Injury, 10, 31-39

31

Printed in Great Britain

Preventable mortality and morbidity after head injury Bryan Jennett and Joseph Carlin University Department of Neurosurgery, Institute of Neurological Sciences, Glasgow THE lecture 'Who cares for head injuries?' (JenneR, 1975a), which I gave to this Institute three years ago, might have carried the subtitle, 'Does it matter?'. I believe that many accident, general and orthopaedic surgeons, who look after most of the head injuries admitted to hospital, often think, even if they don't say it, that it probably doesn't matter too much. They suspect that most mild injuries get better, however little is done for them, and that most severe injuries do badly, however much is done for them. This contrasts with the Hippocratic aphorism, 'No injury is too trivial to be ignored, or too serious to be despaired of'. There is now a growing body of evidence that a positive attitude to head injury care can have a material influence on morbidity and mortality, because some of this is due to secondary brain damage which is potentially preventable. I shall review this evidence and shall indicate how changes in attitude and in organization might improve the care of patients with head injury.

I M P O R T A N C E OF HEAD INJURIES I need not remind this audience of the importance of accidents, which are the leading cause of death in all age groups under 40 years. Head injury features prominently in accidental deaths, accounting for more than half of those from road traffic accidents. Among accident victims admitted to hospital head injuries also loom large. Of admissions to two large accident services (Birmingham and Oxford), head injury was present in about 40 per cent and was the sole diagnosis in about 30 per cent. In a city teaching hospital in

Glasgow head injuries made up 25 per cent of all emergency admissions to male surgical wards (Galbraith et al.. 1977). Severe residual disability after accidents is chiefly due to head injury--of 45 disabled patients resulting from over 2000 road accident admissions, Hoffman (1976) reported that 32 had brain damage, blindness or deafness (71 per cent). Our own studies of severe head injury were largely prompted by a chilling statistic included by Peter London of this Institute in his Hunterian Lecture in 1967. May I remind you that he calculated that some 1500 patients were leaving British hospitals every year with permanent brain damage, their average age being about 30 and a half of them never able to work again. This estimate has subsequently been confirmed by Lewin (1968) from different data. Estimates of the load which head injuries place on accident departments (as distinct from inpatient admissions) were not available until recently. The Scottish Head Injury Management Study, having surveyed all hospitals in that country, now indicates that 10 per cent of all new attenders have a head injury (JenneR, Murray et al., 1977). This means that about one million newly head-injured patients attend accident departments in Britain every year. We can conclude that head injuries are important because they are prevalent, because they occur predominantly in the young and because management may matter; but they are difficult because, although they are so common, only a few are either severe at the beginning or become serious as a result of secondary events.

32

Injury. the BntmhJournal of Accident SurgeryVol. 10/No. 1

Table I Analysis of head-injured patients dying before admission to hospital in Blrmmgham and Glasgow

Cause of injury RTA Window Railway Falls Extracranial injuries* Multiple Chest Cervical spine Brain damage* Disruption Fracture baser Ports

Birmingham (60)

Glasgow (62)

Total (122)

43 (70%) 9 (15%)

17 (27%) 14 (23%)

60 (49%) 23 (19%)

1 5

6 (10%)

7

(6%)

(8%)

11 (18%)

16 (13%)

44 (73%) 21 (35%) 6 (10%)

24 (59%)

68 (68%)

12 (29%) 7 (17%)

33 (33%) 13 (13%)

7 (12%) 34 (57%) 18 (30%)

4 (10%) 23 (56%) 15 (37%)

11 (11%) 57 (57%) 33 (33%)

*Only 41 of the Glasgow subjects are included m this part of the analysis, giving a total of 101 fatahties. 1"Includes pons.

DEATH BEFORE ADMISSION TO HOSPITAL The Field Report (1976) revealed that 60 per cent of deaths from head injury occurred before admission to hospital; 40 per cent were dead at the scene of the accident and 20 per cent died at hospital before admission. This naturally raised the question of whether better care at the site of the accident and during transport to hospital might save some lives. Hospital doctors have no knowledge of these early deaths, but due to the kindness of Dr Whittington (Coroner for Birmingham) and Dr McLay (Chief Medical Officer for the Strathclyde Police), I have been able to scrutinize the pathological reports submitted to them on 122 patients dead before admission to hospital, in whom death was ascribed to head injury (including those with multiple injuries which involved the head). Only half of these cases were injured in road accidents (Table I). Almost all had sustained overwhelming injuries--severe basal skull fracture, brain stem laceration, disruption of the whole brain, fracture/dislocation of the cervical spine or serious chest injury (Table I). One patient died of meningitis from an open injury having signed himself out of hospital two days previously and one patient died of airway obstruction without having impact injury of such severity that it would definitely have been fatal on its own. This confirms that what others have recently reported for accidents in general applies also to

head injury. Yates (1977) reviewed coroners' autopsies over the period 1971-5 in Salford, of which 82 per cent were on victims of traffic accidents; he compared the incidence of airway obstruction in 69 who died at the scene and 105 who died in hospital within 72 hours of admission. He found that airway obstruction contributed to the death of some patients in hospital, but that it could not be indicted as a cause of death before reaching hospital. A review of road traffic accident victims in Teesside for 1971-2 by Hoffman (1976) included 344 deaths and 2392 surviving hospital admissions. This revealed only 2 patients whose lives might have been saved by intubation and assisted ventilation at the scene of the accident: preventable factors (including airway obstruction) occurred more often in patients who had reached hospital alive. A high incidence of preventable factors contributing to deaths in hospital following head injury has previously been reported by us (Rose et al., 1977). It may be, however, that less than fatal degrees of airway obstruction, or of hypotension, in the early stages after injury may add to the brain damage sustained at impact. In head injuries reaching one American neurosurgical unit, mostly within ~-3 hours of injury, a high proportion had hypotension, hypoxia or anaemia (Miller et al., 1978). Many of these patients had multiple injuries, and a correlation was shown between such factors and both mortality and the incidence of persisting disability in survivors.

Jennett and Carlin" Preventable Mortality and Morbidity after Head Injury

Some of these correlations may have been due to the more frequent occurrence of these abnormalities in the more severely injured patients, including those with multiple injuries. Whilst hypotension is rarely due to brain damageper se, hypoxia can occur due to neurogenic influences on pulmonary gas exchange and head-injured patients who are hypoxic are likely to include those with more serious impact damage. A report from the Birmingham Accident Hospital some years ago suggested, however, that recovery from head injury was delayed and sometimes less complete in survivors who had been hypoxic and hypotensive during the first 48 hours after injury (Price and Murray, 1972). In concluding this discussion of preventable factors contributing to death before the patient reaches hospital, I would support the view of Hoffman (1976), who concluded that routine medical attendance at accidents would be wasteful of manpower, but that hospital-based training of selected ambulance personnel in cardiorespiratory resuscitation could improve the primary care of the head injured, particularly in rural areas where more severe injuries are common and distances to hospital are often greater. This issue is discussed later m relation to the evidence that some hospital mortality and morbidity is related to events during transfer between hospitals by ambulance.

N E U R O S U R G E O N S A N D HEAD INJURIES Only a small proportion of head-injured patients in Britain reach a neurosurgeon. In most of Scotland only 4 per cent of hospital admissions for head injury are to regional neurosurgical units (Scottish Home and Health Department, 1975); as in most of England, it is only the more severe or complicated cases that go to the neurosurgeon, They do so usually by secondary referral from a primary surgical ward, which may be general surgical, orthopaedic, accident or paediatric according to local arrangements. This is also the pattern, as far as I can discover, in most of continental Europe, in Australasia, in Japan and in South America. A variation in some European countries is that medical neurologists play a major role in the inpatient care of head injuries. In North America they do things differently, in that a large proportion of head-injured patients are cared for, or at least seen, by a neurosurgeon. That does not mean, however, that they are admitted to what we would recognize as a major neurosurgical unit; most of these patients remain in community hospitals, some of which are quite

33

small but most of which correspond to a British district general hospital. There an attending neurosurgeon will see them, and often take on their care, but he will have no junior staff and may himself work in several other hospitals. The contribution he can make to their care may therefore be limited, and some may question how different in practice this is from the European pattern of care. Some Americans have no doubt that their system of caring for head injuries is better than the British or European way. A senior and robust defender of the American way of neurosurgical life has recently written that 'one of the reasons that patients with head injuries are not receiving optimal care in Great Britain is that there are not enough neurological surgeons'. He comments that the provision of many more neurosurgeons dispersed through the USA has made death from extradural haematoma far less likely (Bucy, 1976). Some neurosurgeons in other countries have expressed concern that head injuries are not as well cared for as they might be. Bushe and Siede (1971) reported that in one German region 90 per cent of head injuries were treated in community hospitals which 'had not the personnel or facilities to provide the essential conditions for optimal treatment'. Frowein (1971) also emphasized the need for many more head injuries in Germany to be treated in neurosurgical or neurological units, where special skills and equipment were available. Sano (1965), commenting on the Japanese scene in 1965, stated that between 10 and 20 per cent of fatalities from head injury in hospital could probably have been prevented by proper neurosurgical treatment; at that time most head injuries were cared for by surgeons with little neurosurgical training. Most British neurosurgeons remain ambivalent on this issue. A few believe that neurosurgeons should be more involved with head injuries and have set up local arrangements to care for the majority of such cases occurring in their own cities. In the rest of Britain, as in most of the rest of the world, neurosurgical units accept only the more severe injuries or those which clearly require intracranial surgery. The number of neurosurgeons per million population is similar in the U K to that in most European countries, and the argument is that with such a limited number there is no possibility of them accepting a wider responsibility for head injuries. Reluctance to recruit more neurosurgeons is based on the doctrine of dilution--the belief that individual neurosurgeons would not then do enough major elective surgery to gain experience or to maintain

Injury: the Brit=sh Journalof Accident Surgery Vol. 10/No. 1

34

Table II. Patients with one or more avoidable factors contributing to

death

No.

Talk and die (G) Coma untd death (G) Survivors after coma with IC haematoma (G) Aberdeen deaths Glasgow deaths

Avoidable Avoidable factor factors wRh certain effect present on outcome

116 50

74% 40%

54% 30%

50 50 *

64% 36% 50%

40% 26% 40%

*Estimated for all head injury deaths in a neurosurgical unit. IC, Intracranial. G, Glasgow.

expertise, a problem which has arisen in many places in North America where there are ten times as many neurosurgeons relative to the population. This is an issue where data are urgently needed rather than declarations of faith, and reason rather than rhetoric. In order to find out how the British system works in practice we have been collecting data for several years in Glasgow on many aspects of head injury care; we have also initiated internationalstudies involvingthe Netherlands and USA in order to make some comparisons possible (Jennett, Teasdale et al., 1977). These studies are still continuing but already we are able to make some suggestions as to how head injury care might be modified in Britain. To what extent these conclusions can be generally applied is for those in other places to decide. These studies have included surveys of accident and emergency departments and of primary surgical wards throughout Scotland, as part of the Scottish Head Injury Management Study (Jennett, Murray et al., 1977). However, I shall begin with data from a neurosurgicai unit because that is where the majority of serious (and fatal) cases eventually come; from a study of these it is possible to consider in retrospect how adequate their care has been.

Lessons from head-injured patients w h o talk and die In 1975 we published an account of the pathological findings in patients who had died in a neurosurgical unit and whom the records showed to have spoken at least a few words in the interval between injury and death (Reilly et al., 1975). Over 90 per cent of these had pathological evidence of having had raised intracranial

pressure, and 75 per cent had an intracranial haematoma; other lesions were ischaemic brain damage, brain swelling and meningitis. Almost one-third of all patients with fatal head injuries in the neurosurgical unit were found to have talked. In the international study of patients with head injuries in coma, about one-third had talked before lasting coma developed (Jennett, Teasdale et al., 1977). In that series of I000 patients, 50 per cent died, and the proportion who talked was similar in the fatal cases to that in the survivors. It seemed to us that a patient who had talked was unlikely to have sustained overwhelming impact damage, and that his subsequent death was therefore probably due to secondary pathological events of one kind or another. Neuropathologists in Glasgow have described in detail the kind of diffuse impact damage which is now recognized as being associated with failure to regain any meaningful cortical function, even though survival for weeks or months is achieved by intensive care. Not one of the patients with this lesion has talked after injury (Adams et al., 1977). From this line of reasoning, we postulated that a certain number of patients who talked before they died might have survived had circumstances been different. We therefore set out to discover avoidable factors in management which might have contributed to death in such cases. A distinction was made between avoidable incidents (or factors) which certainly contributed to death, and those which were judged only possibly to have been significant. Such a judgement could be made only when a full neuropathological autopsy was available, and this meant securing whole brains for adequate fixation before dissection, which in turn required arrangements to be made with forensic pathologists.

Jennett and Carhn : Preventable Mortality and Morbidity after Head Injury Table III. Avmdable factors m 166 deaths from head injury Certain effect Intracranial Delay on treating haematoma 58 Inadequately controlled epilepsy 9 Meningitts 8 Other 8 Total 83 Extracranial Airway obstruction 11 Hypotension 10 Other 16 Total 37 Total inctdents 120 Total patients 78

35

Table IV. Delayed treatment of intracranial haematoma in 164 cases Locatton

%

Total 82 23 8 17 130 35 33 32 100 230 106

Avoidable factors contributing to death In 116 patients with neuropathological autopsy and who had talked, one or more avoidable factors were found in 74 per cent, of which one factor was considered certainly to have contributed to death in 54 per cent (Rose et al., 1977) (Table H). Of incidents which certainly contributed to death, 76 per cent were intracranial and 24 per cent were extracranial (Table III). Avoidable factors are not limited to patients who talk and die. Investigation of 50 patients who died without talking revealed avoidable factors in 40 per cent; and of 50 patients who survived after removal of an intracranial haematoma, but who were in coma at the time of operation, 64 per cent showed avmdable factors (Table H). Including those who talked and those who did not, and those with and those without an intracranial haematoma, we estimate that among all deaths in the Glasgow neurosurgical unit 50 per cent had one or more avoidable factors and 40 per cent had an avoidable factor which certainly contributed to death. Investigation of deaths in another neurosurgical unit (Aberdeen) indicates that the Glasgow experience is not unique (Table H). Pooling all the Glasgow data provides 272 avoidable incidents (145 with a certain effect on death) in 216 patients, and this enables a more detailed analysis of the circumstances in which the most commonly occurring avoidable factors arise. Delayed evacuation of haematoma

This was found to be due to avoidable incidents at one or more of four different locations (Table

Before hospital At A 6" E department In primary surgical ward In neurosurgical untt

23 9 48 20

IV). Some patients failed to seek medical aid soon after injury (or took their own discharge from the accident and emergency department or primary surgical ward). Their late presentation with advanced deterioration could be regarded as due to avoidable delay, but was not the fault of the system. Relatively few significant delays were ascribed to factors operating in the accident and emergency department. However, there were a number of patients who developed fatal complications after having been sent home from the accident and emergency department; this often occurred as a result of failure to recognize the importance of skull radiographs. Delays most often occurred in the primary surgical wards. Galbraith (1976) has already reported that over the 12-year period 1963-74 about a third of intracranial haematomas reaching a neurosurgical unit had been deteriorating for more than 12 hours before transfer. We have confirmed in more recent investigations that this is still a problem; one reason for this was the restricted neurosurgical facilities. It is known that during this period requests for transfer from primary surgical wards were sometimes declined or postponed by the neurosurgeon because of lack of beds, if the case did not seem to be in definite need of urgent surgery; this policy may also have inhibited primary surgeons from asking for transfer until there was definite evidence of deterioration. Whatever the reasons, a high proportion of patients who had earlier been talking had reached the stage of coma, sometimes deep coma, by the time they were operated on, and the mortality and morbidity in these cases was much higher than in those treated at an earlier stage. Delays can occur even after the patient reaches the neurosurgical unit, usually because the extent of deterioration which had already occurred was not realized, often because clinical details of the patient's course in continuity were not available. When angiography was the usual method of investigation there were sometimes delays in arranging for this to be done, especially out of hours, when a radiologist had to be called in.

36

Injury: the Bnt=shJournal of Accident SurgeryVol. 10/No. 1

The availability of computerized tomography (CT or EMI scan) may go some way to resolving this problem, because it is highly reliable in detecting intracranial haematoma and does not require the skills of a neuroradiologist to produce pictures. In some American centres, where head injuries are routinely scanned, it has become apparent that intracranial haematoma is commonly present before the patient markedly deteriorates. The problem to be faced in this country, where scanners are available only in regional neurosurgical units (and general body scanners in a few general hospitals), is to decide who should be scanned. It may be that a higher proportion of patients than at present should be referred to neurosurgery (at least temporarily) for EMI scanning, rather than waiting until clinical deterioration is obvious. Whole body scanners could certainly provide adequate head pictures for this purpose, and at the time of day when head injuries tend to present, the whole body scanner might not be in use. Quite apart from arranging scanning, there is clearly a need for better definition of criteria for transfer and for a local policy to be established between the regional neurosurgical unit and its surrounding hospitals. Those in need of immediate transfer are patients who show any signs of a deteriorating conscious level or the appearance of focal signs. Those who have not regained consciousness within a few hours should also be investigated, because a number will be found to have an intracranial haematoma. Patients with open depressed fractures of the vault or suspected basal fractures can wait several hours (perhaps overnight) before transfer, if necessary. A necessary basis for this dialogue between primary surgeons and neurosurgeons is a common language by which to communicate about the state of patients, and in particular about changes in their conscious level. This was one purpose for which the Glasgow Coma Scale was devised (Teasdale and Jennett, 1974). This is now widely used in Europe and North America, where it is also being taught to ambulance and helicopter personnel responsible for the early care of the injured. It is likewise useful in the primary surgical ward and the neurosurgical ward for the continued monitoring of recently head-injured patients. Epilepsy This is not uncommon soon after head injury and may present two problems. The occurrence of a fit may be accepted as an explanation for a

deteriorating conscious level, when this is in fact due to a serious intracranial complication. When a patient does not rapidly recover from a convulsion there should be a high index of suspioon that some complication is developing, although a fit alone is never the sole evidence of such a complication (Jennett, 1975b). If status epilepticus develops, and is inadequately controlled, it represents a threat in its own right, and every effort must be made to secure control. We have recorded fatal brain damage in children who developed status epilepticus after relatively mild head injury. Intracranial infection This can develop if there is failure to recognize that a patient has an open injury. This may be a depressed fracture of the vault, which has been treated as a simple scalp laceration by the classic casualty cobble--again because of lack of adequate radiographs. We have reported an infection rate for depressed fractures reaching a neurosurgical unit which is many times that achieved for penetrating injuries on the battlefields of Vietnam (Jennett and Miller, 1972). Meningitis can also result from a fracture of the base of the skull, which is much more difficult to show radiologically; but this should be suspected if the patient has bilateral orbital haematoma or a retromastoid haematoma, even if a CSF leak is not yet detected (Leech, 1974). A i r w a y obstruction Although this was a common event, only about one such incident in four was regarded as having definitely contributed to death. About half the incidents of airway obstruction occurred during ambulance transfer between primary surgical wards and the neurosurgical unit. Of patients transferred in coma to a regional neurosurgical unit (including non-traumatic cases), 42 per cent had neither a mechamcal airway nor an endotracheal tube; 61 per cent were supine on the stretcher. Hoffman (1976) reported that 22 per cent of unconscious road accident victims were transported lying on their backs. He found that less than half the casualty officers in the 12 hospitals which he surveyed knew how to pass an endotracheal tube. There appears to be a need to improve care both m hospital and during transfer of comatose patients between hospitals. In the USA training of emergency medical technicians and paramedical personnel is being actively promoted; in the U K anaesthetists in Bristol are pursuing a similar programme (Baskeett et al., 1976).

Jennett and Carhn ' Preventable Mortahty and Morbidity after Head Injury

Hypotension The incidence of hypotension is probably always under-reported because it can be recognized only when blood pressure is recorded, for example on admission or during surgery. The high proportion of hypotensive incidents occurring during operation or in the immediate postoperative period is probably partly due to routine recordings being usual at these times. Several of these incidents were related to cardiac arrest during anaesthesia, others to serious haemorrhage and occasionally unrecognized abdominal Injury was the reason. On occasions cardiac arrest had occurred during an ambulance journey. These extracranial avoidable factors correspond closely to those found by Hoffman (1976) in 28 preventable deaths In hospital following road accidents. He had 12 cases of severe blood loss, 6 of respiratory difficulty and 6 with both; he also reported cases of misdiagnosis, particularly of intra-abdominal bleeding. These factors can be expected to combine with raised intracranial pressure to produce ischaemlc brain damage, which was found in 90 per cent of autopsies in patients dying in a neurosurgical unit (Graham and Adams, 1971). Such damage is also likely to contribute to persisting disability in some headinjured patients who survive.

REDUCING AVOIDABLE INCIDENTS IN M A N A G E M E N T In the face of the data presented, there is no reason to be complacent about the way in which we handle head injuries. Occasional and less detailed reports from other centres, both in Britain and elsewhere, suggest that the mistakes which we have identified in Glasgow also happen to some extent in most places. Views may differ as to what solution should be offered; it may well be that the answer which is appropriate in one place may be less suitable for another. Certain principles can, however, be set down without too much fear of contradiction and these can be related to the four different locations where avoidable factors are apt to occur.

A c c i d e n t and e m e r g e n c y d e p a r t m e n t The need here is for standardized regimes for assessing conscious level, for obtaining satisfactory skull radiographs in patients who are fully conscious and who might be sent home unless a fracture is discovered, for wound excision and closure of scalp lacerations and for deciding which patients to admit for further observation. Our survey of Scottish accident and emergency

37

departments (Jennett, Murray et al., 1977) indicates that some 40 per cent of attenders with head injury have a scalp laceration. Although 20 per cent of all attenders had a history of altered consciousness, only 5 per cent had evidence of impaired consciousness on examination in the accident and emergency department (Strang et al., 1978). Except for open depressed fractures of the vault, it was uncommon to find a skull fracture unless there was some evidence of brain damage (usually an episode of altered consciousness). Almost all patients who presented with serious complications (e.g. intracranial haematoma or infection) after having been sent home from accident and emergency departments, were found in retrospect to have had obvious evidence of actual or potential brain damage when seen in the accident and emergency department. Better selection of patients for admission would probably prevent many of these incidents. This would not entail admitting more patients but admitting the right ones (Strang et al., 1978). Indeed, it should substantially reduce the number of patients who are admitted at present. Such a selective policy would require adequate skull radiography, as it would depend on a reasonable exclusion of skull fracture in patients to be sent home.

Primary surgical wards Most patients admitted are sent home within 48 hours, many within 24 hours. Sometimes they are inadequately examined before discharge and some return with intracranial complications after such a brief admission. Those whose conscious level is impaired require to be systematically monitored, for which the coma scale is useful. Many patients with intracranial haematoma whose transfer had been delayed were mistakenly diagnosed as suffering from alcoholic intoxication or of having had a cerebrovascular accident. Some 90 per cent of patients with extradural haematoma and 75 per cent of those with other intracranial haematomas have a fracture (Jennett and Teasdale, 1979); the finding of a skull fracture is therefore an important guide to the likelihood of intracranial complications. Exclusion of associated injuries should be an active, planned policy, particularly in unconscious patients who cannot complain of their other injuries. Surgery under general anaesthesia for associated injuries is hazardous: if possible this should be postponed for a few days, until the acute stage of the head injury is passed. If intervention is inevitable (e.g. ruptured viscus) then care is needed to protect the brain from

38

untoward physiological and effects of anaesthesia.

Injury. the British Journal of Accident Surgery Vol 10/No. 1

pharmacological

Neurosurgical unit Neurosurgeons should take the problems of head injury more seriously. That does not necessarily entail admitting large numbers of mildly injured patients to neurosurgical units. But it does mean prior consultation with local accident, general, orthopaedic and paediatric surgeons to devise a workable policy, based on a reasonable division of responsibility between primary surgeons and neurosurgeons. Probably more patients should come briefly to the regional unit for EMI scanning; only if these patients can then be returned to the primary surgical ward, once there is no longer a need to stay in the neurosurgical unit, would it be practical for neurosurgeons to deal with a relatively large number of such patients. Earlier transfer of more patients on this basis should mean fewer avoidable factors, reduced mortality and morbidity and therefore less work for the hospital service in the long run. Waiting until suspicion of an intracranial haematoma is high will often mean waiting until it is too late.

Ambulance transfer There is need for more thought about how to reduce the hazard of transporting comatose patients from one hospital to another. In particular the airway should be protected by proper positioning, the availability of portable suction and insistence on an escort who knows what to do in the event of various developments (e.g. vomiting or an epileptic fit). If a patient will tolerate an endotracheal tube, then he should have one. Hoffman (1976) suggested how hospitals might promote better training of ambulance personnel in the care of such patients.

T R E A T M E N T OF SEVERE INJURIES This is largely the concern of intensive care units and neurosurgical departments, and discussion

would be out of place here. In any event, it seems inherently unlikely that marginal advances in the care of severely brain damaged patients would make an impact on mortality and morbidity commensurate with that to be expected from reorganizing early care so as to prevent patients developing severe secondary brain damage. Although the mortality from severe head injury is still about 50 per cent, the degree of recovery of survivors is better than is sometimes realized. Of 376 patients who had been in coma for more than six hours (post-traumatic amnesia > 4 8 hours),

Before

I

Hospital

o..oOoAr.v=

Hospital

/ITa,..o

l

ComaT.,

:i~ll'~ . . ]

&voidable

. .

"~ A v o i d a b l e

Deaths

a,.

~ " I S,~r,,,vo~s .

Factors

Fig. 1. Avoidable factors contributing to death after head injury.

81 per cent were independent a year later and 50 per cent were restored to all normal activities, mental and physical (Jennett et al., 1978).

CONCLUSIONS Head injuries present a major problem to those responsible for the care of accident victims. Avmdable mortality and morbidity is common, and most of it derives from inappropriate management of patients who reach hospital alive (Fig. 1). Policies should be formalized for the care of head-injured patients in accident and emergency departments, in primary surgical wards, in neurosurgical units and during ambulance journeys between these units. Neurosurgeons should play a more active role in formulating these policies, and should give more support to their colleagues in other disciplines. This does not necessarily require them to accept continuing responsibility for a large proportion of head-injured patients. Confidential reporting of avoidable factors contributing to deaths from head injury (as is done for maternal mortality) might prove useful in focusing attention on this problem. A high autopsy rate and good quality neuropathology are essential if the cause of death in individual cases is to be determined, and avoidable factors recognized; liaison with forensic pathologists is necessary to achieve this. REFERENCES

Adams J. H., Mitchell E., Graham D. I. et al. (1977) Diffuse brain damage of immediate impact type: its relationship to 'primary brainstem damage' in head injury. Brain 100, 489. Baskett P. J. E., Diamond A W. and Cochrane D. F. (1976) Urban mobile resuscitation: training and service. Br. J. Anaesth. 48, 377. Bucy P. (1976) Can we learn? (Editorial) Surg. NeuroL 6, 82.

Jennett and Carhn : Preventable Mortahty and Morbidity after Head Injury

Bushe K. A. and Slede H. (1971) The problems of head injury treatment in community hospitals. International Symposium on Head InJuries. p. 153. Field J. H. (1976) Eptdemiology of Head Injuries in England and Wales. (The Fmld Report.) London, HMSO. Frowein R A. (1971) Organization of head injury care ,n the hospital. International Symposium on Head Injuries. p 156. Galbraith S. (1976) Mlsdlagnosls and delayed diagnosis in traumatic intracranlal haematoma. Br. Med. J. 1, 1438. Galbralth S., Murray W. R. and Patel A. R. (1977) Head injury adrmssions to a teaching hospital. Scot. Med J 22, 129 Graham D. I. and Adams H (1971) Ischaemlc brain damage m fatal head injuries. Lancet 1, 265. Hoffman E. (I 976) Mortality and morbidity following road accidents Ann R Coll. Surg. Engl. 58, 233. Jennett B (1975a) Who cares for head injuries? Br. Med. J. 3, 267. Jennett B. (1975b) Epilepsy after Non-missile Head hljuries, 2nd ed. London, Hememann. Jennett B. and Miller J. D. (1972) Infection after depressed fracture of the skull: implications for management of non-missile injuries. J. Neurosurg. 36, 333. Jennett B., Murray A., McMlllan R. et al. (1977) Head injuries in Scottish hospitals Lancet 2, 696. Jennett B. and Teasdale G. (1979) Management of Head Injuries. Phdadelphta, Davies. (In the press.) Jennett B., Teasdale G., Braakman R. et al. (1978) Predicting outcome after head injury. (In the press.) Jennett B., Teasdale G , Galbralth S. et al. (1977) Severe head injuries in three countries J Neurol Neurosurg Psychiatr. 40, 291.

39

Leech P. (1974) Cerebrospinal fluid leakage, dural fistulae and meningxtis after basal skull fractures. Injury 6, 141. Lewm W. (1968) Rehabilitation after head injury. Br. Med. J. 1, 465 London P. S. (1967) Some observations on the course of events after severe injury of the head. Ann. R. Coll. Surg. Engl. 41, 460. Miller J D., Sweet R. C., Narayan R. et al. (1978) Early insults to the injured brain. JAMA. (In the press.) Price D. J. E. and Murray A. (1972) Influence of hypoxia and hypotenslon on recovery from head injury. Injury 2, 218. Reilly P. L., Graham D. I., Adams H. et al. (1975) Patients with head injury who talk and die. Lancet 2, 375. Rose J., Valtonen S. and Jennett B. (1977) Avoidable factors contributing to death after head injury. Br. Med. J. 2, 615. Sano K. (1965) Survey of the organlsation of services for the treatment of acute head injury in Japan. Exerpta Medica hit. Congr. Set. 110, 39. Scottish Home and Health Department (1975) Scottish Hospital hi-patient Statistics, 1974. Edinburgh, Information Services Division of Common Services Agency. Strang I., MacMillan R. and Jennett B. (1978) Head injuries in accident and emergency departments at Scottish hospitals, hljury. (In the press.) Teasdale G. and Jennett B. (1974) Assessment of coma and impaired consciousness. A practical scale. Lancet 2, 81. Yates D. W. (1977) Airway patency in fatal accidents. Br. Med. J. 2, 1249.

Requests for reprints shouM be addressed to Professor B. Jennett, Umverslty Department of Neurosurgery, Institute of Neurologtcal Sctences, Glasgow, G5 [ 4TF.

Preventable mortality and morbidity after head injury.

Injury, 10, 31-39 31 Printed in Great Britain Preventable mortality and morbidity after head injury Bryan Jennett and Joseph Carlin University Depa...
702KB Sizes 0 Downloads 0 Views