This article is part of a series: The impact of COVID-19 on medication reviews in English primary care. An OpenSAFELY analysis

In this blog series, pharmacists Chris Wood and Vicky Speed talk about their research on medication review activity including the launch of structured medications reviews during the COVID-19 pandemic.

In this first blog we will describe how we reported medication review activity from clinical coding in primary care on the OpenSAFELY platform and in the second blog we will describe what we found out.

Why did we do this analysis?

We know that the COVID-19 pandemic had a huge impact on healthcare. Our work and work by others has shown that many routine tasks were impacted. We wanted to see whether medication reviews had been affected and if any particular groups had been affected more so than others. We also wanted to see the uptake and prioritisation of the new structured medication review service (SMR) because it was launched during a challenging period in primary care with competing pressures such as the roll out of the first COVID-19 vaccinations, and with lockdown restrictions still in place.

What are medication reviews and structured medication reviews?

Briefly, medication reviews are one of many routine tasks in primary care. NICE describes a medication review as “a structured, critical examination of a patient’s medicines with the objective of reaching an agreement with the patient about treatment, optimising the impact of medicines, minimising the number of medication related problems and reducing waste”. They can be carried out by a range of health care professionals and range in complexity. You can read more about how NICE describes medication reviews here.

Structured medication reviews (SMRs) are also a form of medication review. The new medication review service was launched by NHS England in September 2020. The new service focuses on offering an SMR to patients at greatest risk of harm from their medications. The new SMR service is led at a practice level by pharmacists with the support of the multidisciplinary team.

How did we do this study?

With the approval of NHS England, we carried out our analyses for this project using the OpenSAFELY platform.

What is OpenSAFELY?

OpenSAFELY is a new secure analytics platform for electronic patient records built by our group on behalf of NHS England to deliver urgent academic and operational research during the pandemic. Analyses can currently run across all patients’ full raw pseudonymized primary care records at 40% of English general practices where TPP EHR software is deployed (OpenSAFELY-TPP), with patient-level linkage to various sources of secondary care data.

To find out more about how OpenSAFELY watch our short explainer video to find out more or visit the website to read more about the platform.

How did we measure medication review activity?

Clinical coding of medication review activity

We used the OpenSAFELY platform to work out how many patients had had medication reviews based on routine clinical coding in primary care.

We had to choose which clinical codes to search for to count the frequency of medication reviews. There were a number of factors we had to consider including:

  • Suitability of existing medication review codelists
  • The large array of codes used to record medication reviews
  • The quality of clinical coding in practice.
  • Differences between EHR systems (TPP vs EMIS for example, although our study only included patients registered by practices using TPP)

We decided to take a pragmatic approach. We wanted to build an inclusive codelist that would capture as much medication review activity as possible. We used the parent terms Review of Medication and Medication Review Done and included all child codes beneath these. To help others doing research on medication reviews and to gain a better understanding of the use of clinical codes entered for medication reviews we reported the frequency of code usage.

What did we find?

The most common code was Medication Review Done, accounting for 59.5% of code use, with all other codes individually accounting for less than 5%.

SNOMED CT code n=35,939,595 %
Medication review done (314530002) 21,382,570 59.5
Review of medication (182836005) 1,651,115 4.6
Medication review with patient (88551000000109) 1,504,035 4.2
Medication review done by clinical pharmacist (1127441000000107) 1,440,845 4.0
Medication review done by pharmacist (719329004) 1,322,265 3.7
Structured medication review (1239511000000100) 1,286,160 3.6
Dispensing review of use of medicines (279681000000105) 939,180 2.6
Medication review of medical notes (93311000000106) 884,945 2.5
Asthma medication review (394720003) 844,270 2.3
Medication review without patient (391156007) 730,365 2.0

SMRs were straightforward to count, as we were able to look for the nationally mandated SMR code (1239511000000100).

For more on which codes were used to report medication review activity in this study, check out our paper and the supplementary material.

More Analytic Choices

Timeframes

We needed to pick a time frame in which to report the frequency of medication reviews. As a team we decided to report the results in two ways.

i) Individuals with a medication review coded each month. This measure allowed us to see, in detail, changes that were happening at different points in the pandemic.

ii) Individuals with a medication review coded in the last 12 months. We selected this measure since it is generally accepted that patients who are prescribed medications for long-term conditions should have an annual review.

Sometimes individuals had multiple codes for medication reviews. In this scenario, we used only the latest record to calculate the measure.

Breakdowns - regional, demographic and clinical

We then broke down these counts into regional, demographic and clinical subgroups to see if there was any variation in frequency based on these factors. Using the OpenSAFELY platform, all of the codelists and code for these important breakdowns are available for review and reuse. In particular, existing codelists to extract patients with a primary care record of learning disability, and/or living in a nursing/care home and breakdowns by ethnicity (6-level or 16-level) were of high relevance to this study.

Breakdowns - high risk drugs

We wanted to see the frequency of medication reviews in patients who would be likely to benefit the most/be at most risk of harm from their medicines. We pragmatically identified selected groups of patients prescribed high-risk medications. Patients were reported as prescribed a high-risk medication if they had received 2 or more issues of medication(s) within a subgroup in the previous 12 months. The codelists for each group are available for review and reuse.

High risk drug subgroup Drug class Codelists
Potentially addictive medication High dose long acting opioids
Z-drugs (zopiclone/zolpidem)
Gabapentinoids
Benzodiazepines		 OpenCodelists: High dose long acting opioids (OpenPrescribing) - dm+d OpenCodelists: Addictive medicines
DMARDs Azathioprine
Mercaptopurine
Sulfasalazine
Hydroxychloroquine
Ciclosporin
Methotrexate
Penicillamine
Leflunomide
Mycophenolate mofetil		 OpenCodelists: DMARDs
Teratogenic medications Carbimazole
Sodium Valproate / Valproic acid
Pregabalin
Modafinil
Topiramate		 OpenCodelists: Teratogenic medicines

Summary

We took a pragmatic approach to explore the impact of COVID-19 on medication reviews and structured medication reviews.

You can view all of our analyses and codelists at Github opensafely/medication-reviews: Medication reviews during the COVID-19 pandemic, including structured medication reviews. (github.com). We hope by sharing our code this will help other researchers working in this area in the future.

To read about the results of our study, check out part 2 of this blog series or the paper itself.