Xiaomin Zhong, Ya-Ting Yang, Dr Ali Fahmi, Dr Victoria Palin and Professor Tjeerd Van Staa from Manchester University have been using OpenSAFELY for analyses focused on the impact of COVID-19 on antimicrobial resistance since 2021. In this guest blog they describe some of their recent papers that have been published in peer review journals.

Overview

Our research team from the University of Manchester has been diligently examining alterations in primary care treatment pathways for common infections using OpenSAFELY platform. We have also analysed how exposure to prescribed antibiotics affects the severity of these commonplace infections as well as COVID-19. Our findings have been recently featured in reputable publications including The Lancet Regional Health – Europe, Eclinicalmedicine, Journal of Infection, and Antimicrobial Resistance & Infection Control.

Understanding Antimicrobial Resistance During the Pandemic

Antimicrobial Resistance (AMR) arises when microbes adapt, rendering standard treatments ineffective. This resistance prolongs illnesses, elevates healthcare expenses, and increases mortality risks, particularly during a pandemic. Responsible and limited use of antimicrobials is crucial to mitigate the emergence of resistant strains, thereby safeguarding public health now and for future generations.

Our Research Focus

We aimed to discern how COVID-19 has modified primary care approaches to treating prevalent infections, including respiratory and urinary tract infections, coughs, and colds. Our research also explores factors exacerbating these changes, such as socio-economic disparities and variations in practices.

Furthermore, we used the OpenSAFELY-TPP data to construct a series of risk prediction models, offering personalised risk scores calculated based on attributes like age, gender, clinical and medication risk factors, ethnicity, and socioeconomic status. These refined prediction models were integrated into the Knowledge Support System (KSS). The KSS provides individualised information, including the risks of developing infection-related complications leading to hospital admission, resistance based on the number of antibiotics prescribed in the past year, the probability of antibiotic failure necessitating another prescription within 30 days, and severe outcomes like renal failure admissions.

Additionally, we have investigated how past antibiotic exposure - considering types and dosage - affects COVID-19 severity and outcomes.

How did we do these studies?

Our studies encompass various observational methods, including cohort and matched case-control studies, involving approximately 24 million patients in England registered with practices using the TPP (SystmOne) software. We analysed 79 antibiotic types, as defined by the NHS Dictionary of Medicines and Devices (dm+d) and identified based on the British National Formulary’s Chapter 5.1.1 (Antibacterial drugs).

What Did We Find Regarding Infection-Related Consultations?

We observed a substantial decline in the rate of infection-related consultations, with a mean reduction of 39% in 2020 compared to the mean rate in 2019. Interestingly, the rate of consultations for urinary tract infections (UTIs) remained stable. Lower respiratory tract infections (LRTIs) experienced the most significant reduction, underscoring the varied impact on different infection types.

How Have Antibiotic Prescriptions Changed?

Our research found shifts in antibiotic prescribing patterns, with over 32 million prescriptions analysed. Initially, there was an increase in broad-spectrum antibiotic prescriptions, accounting for 8.7% of the total. Despite a subsequent decrease in overall antibiotic use during the pandemic, the broad-spectrum prescriptions increased initially, highlighting an urgent reliance on these drugs during the pandemic’s early stages.

How Do Socio-Economic Factors Influence Prescriptions?

A deeper look into the data revealed that patients from more deprived backgrounds were more likely to receive broad-spectrum antibiotics. This discrepancy in prescribing patterns, influenced by socio-economic factors, remained consistent over time, warranting further exploration and address to ensure equitable healthcare provision.

What About Repeat and Inappropriate Prescriptions?

Approximately 29.1% of all antibiotic prescriptions were repeats. Individuals with incident infections coded on the same day in electronic health records (EHRs) showed considerably lower rates of repeat prescriptions (18.0%). Moreover, 8.6% of prescriptions were potentially inappropriate choices for incident infections, as identified against the most recent guidelines provided by NICE and PHE.

Is There a Link Between Antibiotic Exposure and COVID-19 Severity?

Our findings indicate a concerning correlation between the frequency and diversity of antibiotic exposure and adverse COVID-19 outcomes. Patients with more frequent antibiotic exposure in the past three years faced higher odds of severe COVID-19 outcomes, including hospital admission and mortality within 30 days. Moreover, exposure to a diverse range of antibiotics was linked to higher rates of COVID-19 hospital admissions. In particular, patients with a history of extensive antibiotic use were 1.8 times more likely to be hospitalized, emphasizing the need for careful consideration of antibiotic prescribing practices.

What Implications Do These Findings Have for Future Practice?

These findings are not only pivotal for understanding the immediate implications of antibiotic use during the COVID-19 pandemic but also crucial for future antibiotic stewardship efforts. By gaining insight into the adverse effects of frequent and diverse antibiotic use, healthcare professionals can better navigate prescribing practices to mitigate risks associated with both antibiotic resistance and severe COVID-19.

Get in touch

All of the code used for our analyses is open and reusable so that other researchers interested in similar work can readily adopt these methods in their own work. If you have more questions about any of our research please get in touch via xiaomin.zhong@manchester.ac.uk or @BillyZhong229.