Clinical Trial Details
— Status: Not yet recruiting
Administrative data
| NCT number |
NCT05675540 |
| Other study ID # |
308657 |
| Secondary ID |
|
| Status |
Not yet recruiting |
| Phase |
|
| First received |
|
| Last updated |
|
| Start date |
June 1, 2023 |
| Est. completion date |
March 30, 2024 |
Study information
| Verified date |
December 2022 |
| Source |
Guy's and St Thomas' NHS Foundation Trust |
| Contact |
n/a |
| Is FDA regulated |
No |
| Health authority |
|
| Study type |
Observational
|
Clinical Trial Summary
The investigators have worked with software designers to develop a software that allows us to
analyse current adherence to guidelines on Ophthalmic conditions such as Age related Macular
Degeneration (AMD), Diabetic Macular Edema (DMO) and Retinal vein occlusion (RVO). National
guidelines state that those patients with fluid accumulation in their central macular,
meeting criteria, are eligible for injections into the vitreous cavity of the eye
(intravitreal).(1) As these condition are common the trial is relevant to the public and
patients as future management may be affected by the outcomes of this trial. The
investigators will trial the software which uses Artificial Intelligence (AI) algorithms to
determine the most suitable review required for patients being managed in clinics, based on
'Vision' and 'Retinal Thickness' demographics. This will be done prospectively, in real time.
The question to be addressed is 'Can medical and non-medical practitioners accurately
determine treatment and follow-up for patients assisted by an AI clinical decision support
system, for the three most common chronic macular diseases - Wet Age-Related Macular
Degeneration (wAMD), Diabetic Macular Oedema (DMO) and Retinal Vein Occlusion (RVO) - in a
safe and clinically cost effective way?' Patients undergoing treatment for at least 12 months
are eligible to participate, so long as they are able to provide consent for their data to be
used. Participants will have no change to their care during the trial. The study, will take
place at Guy's and St. Thomas' NHS FT (GSTT) from where participants will be recruited, and
will last approximately 6 months of data collection.
The software will be used by the research Fellow, alongside the masked consultant. Therefore
the patient pathway and management will not be impacted by this trial. Patients will be
consented for data use.
Description:
Gulliksen (2017) describes digitalization as the biggest societal transformation process,
including for the health-care sector. Since the challenges of the pandemic, digitalization
has particularly taken place during changes to how healthcare is provided. The current Global
Healthcare situation dictates flexible and innovative steps to practically adapt working
methods in clinic to meet the demands of specialities, such as Ophthalmology, in safe and
efficient ways. Macular disease causes devastating sight loss. It is the most common cause of
preventable sight loss in the UK and the developed world, leading to unnecessary and
debilitating loss of vision and blindness if not monitored and treated regularly. In the UK,
1.8M people have chronic macular disease, of which approximately 300,000 reside in London.
Allied health professionals are trained to give the injections but at the moment there is no
formal structure for AHPs to be trained to assess patients data (vision and OCT scan).
The investigators therefore suggest it is an important area of research, to develop a
management aid to allow standardization of care, and allow more patients to be seen by
increasing the clinical workforce. This new AI software enables the mobilization of the
broader non-medical workforce of over 14,000 optometrists as well as nurses, allied health
professionals and technicians who can be up-skilled with AI tools and training, and operate
from community settings to expand resources, capacity and the quality of care. Whilst the
algorithm is not aimed to improve knowledge of the diseases, there will be an element of
learning of good management by the end user. There are currently many examples of automated
software in literature, however they all end at the validation stage. They all address the
use of AI to diagnose conditions, but none have addressed the issue of implementation, as
highlighted by many of the papers. There are also no trails aimed at the management of these
patients, subsequent to diagnosis.
This study will compare gold standard clinical care of patients receiving treatment for the
three main macular diseases (treatment decisions made by an ophthalmologist) with an
assessment and treatment recommendation made by the MacuSense software. The patients'
treatment will follow the ophthalmologists decision at all times, and the result of the study
will compare this with the MacuSense output. The ophthalmologist will be blind to the
MacuSense output throughout the study.
The research participant will attend their clinic appointment, with measures of vision and a
retinal (OCT) scan as normal. The ophthalmologist will review the patient and scan with a
treatment and follow up interval decided. Treatment will be given at that visit.
The data collected will be entered into the MacuSense software by the principal investigator
(Dr Zakri). The output of the software gives a follow up interval. The result of the follow
up interval decided by the clinician and the follow up interval indicated by the Macusense
software will be compared for each patient. The ophthalmologist will not have access to
Macusense software or the output of the software, so that their treatment decision is not
affected. The output of the Macusense software will not be used to change the clinical
decision.
The output of the study is therefore to compare the clinician treatment and follow decision
with the Macusense result, for research purposes only. The aim is to test the accuracy of the
Macusense algorithm against gold standard care. The null hypothesis states that there is no
difference between outcomes in standard clinical care and macusoft software.
The investigators anticipate from previous work that there is deviation from NICE guidelines
on a number of occasions. Therefore The investigators suggest an alternative hypothesis that
states there the outcomes will show a difference between standard care and outcomes presented
from macusoft algorithm software.
Study timetable
The study is estimated to take 9 months, broadly as follows:
- Months 1-6: Research set-up, patient recruitment, data preparation. Collect data on
decisions made with and without MacuSense (the algorithm will not be altered during the
study)
- Months 7-9: Statistical analysis and validation of observational evidence
Study details:
How the study will operate
The study will operate as follows:
1. Patient recruited and consented when attending clinic for a regular follow-up
appointment.
2. Consultant ophthalmologist makes a decision regarding intravitreal treatment interval
following current practice.
3. Patient's historical diagnostic data will be analysed by the computer program,
MacuSense. This pseudonymiseddata will be able to view by Dr Zakri and the macusoft
design team.
4. The consultant ophthalmologists' original decision will be compared to the analysis made
with MacuSense. No impact on the clinical care given will be made.
