Clinical Decision Support Systems (CDSS)

Clinical Decision Support Systems (CDSS)

Learning Outcomes

• Recognize the value of Clinical Decision Support Systems (CDSS) for improving patient care.
• Describe how a typical CDSS works.
• Suggest situations where CDSS are likely to be helpful, and unhelpful.
• Be aware of how badly designed CDSS can hamper good patient care.

Definition of CDSS

• "Clinical decision support systems link health observations with health knowledge to influence health choices by clinicians for improved health care."

How CDSS Works

• Utilizes computer algorithms.
• Executes rules based on these algorithms.
• Prompts clinicians about ways to improve patient care.

Pyramid of Evidence

• Illustrates the hierarchy of evidence, from single studies to systems, for informing clinical decisions.
  • Systems
  • Summaries
  • Synopses of Syntheses
  • Syntheses
  • Synopses of Single Studies
  • Single Studies

Example Problem: Blood Pressure Control in Diabetes

• Good blood pressure control is critical for people with diabetes.
• However, it is often forgotten, not measured, or not treated adequately in diabetic patients.

Paper Records vs. Electronic Records

• Paper Records:
  • Example: "3/12/2021 o/e 142/92…"
• Electronic Records:
  • Plain Text: "BP today 140/92, advised to recheck in 2 weeks, increase medication as needed."

Electronic Records (Coded)

• Utilizes structured data for blood pressure readings.

  • Systolic:

    • Property: Pressure
    • SCTID: 271649006

  • Diastolic:

    • Property: Pressure
    • SCTID: 271650006

SNOMED CT Example

• Example: Cannot remember wedding anniversary (finding)
  • SCTID: 247614002

Example Problem (2): Obesity

• Obesity is an important risk factor for many conditions.
• Many weight loss services exist but have low rates of uptake.

Electronic Records (Coded) - Weight

• Weight recorded as a numerical value with units.
  • Example: 96 kg

CDSS Solution for Obesity

• Identify overweight individuals and offer them a referral to a weight management service.

Executing Rules

• Alert Path Diagram:
  • Checks conditions like BMI > 30 after a specific date.
  • Checks if 'Weight management referral' is recorded.
  • Checks if 'Weight management referral declined' is recorded.
  • Shows alert 'Refer for weight management' if criteria are met.

Alerts Examples

• COVID-19: Positive PCR test
• Eligible for Weight Management ES
• Named GP not informed
• 1st MMR vaccination
• Patient on QOF Registers

Prescribing - Drug Information and Warnings

• Example Using Nifedipine:

  • High Severity Warnings:

    • Contra-Indication: Nifedipine is contra-indicated within 1 month of myocardial infarction.

  • Medium Severity Warnings:

    • Caution: Use Nifedipine with caution in diabetes mellitus.
    • Caution: Use Nifedipine with caution in diabetes mellitus when Haemoglobin A1c level - IFCC standardised is high.

Prescribing - Override Warnings

• The system allows overriding warnings with caution.

Prescribing - Best Practice Guidance

• Example: Nifedipine 12-hour modified-release preparations should be prescribed by brand.
• Disclaimer: Healthcare professionals retain full responsibility for deciding treatment.

Prescribing - Display Text

• Example: "Renal Profile not found in the last 12 months, request test."

Prescribing - Protocol Prompt

• Presents a list of items (e.g., Patient guide, Steroid Emergency card) to select and run.

Alert Fatigue

• Too many alerts can lead to alert fatigue which might cause the doctor to dismiss important alerts.

Key Points

• Clinical decision support systems are used widely.
• They aim to improve patient care by integrating health data with computable knowledge.
• The current generation of systems rely on structured health data, execute rules, and prompt clinicians to take evidence-based decisions.
• 'Alert fatigue' and difficult to use systems can lead to CDSS wasting time, confusing clinicians, and potentially worsening patient care.