(085) Inhaled corticosteroid prescription patterns in UK primary care before and during the COVID-19 pandemic in patients with COPD

Publication Number: 422

Background: Researchers make assumptions about how patients use medication to ascertain exposure using prescription data from electronic health records. Risk of exposure misclassification increases when irregularities such as stockpiling occur or when consulting patterns are affected by population-level healthcare disruptions, both of which increased during the COVID-19 pandemic.

Objectives: To describe prescription patterns of inhaled corticosteroids (ICS) in patients with chronic obstructive pulmonary disease (COPD) between March 2019 - March 2021 in the UK and evaluate the impact of different discontinuation definitions on exposure ascertainment.

Methods: We selected patients aged ≥35 years with COPD and ≥12 months’ registration history from the Clinical Practice Research Datalink (CPRD) Aurum to construct two cohorts followed until 12 months after entry or death (whichever came first) from 1 March 2019 and 1 March 2020 respectively. Patients with asthma within 3 years before or another chronic respiratory disease at any point before this date were excluded.
We examined measures of ICS use for each calendar month. The denominator was the number of people with COPD by the 15th each month. We calculated the monthly total of ICS prescriptions, the proportion of people with prevalent ICS exposure ( < 6 months since last ICS prescription), the proportion of people initiating ICS (new ICS prescription using 365-day washout period), and the number of people discontinuing ICS (not receiving a new ICS prescription (i) within 6 months of the last prescription date, (ii) within 60 days of the end of the calculated exposure period).

Results: Each cohort included around 175,000 people [2019 cohort: median age: 71.7 (IQR 62.7 - 78.7), 53.8% male; 2020 cohort: median age: 70.7, (IQR 62.7 – 78.7), 53.6% male]. The rate of ICS prescribing was higher in March 2020 (37.3 per 100 person-months) than in March 2019, February 2020, or February 2021 (29.7, 29.3 and 29.7 per 100 person-months respectively). The proportion of people estimated to be prevalent ICS users each month stayed relatively stable in the year before the pandemic and increased steadily during the pandemic. There was a spike in initiations in March 2020 (0.66 %) compared with March 2019 (0.49%) and February 2020 (0.42 %), followed by a sharp drop in initiations which remained low during the pandemic. Discontinuations decreased early in the pandemic compared with the months before the pandemic and increased in September 2020 (0.82 %) when defining discontinuations as no new prescription within 6 months of the last ICS prescription. An increase in discontinuation was seen in July 2020 when defined as not receiving a new ICS prescription within 60 days of the end of the calculated exposure period (1.72%).

Conclusions: We observed a small spike in ICS prescriptions and initiations in March 2020. However, evidence of significant stockpiling was limited.