This site is intended for health professionals only

At the heart of general practice since 1960

pulse june2020 80x101px
Read the latest issue online

GPs go forth

GPs 'overestimate harms and benefits of drugs', say researchers

GPs may be overestimating the risks and benefits of different drugs for long-term illnesses, meaning they are relying on their broad understanding rather than detailed knowledge of different treatments, a study has found.

The Oxford-based researchers said that the discrepancies are of a magnitude that is likely to ‘meaningfully’ affect clinical decision-making. 

The study, published in the British Journal of General Practice, interviewed just under 450 GPs about the effects of different drugs for conditions like osteoporosis, diabetes and atrial fibrillation.

They found that almost 90% of GPs overestimated the extent to which a patient would benefit from a drug, by reducing their risk of their condition worsening, or be at risk of side-effects.

They also found that only 23% of respondents gave correct answers when asked to estimate the effects of drugs like aspirin for stroke prevention, statins and alendronate.

The authors said in the paper: ’Dependent on clinical context and patient preferences, these inaccuracies in understanding could have negative implications for shared decision-making.’

Lead author Dr Julian Treadwell, a GP and doctoral research fellow in Oxford University’s Nuffield Department of Primary Care Health Sciences, said: ‘No-one would expect us to have encyclopaedic knowledge, but our survey found the range of GPs’ estimates of the benefits and harms of treatment to be highly varied and often inaccurate.

'When we are sitting in our surgeries, this is likely to affect the choices we make with our patients, particularly when faced with deciding which combination of several options will suit them best.’

But he stressed this 'isn’t the fault of individual GPs not keeping up to date'.

He said: 'It is a system-wide failure in how we receive up-to date information about treatments. This kind of detailed information isn’t available to us via clinical guidelines. There is so much research being produced, it is impossible for healthcare professionals across the world to read original research papers to extract the figures they need.'

Readers' comments (5)

  • Ivan Benett

    If only clinical decision making were a simple as feeding a figure into an algorithm. Set aside the issue that risk is an estimate based on (usually) 95%confidence intervals, themselves based on study populations that may not reflect our patients, and all the other weaknesses of studies.
    Perception of risk varies from one to another, the value placed on harms varies as does the value of a potential benefit. Societal norms, and health beliefs and peer group or family pressure also play a part. The JW's rejection of blood transfusions has little to do with an assessment of harm or benefit.
    Whole patient medicine means taking all these factors into account and coming up with a join decision with the individual.
    But it would help if Pharma, Government and Researchers would come up with better ways of informing us accurately, rather than telling us we're getting it wrong, and undermining our confidence and credibility.

    Unsuitable or offensive? Report this comment

  • fully paid study leave to keep up to date for all the mandatory training and other stuff would be helpful. NICE guidelines up to date quicker. Time in the day - fully protected - to read this stuff. otherwise you get what, in this case, is what you don't pay anything for. which is nothing much. you want better then change the system and pay for it because it currently is failing GPs en masse.

    Unsuitable or offensive? Report this comment

  • It matters very little what individual GPs know about exact figures when we have QOF/NICE mandating that we prescribe according to their didactic "guidelines" (for which read inviolate rules). The statin debacle goes on and on despite minimal risk reduction compared to stopping smoking for example, and yet smoking cessation services are all but dead in the water in many places. We all need to follow the guidelines like good little worker ants or risk castigation...

    Unsuitable or offensive? Report this comment

  • Caroe vinum 12:55 spot on!

    CQC: your practice above average QOF exceptions for one domain
    Me: 50% of practices will be above average
    CQC: why your practice above average?
    Me: we look after one of oldest patient cohorts in Europe. In top 1% of all practices for age related chronic disease prevalence, huge housebound, severe frailty and eolc cohort
    CQC: but why your practice above average exceptions?
    Me: computerised records since 1980’s, many historical codes, now long irrelevant as frailty deprescribing, and holistic management are main drivers of care. Robust systematic recall systems with good recording of patients informed decisions to decline inappropriate therapies and interventions
    CQC: but why your practice above average exceptions?
    Me: - silence- gone to boil head

    Unsuitable or offensive? Report this comment

  • Lots of misconceptions about this research.

    I am proud to have been one of the GPs involved with this study and I'm certain I have over estimated benefits of treatment.

    This research is a forerunner to actually doing something constructive about the problem.

    The problem is that there are no up to date, system-wide tools to assist clinicians with making treatment choices in partnership with patients.

    Ivan - of course it is about making decisions in partnership with the patient as there are 2 types of expertise - the doctor and the patient. But as this research shows, sometimes the doctor's expertise is factually incorrect.

    As for being bulldozed into treatment by NICE - this is also incorrect if you read the first few pages of every guideline, which emphasises the importance of a patient centred approach.

    Shared decision making with proper documentation actually helps you with managing risk and doesn't increase it

    Unsuitable or offensive? Report this comment

Have your say