Decoding Digital Healthcare: An Introduction to Clinical Informatics
Exploring the challenges and opportunities at the intersection of healthcare and information technology.
This blog is based on topics covered in Episode 5: Digital Health and Career Reflections at Microsoft with Dr Simon Kos.
If you haven’t listened to it yet, check it out on Apple, Spotify, Substack or wherever you get your podcasts.
Introduction.
As a Clinician or a patient, you’ve likely experienced frustration when using IT systems in healthcare. From clunky interfaces to inaccessible or unreliable information, these issues seem perennial.
But why is this the case? Especially when technology in other areas, like our smartphones, is so seamless.
Combining the vast intricacies of healthcare and information technology has been a formidable challenge. For decades healthcare and information systems have been intertwined leading to a complex legacy, resulting in a stifling oligopoly of electronic health records in many places1.
Layer in the rapid pace of change in medical practice, research, regulation, reporting, patient experience and technology advancements - and the result is a checkered history.
But it’s not for lack of trying. Efforts to crack these problems can be elusive when you look at the many expensive private and public failures in the graveyard of healthcare IT23456.
So when delving into this topic, I recommend adopting an attitude of humility and curiosity. Humility to learn from the past and the experiences of others, while fostering curiosity to investigate new approaches to old problems.
This piece will look to highlight some ideas and principles for clinicians when approaching the field of Clinical Informatics.
Clinical Reasoning and the Patient Journey.
A key part of being a clinician in this space is understanding how clinical reasoning and behaviour take place in the real world and the digital world. This is because translating these actions into processes enabled by digital tools requires careful instruction.
Progressing from a student, all the way up to an experienced clinician, decision-making moves from a conscious, effortful process to an intuitive one. However translating this can be difficult, especially if that very medical practice is tied to specific cultural or social norms within the field.
This reflection can expose us to our own deficiencies in logic, understanding or evidence base. While uncomfortable, this scrutiny is ultimately necessary for us to translate it into a digital process. This may be straightforward if there are high-quality clinical guidelines available, however, it can be particularly disorientating when approaching avenues of uncertainty.
Particularly with the added stakes of this decision being scaled and magnified across entire digital systems.
When brushing up against the limits of our understanding, utilising good clinical governance is essential.
Identify The Problem.
It might sound obvious, but solving the right problem is paramount. You’d be surprised how easy it is to get distracted on fixing something which was unnecessary to begin with.
Zooming in and out on the issue at hand can help with further context and see the relationships connected to the issue. Gathering multiple perspectives, including patients, multidisciplinary clinical teams and non-clinical stakeholders (administrators, operations experts, researchers etc) will further illuminate your understanding of the problem.
This can be facilitated through collaborative workshops, prioritisation matrixes and other tools.
Garbage in, Garbage out.
Data is central to any IT system and as it gains context and utility it becomes Information. If the quality of the data is poor, it doesn’t matter what you do with it, it is “Garbage in, garbage out”. So a premium is placed on ensuring data is relevant, timely, accurate, coherent, interpretable and accessible.
Using standardised terminology, such as SNOMED-CT is an important way to ensure data quality. It enables powerful capabilities such as decision support, and analytics - not to mention individual patient and provider benefits.
All too often healthcare, clinicians name the same thing in five different ways. This means when it comes to using this data, it’s a mess.
The use of standardisation allows for more accurate data collection ensuring that clinical records a more usable and better individual and population decisions can be made.
Data has a lifecycle. This means consideration is placed on the planning, obtaining, storing, retrieving, sharing, maintaining, applying and disposing of data. Being that the data is health data, the sovereignty, privacy and security of the data also is essential for maintaining the integrity and trust in the system.
Storage and Retrieval.
Rather than just pouring all the data into a big bucket, data is often compartmentalised into distinct information layers to improve its utilisation. This is because specific uses relate to how data is stored and retrieved.
One example of this type of schema or architecture has four layers.
First, the foundation or source of the record layer. This is where key data is collected once - such as name or date of birth, and then reused during the user’s journey in the system. This is so information here can be stored once and used multiple times. No more repeating your name each time you enter the system.
Next is the information exchange layer. This is the connector of all other layers through standardised application programme interfaces (APIs) with a common standard called FHIR (Fast Healthcare Interoperability Resources). This is key, as it standardises how the data is structured, and therefore shared and accessed via secure interfaces within your systems and outside your system. Publishing these standards publically allows for a more joined-up digital healthcare ecosystem.
Next, you have the analytics layer, where population health, operations, medical research, resource planning and strategy can be done.
And finally, we have the engagement layer. This is where the patients and clinicians typically experience and navigate the system. Here user interfaces and user experience are of particular importance.
An interesting and popular open standard specification is called openEHR. This offers a paradigm shift moving to a more flexible data-centric architecture, dropping silos and enabling a way to maintain data throughout clinical systems.
Developing and Implementing Solutions.
The development of new solutions can vary in its approach, from sequential waterfall to the iterative agile approach. Both have strengths and weaknesses. Agile has become a favoured approach due to its ability to deliver value early and respond to customer needs responsively.
For implementation, this too has many different strategies. But critical to the success of any implementation is a thoughtful change management process that builds trust and empowers users.
Navigating Paradoxes.
There is an inverse relationship between the depth of understanding of the digital requirements and the flexibility to respond to them. This means that early on in projects when defining what the system needs, there can be a lot of freedom. However, as the development progresses earlier decisions ‘lock in’ the degrees of freedom later in the process.
Often this is not noticed by stakeholders until it starts to but up against the technical, budgetary, time and resource constraints.
Another common challenge is how the benefits of a new system may accrue at different levels. Just as it sounds, benefits in systems may only apply to different stakeholders of an organisation. This can be a cause of conflict. For example; a new feature that automatically logs clinicians out of systems to enhance privacy and security may harm clinical workflows in emergencies, creating workarounds.
Summary.
Getting these things right (or wrong) can be a major factor in the success or failure of IT projects in healthcare.
As healthcare continues its digital evolution, these concepts and skills will become increasingly essential for modern clinical leaders and innovators.
https://www.beckershospitalreview.com/ehrs/is-epic-s-dominance-good-for-healthcare.html
https://www.fastcompany.com/90825288/heres-why-big-tech-has-failed-to-disrupt-healthcare
https://www.sciencedirect.com/science/article/abs/pii/S1386505622001824
https://techcrunch.com/2023/08/31/the-fall-of-babylon-failed-tele-health-startup-once-valued-at-nearly-2b-goes-bankrupt-and-sold-for-parts/
https://kffhealthnews.org/news/death-by-a-thousand-clicks/
https://www.newyorker.com/magazine/2018/11/12/why-doctors-hate-their-computers