Closed for comments This consultation ended on at Request commenting lead permission
Four artificial intelligence (AI) technologies can be used in the NHS during the evidence generation period as options to aid the opportunistic detection of vertebral fragility fractures (VFFs). The technologies are:
BriefCase-Triage
CINA-VCF Quantix
HealthVCF
IB Lab FLAMINGO.
These technologies can only be used:
according to their indications as outlined in their instructions for use and with consideration of the risk groups as recommended in NICE's guideline on osteoporosis: assessing the risk of fragility fracture
if the evidence outlined in the evidence generation plan is being generated
as long as they have appropriate regulatory approval including NHS England's Digital Technology Assessment Criteria (DTAC) approval.
The companies must confirm that agreements are in place to generate the evidence. They should contact NICE annually to confirm that evidence is being generated and analysed as planned. NICE may revise or withdraw the guidance if these conditions are not met.
At the end of the evidence generation period (3 years), the companies should submit the evidence to NICE in a format that can be used for decision making. NICE will review the evidence and assess if the technology can be routinely adopted in the NHS.
What this means in practice
The 4 technologies listed in recommendation 1.1 can be used as an option in the NHS during the evidence generation period (3 years) and paid for using core NHS funding. During this time, more evidence will be collected to address any uncertainties. Companies are responsible for organising funding for evidence generation activities.
After this, NICE will review this guidance and the recommendations may change. Take this into account when negotiating the length of contracts and licence costs.
Potential benefits of use in the NHS with evidence generation
Clinical benefit: Clinical evidence suggests that AI technologies can help opportunistically detect VFFs that would otherwise have been missed. This could help identify more people with a VFF who need treatment, to improve their quality of life and reduce the risk of future fractures.
Resources: By reducing the risk of further fractures, earlier detection and treatment of VFFs could reduce the demand on other costly services, such as those needed to manage hip fracture.
System benefit: Using AI technologies can help reduce variation in clinical practice and help healthcare professionals to implement the Royal College of Radiologists' guidance for the recognition and reporting of osteoporotic vertebral fragility fractures.
Managing the risk of use in the NHS with evidence generation
Clinical subgroups: There is no evidence to show whether the AI technologies are as clinically effective across all age groups. Older age is a risk factor, but there are other risk factors independent of age. It is uncertain whether the opportunistic detection of VFFs in all subgroups represents value for money in the NHS.
Resources: Implementing the AI technologies could have a big impact on radiology services, such as increasing the number of radiographs that need to be reviewed by a radiologist and the number of follow-up dual-energy X-ray absorptiometry (DEXA) scans that need to be done.
Costs: Early results from the economic modelling show that the technology could be cost effective. But, there is uncertainty around the cost of some of the technologies and the true cost of implementation and ongoing post-market surveillance. Trusts should take into account the costs of the AI technologies used in this assessment when implementing the technologies. When negotiating with companies, trusts should also consider the upfront costs for implementing a technology, and should monitor costs associated with its use in populations with a lower prevalence of osteoporosis.
Clinical risk: Using AI technologies to help detect VFFs on radiographs is considered to have a low clinical risk. This is because the technologies are used in addition to standard care in which healthcare professionals make treatment decisions. AI technologies do not replace the definitive radiology review.
Implementation guidance: Clear local protocols will need to be in place when using AI technologies. This is to ensure that healthcare professionals refer people with a newly identified VFF to the appropriate services.
Equality: There is a risk that the AI technologies may have reduced diagnostic accuracy in different populations, such as younger people who may have risk factors for VFF, people from ethnic minorities, and other groups that may have been underrepresented in the AI training set.
More research is needed on the following AI technologies that aid the opportunistic detection of VFFs before they can be funded by the NHS:
Annalise Enterprise CXR/Annalise Container CXR
BoneView
HealthOST
TechCare Spine.
What this means in practice
There is not enough evidence to support core NHS funding for the 4 technologies listed in recommendation 1.4 in the NHS.
Access to the technologies should be through company, research or non-core NHS funding, and clinical or financial risks should be managed appropriately.
Evidence generation and more research are needed on:
the diagnostic accuracy of the technologies compared with current NHS standard care, including in key subgroups such as people under 50 years and those at a higher risk of a VFF
the failure rates of the technologies and the reasons for failure
the impact on referral and treatment rates of identifying additional VFFs
the impact on the workload of healthcare professionals of introducing the technologies
the short-term impact on quality of life of identifying and managing a VFF.
The evidence generation plan gives further information on the prioritised evidence gaps and outcomes, ongoing studies and potential real-world data sources. It includes how the evidence gaps could be resolved through real-world evidence studies.
AI technologies can help healthcare professionals spot VFFs on X-ray images and CT scans involving the spine that are done for unrelated conditions (opportunistic detection). Treatment can reduce symptoms and the risk of future fractures, so detecting VFFs early has clear benefits. Preventing future fractures can also reduce the demand on radiology services and save money elsewhere in the NHS.
BriefCase-Triage, CINA-VCF Quantix, HealthVCF and IB Lab FLAMINGO are designed to help detect VFFs on CT scans. Diagnostic accuracy evidence comparing them with a reference standard suggests that they can help detect moderate to severe VFFs. Early economic evidence shows that they could be cost effective. So, these 4 technologies are recommended for use with evidence generation.
Annalise Enterprise CXR/Annalise Container CXR, BoneView and TechCare Spine are designed to help detect VFFs on X-ray images. Clinical evidence for Annalise Enterprise CXR/Annalise Container CXR is uncertain because it is based on studies that mostly included side-view chest X-ray images. These are not commonly used in UK clinical practice and so the evidence may not be generalisable to the NHS. The usefulness of BoneView and TechCare Spine is uncertain because they only analyse spine X-ray images, which are usually done for back pain and include a thorough review of the spine. This means VFFs are less likely to be missed on these X-ray images and so the technologies may not offer additional benefit. There is no clinical evidence for HealthOST and TechCare Spine. So, these 4 technologies can only be used in research.
Evidence on the diagnostic accuracy of the technologies compared with standard care in the UK is limited. So, more data should be collected to show how much better they are at detecting additional VFFs in clinical practice. More evidence is also needed on the downstream effects of the technologies for both people having X-rays and CT scans and healthcare professionals, including the effect of the technologies on referral and treatment rates, quality of life and healthcare professional workload. Companies should also address gaps in the evidence around how often their technologies are unable to process an image (the failure rate).
How are you taking part in this consultation?
You will not be able to change how you comment later.
You must be signed in to answer questions
Question on Consultation
Question on Consultation
Question on Consultation
Question on Consultation