Polypharmacy prescribing guidance - draft: consultation

17. Appendix E: Monitoring the effect of Polypharmacy medication reviews

The ultimate aims of polypharmacy medication reviews are to reduce drug-related harm and to achieve therapeutic objectives in line with individuals’ preferences, rather than simply reducing the numbers of medicines patients are taking.

The implementation of polypharmacy reviews can be monitored from both process measures and clinical outcome measures.

1. Process measures

To facilitate evaluation on the impact of polypharmacy reviews and outcomes, an electronic template is available for both VISION and EMIS that can be used in the consulting system and will automatically read code the reviews and the interventions made. This can be found in Appendix O with a link to instructions on how to implement this.

When reviews are undertaken without the above template, the reviews should be digitally recorded, adding read code 8B31B for primary care records.

Monitoring the number of individuals with this read code will support evaluation of the implementation of polypharmacy reviews.

2.1 Clinical Outcome measures

When considering polypharmacy reviews, establishing whether clinical outcomes are attributable to drug therapy or other underlying causes is not realistically possible at scale. Monitoring of the effect of polypharmacy medication reviews therefore requires the use of proxy outcome measures that can be implemented in routine data sources available at national level. These measures fall into two categories: drug utilisation and hospital admissions.

2.2 Clinical Outcome - drug utilisation measures

It is recommended that the high-risk medication Case Finding indicators listed in Appendix D1 of this guideline are used as a basis to monitor the effect of polypharmacy medication reviews. Given the large number of indicators, it is impractical to consider each indicator separately. The following strategies can be used to reduce the number of drug utilisation measures used:

• Measure the average number of high-risk medication Case Finding indicators triggered per person in the target population (as defined in Section 4.18 of this guideline, with further detail in Appendix D1)
• Measure the proportion of patients triggering on any high-risk Case Finding medication indicator (overall composite)
• Measure the proportion of patients triggering on any high-risk Case Finding medication indicator targeting the same adverse event (event specific composites)

Since 2018, Scotland has used a suite of 17 quality and safety indicators to monitor the appropriateness of polypharmacy. For this updated guidance, these have been peer reviewed with evidence and there are now 27 national indicators. These have been developed to include a sub analysis for care home residents. The full suite is found in Appendix C. These indicators provide prescribing measures, which are closely related to the high-risk medication Case Finding indicators and may be used to monitor Clinical Outcomes. A clear advantage of taking this approach is that they are mostly already built into the GP clinical systems, prescribing support tools and national dashboards.

The NTI dashboards identify where GP practices, clusters or health boards are in relation to key prescribing indicators for polypharmacy. The three-point timeline allows benchmarking of cluster or health board region against similar clusters, or the Scottish average, in a range of prescribing areas. Practices can compare their data to others in the cluster or similar practices to help account for demographic or regional variation in practice. One of the new indicators developed is the mental health triple whammy, shown below

National Therapeutic Indicators and dashboards are online with open access.

2.3 Clinical Outcomes - hospital admissions reduction measures

Hospitalisation data is routinely collected at national level in the Scottish (SMR01) record. Although the outcome of interest are hospital admissions that are explicitly drug-related (which is unfeasible to measure at national scale), for some types of hospital admissions, a drug-related aetiology may be sufficiently common to attribute the admission to that cause. Improvements in drug utilisation may be reflected in an overall reduction in these admissions among those targeted for polypharmacy medication reviews. Although the specific criteria used by different health boards to prioritise patients for review may differ, it is suggested that outcomes are measured among the subpopulation of patients aged 75 years older. Since the case finding criteria (nursing home residents, those on 10 or more drugs and those triggering high-risk medication indicators) are most commonly met in this subpopulation, and any effects of polypharmacy medication reviews are therefore likely to be most visible. The following Clinical Outcome hospital admission measures are recommended:

Proportion of patients 75 years or older with an emergency admission for:

• gastro-intestinal bleeding
• bleeding of any cause
• heart failure
• acute kidney injury
• falls and fractures
• stroke
• delirium
• clostridium difficile infection
• hypoglycaemia
• hyperglycaemia
• asthma
• COPD

Combining the measurement of specific types of hospital admissions with drug utilisation patterns may enhance the interpretation of any observed changes. For example, if a reduction in hospital admissions for gastro-intestinal bleeding was accompanied by a reduction in the prevalence of patients triggering high-risk medication use indicators targeting gastro-intestinal events, this would increase the confidence that the observed changes in clinical outcomes are attributable to improved medication use processes.

2.4 Clinical Outcomes - undesirable increase in specific hospital admissions- balancing measures

All plausibly beneficial health care interventions have the potential to have unintended consequences. Where such potential consequences can be identified, it is good practice to measure them to enable a balanced accounting of intervention effects. As part of polypharmacy medication reviews, patients and practitioners are encouraged to have informed discussions about omitting, discontinuing or de-intensifying prophylactic treatments (such as blood pressure lowering or antidiabetic treatment) that have doubtful benefits over the patient’s likely remaining life span. The following balancing measures are therefore recommended to provide reassurance that reviews do not adversely impact on the incidence of cardiovascular events:

Among patients 75 years or older:

• the proportion of patients with an emergency hospital admission for myocardial infarction
• the proportion of patients with an emergency hospital admission for stroke
• the proportion of patients with an emergency hospital admission for diabetes/hyperglycaemia

2.5 Clinical Outcomes - reductions in all cause health care utilisation

In addition to reductions in specific hospital admissions it is possible that polypharmacy medication reviews also impact on unscheduled health care utilisation more generally. However, ‘all cause health care utilisation’ may be more commonly influenced by non-drug-related causes and it is therefore likely to be a measure that is less responsive to the impact of polypharmacy medication reviews than hospital admissions for commonly drug-related causes. Nevertheless, the following outcome measures may usefully supplement the outcome measures specified in sections 2.1 and 2.2:

Among patients 75 years or older:

• the proportion of patients with an emergency hospital admission of any cause
• the number of unscheduled occupied bed days
• the proportion of patients discharged into dependent care