Skip to main content

Clinical Decision Support & Clinical Informatics by Sarah Rosetti

An overview of some rather 'timeless' foundational principles, beginning with Learning Health Systems. An loop between Knowledge, Practice, and Data. This is essentially (well, roughly) the Agile Methodology applied to health systems. Tight iteration cycles, ability to pivot to better interventions and policies, and so on. This is monstrously hard in healthcare in general.

Clinicians are busy and burdened in general but they get really tuckered out documenting patient care1!

Clinical Decision Support (CDS)​

(Clinical Knowledge + Patient Info)
  -> Filter Intelligently
  -> Present at appropriate time = Enhanced Patient Care 😎

Why would you do this?

  • Nudge clinican to engage in desirable behaviours ("don't order that, order that instead (it's cheaper)")
  • Patient engagement
  • Guide treatment
  • PopHealth Improvements
  • Improve Knowledge 🧠

Some commonsensical things apply when you design a CDS. Transparency, privacy, communication, usability, all that good stuff.

We want a CDS to "enhance decision-making" and "drive evidence-based practice". You do this by thinking about how you go from Data -> Information -> Knowledge -> Wisdom 🧘‍♂️

Basic CDS Things​

Examples:

  • Some trigger that causes a Decision Support Rule to be invoked.
  • Input data: The data used by the Rule to make inferences.

I asked how the rules are encoded (e.g. "Don't prescribe these two drugs at once"). This is configuration mostly and not code. It's done by analysts who work at the hospital who work with experts to encode the rules. Note that Epic provides the software foundation for these configurations/encodings to be expressed. The analysts mentioned simply use Epic's foundation.

Five Rights of CDS​

  • Right Information
  • Right Person
  • Right Intervention (e.g. Alerts, Order Sets, Protocols)
  • Right Channel (e.g. EHR, Patient Portal, App)
  • Right Point in Workflow

👉 Find the right information at the right time for the right clinician to make the right decision. And whither the patient in this chain? Well, patient-centered CDS (the hallowed goal) involves some agency of the patient!

The Ten Commandments for Effective Clinical Decision​

See this paper.

  1. Speed is everything.
  2. Anticipate needs and deliver in real‐time.
  3. Fit into the user’s workflow.
  4. Little things can make a big difference.
  5. Recognize that physicians will strongly resist stopping.
  6. Changing direction is easier than stopping.
  7. Simple interventions work best.
  8. Ask for additional information only when you really need it.
  9. Monitor impact, get feedback, and respond.
  10. Manage and maintain your knowledge‐based systems.

Grand Challenges + CDS Checklist​

There are a lot of them. Refer to slides. Plenty of opportunity 🤩 Make sure you see the "CDS Exercises" to get your juices flowing (and to make sure you're doing a 'proper' CDS).

Miscellaneous​

Here's a graphical slide about how "interrupt-driven" clinicans daily schedules are. It's pretty crazy. Last thing you want for these folks to struggle with CDS's and their interfaces.

It's vital to remember that you need to see this in a 'wholesome' way. A fantastic UI will not save terrible information architecture.

EHR data is very noisy! It's a "flipped iceberg" where there's more Noise than Signal. People and systems capture a lot of data for regulatory or FOMO reasons.

Any tool takes a year to a year-and-a-half to go live! This is not Tech; calibrate your expectations.

Read this.

Footnotes​

  1. Based on a TrenBurden survey in April 2024. ↩