AI-guided deep vein thrombosis (DVT) diagnosis in primary care: protocol for cohort with qualitative assessment

ISRCTN	ISRCTN14795960
DOI	https://doi.org/10.1186/ISRCTN14795960
IRAS number	332800
Secondary identifying numbers	IRAS 332800, CPMS 58280

Submission date: 30/06/2024
Registration date: 04/07/2024
Last edited: 02/08/2024

Recruitment status: No longer recruiting
Overall study status: Ongoing
Condition category: Circulatory System

Prospectively registered

Protocol

Statistical analysis plan

Results

Individual participant data

Record updated in last year

Plain English summary of protocol

Background and study aims
In this study, we aim to investigate a new approach with the overarching goal of enhancing the detection of deep vein thrombosis (DVT). DVT is a medical condition characterised by the formation of blood clots within the deep veins, most commonly occurring in the legs. The current standard for diagnosing DVT involves utilising the Wells Score, a scoring system, in conjunction with a blood test called D-dimer, followed by an ultrasound scan (USS) administered by specialised healthcare professionals (HCP).

However, this study explores the potential of a transformative technology that could change this paradigm. A scenario has been envisioned in which ordinary individuals, without specialised ultrasound (US) training, could employ a smartphone application (app) to perform a DVT scan of the leg. This innovative app serves as a comprehensive guide, leading users through the steps necessary to conduct a leg US examination.

The significance of this research becomes apparent when considering the potential implications. If successful, this approach has the capacity to revolutionise DVT diagnosis, enabling patients to receive prompt assessments in the convenience of their primary care provider's office or even within their own homes. Such a development could substantially expedite the diagnosis process, offering enhanced convenience to patients and potentially reducing the number of emergency department (ED) visits. Additionally, this innovation holds the promise of alleviating some of the financial strain on healthcare systems like the National Health Service (NHS) by optimising DVT diagnostic services. In essence, this study seeks to pave the way for a future in which DVT diagnosis becomes more accessible, efficient, and financially sustainable for HCPs.

Who can participate?
GP Care DVT Clinic eligible patients who have the capacity to consent can participate in this study.

What does the study involve?
During an appointment at the GP Care DVT clinic, patients who agree to participate will be consented and receive an additional ultrasound scan with the AI-guided handheld probe operated by the GP Care HCA. Participants are prompted to complete a participant satisfaction survey after the index and reference scan. A subgroup of 20 out of 562 participants will be invited for semi-structured interviews.

What are the possible benefits and risks of participating?
Benefit
- Participants may benefit from a quicker and more convenient assessment of their DVT risk.
- This method could reduce the need for immediate visits to specialised healthcare providers or emergency departments, saving time and potentially reducing healthcare costs.
- By participating, individuals contribute to research that could make DVT diagnosis more accessible and efficient in the future.
Risk
- As with any medical examination, there is a small risk of discomfort during the ultrasound scan.
- There is a potential risk that the app-guided scan might not be as accurate as a scan performed by a specialist, which could lead to either false reassurance or unnecessary concern. However, all app-guided scans will be reviewed by healthcare professionals to mitigate this risk.

Where is the study run from?
Keele University (UK)

When is the study starting and how long is it expected to run for?
May 2023 to May 2026

Who is funding the study?
NHS Bristol, North Somerset and South Gloucestershire Integrated Care Board (UK)

Who is the main contact?
kerstin.nothnagel@bristol.ac.uk
Alastair.Hay@bristol.ac.uk
michelle.spano@bristol.ac.uk

Contact information

Mrs Kerstin Nothnagel
Principal Investigator

Canynge Hall
Clifton Bristol
Bristol
BS8 2PN
United Kingdom

ORCID ID	0009-0004-5767-8346
Phone	+44 1174559936
Email	Kerstin.Nothnagel@bristol.ac.uk

Prof Alastair Hay
Scientific

Canynge Hall
Clifton Bristol
Bristol
BS8 2PN
United Kingdom

ORCID ID	0000-0002-4789-924X
Phone	+44 117 928 7279
Email	Alastair.Hay@bristol.ac.uk

Miss Michelle Spano
Public

Canynge Hall
Clifton Bristol
Bristol
BS8 2PN
United Kingdom

ORCID ID	0000-0001-7100-7099
Phone	+44 117 928 7279
Email	michelle.spano@bristol.ac.uk

Study information

Study design	Diagnostic cross-sectional study coupled with a qualitative evaluation
Primary study design	Observational
Secondary study design	Cross sectional study
Study setting(s)	GP practice
Study type	Diagnostic
Participant information sheet	45715 Patient Information Sheet Quantitative 09.10.23.pdf
Scientific title	AI-guided point-of-care ultrasound for deep vein thrombosis diagnosis in primary care: protocol for a diagnostic accuracy and qualitative assessment study
Study acronym	AID-DVT
Study objectives	Primary Hypothesis: AI-guided point-of-care ultrasound (POCUS) performed by non-specialists in primary care settings is as accurate as traditional ultrasound scans (USS) conducted by sonographers for the diagnosis of deep vein thrombosis (DVT). Secondary Hypothesis: AI-guided POCUS for DVT diagnosis will be acceptable to patients and healthcare assistants (HCAs) in primary care settings, as measured by patient satisfaction surveys and semi-structured interviews.
Ethics approval(s)	Approved 07/12/2023, London Bridge REC (2 Redman Place, Stratford, Stratford, E20 1JQ, United Kingdom; +44 2071048387; londonbridge.rec@hra.nhs.uk), ref: 23/PR/1242
Health condition(s) or problem(s) studied	Deep vein thrombosis
Intervention	AI-guided point-of-care ultrasound (POCUS) for the diagnosis of deep vein thrombosis (DVT) performed by healthcare assistants (HCAs) in primary care settings. This intervention is being compared to the traditional standard of care, which involves specialized ultrasound scans (USS) conducted by sonographers. 1. Enrolment: - Participants with suspected DVT are enrolled at primary care DVT clinics. 2. Diagnostic Test Accuracy (DTA) Study: - Participants undergo an AI-guided point-of-care ultrasound (POCUS) conducted by healthcare assistants (HCAs). - This is followed by a standard ultrasound scan (USS) performed by sonographers. 3. Patient Satisfaction Survey (PSS): - After completion of both the index (AI-guided POCUS) and reference (standard USS) scans, participants are prompted to complete a patient satisfaction survey. 4. Qualitative Semi-Structured Interviews: - A subgroup of the participants will be invited to participate in semi-structured interviews to explore the acceptability of AI-guided DVT diagnosis in primary care. Duration: - Participants in the DTA will be in the study for 20-30 minutes. - The subgroup of participants who agree to participate in semi-structured interviews will have their interviews within 1 to 3 weeks of participating in the DTA. - There is no additional follow-up required after the qualitative interviews.
Intervention type	Mixed
Primary outcome measure	Data collected during appointment: 1. Accuracy of HCA-reported AI-guided DVT diagnosis. Data Collected: Diagnosis of proximal DVT (yes/no). Method of Measurement: Comparison with reference standard of sonographer scan 2. Accuracy of AI-provided DVT diagnosis. Data Collected: Diagnosis of proximal DVT (yes/no). Method of Measurement: Comparison with reference standard of sonographer scan 3. Accuracy of remote specialist diagnosis using AI-guided acquired ultrasound images. Data Collected: Diagnosis of proximal DVT (yes/no). Method of Measurement: Comparison with reference standard of sonographer scan
Secondary outcome measures	Data collected after participant received index and reference scan: 1. Proportion of HCA-acquired images of adequate quality for renovate diagnosis. Data Collected image quality (adequate/inadequate according to ACEPO Score). Method of Measurement: Evaluation by 5 independent specialist (sonographer/ radiographer 2. Factors associated with inadequate image quality. Data Collected: Factors such as HCA experience, patient BMI, and machine settings. Method of Measurement: collected by CRN staff during patient recruitment at GPCare DVT Clinic 3. Patient satisfaction with AI-guided DVT diagnosis. Data Collected: Satisfaction score. Method of Measurement: Patient Satisfaction Survey Qualitative Objectives: 1. Explore patient reservations, obstacles, and confidence in relation to AI-guided point-of-care US undertaking as a diagnostic scan for DVT 2. Explore the confidence of HCAs in the index scan 3. Explore the experience of HCAs in delivering the scan Data Collected: Patient interviews, including demographic information (gender, age, postcode) Method of Measurement Semi-structured interviews Time-point 1-3 weeks post-scan
Overall study start date	01/05/2023
Completion date	01/05/2026

Eligibility

Participant type(s)	Patient
Age group	Adult
Lower age limit	18 Years
Upper age limit	105 Years
Sex	Both
Target number of participants	500
Total final enrolment	562
Key inclusion criteria	1. Eligible for GP Care DVT Services †1 flowchart 2. Participant has capacity to consent to study participation
Key exclusion criteria	1. Unable to communicate in English (spoken or written) and there is no translator available. 2. For the qualitative study, patients who have a hearing impairment that would hinder the feasibility of an e-interview.
Date of first enrolment	26/02/2023
Date of final enrolment	14/06/2024

Locations

Countries of recruitment

England
United Kingdom

Study participating centre

GP Care DVT Clinic in NHS Bristol, North Somerset and South Gloucestershire

160 Aztec West
Bristol
BS32 4TU
United Kingdom

Sponsor information

Keele University
Research organisation

School of Primary Care Research (SPCR), School of Medicine
Newcastle-under-Lyme
ST5 5BG
England
United Kingdom

Phone	+44 1865 289300
Email	spcr@keele.ac.uk
Website	https://www.spcr.nihr.ac.uk/
"ROR"	https://ror.org/00340yn33

Funders

Funder type

Hospital/treatment centre

NHS Bristol, North Somerset And South Gloucestershire Integrated Care Board

No information available

Results and Publications

Intention to publish date	05/01/2026
Individual participant data (IPD) Intention to share	Yes
IPD sharing plan summary	Stored in non-publicly available repository
Publication and dissemination plan	The protocol of this study will be published in BJGP Open. Upon completion of the study, we aim to publish the findings in Nature. The data supporting the findings of this study will be made available upon reasonable request to the corresponding author, following publication of the results. This will include de-identified participant data, study protocols, and statistical analysis plans to ensure transparency and reproducibility of our research.
IPD sharing plan	The dataset generated during the current study will be stored and available on the University of Bristol Data Repository.

Study outputs

Output type	Details	Date created	Date added	Peer reviewed?	Patient-facing?
Participant information sheet		09/10/2023	04/07/2024	No	Yes
Protocol file		09/10/2023	04/07/2024	No	No

Additional files

45715 Patient Information Sheet Quantitative 09.10.23.pdf
45715 Protocol 09.10.23.pdf

Editorial Notes

02/08/2024: Internal review.
31/07/2024: Contact details corrected.
01/07/2024: Trial's existence confirmed by NHS HRA.