Antibiotic treatments 'fail' 15% of the time

“Antibiotic treatments from GPs 'fail 15% of the time’,'' BBC News reports. In one of the largest studies of its kind, researchers estimated that just under one in seven antibiotic prescriptions in 2011 "failed".

This study examined the failure rates of antibiotics prescribed by GPs in the UK for common infections over a 21-year period – from 1991 to 2012. Most of the failures (94%) were cases where a different antibiotic needed to be prescribed within 30 days, suggesting that the first antibiotic had not worked.

In general, the overall failure rate remained fairly static over the course of three decades; 13.9% in 1991 only increased to 15.4% by 2012.

When considering specific types of infection in combination with specific classes of antibiotics, there were notable changes in failure rates. For example, when the antibiotic trimethoprim was prescribed for an upper respiratory tract infection, failure rates increased from 25% in 1991 to 56% in 2012. Reassuringly, failure rates with commonly prescribed antibiotics (such as amoxicillin) currently remain fairly low.

The study did not look at the reasons for antibiotic failure, but one reason could be antibiotic resistance – an increasing problem worldwide.

If you are prescribed an antibiotic, you can increase the chances of it working and decrease the risk of antibiotic resistance by ensuring that you take the full course as prescribed by your GP, even when you start to feel better.

Where did the story come from?

The study was carried out by researchers from Cardiff and Oxford universities, and Abbott Healthcare Products in the Netherlands, who also funded the study.

The study was published in the peer-reviewed British Medical Journal (BMJ) on an open access basis, so it is free to read online.

While the overall reporting by the UK media was broadly accurate, many of the headlines were not.

The Daily Telegraph claimed that “Up to half of antibiotics 'fail due to superbugs'”. 

We don't actually know the reason for needing another antibiotic prescription, as this was not examined in this study. Therefore, we don't know that any of these apparent antibiotic failures were due to “superbugs” as no laboratory data was available.

The Daily Mail claims that, “Now one in seven patients cannot be cured using antibiotics”, which is also not correct. It could well be the case that many patients were cured through the use of alternative antibiotics.

What kind of research was this?

This study examined the failure rates of antibiotics prescribed by General Practices in the UK over a 21-year period – from 1991 to 2012. Antibiotic resistance is a problem that has been increasing over the past few decades. As the World Health Organization (WHO) has declared, this is becoming worldwide public health crisis, as previously effective antibiotics become ineffective at treating certain infections. Though many people may think of antibiotic resistance as a problem predominantly found in hospital care (e.g. patients becoming ill with resistant “superbugs”), resistant bugs are just as much a problem in the community. As the researchers say, recent antibiotic treatment in primary care puts a person at risk of developing an infection that is resistant to antibiotics.

This study used a large general practice database to assess the failure of first-line (initial) antibiotic treatments prescribed in the UK over a 21-year period, alongside looking at general antibiotic prescription patterns.

What did the research involve?

This study used the UK Clinical Practice Research Datalink (CPRD) – an anonymised database collecting data from more than 14 million people attending almost 700 general practices in the UK. The database contains well-documented medical records and information on prescriptions, and these were examined between 1991 and 2012.

The researchers decided to look at antibiotics prescribed for four common classes of infection:

  • upper respiratory tract infections (e.g. sore throats, tonsillitis, sinusitis)
  • lower respiratory tract infections (e.g. pneumonia)
  • skin and soft tissue infections (e.g. cellulitis, impetigo)
  • acute ear infection (otitis media)

They looked at whether these infections had received treatment with a course of a single antibiotic (termed monotherapy, rather than two antibiotics in combination, for example). An antibiotic was considered as the first-line treatment if there had been no prescriptions for other antibiotics in the preceding 30 days.

They assessed the proportion of antibiotic courses resulting in treatment failure. As the researchers say, there is no specific definition of treatment failure, but based on previous research findings they considered treatment failure as:

  • prescription of a different antibiotic within 30 days of the first antibiotic prescription
  • GP record of admission to hospital with an infection-related diagnosis within 30 days of  prescription
  • GP referral to an infection-related specialist service within 30 days of prescription
  • GP record of an emergency department visit within three days of prescription (the shorter time window being selected to increase the probability that the emergency was related to the infection, rather than another cause)
  • GP record of death with an infection-related diagnostic code within 30 days of prescription

For each year, from 1991 to 2012, the researchers determined antibiotic treatment failure rates for the four infection classes and overall.

What were the basic results?

The database contained records of almost 60 million antibiotic prescriptions prescribed to more than 8 million people.

Almost 11 million prescriptions were the first-line single antibiotic treatment of the four groups of infection being studies: 39% for upper and 29% for lower respiratory tract infections, 23% for skin and tissue infections, and 9% for ear infections.

Overall, GP consultation rates for the four common infection groups decreased over time, but the number of consultations for which an antibiotic was prescribed marginally increased: 63.9% of consultations in 1991 and 65.6% in 2012. Across the whole 21 years, the proportion of consultations where an antibiotic was prescribed was 64.3%. However, within infection groups, there were more significant changes: prescriptions for lower respiratory tract infections decreased (59% in 1991 to 55% in 2012) while those for ear infection went up considerably (63% in 1991 to 83% in 2012).

The most commonly prescribed antibiotics were amoxicillin (42% of all prescriptions), and most upper respiratory tract infections received this antibiotic.

Most antibiotic treatment failures (94.4%) were cases where an alternative antibiotic had been prescribed within 30 days of treatment.

The overall antibiotic treatment failure rate for the four infection classes was 14.7%. The rate was 13.9% in 1991 and 15.4% in 2012, but there was not a clear linear increase in the rate over the time period. For each year, the highest failure rates were seen for lower respiratory tract infections (17% in 1991 and 21% in 2012).

Within the infection classes, individual antibiotics were associated with different failure rates. There were some particularly high rates of failure. For example, when the antibiotic trimethoprim (most often prescribed for urine infections) was prescribed for an upper respiratory tract infection, it failed 37% of the time overall, increasing from 25% in 1991 to 56% in 2012. For lower respiratory tract infections, failure rates were highest for a group of broad-spectrum antibiotics called cephalosporins (including antibiotics like cefotaxime and cefuroxime), with failure rates increasing from 22% in 1991 to 31% in 2012. 

In 2012, despite its high prescription rate for upper respiratory tract infections, amoxicillin had quite a low failure rate (12.2%).

How did the researchers interpret the results?

The researchers conclude that, “From 1991 to 2012, more than one in 10 first-line antibiotic monotherapies for the selected infections were associated with treatment failure. Overall failure rates increased over this period, with most of the increase occurring in more recent years, when antibiotic prescribing in primary care plateaued and then increased”.


Overall, this is a highly informative study of GP antibiotic prescribing for common infections in the UK. The overall antibiotic treatment failure rate was 15% over the course of the study period; these were mainly cases where there was a need to prescribe a different antibiotic within 30 days. There was a slight increase in failure rate, from 13.9% in 1991 to 15.4% in 2012. Within the infection classes, particular antibiotics had notable changes in failure rates, while others remained fairly stable. Reassuringly, amoxicillin and other commonly prescribed antibiotics currently still have fairly low failure rates.

However, despite this study using a wealth of data from a reliable GP database, there are some limitations to bear in mind.

Importantly, as the researchers say, there was no specific definition of treatment failure for them to use, so they had to use various proxy measures. They had no laboratory data available on the resistance of organisms to different antibiotics, so the study is not able to definitely say that antibiotic resistance was the reason for treatment failure. The most common indication of “treatment failure” in this study was the need for prescription of another antibiotic within 30 days, but it may not mean that the organism was resistant to the first antibiotic. – e.g. the person may not have taken the full prescribed treatment course, or the antibiotic may not have turned out to be appropriate for the type of bacteria the person had.

There is also the possibility of incorrect coding within the database, or the antibiotic not being prescribed for the indication that it was assumed to be.

However, antibiotic resistance is an increasing global problem, and is likely to have contributed to the failure rates. As a patient, it is important to be aware that many common respiratory infections can be self-limiting viral infections that do not need an antibiotic. If you are prescribed an antibiotic, you can help decrease the risk of the bug being developing resistance to the antibiotic by ensuring that you take the full course as prescribed by your GP, even when you start to feel better.

NHS Attribution