This article is a follow-on from a presentation done at a recent SNOMED International webinar, where we discussed our work about integrating clinical trials with electronic patient records. It outlines our work on Keytrials, an open-source clinical trials discovery platform, that allows users to answer the following question:
How do you enable a clinician to know what clinical trials the patient in front of them is eligible for, at the click of a button within their EHR system’…
As we all know making clinical research/trials accessible/discoverable to practising clinicians and patients is not always straightforward. So the backstory to this started on a ward round in University College London Hospitals (UCLH), where a consultant thought – `we are one of the largest centres for pre-clinical trials in Europe, but how come there isn’t a straightforward way of finding out what trials a patient is eligible for in the clinical or during multi-disciplinary team (MDT) meetings`. Two years down the line, with the support from UCLH and the UCL Biomedical Research Centre (BRC), it is now possible with a single click:
- View all trials that a patient is eligible for based on their condition, age & gender
- Browse all these trials without ever leaving the interface of the EHR system
All this is a enabled by an open-source platform called Keytrials. When I first started creating Keytrials, it was just a just a `side project`, but four months later my dad was diagnosed with cancer and things changed. I suddenly realised that as an ex-doctor, even I was finding it hard to get access to all possible treatment options! Following his first round of treatment, he developed a `growing haziness` in his lung which doctors could not tell for sure if it was a metastasis or a primary lung cancer! It was incredibly frustrating going through the experience and I started to wonder if there are any known genetic predispositions that could cause two different cancers in such a short time. Trying to search for such studies or any options for cancers was like hitting a brick wall! Very little information about any on-going research was publicly available, even for me to discuss with the doctors.
Bingo, I was in the shoes of the clinician in the ward, wondering what the treatment options (trials) that my dad (a patient) was suitable for!
What makes this a problem worth solving is the impact it can have on millions! Having access to information to choose the best treatment option makes a huge difference – patients and family members who have gone through it can attest to this! But we don’t just want patients to have access to on-going studies, we want all clinicians (doctors, etc) involved in those decisions to have access to the same information! This is exactly what Keytrials does:
- It makes all clinical trials/studies in a hospital/trust accessible accessible via a browser
- It allows clinicians to access this information not only from browsers, but also directly from their EHR system interface
- It allows clinicians to use smarter/semantic search – when someone is looking for lung cancer trials, it also brings back results for `oat-cell carcinoma of lung`, even if it does not have the word `lung cancer` anywhere in its description!
Keytrials supports searching/matching based on meaning of words, not just the words – we call this semantic search
So how does Keytrials enable this?
While a much longer version of this is provided in the attached presentation, here are the outline steps:
1. Import trials
Import all trials from the local trial registry into Keytrials platform – which comes with built-in support for importing from ClinicalTrials.gov and Edge a popular trial registry software used in UK.
2. Annotate with SNOMED CT
To enable smart searching, we rely on the power of SNOMED CT – a recognised healthcare standard. It contains all the knowledge needed to infer that `oat-cell carcinoma of lung` is a type of `lung cancer`. To associate trials with SNOMED CT codes, we use Natural Language Processing, that annotates trials with their associated conditions (coded in SNOMED CT). We then use a terminology server, that allows us to access SNOMED CT hierarchical relationships on the fly as part of searching for eligible trials.
3. Integrate with EHR
To enable integration between the EHR and Key trials, we rely on yet another standard called HL7 InfoButton. This allows us to access the trials in Keytrials via a standardised HL7 InfoButton query. Don’t worry if you don’t know what that query needs to look like – that’s the beauty of the standard as major EHR systems come with support for it built in. So all you need to know is a few basics and you are in business!
In fact, it only took a clinician 15-20 minutes to connect the UCLH instance of Keytrials to their EHR system (Epic)! I am not making this up… 🙂
Of course, there might be other solutions that offer similar functionality. However, we designed this solution from day 1 to be open-source and use healthcare standards – SNOMED CT, ICD10 and HL7 InfoButton. What this allows you to do is to make all clinical trials accessible to patients and clinicians. You don’t need to have an EHR system or use those standards since Keytrials comes with blazingly fast search and filtering built in! Of course, if you already have SNOMED CT and an InfoButton compliant EHR system, then you can integrate your clinical trials with your EHR system without too hassle – since Keytrials platform does most of the work for you!
We have this integration between EHRs and clinical trials working UCHLH, in one of the largest hospitals in London – who also happen to use Epic. They use the power of Keytrials to make all their on-going clinical trials publicly visible to everyone. There is of course a mobile app that allows users to browse the same instance from their phones! This is the power of platforms – standards and APIs!
We would not have been able to create this without support of UCLH and UCL BRC! But now thanks to them, we are exploring ways of making Keytrials a shared research platform for multiple trusts/hospitals to share/publish all their research studies in a single place. Needless to mention, there is this cool standard called HL7 FHIR – which allows us to expose these SNOMED CT annotated clinical trials in Keytrials via a FHIR interface!
We would love to collaborate and get feedback with anyone working in this space or with an interest in making clinical trials more accessible. If this sounds interesting (or full of ^*&%*£), then please feel free to let me know… 🙂
This article is a cross-posting from a LinkedIn article published at https://www.linkedin.com/pulse/integrating-electronic-health-records-clinical-trials-jay-kola