Anxiety is a common problem in the UK, and medication can be prescribed by GPs to help patients with their symptoms. Antidepressants are the main recommended drug treatment for anxiety, although other drugs such as benzodiazepines, antipsychotics or beta-blockers, may be used. However, benzodiazepines, antipsychotics and beta-blockers can also be prescribed for other things. For example, antipsychotics may be prescribed for patients with psychosis, and beta-blockers (such as propranolol) may be prescribed for hypertension. Between 2003 and 2018, the number of antipsychotic and beta-blocker prescriptions for patients with anxiety increased. During the same period, benzodiazepine prescriptions also increased in young adults with anxiety, and 44% of prescriptions in 2017 (for all adults) were longer than the NICE recommended four weeks. Although all of these patients had anxiety, we do not know how many of these prescriptions were actually for the treatment of anxiety, and how many were for the treatment of other conditions. Therefore, our aim is to investigate how many patients with anxiety have been prescribed benzodiazepine, antipsychotic or beta-blocker treatment specifically for their anxiety symptoms.
This study will investigate prescribing of benzodiazepines, antipsychotics and beta-blockers for the indication of anxiety, between 2003-2018. The study will use a retrospective cohort design, including patients aged 18+.
For the benzodiazepine analysis, patients will be excluded if they ever have a recorded epilepsy or muscle spasticity code, or an alcohol withdrawal code in the six months before the prescription. For the antipsychotic analysis, patients will be excluded if they have a psychosis code in the five years before the prescription. A sensitivity analysis will exclude those who ever have a psychosis code. For the beta-blocker analysis, patients will be excluded if they have a neurological code in the year before the prescription, with a sensitivity analysis excluding those where it occurs in the three years prior. Again, for the beta-blocker analysis, patients will be excluded if they have a cardiovascular code in the five years before the prescription, with a sensitivity analysis excluding those who ever have a cardiovascular code.
The prescription must have occurred within the 3 months before an anxiety code date, or the 6 months afterward.
For each drug class, we will calculate the number of patients who received at least one benzodiazepine, antipsychotic or beta-blocker prescription (monotherapies and combination treatment), and the number of patients who started such medication in each calendar year. Person-years-at-risk will be the denominator. Incidence rates and 95%CI will be calculated using Poisson regression. Data will be plotted to examine patterns of prescribing over time. Changes in trends over time will be examined using joinpoint regression. Data will be stratified by age and gender. We will calculate average treatment duration for incident cases to examine whether long-term prescribing has changed. This work will enable understanding of how prescribing trends have changed over 16-years for the large number of patients with anxiety.
This project will explore trends in benzodiazepine, antipsychotic and beta-blocker prescribing for anxiety, in the UK between 2003 and 2018.
Charlotte Archer - Chief Investigator - University of Bristol
Charlotte Archer - Corresponding Applicant - University of Bristol
David Kessler - Collaborator - University of Bristol
Nicola Wiles - Collaborator - University of Bristol
Stephanie MacNeill - Collaborator - University of Bristol