Study type

Study type

Non-interventional study

Scope of the study

Assessment of risk minimisation measure implementation or effectiveness
Drug utilisation
Effectiveness study (incl. comparative)
Non-interventional study

Non-interventional study design

Cohort
Cross-sectional
Study drug and medical condition

Anatomical Therapeutic Chemical (ATC) code

(R03AK) Adrenergics in combination with corticosteroids or other drugs, excl. anticholinergics
Adrenergics in combination with corticosteroids or other drugs, excl. anticholinergics
(R03BA) Glucocorticoids
Glucocorticoids
(R03BA01) beclometasone
beclometasone
(R03BA02) budesonide
budesonide
(R03BA05) fluticasone
fluticasone
(R03BA07) mometasone
mometasone
(R03BA08) ciclesonide
ciclesonide
(R03BA09) fluticasone furoate
fluticasone furoate

Medical condition to be studied

Asthma
Chronic obstructive pulmonary disease
Population studied

Age groups

Adults (18 to < 46 years)
Adults (46 to < 65 years)
Adults (65 to < 75 years)
Adults (75 to < 85 years)
Adults (85 years and over)

Estimated number of subjects

1000000
Study design details

Main study objective

1. To describe prescription patterns of ICS among patients with respiratory diseases before and during the COVID-19 pandemic. 2. To investigate the association between prevalent ICS use and COVID-19 related outcomes among patients with respiratory diseases. 3. To develop approaches and apply methods of QBA to account for potential biases arising in these estimates of association.

Outcomes

1. SARS-CoV-2 infection (positive test for SARS-CoV-2 in the primary care record sourced from the Second Generation Surveillance System). 2. hospitalisation with COVID-19 (admission to hospital with an ICD-10 code for COVID-19, ascertained using HES data). 3. death with COVID-19 (ICD-10 code for COVID-19 listed as an underlying or contributing cause of death in the ONS death registry).

Data analysis plan

We will conduct descriptive analyses to assess characteristics of the patients in each cohort, stratified by exposure group at baseline. Time to each outcome will be presented using Kaplan-Meier plots using time in study as the time scale. Propensity scores will be generated using logistic regression to estimate likelihood of ICS prescription based on baseline characteristics. All pre-specified covariates will be included in the logistic regression. The association between ICS prescription and each outcome will be estimated using a Cox regression model to calculate hazard ratios and 95% CIs, using time in study as the time scale. Univariable models, models adjusted for age and sex, and propensity score weighted models will be presented. If a time-updated exposure definition is used, we will account for potentially informative censoring using inverse-probability of censoring weights. Propensity scores will also be time-updated confounding.