You are on the Natcen site

Click here for Scotcen

natcen map

You are on the Natcen site

Click here for Scotcen

natcen map

Prescribing patterns in dependence forming medicines 2000-2015

Published: September 2017

This research uses national GP data to look at the number and characteristics of patients being prescribed dependence forming medicines.


This study analysed prescribing data to examine the:

  • Prevalence of and trends in primary care prescribing of four types of medicines with the potential to be dependence forming (DFM): benzodiazepines, Z-drugs, opioids and GABAergic medicines. 

  • Characteristics of the prescriptions issued: the type of medicine, the number of days it was prescribed for, and the dose. 

  • Characteristics of people prescribed to: their age, sex, region, deprivation level of their neighbourhood,
and the symptoms and diagnoses present when the medicine was prescribed. 


A report has been published on the Public Health Research Consortium's website. 


  • DFM are widely prescribed in primary care. In 2015, opioids were prescribed to 5% of all patients on the Clinical Practice Research Datalink (CPRD). Opioids were about twice as likely to be prescribed as benzodiazepines (2%), Z-drugs (2%), or GABAergic medicines (2%).
  • Overall, DFM prescribing has been increasing: 6% of patients on the CPRD were prescribed at least one of these four types of DFM in 2000; the rate in 2015 was 9%. Benzodiazepines were the only DFM with a long-term fall in prescribing (3.5% of patients in 2000, to 2.5% in 2015).
  • There has been an upward trend in how long DFM are prescribed to people for, especially for opioids which increased from 64 days in 2000 to a peak of 102 days in 2013/2014.
  • Benzodiazepines, Z-drugs, and opioids have tended to be prescribed for longer to people living in the most deprived neighbourhoods.
  • Clinical guidance recommending more cautious and time limited prescribing of benzodiazepines may well have been effective in changing practice, with reductions in the extent and length of benzodiazepine prescribing. Close and ongoing monitoring of opioid prescribing is needed.


We used a sample of 49,999 patients prescribed at least one of four types of DFM between 2000 and 2015. This was drawn from the Clinical Practice Research Datalink (CPRD), and included information from over 13 million consultations.

The CPRD contains data about prescriptions issued by GPs (including the length and size of prescription) and characteristics of the patients prescribed to (such as their age, sex, and area where they live).

A profile of the whole CPRD population was also used, so that overall rates of prescribing could be produced.

This was independent research commissioned and funded by the National Institute for Health Research (NIHR) Policy Research Programme (PHPEHF50/14). The views expressed in the publication are those of the authors and not necessarily those of the NIHR or the Department of Health. Information about the wider programme of the PHRC is available from