Survey of intensive care of severely head injured patients in the United Kingdom

BMJ 1996; 312 doi: (Published 13 April 1996) Cite this as: BMJ 1996;312:944
  1. D R Jeevaratnam, senior registrara,
  2. D K Menon, lecturer and honorary consultanta
  1. a Department of Anaesthesia, University of Cambridge Clinical School, Addenbrooke's Hospital, Cambridge CB2 2QQ
  1. Correspondence to: Dr Menon.
  • Accepted 29 February 1996


Objectives: To study practice in intensive care of patients with severe head injury in neurosurgical referral centres in United Kingdom.

Design: Structured telephone interview of senior nursing staff in intensive care unit of adult neurosurgical referral centre.

Setting: 39 intensive care units in hospitals that accepted acute head injuries for specialist neurosurgical management, identified from Medical Directory and information from professional bodies.

Main outcome measures: Details of organisation and administration of intensive care and patterns of monitoring and treatment for patients admitted with severe head injury.

Results: Patients were managed in specialist neurosurgical intensive care units in 21 of the centres and in general intensive care units in 18. Their intensive care was coordinated by an anaesthetist in 25 units and by a neurosurgeon in 12. Annual caseload varied between units: 20 received >100 patients, 12 received 50-100, and seven received 25-49. Monitoring and treatment varied considerably between centres. Invasive arterial pressure monitoring was used routinely in 36 units, but central venous pressure monitoring was routinely used in 24 and intracranial pressure was routinely monitored in only 19. Corticosteroids were used to treat intracranial hypertension in 19 units. Seventeen units routinely aimed for arterial carbon dioxide pressure of 3.3-4.0 kPa, and one unit still used severe hyperventilation to a pressure of <3.3 kPa.

Conclusion: The intensive care of patients with acute head injuries varied widely between the centres surveyed. Rationalisation of the intensive care of severe head injury with the production of widely accepted guidelines ought to improve the quality of care.

Key messages

  • Key messages

  • We conducted a structured telephone survey of senior nursing staff in intensive care units in 39 neurosurgical referral centres

  • The intensive care of patients varied widely, with only half the centres surveyed routinely monitoring intracranial pressure in comatose patients

  • Moderate hyperventilation and treatment with corticosteroids were still used by several centres despite increasing evidence of their lack of efficacy and potential for causing harm

  • There is a strong argument for establishing national minimum standards of care for the intensive care of patients with severe head injury


Half a million patients with head injuries are seen by the health care system in the United Kingdom each year1; a fifth of these are admitted to hospital,2 and 10% of admissions are for severe head injury (defined as a Glasgow coma score of less than 83). Secondary physiological insults contribute to the extent of neurological injury,4 5 and the quality of intensive care can be a major determinant of outcome. Recent research has re-evaluated some treatment methods that were commonly used in the past.6 7 However, a recent paper has shown wide variations in the management of severe head injury in the United States, with some centres still using treatments that were not supported by available findings from clinical research.8 We report the results of a structured telephone survey of intensive care of severe head injury in the United Kingdom.

Materials and methods


We identified 39 neurosurgical units from data in the Medical Directory and from neurosurgical and anaesthetic professional bodies. We conducted our survey by telephone. Clinical nurse specialists or staff nurses working in the units were interviewed by a single interviewer, who asked for the senior nurse on duty. The survey was in the form of structured questions with a set of defined answers, from which the interviewee chose one, except where a specific volunteered response was clearly appropriate. We encouraged respondents to consult medical and nursing colleagues and offered to recontact them after a short period of consultation and data collection if they wished. Information on use of corticosteroids was accepted only after we had emphasised that our survey was specific to head injury and asked the respondent to exclude reference to patients with other diagnoses, including intracranial tumours.

All 39 centres participated in the survey. We assessed the reliability of the data by repeating our survey of 20 of the centres the following week, when a different nurse was asked the same questions.


In 31 of the centres the intensive care unit and the neurosurgical referral unit were attached to a multidisciplinary hospital, while the remaining eight were either within free standing neurosurgical units or were in hospitals with one or two other specialised units (for example, plastic surgery). The respondent was a senior staff nurse in 20 units, a sister in 18, and a clinical nurse specialist in one. Our repeat survey of 20 intensive care units produced results that showed excellent concordance with those obtained in the first interview, with no changes in any of the questionnaire items except for a rebanding of the percentage of patients receiving corticosteroids (from >50% to 25-50% in two centres).

Tables 1 and 2 show the results of our survey. The use of specific monitoring procedures or treatments was unrelated to the type of intensive care unit, estimated annual case load, or speciality of the unit's director.

Table 1

Characteristics of 39 intensive care units that accepted acute head injuries for specialist neurosurgical management

View this table:
Table 2

Care of patients with acute head injury in 39 intensive care units: monitoring of haemodynamics and intracranial pressure and treatment of intracranial hypertension

View this table:

While we wished to determine whether all severely head injured patients were admitted to intensive care units, reliable information on this issue was difficult to obtain. We were unable to estimate the referral rate from peripheral hospitals, and the format of the survey did not allow us to investigate this issue further.



Telephone surveys are more effective than postal surveys in achieving complete participation since posted forms may not be received by the appropriate person or may be mislaid or ignored. While responses to written questionnaires are likely to be more considered, and hence more accurate, this cannot be guaranteed and there is no opportunity to discuss responses or allow interactive confirmation of the data obtained. However, telephone surveys suffer from several disadvantages. Responses are based on impressions rather than accumulated data, and the accuracy of the information obtained will vary with the training, seniority, and experience of the respondent. To minimise these effects we spoke to the most senior nurse on duty and encouraged respondents to consult colleagues and offered to recontact them after a short period of consultation and data collection if they wished. This option was taken up by five centres. We also provided them with a range of specified responses rather than asking them to volunteer quantitative information. The excellent concordance that we obtained on the repeat survey of 20 of the units shows the reproducibility of our method of data collection. We chose to interview senior nursing staff because they probably provide the most objective source of information about actual (rather than planned) clinical practice.


The results of our survey highlight several important issues that are at odds with an emerging consensus about the management of patients with severe head injury.2 3

Clinical signs cannot be used to detect neurological deterioration in a sedated and paralysed patient, and isolated imaging studies cannot replace monitoring of intracranial pressure.9 Intracranial and cerebral perfusion pressures have been shown to correlate strongly with outcome in several studies,4 5 10 11 12 13 14 15 16 17 18 and many treatments are designed to optimise these variables. Clearly, in the absence of continuous monitoring these interventions may be underused or used blindly and, in some cases, inappropriately.

Induced hypocapnia can reduce cerebral blood volume and intracranial pressure. However, severe hypocapnia (<3.3 kPa) can reduce cerebral blood flow to dangerous levels19 20 and result in cerebral venous oxygen desaturation,21 22 which is known to worsen outcome.22 There is less information on the effects of moderate hyperventilation (arterial carbon dioxide pressure 3.3-4.0 kPa), but routine prolonged hyperventilation was shown to worsen outcome in one study.6 These findings provide a rational basis for avoiding severe hypocapnia and using moderate hypocapnic ventilation with caution.

Corticosteroids are effective in reducing oedema in intracranial malignancies but are ineffective in head injury,23 24 25 26 where they may worsen outcome27 28 perhaps via metabolic effects. In several units corticosteroids were used by a single consultant, rather than as part of a unit's protocol.


The variations in clinical practice that we observed, both between centres and between the quality of care seen in the survey and that which might be described as the best possible standard of care, have important implications. We do not think that our findings are the consequence of justifiable therapeutic nihilism. There is little doubt that the combination of early surgery and good intensive care can result in a 10-20% improvement in outcome in severe head injury.29 Equally, we do not believe that many of these variations arose because of a lack of consensus among experts in the specialty. Many studies, only a small proportion of which are referenced in this paper, have demonstrated the need to monitor and control intracranial and cerebral perfusion pressures in patients with severe head injury. While costs and funding may be an important issue, a recently published survey in the United States, where spending on intensive care is higher, showed similar results.8

The findings of our survey provide a rational basis for a more detailed study, but there seems to be a strong case for producing nationally accepted guidelines on minimum standards of care for patients with severe head injury. Such guidelines would not only address the issues highlighted in this paper but could also provide guidance on the need for referral from receiving hospitals and the necessary levels of care in neurosurgical units for individual patients, depending on the severity of their head injury.

We thank Colette O'Kane and Liz Fahie, University Department of Neurosurgery, Addenbrooke's Hospital, for help with identifying neurosurgical centres.


  • Funding DKM was supported by a grant from the Critical Care Trust, Leeds.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
View Abstract