cancer screening risk is focusing on individuals at the greatest risk of developing cancer. In the TLHCs, eligible individuals are 55 to 74 years old and have a smoking history. The programme incorporates two risk prediction models. But, John Field, Director of Research at the Roy Castle Lung Cancer Programme at the University of Liverpool, UK, urges caution:

“Whatever model is used, it needs to be remembered that risks for developing lung cancer vary over time, due to changes in risk factors, like age. Consequently, people who have been deemed ineligible may become eligible, underlying the importance of regular risk assessments.”

Population insights are central to highlighting locations where people are at a higher risk of developing lung cancer. This is another indicator that the quality dashboard can provide for lung cancer screening.

Just getting started

We are aware that there is more to a lung cancer screening programme than radiology. To optimise the TLHCs, we should also look at decision-making down the patient pathway. We are advocating for a step-by-step approach, starting with radiology, which plays a central role in screening initiatives.

We’re building on the TLHCs as an example, but as more and more countries in Europe are piloting lung cancer screening, we can expand our learnings to a concerted European approach. We have started some great conversations in Italy and Israel, for instance.

Would you like to join us? Get in touch!


About Jeroen

Jeroen van Duffelen is co-founder and Chief Business Officer at AidenceJeroen's entrepreneurial spirit led him to teach himself software engineering and start his own company commercialising an online education platform. He then tried his hand in the US startup ecosystem, where he joined a rapidly scaling cloud company. Jeroen returned to Amsterdam to run a high-tech incubator for academic research institutes. Here, Jeroen first got his taste for applying AI to healthcare. In 2015, he founded Aidence together with Mark-Jan Harte.

Connect with Jeroen on

Aidence is now DeepHealth. Learn more about our AI-powered health informatics portfolio on deephealth.com


Book a demo

    * Required fields.
    For more information on how to unsubscribe, our privacy practices, and how we are committed to protecting and respecting your privacy, please review our Privacy Policy.
    This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.