Polypharmacy prescribing guidance - draft: consultation

15. Appendix D: Case Finding Indicators to prioritise patients for review

Development of polypharmacy case-finding indicators: driving Scottish Therapeutics Utility (STU) reports and decision support integrated with primary care systems

1. Purpose of appendix

This appendix sets out:

a) A definition and description of the polypharmacy case-finding indicators

b) How these indicators work to underpin:

• The Scottish Therapeutics Utility (STU) reports

• The Right Decision Service (RDS) computerised decision support prompts, integrated with primary care systems

c) The methodology used to develop, update and test the polypharmacy case-finding indicators, and the associated recommendations for action

d) The resulting list of indicators and recommendations, mapped to the relevant evidence sources.

1. What are case-finding indicators?

In the context of the national polypharmacy guidance, a case-finding indicator is a set of criteria which enable clinicians to identify individuals potentially at risk of medicines-related harm. Each indicator is associated with:

• A description which summarises the high-risk criteria
• A statement of the type of risk – e.g. bleeding, cardiovascular event
• Recommendations for clinical action

This information about potentially high-risk prescribing is presented to the clinician through two tools that support decision-making:

• The Scottish Therapeutics Utility (STU). STU enables clinicians to run reports which identify all individuals in their practice who match specified indicators. The clinician can then select the highest risk or most urgent cases for immediate medication review, and to inform improvement work. The indicators can also be used as a basis to monitor the effect of polypharmacy medication reviews or other interventions which aim to improve safe and appropriate prescribing.

• The Right Decision Service high risk prescribing decision support system. The RDS is the national decision support platform for Scotland’s health and social care that works in the GP prescribing systems. When a clinician opens a patient record, RDS uses the polypharmacy indicators to present patient-specific alerts to high risk prescribing issues. The clinician can click on the alert to view details of the clinical risk involved, any relevant test results, and recommendations for action. The clinician is able to record decisions, which can then be copied back to the patient record.

3. How are polypharmacy case-finding indicators and recommendations developed?

3.1 Baseline set of polypharmacy indicators (2012-2018)

An initial set of polypharmacy indicators were identified for the 2018 Polypharmacy guidance25 through methodology described in that guidance and in Dreischulte et al. ^[360]

3.1.1 Literature review

A list of candidate criteria was generated based on a structured literature review of UK national clinical guidelines, prescribing advice and safety alerts, supplemented by European or other clinical guidelines, and targeted primary literature review in selected areas.

From this, candidate criteria were identified which either described potentially beneficial medication use (’quality’) or the use of potentially harmful medication/combination of medicines (’safety’).

Candidate ‘quality’ criteria targeted common conditions where there were compelling indications for drug therapy based on UK and European guidelines.

Candidate ‘safety’ criteria initially targeted the medicine groups reported to be most frequently implicated in preventable medicines-related hospital admissions, often identified from systematic reviews and large-scale studies. For each medicine or medicine group identified, a more extensive literature search was conducted to identify person and/or treatment related risk factors. Standard medicines information resources, the primary research literature and published medication assessment instruments - Beers ^[361] and STOPP-START362 - were considered. Safety alerts in the British National Formulary, and the Medicines and Healthcare products Regulatory Agency (MHRA) were examined to identify prescribing less commonly implicated in drug-related hospital admissions but associated with severe harm.

3.1.2 Consensus process – for initial set of indicators25

These candidate criteria were then refined and prioritised by a multidisciplinary expert panel from across primary and secondary care. A four-step consensus process was employed, based on a modified RAND/ UCLA (University College of Los Angeles) Appropriateness Method (RAM).360

3.1.3 Consensus process – for indicators published in 2018 polypharmacy guidance

The criteria published in the original research360 were further refined for the 2018 polypharmacy guidance through a modified Delphi process. This involved five steps:

1. A list of candidate indicators was compiled based on previously published indicator sets

2. In the first round, panel members rated each candidate indicator on a 5-point scale reflecting their level of agreement with the statement ‘It is necessary that a patient triggering on the respective indicator receives a medication review as soon as possible and it would be inappropriate to wait until the next routine medication review’

3. Panel members met in person for a discussion of first round ratings, informed by a presentation of current evidence and guidance, subsequent to which all candidate indicators were rerated

4. Candidate indicators, for which there was disagreement in the second rating round (defined as >30% of panellists agreeing or strongly agreeing and >30% of panellists disagreeing or strongly disagreeing with the statement) were rerated

5. Indicators that achieved a median rating of 4 or higher without disagreement after three rating rounds were accepted as case finding criteria for the polypharmacy guidance.

3.2 Review and updating of baseline indicators

The initial set of indicators was reviewed and updated in 2022-23, and new recommendations for indicators defined. A three-step methodology was adopted:

Step 1. Evidence review

The following sources were checked for up to date evidence for each indicator:

BMJ Best Practice, and DynaMed. These providers carry out comprehensive, systematic and frequent searches for the latest guidelines and primary evidence. All evidence is critically appraised and used as a basis for recommendations for practice. Topics are generally up to date to within three to four months of the date of access. These are therefore key sources of up to date evidence and signpost to relevant health services and professional guidelines for many indicators.

NICE sources - guidelines, technology assessments, Clinical Knowledge Summaries. NICE guidelines and technology assessments are usually referenced within BMJ Best Practice and DynaMed but were also retrieved independently to identify specific recommendations.

SIGN guidelines and Scottish Health Technology Assessments. As for NICE, these are often referenced within the evidence summaries but were also retrieved directly to pull out specific recommendations.

Cochrane systematic reviews and MEDLINE systematic reviews. Any reviews published since the date of the last BMJ Best Practice and DynaMed summaries were identified. It was considered unlikely that anything major will be identified in this stage, due to the comprehensive updating strategies used by BMJ Best Practice and DynaMed.

Key medicines information sources:

• Scottish Medicines Consortium
• European Medicines Consortium (EMC) Statements of Product Characteristics
• British National Formulary
• Martindale
• Stockleys
• Medicines and Healthcare Products Regulatory Authority (MHRA) safety alerts
• Food and Drug Agency (FDA)

• Beers361 and STOPP-START ^[362] criteria

Cut-off publication date was normally 2016, but sometimes the most recent drug monographs had earlier dates.

Key statements were extracted from evidence sources and a lead knowledge specialist put forward suggestions for updates to the original indicators and recommendations. Key issues and questions were also highlighted at this stage.

Step 2: Expert review and consensus

A multidisciplinary expert panel of sixteen individuals was convened, comprising:

• Clinicians from across primary and secondary care with specialist knowledge in polypharmacy
• Health informaticians with expertise in knowledge and decision support and in the operation of EScro and STU.

For each indicator, one or two lead clinical reviewers were invited to review the evidence retrieved through expert searching and the suggested updates.

Those reviewers then put forward their recommendations for change at a group meeting. Consensus was arrived at through facilitated discussion.

Updated indicators and recommendations post-discussion were re-circulated to the group for final sign-off.

Step 3: Mapping of updated indicators to clinical codes

The original technical specifications which mapped the original indicators to clinical codes (as outlined in Section 5 below) were updated to reflect the amendments to the indicators and the supporting evidence. This updating process highlighted points for clarification within the indicator definitions. Examples included inclusion or exclusion of particular medicines or delivery modes within a given medication class, and how to address gaps in the clinical coding system.

These were reviewed by the clinical lead for the STU team and the original reviewer on the expert panel. Where any uncertainty remained, the proposed response was circulated to the expert panel for sign-off.

Step 4: Technical deployment and testing

The updated indicators and associated recommendations were implemented within EScro. Their operation as decision support alerts was then tested rigorously by deploying the Right Decision Service platform to operate with EScro and the electronic health record systems. Testing was conducted using a suite of dummy patient records corresponding to each indicator.

This testing process highlighted several areas where the indicators needed refined =to generate accurate alerts and STU reports.

Once signed-off through technical testing, the indicators were ready to use in STU reports and in live implementation of the decision support system.

4. List of evidence-based indicators and recommendations

The associated supplementary table provides details of all the indicators, recommendations and underpinning evidence resulting from this process.

5. Technical delivery - how do the indicators work to deliver STU reports and decision support alerts?

A team of pharmacists and pharmacy technicians produce a specification for each indicator. This specification maps each element of the criteria within the indicator to clinical codes used in the primary care patient record – read codes (which in turn map to SNOMED-CT codes) and the Dictionary of Medicines and Devices (DM&D).

The specifications are stored in a database called EScro (Enhanced Services contract reporting options), hosted by NHS National Services Scotland. On a nightly basis, every patient record from the primary care electronic health record systems used in Scotland is downloaded to the local practice version of the EScro database. This is mapped to the polypharmacy indicators and flags those individuals identified as matching any of the high-risk criteria.

Clinicians can access these flagged patients through two routes, as noted in Section 2:

• By running STU reports on indicators of interest
• At point of opening a patient record in EMIS or Vision. Figure 11 illustrates how the various components interact to highlight individuals at risk to the clinician