A pragmatic randomised-controlled trial (RCT) comparing the clinical- and cost-effectiveness of artificial intelligence (AI)-guided ultrasound (US) performed by non-specialists in the primary care setting to usual care Deep Vein Thrombosis (DVT) diagnostic pathway.

Funding:

BNSSG ICB Research Capability Funding.

 

What is the problem?

DVT is a blood clot which usually occurs in the deep veins of the leg (1). It is a common vascular disease. DVTs are potentially fatal. Parts of the clot can break off, travel through the blood vessels, and become lodged in the lungs, causing a Pulmonary Embolism (PE). Every 37 seconds someone in the western world dies from a DVT and its complications. That accounts for circa 12 640 deaths per annum in the United Kingdom (UK) alone (1).

The current DVT diagnostic pathway is laborious, time-consuming and is a financial burden for the National Health Service (NHS) (7). Patients with pain, redness and swelling of the leg usually present first at their GP. If the GP has the suspicion of a DVT, in most parts of the UK the GP sends the patient to an emergency department (ED) and from there to a vascular clinic (3).

There have been attempts to improve the DVT diagnostic pathway efficiency in various parts of the UK (11). In the Bristol, North Somerset, and South Gloucestershire (BNSSG) GP Care, a social enterprise, was awarded the Community DVT Service in May 2019. Since then, GPs can refer patients with DVT-symptomatic to a GP Care DVT Clinic. DVT diagnostic and initiation of treatment can then be performed at the DVT clinic.

Nevertheless, GP Care does not provide services for patients with mobility issues, housebound patients, patients with other significant health problems, patients which are suspected to have an upper leg DVT and patients under the age of 18.

Those typically affected by a DVT are elderly patients with reduced mobility, who are living with multiple long-term conditions and are often isolated in their homes. (4) Those people are underserved by the current diagnostic pathway. Diagnosis is often delayed, causing increased complications such as possible life-threatening PE.

While waiting for diagnostic scans, GPs are recommended to prescribe blood thinning treatment. This also increases the risks of bleeding, which can be serious if affecting the brain or the gut, particularly in the frail elderly. Earlier diagnostic scans in primary care settings could avoid the need for unnecessary blood thinning treatment in patients without DVT. (5)

What is the aim of the research?

The primary aim is to develop a funding application for a RCT comparing AI-guided US performed by non-specialists in primary care to usual care, highlighting potentially avoided hospital admissions and improved cost-effectiveness.

The intended impact of the research is:
· To reduce ED attendances and hospital admissions.
· To reduce workload on specialists.
· To simplify DVT diagnostic pathways providing easier access for
underserved population such as elderly, housebound, and ethnic minorities.
· To reduce costs for the NHS.
· To be more inclusive in primary care.

 

How will this be achieved?

The RCF funding will be used to:

Train HCAs working for GP Care in the use of AutoDVT. Patients attending GP Care DVT clinics will be informed of the study and be invited to participate. Consented participants would receive an AutoDVT scan first (done by the GP Care HCA) which will be followed by the GP Care standard scan performed by the sonographer. Patients will be invited to complete a patient satisfaction survey after the scan. HCAs will be invited to complete a user-friendliness survey after the last participant if the DVT clinic of the day.

A subset of patients, clinicians, carer and AutoDVT operators will be recruited to qualitative interviews (15 patients, 5 GPs and ED physicians, 5 nurses and carers and 5 HCAs as indicated in costing plan).

The research will be divided into three parts:
1. A DTA study to confirm the sensitivity and specificity of the AutoDVT device in 500 patients with suspected DVT
2. A survey to explore the user-friendliness, and patients’ satisfaction.
3. Qualitative research using semi-structured interviews with patients and clinicians to explore the acceptance and resistance towards AI-guided DVT diagnosis.

 

Who is leading the research?

Kerstin Nothnagel (PhD student, University of Bristol) alongside co-chief investigator Professor Alastair Hay (Professor of Primary Care, University of Bristol).

Further information

About Kerstin Nothnagel.

About Professor Alastair Hay.

For more information or to get involved with this project, please contact bnssg.research@nhs.net.