Intended for healthcare professionals

Dr Foster's Case Notes

How often are adverse events reported in English hospital statistics?

BMJ 2004; 329 doi: (Published 12 August 2004) Cite this as: BMJ 2004;329:369

This article has a correction. Please see:

  1. Paul Aylin,
  2. Shivani Tanna,
  3. Alex Bottle,
  4. Brian Jarman
  1. Dr Foster Unit, Imperial College

    About 850 000 medical errors occur in NHS hospitals every year, resulting in 40 000 deaths.w1In the United Kingdom, the National Patient Safety Agency (NPSA) was created to learn from patient safety incidents occurring in the NHS.w2 In February 2004, it launched a new patient reporting system, drawing together reports of patient safety errors and systems failures provided by health professionals across England and Wales.w3 An adverse event can be defined as “an unintended injury caused by medical management rather than a disease process, resulting in death, life threatening illness, disability at the time of discharge, admission to hospital, or prolongation of hospital stay.” w4 A medical or surgical misadventure is an adverse event that might have been avoided if the patient had received ordinary standards of care. We look at four years of hospital episode statistics to examine patterns in the recording of adverse events within this routinely collected source of data and ask whether it could be of use in monitoring this problem.

    The bottom line

    • 2.2% of all hospital episodes contain a mention of an adverse event

    • Nearly 4000 misadventures are recorded each year

    We examined four years of hospital episode statistics (1999-2000 to 2002-3) comprising 50 215 687 episodes of care. We identified 41 three-digit ICD-10 diagnosis codes with some indication of an adverse event (see table on Eight of these codes (Y60-Y69) related directly to medical or surgical misadventures. We examined the proportion of all episodes with a mention of one or more of the 41 codes by age, sex, method of admission, year, and trust.

    We found, on average, 2.2% of all episodes (276 514 per year) included a code for an adverse event. Events were more likely to occur in men, in elderly people, and in emergency admissions. The differences may be due in part to the severity of underlying disease in the different groups and the length of time people are in hospital. Other studies have found overall rates of 0.97%w5 and 36%,w6 but a study using routine hospital data in Australia found a rate of 4.75%.w7 Studies using routine data tend to have lower estimates than those based on casenote reviews or purpose designed systems. Some trusts reported zero levels of adverse events, which seems unlikely. Therefore, adverse events may be under-recorded within hospital episode statistics. We will have missed some conditions arising as a complication of treatment not specifically coded as an adverse event (for example, pulmonary embolus following surgery or stroke following carotid endarterectomy). Hospital acquired infections are also poorly represented within ICD-10 (there is no specific code for methicillin resistant Staphylococcus aureus, MRSA). We have not included obstetric complications, and there may be additional codes that might be used. We have demonstrated that adverse events are recorded within hospital episode statistics; for these statistics to accurately monitor adverse events, hospitals should be encouraged to improve the recording of events on their systems.

    Episodes of care (number (rate per 1000)) with mention of adverse events and misadventure. Source: Hospital Episode Statistics 1999-2000 to 2002-3

    View this table:

    The basic figures

    • On average 2.2% of all episodes (about 27 500 per year) included a code for an adverse event

    • Misadventures were mentioned in 0.03% of episodes (3980 per year)

    • Incidence did not change over time

    • The rate of adverse event recorded in each trust ranged from 0% to 15% and the rate of misadventures recorded ranged from 0% to 1.02%

    • Adverse events were mentioned more frequently in emergency than elective admissions (2.82 v 2.16%, P < 0.001), but misadventures were similar

    • Inpatient episodes were more likely to mention an adverse event than day cases (2.77% v 0.83%, P < 0.001)

    • Adverse events were recorded more commonly in men (2.5%) versus women (2.50% v 1.97%, P < 0.001)

    • Adverse events were more common in elderly people


    • Embedded ImageThis month's Dr Foster's case note was compiled by Paul Aylin, Shivani Tanna, Alex Bottle, and Brian Jarman at the Dr Foster Unit at Imperial College. Dr Foster is an independent research and publishing organisation created to examine measures of clinical performance.

    • Embedded ImageReferences, full methodological details, and a table of ICD-10 codes are on and

    View Abstract