Intended for healthcare professionals

Views & Reviews Personal View

What is the main cause of avoidable harm to patients?

BMJ 2010; 341 doi: (Published 09 September 2010) Cite this as: BMJ 2010;341:c4593
  1. Gordon Caldwell, consultant physician, Worthing Hospital, Western Sussex Hospitals NHS Trust
  1. Gordon.Caldwell{at}

    As a hospital consultant I lead a team that helps patients to recover from illnesses such as myocardial infarction, pulmonary embolism, and gastrointestinal haemorrhage. I do not want to harm patients, make their illnesses worse, or unnecessarily prolong their stays in hospital.

    Since attending the International Forum on Quality and Safety in Healthcare in Berlin in March 2009 I have been thinking and working hard to improve the care that my team of doctors and nurses give to inpatients. We have used a “considerative checklist” to ensure that we do everything we can to reduce the chances of pulmonary embolism, septicaemia from meticillin resistant Staphylococcus aureus, and diarrhoea from Clostridium difficile, among several other types of avoidable harm. We have greatly improved our attention to detail and safety in writing prescription charts. All this has helped us to be more careful to avoid harming patients during treatment.

    However, I believe that there is an even more important avoidable harm, one generally unacknowledged: incorrect diagnosis. When an obviously ill patient is admitted to hospital the medical team formulates a “working diagnosis.” At this point the diagnosis is uncertain, but the patient is treated as if the working diagnosis is correct. For example, if it is bronchopneumonia, antibiotics, chest physiotherapy, and oxygen are prescribed. If over the next days the patient gets better the working diagnosis is confirmed and becomes the diagnosis. But if the patient doesn’t improve we think again and consider whether the working diagnosis was wrong.

    Bronchopneumonia can easily be confused with pulmonary embolism or cardiac failure, and the treatments are quite different. Treating all three conditions at once is suboptimal and potentially dangerous, so we try to limit the treatments. The consequences can be significant. A patient treated according to a working diagnosis of pulmonary embolism who actually has pneumonia may die of untreated pneumonia.

    The time taken to reach the correct diagnosis may be crucial for the patient’s chance of survival. Over my career I have seen many errors in the working diagnosis causing harm to patients and even death.

    Little consideration seems to have been given to how doctors make and refine a patient’s working diagnosis and treatment plan. The working diagnosis is reached by deliberating on information from interviewing and examining the patient, old notes, referral letters, drugs lists and test results, and the time course of the illness. Even if the principal working diagnosis is straightforward, the patient may have multiple pathologies. The main working diagnosis may be a fractured hip, but the patient may also have atrial fibrillation, be taking warfarin, have hyperglycaemia, and be known to have mild dementia and to have been treated for breast cancer. How do we balance all these factors and swiftly get the patient to theatre for hip surgery?

    I believe that we have not thought about the best places, the physical and psychological environments, in which doctors should do this complex clinical thinking. Often it occurs in small hot rooms subject to constant interruption or even in ward corridors without easy access to laboratory results.

    Serious errors in working diagnoses made among patients under my care occurred because the old medical notes were unavailable or very disorganised; it was difficult to see important blood test results at the same time as hearing the history from the junior doctor; an old electrocardiograph was not easy to find; or there was too much pressure to get the round done and to get patients out of hospital. I have seen myself on video miss vital information in what a doctor was saying during a post take ward round because I was interrupted by another doctor asking a trivial question.

    Through fear of litigation and losing face and simply because of the difficulty of explaining the complexity of what we do every day, we have failed to let our patients and society know about this very important problem. We must design our working spaces and information systems to maximise doctors’ ability to see, understand, and deliberate on the information needed for more precise diagnosis. We must allow clinicians enough time to be careful in diagnosis, treatment planning, and treatment review. We must urgently consider how to provide rooms, time, and information for doctors to do the most difficult part of their job and the part most prone to error: the clinical thinking in making the working diagnosis and treatment plan.

    Perhaps we need to be like pilots and have a “diagnostic cockpit”? Our work is more dangerous than theirs.


    Cite this as: BMJ 2010;341:c4593

    View Abstract