Analysis of primary care prescription trends in England during the COVID-19 pandemic compared against a predictive model

By John Scott Frazer, Glenn Ross Frazer

Reference: Frazer JSFrazer GR, Analysis of primary care prescription trends in England during the COVID-19 pandemic compared against a predictive model,


The COVID-19 pandemic has resulted in profound changes in patient and prescriber behaviour. Increased health anxiety has been a feature of the pandemic, with fears around leaving the home, social distancing, hand hygiene, and shortages of goods and medications due to panic buying commonplace. In our article entitled “Analysis of primary care prescription trends in England during the COVID-19 pandemic compared against a predictive model”, we attempt to quantify the effects of the COVID-19 pandemic on primary care medication prescribing and dispensing, addressing a gap in the literature. In the United Kingdom (UK), primary care practitioners are the first-line clinicians which most patients encounter, providing initial assessment and treatment of a wide variety of conditions. Over one billion prescriptions per year are signed in the UK by these general practitioners (GPs) in primary care. We sought to analyse how prescribing trends have changed over time, utilising a dataset containing all prescriptions signed and dispensed in the UK by primary care facilities from January 2014 to November 2020. We used the first six years of the dataset to form a predictive model using Holt-Winters forecasting (a time series predictive method based on trends and seasonality), and then compared the observed data from November 2019 to November 2020 with the model to assess the impact of the COVID-19 pandemic on prescribing.

Increased understanding of changes to prescribing habits as a result of COVID-19 will allow primary care practitioners to better anticipate high-demand medications, and to anticipate which patient groups may experience disproportionate hardship due to the pandemic. Increased prescribing of certain groups of medications, such as insulin, immunosuppressant drugs, and medications for long-term respiratory illnesses, may hint that these patient groups experienced disproportionate anxiety over medication procurement during the pandemic. Despite the spike in prescribing of insulins, oral antidiabetic medications were not similarly affected. Respiratory drugs, including many inhaled medications, exhibited dramatically increased prescribing during March 2020, with some evidence from the types of inhalers involved that this was driven primarily by patients with COPD rather than asthma. In addition to this, drugs included in ‘COPD rescue packs’ (doxycycline and prednisolone, for example), also exhibited spikes relative to the model.

With prescribing and dispensing of some medications increasing as much as 50% above forecast levels during the early months of the pandemic, results from such databases may have additional implications for planning of pharmaceutical supply chains, as well as for individual prescribers. Echoes of the societal impacts of the COVID-19 pandemic are also seen amongst our results; reduced travel during the pandemic was manifest in reduction travel vaccine and malaria prophylaxis prescribing, reduced social contact was evidenced by reduced prescribing of contraceptives, and reduced face-to-face GP visits revealed by a fall in injectable medications and local anaesthetic agents. Local anaesthetics decreased 50-80% following the beginning of the pandemic, while the intramuscular vitamin B12 injection decreased dramatically during this period, and was supplemented by a corresponding increase in the oral form of the medication. It is interesting, however, that despite an overall reduction in dispensing and prescribing of medications and devices requiring face-to-face visits, the influenza vaccine was unaffected, and the pneumococcal vaccine experienced record prescribing towards the end of 2020.

Prescribing and dispensing of drugs involved in elderly care were also affected by the pandemic; drugs for treatment of dementia or Parkinson’s disease experienced a slight attrition during 2020 rather than the steady increase that they had been experiencing hitherto. This could suggest an overall choice in this patient group to avoid interaction with healthcare professionals, or could reflect an inability to seek care brought about by reduced public transport services, reduced family support, or fears of leaving the home. We postulate that the slight decrease in prescribing and dispensing of these medications during the pandemic is likely to represent deaths amongst this group. The starkest results were observed in the end-of-life medications, typically prescribed to alleviate symptoms for patients in the last days or weeks of life. Prescribing and dispensing of this drug class spiked dramatically during April 2020, corresponding with the peak number of overall deaths in the United Kingdom. This in particular highlights the profound adversities experienced by community healthcare services during the pandemic, as an increasing number of were cared for at home during their last days of life.

Thus, our results indicate clear shifts in prescribing and dispensing trends during the period of the COVID-19 pandemic. Multiple drugs saw increased prescribing at specific points, while others were clearly reduced. These observed changes could conceivably be due to alterations in the primary care consultation model during this time, with face-to-face visits much reduced, or could indicate changing patient attitudes to healthcare visits or medication stockpiling. In addition, increased prescribing of a multitude of medications highlights groups particularly affected by the pandemic. Of note, prescribing for many long-term health conditions such as type 2 diabetes mellitus, hypertension, and mental health disorders, was not significantly altered relative to the 2020 predictive model, suggesting that the measures taken by pharmacies and primary care practices during this time to maintain care for these patients were effective. Overall, this work highlights possible patient groups who may benefit from targeted follow-up as conditions ease, thus helping to ease the effects of the wake of the pandemic, even for those who did not directly suffer from COVID-19 infection.


Conflict of Interest: None declared

(Visited 168 times, 1 visits today)

Leave a Reply

Your email address will not be published.

Name *

This site uses Akismet to reduce spam. Learn how your comment data is processed.