Polypharmacy Guidance: appropriate prescribing, making medicines safe, effective and sustainable 2026 - 2029

11. Appendix A: Updating the Drug Efficacy (NNT) Table

To update the NNT table in the 2018 Polypharmacy Guideline, more recent reports of NNT were sought (from 2017 onwards). Where there was not a more recent NNT report, the 2018 reported figures were kept and labelled as “in prior guidance”.

Cochrane SR are defined as the primary source, because of their methodological reliability and reporting of explicit NNT data. When no Cochrane Review provided NNT data, evidence was sought from other systematic reviews. Therefore, the unit of analysis was the report of primary studies from a secondary source (SR and GL) and not trials themselves. The previous version of the NNT analysis used RCT as the primary source, and where there was no updated evidence, these were kept in.

Aim

To update the drug efficacy (NNT) tables with a search from recent evidence (2017 onwards), to provide the relative clinical efficacy of frequently prescribed medicines in primary care.

Methods

Defining the medicine, intervention, and clinical outcome of relevance

We included medicines covered in NNT tables in previous editions of the Scottish Government Polypharmacy Guidance 2015 and 2018 because of their relevance to clinical practice. Reflecting advances in clinical care, we introduced new medicines, for example, for the treatment of diabetes. We have only used the generic name of each medicine. We noted the intervention as reported, i.e. the strength of the dose and quantity per day (if reported). The clinical context in which the medicine was likely to be used tended to determine the outcomes reported; for example, an antiplatelet can be prescribed for primary prevention of a serious vascular event or a secondary prevention following an event.

Search strategy

We carried out an iterative, systematic literature search during January-February 2023 to identify systematic reviews, using the following databases/resources:

• BMJ Best Practice
• Cochrane Database of Systematic Reviews
• Embase
• Google
• Guidelines International Network
• Medline
• National Institute for Health and Care Excellence (NICE) [guidelines]
• NICE Evidence Search, NIHR Journals (Health Technology Assessment, Health and Social Care Delivery Research, and International HTA database)
• Scottish Intercollegiate Guideline Network (SIGN)
• The Knowledge Network
• Turning Research into Practice (TRIP) Medical Database.

Because separate searches were done for each drug class, different search strategies were used. An example of a search strategy for antidepressants was fluoxetine; sertraline; citalopram; duloxetine; venlafaxine; amitriptyline; clomipramine; dosulepin; imipramine; lofepramine; nortriptyline; numbers needed to treat; numbers needed to harm (NNH).

Inclusion criteria for NNT searches

• Population: adults taking two or more medicines (polypharmacy)
• Intervention: the specific medication, e.g. antithrombotic
• Control: standard treatment or placebo
• Outcome:
• Primary: NNT
• Secondary: NNH
• Settings: any
• Study design: Cochrane reviews or other systematic reviews

Exclusion criteria for NNT searches

• Non-English language.
• 2016 or before, because we were seeking to update the 2018 evidence, we looked to find more recent reports of NNTs for these drugs. Where we did not find a more recent NNT report, we kept the 2018 reported NNT, which we have labelled as ‘in prior guidance’.

Study selection and data extraction

One reviewer sifted the titles and abstracts, and then the full texts of the retrieved articles. One reviewer extracted information and data. A second Health Service Researcher independently checked the data extraction and they discussed differences. No quality assessment was conducted because Cochrane Reviews were used and are considered ’gold standard’ for their reliability. For each study, the following information was extracted:

• Medicine/intervention: see above
• Comparator: placebo, no treatment, or control drug
• Condition/population: age, condition(s), at risk of what disease
• Outcome aim to prevent: primary/secondary
• Duration: time during which the studies ran
• NNT: the reported number needed to treat (to benefit)
• Annual NNT
• Participants, study design, numbers, Quality: as reported by the SRs.
• Comments: free text box
• NNH: the reported number needed to harm, if reported

Annual NNT calculation

This is computed this by multiplying the NNT by the number of years for which the studies ran in the review.

Analysis

The relevant information from each review was captured using a modified version of the 2018 Guideline NNT Efficacy Table.