ClinTech - Doing Better (1) - Trial Information

I read with great sadness this morning of the news from his brother Sam Seliger , that Jake Seliger has passed after an 18+ month battle with an aggressive form of cancer.

Jake and his wife Bess Stillman, M.D. were an inspiration to people like me. Rather than writing about the great inspiration that Jake indeed was, instead I am going to focus on the future. In particular, in my own space, how we can do better with technology to help patients like Jake.

My initial post from November 2023 made some suggestions - please take a look at these first. The posting below builds on them. I will create further posts on the topic.

Information Sharing

Information exists across investigator sites and healthcare environments, but, it is not shared. It is this lack of public sharing of this information that leads to the challenges that patients face. If you have a chronic condition and are urgently seeking a clinical trial, then you need to be prepared to do your own work to boil down a set of potential options down to those that are currently relevant for you. If it takes between 2 and 12 months for up to date information to appear on ct.gov, then it is often too late for cancer patients.

Sponsors and Investigators wish to fill their studies with relevant patients. We therefore have the required incentives from the funders and supporters to provide this information and yet, we do not see change.

Technology Bottlenecks

The current problem we have is with the tech that is typically used, and, the processes surrounding the tech that mean the data is late and of poor quality. I am going to drill into this to help explain.

In a typical clinical trial, you might use 5 systems. These systems may or may not be connected with each other. If they are not connected, then it is typically only when reconciliation occurs (manual) that the information can be considered clean and trustworthy. It is only when that reconciliation occurs that the information recorded can be made public - in somewhere like ct.gov.

In addition, the weaknesses of some of these clinical trial systems mean that Excel spreadsheets are used to review / clean data. I know - it is tragic, but the industry lacks proper incentives and drive to wipe out Excel and improve modern tools.

Answers to timely information sharing

I am going to skip the details and move directly onto how I believe we can force change

1. Enforce timelines - source data. Today, it is ok for a site to record information on paper, and for that paper to be the source of information the status of a patient in a clinical trial for days / weeks - that cannot happen. While this information is only on paper, it cannot be used or shared. We need to enforce a maximum time for data that is acquired in a clinical trial to be digitally stored centrally - this is a first enabler. We use technology in clinical trials, lets not bind ourselves to paper as the 'baseline' performance measures.
2. Integration - it is 'ok' today to have 5 different clinical trial technologies used on a clinical trial, and for them to NOT be integrated. This creates lag, lag, lag. This has the effect of creating a 'Garbage in' with ct.gov... Integration of these systems is not a nice to have, it must be a mandate. We need regulations that made a lack of integrated systems a risk and a potential compliance issue.
3. Realtime data cleaning. It typically takes 4->8 weeks for data to be considered 'clean' when recorded in a clinical trial. Clean means it can be trusted. This may also be the time when it may be shareable. This is about 4-8 weeks too long. We already have tools and methods, soon to be enhanced with AI to be able to achieve an auto-risk assessed state of 'clean' in under 24hrs.
4. Standards compliance. Companies think that they can avoid complying with standards, especially if they are 'big'. The non compliance is leveraged for competitive advantage. Compliance with standards such as CDISC ODM and improvements with a common syntax for RESTful API's must be made a regulatory requirement in order to achieve (2) above and to provide consumers with choice.
5. Technology Quality Audits. All software that is used in clinical research must be validated. However, to be validated, you simply need to tick a few boxes. I have seen some very bad software that flys through audits. Today, passing an audit means next to nothing with regards software quality. Regulators need to tighten up the guidelines on good tech for clinical trials.
6. Information sharing regulations. There are competitive advantage reasons for information on clinical trials not being shared. We have clinical trial registers like ct.gov, but we don't have patient registers. A patient should be able to register their own existence, condition and state for clinical trials to find them. Investigators and Sponsors should be encouraged (required) to share details of patients that become available for other trials. Today, it is difficult if not impossible for software vendors to develop these types of systems (though I personally tried).

I will stop for today. More to follow...