Acknowledgement of “no fault” medical injury: review of patients' hospital records in New Zealand

BMJ 2003; 326 doi: (Published 11 January 2003) Cite this as: BMJ 2003;326:79
  1. Peter Davis, professora (peter.davis{at},
  2. Roy Lay-Yee, analystb,
  3. Alastair Scott, professorc,
  4. Robin Briant, clinical directorb,
  5. Stephan Schug, professord
  1. a Department of Public Health and General Practice, Christchurch School of Medicine and Health Sciences, University of Otago, PO Box 4345, Christchurch, New Zealand
  2. b Division of Community Health, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand
  3. c Department of Statistics, University of Auckland
  4. d Department of Anaesthesia, University of Western Australia, Perth, Australia
  1. Correspondence to: P Davis
  • Accepted 26 July 2002

Investigations of the epidemiology of adverse events have advanced the safety of patients in hospital.1 These studies, however, were done in tort jurisdictions, where the fear of litigation may have inhibited frank and open discussion.2 New Zealand abolished tort liability in 1972, instead providing an administrative system of compensation without the need to prove fault.3 We analysed data on adverse events in hospitals in New Zealand and the extent to which medical injury is acknowledged in patient records.

Participants, methods, and results

We took data on patient admissions from a representative sample of 13 from the 20 public hospitals with 100 or more beds. The survey population comprised all patients admitted in 1998 (excluding day patients, psychiatric patients, and patients attending just for rehabilitation). We reviewed the records of sampled patients retrospectively in two stages. To qualify as an adverse event, an incident had to have occurred or been detected by a healthcare professional during the sampled admission.1

Acknowledgement of medical injury in patients' hospital records in New Zealand (n=883*)

View this table:

We defined an adverse event as an unintended injury resulting in disability that was likely to have been caused by healthcare management rather than the underlying disease. We defined an acknowledgement as an annotation in a patient's record indicating or suggesting that healthcare management had caused the medical injury.

Of the 6579 admitted patients who were screened according to set criteria (see, the records of 4119 were reviewed by doctors using a structured protocol. Doctors judged 883 patients as having unintended injuries and resulting disabilities, and they assessed whether healthcare management had caused these injuries. Reviewers considered whether any note in the medical records indicated or suggested that healthcare management had caused the injuries.

After adjusting for sample design, reviewers classified 672/717 (94%) patients with records acknowledging injury as having had an adverse event compared with 81/166 (47%) patients whose records did not have such acknowledgement (relative risk 2.01; 95% confidence interval 1.75 to 2.32). We did similar calculations for subsets of adverse events that occurred in hospital (table). We estimated relative risks using the Mantel-Haenszel method and adjusted for the sample design (stratified cluster). Relative risks were greater for higher impact incidents and for “non-preventable” events.

For almost 672/753 (90%) adverse events, an annotation in the patient's record acknowledged medical injury. More than 148/181 (80%) adverse events involving systems failure in hospital were annotated.


Annotations in patients' records were a good predictor that a medical injury had been caused by healthcare management, regardless of clinical context. Fear of litigation may be an obstacle to reporting error—particularly for high impact, preventable, and systemic events. Our results show that the level of acknowledgement of medical injury in patients' records can be remarkably high in a no fault jurisdiction and strongly predictive of such occurences.

Doctors in many countries are discouraged from reporting medical errors,4 yet litigation in tort jurisdictions is becoming more common.5 In no fault jurisictions, the relatively high level of annotation in patient records that we found could provide a basis for more vigorous error reporting.


Contributors: PD designed the study and wrote the paper. RL-Y undertook the analysis, advised by AS. RB organised and oversaw the clinical assessments. SS provided technical advice. PD is guarantor.


  • Funding Health Research Council of New Zealand.

  • Competing interests None declared.

  • Embedded Image Screening criteria are on


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract