Ipsos
Sally Mouland, Michael Lawrie, Ellis Akhurst, Stephen McSwiney, Dr Anita Patel, Chris Hale, Katherine Fisher & Laura Tuhou (Ipsos) and Mike Woodall & Craig Parylo (Strategy Unit)
Conclusions
The report in short summary
The TLHC programme (2019-2024) was established to test lung health checks in real-world settings to diagnose lung cancers earlier. It followed promising results from several small-scale trials and pilots. Pilot areas were selected based on lung cancer mortality rates. The pilot stage of the programme was delivered across three phases comprising more than 25 sites. It set out to achieve a range of outcomes. Most prominently, the programme sought to replicate the early diagnostic staging results seen at small scale in these larger scale, real world projects.
Phase 1 and 2 projects issued 1.22 million invites to 735,000 people. The uptake rate of Lung Health Checks (LHC) was 44%, which meant 324,000 people attended a LHC. This was higher than observed in the Manchester pilot study (a key benchmark), although there remain opportunities to increase this. Following this, 163,000 people had at least one CT scan (50% of LHC attendees). There are significant levels of attrition at each stage of this pathway, associated with demographic factors, and variants in the model. Most notably, despite the projects reporting a range of outreach activity, and inviting these groups at a disproportionately high rate, uptake of LHCs by people living in areas of high deprivation, and from other ethnic demographic groups, was lower than average.
A total of 2,748 people had a lung cancer diagnosis (a conversion rate of 1.7%) through the programme. Around three-quarters of these cancers were diagnosed at stage 1 and 2, which aligns with an important policy benchmark set out in the NHS LTP.
Projects identified incidental findings (such as coronary calcification) in three-quarters of those who had a CT scan. In addition to this, in a smaller cohort of sites submitting patient level data (14 projects, accounting for 71% of first invites sent), 1,697 other cancers were diagnosed among 1,673 TLHC participants who received a LHC and were deemed high risk. Over one-quarter (28%) of these cancers were diagnosed within three months of participants’ last interaction with the programme.
A robust quantitative impact analysis, comparing the diagnostic outcomes of pilot areas to those of carefully designed counterfactual areas, was conducted. This shows that the TLHC programme was effective in meeting its objectives relating to additional lung cancer diagnoses (as show by the divergence of the green and orange trend lines in the chart); and the number and share of lung cancers diagnosed at an earlier stage. It is estimated that an additional 781 lung cancers were diagnosed at stage 1 or 2 in pilot areas between 2019 and 2022. These are the cancers that would not have otherwise been diagnosed until a later date or not diagnosed at all in the absence of the programme. The number of additional lung cancers diagnosed begins to fall three years following the introduction of the pilot.
The results of the impact analysis on staging are in line with expectations. The conversion rate (1.7% of those scanned) is within the range established by previous studies. The staging results as evidenced through the robust impact analysis also broadly replicate the results shown in smaller scale pilots – a key aim of the programme. The absence of an effect on mortality is consistent with clinical interpretation at this stage of the programme.
However, there are several areas for further study and development as the programme rolls out more widely. Improvements in earlier diagnostic staging were predominantly concentrated on individuals identifying as white British. Uptake in deprived areas and among ethnic minorities was lower than average, missing high-risk individuals, although the programme as a whole was delivered in some the most deprived areas of England and those from the most deprived areas have therefore benefited disproportionately. A detailed examination of the real-world costs of delivering the LHCs and scans found the LHCs generated high costs largely driven by the significant labour requirements. This was supported by testimony gathered from projects across the evaluation. LHC projects are challenging to set up and projects spent resource on engaging primary care, data sharing, clinical oversight and community engagement.
As well as demographic factors, variants of the model affected uptake too. Most significantly, while opt-out models for LHC invitation help in driving LHC uptake, they do not appear to result in the highest-risk attendees or the highest rates of lung cancer detection.
Disclaimers:
Chest Suite comprises multiple applications including Veye Lung Nodules, Veye Reporting, DeepHealth Chest CT AI, DeepHealth Viewer, DeepHealth Reporting, DeepHealth Worklist and HealthCCSng. Veye Lung Nodules and Veye Reporting are manufactured by Aidence B.V. and distributed by DeepHealth Inc. DeepHealth Viewer is manufactured by eRAD, Inc. and distributed by DeepHealth, Inc, HealthCCSng is manufactured by Nanox AI Ltd. and distributed by DeepHealth, Inc, DeepHealth Chest CT AI is 510(k) pending.
Any claims made about Chest Suite may reference claims associated with its individual components.
Not all products and functionalities are commercially available in all countries. For clearance and commercial availability in your geography of functionalities listed and compatibility with other systems, please contact your account manager.