I
Volume 69 February 1976
137
Section of Neurology President J W D Bull MD
Meeting 6 February 1975
Sequele of Closed Head Injuries Dr A H Roberts
(Regional Neurological Centre, Brook Hospital, London SE18) Long-term Prognosis of Severe Accidental Head Injury There are still few published accounts of the ultimate outcome of severe head injury in civil life (Miller & Stern 1965, Rowbotham 1949, Fahy et al. 1967), and fewer attempts to examine the pattern of neural lesions and the personality and intellectual defects these head injuries cause (Russell 1966). I have therefore studied, from a consecutive series of 7000 cases of head injury admitted between 1948 and 1961 to the Radcliffe Infirmary, Oxford, every patient who was amnesic or unconscious for a week or more and had not had an intracranial infection or a spinal cord injury. There were 468, 6.7% of the total. Eleven were lost to the survey, and I examined the remaining 291 survivors ten to twenty-five years after their injury. In addition I studied a selected group of 68 patients from the Army and from Addenbrooke's Hospital, Cambridge, who had been unconscious for longer than a month. A close relative was interviewed in 80 %0 of cases.
Principal Patternis of Residutal Brain Damage Decerebrate demenitia: In this pattern decerebrate reflexes are the principal response to stimuli though semi-purposive movement may develop on one side after many months. There is usually
Athetoid pseudobulbar: In this pattern there is evidence of severe bilateral pyramidal damage with postural dystonia and often striking bradykinesia and fragmentary athetosis. There were 5% of these in the consecutive series. In several patients there was a remarkable discrepancy between the severity of the physical disability and the relative preservation of personality and intellectual function, suggesting, contrary to recent assumptions (Ommaya & Gennarelli 1974), that there may be a considerable contribution to the residual deficit by focal secondary infarction rather than diffuse traumatic damage. The brain stem pattern: In this the evidence of asymmetrical cerebellar and pyramidal lesions is the most striking feature. Although the pathology is clearly not confined to the brain stem it is certain that this condition reflects the lesions found extensively throughout the brain stem due either directly to trauma or secondary to compression from brain swelling (Tomlinson 1970). Such patients comprised 20% of the consecutive series. Hemiparetic: This was the most frequent residual pattern and it was found in 400%, but the signs of pyramidal damage were slight in the majority. A dense hemiparesis was usually associated with the preceding three patterns. Of the consecutive series 25 % had no detectable neural lesion, and 10% did not fall into any of these groups.
no detectable intellectual function but there may
ultimately be some form of emotional response to the spoken word or gesture. This condition is therefore more inclusive than the 'persistent vegetative state' of Jennett & Plum (1972). Prolonged survival is uncommon, though occasional cases set memorable records. In the consecutive series not one was still alive ten years after injury.
Patients whose level of responsiveness deteriorated after admission to hospital, or who developed a new or worsening hemiparesis comprised twothirds of the brain stem group, a third with the decerebrate dementia and athetoid pseudobulbar syndrome, but less than a quarter of the hemiparetic. The severity of the brain stem syndrome was greater among those who had deteriorated,
138 Proc. roy. Soc. Med. Volume 69 February 1976 5
N.8 L.. I.-
-.1
3.
00
I
2-
1.-
I
I
_
1 day lweek 1month 6 lyear 5 10 20 1 Fig Rate of Recovery ofpatients aged 16-25 with injury uncomplicated by brain penetration, traumatic or surgical, or compression. Coma less than a month. 0, patients who ultimately had no neural disability. *, patients who ultimately had disability graded 1.5 or worse. 5, decerebrate; 4.5, purposive or semi-purposive response; 4, bed-bound confusion, 3.5, walking assisted; 3, neural disability precluding normal activities; 2.5, 2 and 1.5, limiting these seriously moderately and slightly; 1, minimal disability; O.5, just demonstrable abnormal signs
2
initially decerebrate survived, but of those who did none was severely disabled by neural lesions. Children aged 5 to 15, even though unconscious longer than a month, continued to improve over many years and the final disability was often mild (Fig 2). Combining two age groups, 16 to 25 and 26 to 45 years, in whom residual disability is negligible or minimal, Fig 3 shows that rapid improvement in the first month accurately predicts the ultimate outcome, though there may be moderate disability for as long as two years. Patients with extradural or subdural hkmatoma, or brain penetration, whether traumatic or surgical, may become decerebrate for short periods before their operation, but ultimate disability is minimal if there is no secondary infarction. The time-scale of recovery is lengthened but again it is the rate at which this takes place that is most closely related to the final degree of disability. Cerebral infarction due to prolonged compression no doubt accounts for the fact that more in each age group remain severely disabled than in those whose injuries were uncomplicated. The relationship between length of posttraumatic amnesia (PTA) and ultimate physical disability (Russell & Smith 1961) is in general as expected (Fig 4). The proportion of patients disabled and the degree of disability was related to the PTA, and to a lesser degree to age. However, the decline in the level of employment is due in two-thirds of cases to impairment of personality or intellectual function, and in only one-third to other neural disability.
suggesting the effects of potentially treatable brain swelling rather than trauma itself was the cause.
Disability profiles were scored for various components of the neural lesion: akinesia, imbalance, ataxia, paresis and sensory deficit. The same ten-point scale was used to assess the initial level of neurological responsiveness, progress and final disability (see Fig 1). In the consecutive series the early state of those patients who were later left with no disability was compared with those who were left with severe disability (Fig l). Although prognosis depends to some extent on whether there was decerebration or only semi-purposive movement, or merely bed-bound confusion (Overgaard et al. 1973), there is overlap, and it is the rate of recovery in the first months that most clearly predicts outcome. These patients were injured between 16 and 25 years of age and were unconscious for less than a month. Very few were left with any real disability. Among those who were ten to twenty years older at the time of injury few made complete recovery and a substantial number had severe disability. Over the age of 45 no one
LI.~~~~~~~~~~~~~~~~~J U,
1
Io
iweek Imonth 3 6 1year 5 10 20 Iday Fig 2 Rate of recovery in neural disability in patients aged 5-15 with uncomplicated injury and coma lasting one month or longer. Each line represents a single patient
3
Sectionz of Neurology
139
Disability profiles were also used to quantify 100 C rade personality change and memory defects. Ultimate n25+ outcome was scored in terms of the domestic, 1- 5 - 2 0 -1 social or occupational disability these caused. In the most commonly recurring pattern of personality change there was disabling frontal U) euphoria, disinhibition or anergia, which was usually associated with intense irritability as 50 noted by Kremer (1944). Uncontrollable outbursts of rage of the kind described by Hooper -0 et al. (1945) were often the reason for institutional commitment rather than dementia, though severely defective memory was usually associated when this type of personality change was marked. A less common pattern, in which there was 5- 25 26+ in years frontal personality change without irritability Age < 4 >4 Weeks P.T.A. < 4 >4 but with memory deficit, occurred frequently in No.of patients those younger than 35 at the time of their injury. Fig 4 Residual neural disability in In 10 of these the injury produced improvement patients with injury uncomplicated by in personality, as Storey (1967) and Logue et al. brain penetration, traumatic or (1968) .found in a few cases after hemorrhage surgical, or compression from anterior cerebral aneurysms. It has long been thought that there was a A combination of mild defect in memory with disabling anxiety, often as phobic imbalance relationship between the side of impact on the (Pratt & McKenzie 1958, Roberts 1963), was a head and the side of hemiparesis (Breasted 1930). common sequel to injury after the age of 35. In the consecutive series with uncomplicated Disability tended to decline over two years or so, injury, in whom the site of impact to the head but there was often considerable restriction of could be determined, 80 % who had a PTA of less normal activities which persisted and this without than four weeks had an ipsilateral paresis. The proportion declined as the PTA lengthened. In any other evidence of a neural lesion. injury complicated by surface compression only half had an ipsilateral paresis. This may be further evidence of the contribution made to 5 neural damage by secondary infarction. In the consecutive uncomplicated series of 214 the site of impact to the head could be identified in 175, N-35 most by fractures (Gurdjian & Webster 1958), and in 104, significantly more than half, impact N-33o 4. was on the right. Since the criterion for inclusion in the study was PTA of a week or more, it appears that the length of PTA or its assessment is to some extent determined by the side the head 3 J;t 2 is struck. Verbal and non-verbal cognitive functions were significantly more affected in those struck v) on the right side of the head than on the left I 2-i (Smith 1974), but there is no clear relationship between the side of head impact and personality uq Q2 change. cx Post-traumatic epilepsy occurred in 75 patients. 1In only 4 of these was the epilepsy disabling, though another 4 died of their epilepsy, and the cc attacks in most remitted or decreased to negligible numbers over many years. This is a better outcome than was found in a larger series followed 6 lyear 5 10 20 iday lweek lmth. 3 for a shorter time (Jennett 1962). Fits often Fig 3 Rate of recovery in patients with occurred for the first time after a substantial uncomplicated injury and minimal or negligible delay, in 13 % after five years, and in one case residual neural disability. * patients aged 16-25; after as long as fifteen years. o, patients aged 26-45 I.,J
-0
I..
-
U)
I
I
I
I
I
140 Proc. roy. Soc. Med. Volume 69 February 1976
4
Progressive deterioration is a feature in some Professor Bryan Jennett cases of repeated minor head injury in boxers (Department of Neurosurgery, Institute of (Roberts 1969), but remains a questionable Neurological Sciences, Glasgow G51 4TF, entity after a single severe injury (Corsellis & and the University of Glasgow) Brierley 1959). Fifteen patients in this series had progressive intellectual deterioration which could Predicting Outcome after Head Injury have been due to normal ageing superimposed on [Abstract] traumatic defect. Three others demented after improvement and could have had Alzheimer's We have heard from other speakers what can disease unrelated to the injury. In 3 paretic happen after head injury, but the doctor dealing limbs became more spastic after many years. Five with the head injured patient wants to know more, all with complicated injuries, developed what will happen. He must decide whether or not what appeared to be a communicating hydro- to admit the mild case to hospital and whether or cephalus after a delay of some years, and in 4 of not to persist with all-out intensive care for a severe injury. Deprived of reliable predictive data these after a further head injury due to falls. In almost 80 % of the consecutive series deaths doctors admit thousands of mild injuries briefly due to head injury took place within six months. to hospital every year, although only a tiny proIn those unconscious for longer than a month portion develop complications, whilst an undue who died of their injury there were some pro- proportion of effort on severe injuries is liable longed survivals, most of these mute, akinetic to be expended on those who either do not and decerebrate. It is to be hoped that earlier survive or are left very severely disabled. If we and more accurate prediction in the future may could recognize soon after injury which mildly injured patients were at risk from complications,, limit the number of these. and which severely injured patients had a reasonAcknowledgments: I am indebted to Mr Walpole able chance of recovery, we could concentrate Lewin, whose punchcard records and continuing our maximum effort on those patients. Reliable interest in head injury made this study possible; predictions would also make it possible to assess to Mrs D M Weir, Psychiatric Social Worker, for the efficacy of newly proposed methods of her invaluable assistance; and to Miss M Har- management. Prognosis is a probability statement which greaves for tracing many of the patients. The survey was financed by a research grant from the assumes a logical relationship between outcome and certain items of antecedent data. It requires Department of Health aqd Social Security. rigorous definition of outcome, and then identiREFERENCES fication of which items of data have predictive Breasted J H (1930) The Edwin Smith Surgical power and what is their relative influence alone Papyrus. University of Chicago Press; p 203 and in combination. The simplest kind of preCorsellis J A N & Brierley J B (1959) Journal of Mental Science 105, 714 diction is of a specific complication, such as Fahy T J, Irving M H & Millac P (1967) Lancet ii, 475 epilepsy or intracranial hmematoma, because Gurdjian E S & Webster J E (1958) Head Injuries, Mechanisms, Diagnosis and Management. Little, Brown, Boston; p 62 there are only two outcomes - either the compliHooper R S, McGregor J M & Nathan P W cation appears or it does not. The probability (1945) Journal of Mental Science 91, 458 that traumatic epilepsy will develop can now be Jennett B & Plum F (1972) Lancet i, 734 Jennett W B (1962) Epilepsy after Blunt calculated accurately on the basis of very few Head Injuries. Heinemann, London items of clinical data, and patients in the low Kremer M (1944) Proceedings of the Royal Society of Medicine 37, 564 risk category can be reassured and the others Logue V, Durward M, Pratt R T C, Piercy M & Nixon W L B given anticonvulsants (Jennett 1975). It is hoped (1968) British Journal ofPsychiatry 114, 137 Miller H & Stern G (1965) Lancet i, 225 that the risk of acute intracranial hzematoma may Ommaya A K & Gennarelli T A likewise prove predictable. Preliminary studies (1974) Brain 97, 633 Overgaard J, Christensen S, Haase J, Hern 0, Havid-Hansen 0, indicate that the risk in a fully conscious adult Land A, Pederson K K & Tweed W A (1973) Lancet ii, 631 without a skull fracture, even if he has been Pratt R T C & McKenzie W (1958) Lancet ii, 347 briefly unconscious, is very small indeed (GalRoberts A H (1963) British Journal ofPsychiatry 110, 191 braith 1973). Under the present system of large (1969) Brain Damage in Boxers. A Study of Prevalence of scale admission of mild head injuries to hospital, Traumatic Encephalopathy among Ex-professional Boxers. Pitman, London hmmatomas are frequently overlooked for many Rowbotham G F (1949) Journal of Mental Science 14, 336 hours. This may derive from an inability to mainRussell W R (1966) Proceedings of the Royal Society of Medicine 59,266 tain a high state of vigilance when the complicaRussell W R & Smith A (1961) Archives of Neurology 5, 4 tion rate is so low. More selective admission Smith E might not only reduce demand on acute surgical (1974) Journal ofNeurology, Neurosurgery and Psychiatry 37, 719 Storey P B (1967) British Medical Journal iii, 261 beds but might result in more effective manageTomlinson B E ment of those who are taken in. (1970) Journal oJ Clinical Pathology 23, Suppl 4, 154
Volume 69 February 1976
137
Section of Neurology President J W D Bull MD
Meeting 6 February 1975
Sequele of Closed Head Injuries Dr A H Roberts
(Regional Neurological Centre, Brook Hospital, London SE18) Long-term Prognosis of Severe Accidental Head Injury There are still few published accounts of the ultimate outcome of severe head injury in civil life (Miller & Stern 1965, Rowbotham 1949, Fahy et al. 1967), and fewer attempts to examine the pattern of neural lesions and the personality and intellectual defects these head injuries cause (Russell 1966). I have therefore studied, from a consecutive series of 7000 cases of head injury admitted between 1948 and 1961 to the Radcliffe Infirmary, Oxford, every patient who was amnesic or unconscious for a week or more and had not had an intracranial infection or a spinal cord injury. There were 468, 6.7% of the total. Eleven were lost to the survey, and I examined the remaining 291 survivors ten to twenty-five years after their injury. In addition I studied a selected group of 68 patients from the Army and from Addenbrooke's Hospital, Cambridge, who had been unconscious for longer than a month. A close relative was interviewed in 80 %0 of cases.
Principal Patternis of Residutal Brain Damage Decerebrate demenitia: In this pattern decerebrate reflexes are the principal response to stimuli though semi-purposive movement may develop on one side after many months. There is usually
Athetoid pseudobulbar: In this pattern there is evidence of severe bilateral pyramidal damage with postural dystonia and often striking bradykinesia and fragmentary athetosis. There were 5% of these in the consecutive series. In several patients there was a remarkable discrepancy between the severity of the physical disability and the relative preservation of personality and intellectual function, suggesting, contrary to recent assumptions (Ommaya & Gennarelli 1974), that there may be a considerable contribution to the residual deficit by focal secondary infarction rather than diffuse traumatic damage. The brain stem pattern: In this the evidence of asymmetrical cerebellar and pyramidal lesions is the most striking feature. Although the pathology is clearly not confined to the brain stem it is certain that this condition reflects the lesions found extensively throughout the brain stem due either directly to trauma or secondary to compression from brain swelling (Tomlinson 1970). Such patients comprised 20% of the consecutive series. Hemiparetic: This was the most frequent residual pattern and it was found in 400%, but the signs of pyramidal damage were slight in the majority. A dense hemiparesis was usually associated with the preceding three patterns. Of the consecutive series 25 % had no detectable neural lesion, and 10% did not fall into any of these groups.
no detectable intellectual function but there may
ultimately be some form of emotional response to the spoken word or gesture. This condition is therefore more inclusive than the 'persistent vegetative state' of Jennett & Plum (1972). Prolonged survival is uncommon, though occasional cases set memorable records. In the consecutive series not one was still alive ten years after injury.
Patients whose level of responsiveness deteriorated after admission to hospital, or who developed a new or worsening hemiparesis comprised twothirds of the brain stem group, a third with the decerebrate dementia and athetoid pseudobulbar syndrome, but less than a quarter of the hemiparetic. The severity of the brain stem syndrome was greater among those who had deteriorated,
138 Proc. roy. Soc. Med. Volume 69 February 1976 5
N.8 L.. I.-
-.1
3.
00
I
2-
1.-
I
I
_
1 day lweek 1month 6 lyear 5 10 20 1 Fig Rate of Recovery ofpatients aged 16-25 with injury uncomplicated by brain penetration, traumatic or surgical, or compression. Coma less than a month. 0, patients who ultimately had no neural disability. *, patients who ultimately had disability graded 1.5 or worse. 5, decerebrate; 4.5, purposive or semi-purposive response; 4, bed-bound confusion, 3.5, walking assisted; 3, neural disability precluding normal activities; 2.5, 2 and 1.5, limiting these seriously moderately and slightly; 1, minimal disability; O.5, just demonstrable abnormal signs
2
initially decerebrate survived, but of those who did none was severely disabled by neural lesions. Children aged 5 to 15, even though unconscious longer than a month, continued to improve over many years and the final disability was often mild (Fig 2). Combining two age groups, 16 to 25 and 26 to 45 years, in whom residual disability is negligible or minimal, Fig 3 shows that rapid improvement in the first month accurately predicts the ultimate outcome, though there may be moderate disability for as long as two years. Patients with extradural or subdural hkmatoma, or brain penetration, whether traumatic or surgical, may become decerebrate for short periods before their operation, but ultimate disability is minimal if there is no secondary infarction. The time-scale of recovery is lengthened but again it is the rate at which this takes place that is most closely related to the final degree of disability. Cerebral infarction due to prolonged compression no doubt accounts for the fact that more in each age group remain severely disabled than in those whose injuries were uncomplicated. The relationship between length of posttraumatic amnesia (PTA) and ultimate physical disability (Russell & Smith 1961) is in general as expected (Fig 4). The proportion of patients disabled and the degree of disability was related to the PTA, and to a lesser degree to age. However, the decline in the level of employment is due in two-thirds of cases to impairment of personality or intellectual function, and in only one-third to other neural disability.
suggesting the effects of potentially treatable brain swelling rather than trauma itself was the cause.
Disability profiles were scored for various components of the neural lesion: akinesia, imbalance, ataxia, paresis and sensory deficit. The same ten-point scale was used to assess the initial level of neurological responsiveness, progress and final disability (see Fig 1). In the consecutive series the early state of those patients who were later left with no disability was compared with those who were left with severe disability (Fig l). Although prognosis depends to some extent on whether there was decerebration or only semi-purposive movement, or merely bed-bound confusion (Overgaard et al. 1973), there is overlap, and it is the rate of recovery in the first months that most clearly predicts outcome. These patients were injured between 16 and 25 years of age and were unconscious for less than a month. Very few were left with any real disability. Among those who were ten to twenty years older at the time of injury few made complete recovery and a substantial number had severe disability. Over the age of 45 no one
LI.~~~~~~~~~~~~~~~~~J U,
1
Io
iweek Imonth 3 6 1year 5 10 20 Iday Fig 2 Rate of recovery in neural disability in patients aged 5-15 with uncomplicated injury and coma lasting one month or longer. Each line represents a single patient
3
Sectionz of Neurology
139
Disability profiles were also used to quantify 100 C rade personality change and memory defects. Ultimate n25+ outcome was scored in terms of the domestic, 1- 5 - 2 0 -1 social or occupational disability these caused. In the most commonly recurring pattern of personality change there was disabling frontal U) euphoria, disinhibition or anergia, which was usually associated with intense irritability as 50 noted by Kremer (1944). Uncontrollable outbursts of rage of the kind described by Hooper -0 et al. (1945) were often the reason for institutional commitment rather than dementia, though severely defective memory was usually associated when this type of personality change was marked. A less common pattern, in which there was 5- 25 26+ in years frontal personality change without irritability Age < 4 >4 Weeks P.T.A. < 4 >4 but with memory deficit, occurred frequently in No.of patients those younger than 35 at the time of their injury. Fig 4 Residual neural disability in In 10 of these the injury produced improvement patients with injury uncomplicated by in personality, as Storey (1967) and Logue et al. brain penetration, traumatic or (1968) .found in a few cases after hemorrhage surgical, or compression from anterior cerebral aneurysms. It has long been thought that there was a A combination of mild defect in memory with disabling anxiety, often as phobic imbalance relationship between the side of impact on the (Pratt & McKenzie 1958, Roberts 1963), was a head and the side of hemiparesis (Breasted 1930). common sequel to injury after the age of 35. In the consecutive series with uncomplicated Disability tended to decline over two years or so, injury, in whom the site of impact to the head but there was often considerable restriction of could be determined, 80 % who had a PTA of less normal activities which persisted and this without than four weeks had an ipsilateral paresis. The proportion declined as the PTA lengthened. In any other evidence of a neural lesion. injury complicated by surface compression only half had an ipsilateral paresis. This may be further evidence of the contribution made to 5 neural damage by secondary infarction. In the consecutive uncomplicated series of 214 the site of impact to the head could be identified in 175, N-35 most by fractures (Gurdjian & Webster 1958), and in 104, significantly more than half, impact N-33o 4. was on the right. Since the criterion for inclusion in the study was PTA of a week or more, it appears that the length of PTA or its assessment is to some extent determined by the side the head 3 J;t 2 is struck. Verbal and non-verbal cognitive functions were significantly more affected in those struck v) on the right side of the head than on the left I 2-i (Smith 1974), but there is no clear relationship between the side of head impact and personality uq Q2 change. cx Post-traumatic epilepsy occurred in 75 patients. 1In only 4 of these was the epilepsy disabling, though another 4 died of their epilepsy, and the cc attacks in most remitted or decreased to negligible numbers over many years. This is a better outcome than was found in a larger series followed 6 lyear 5 10 20 iday lweek lmth. 3 for a shorter time (Jennett 1962). Fits often Fig 3 Rate of recovery in patients with occurred for the first time after a substantial uncomplicated injury and minimal or negligible delay, in 13 % after five years, and in one case residual neural disability. * patients aged 16-25; after as long as fifteen years. o, patients aged 26-45 I.,J
-0
I..
-
U)
I
I
I
I
I
140 Proc. roy. Soc. Med. Volume 69 February 1976
4
Progressive deterioration is a feature in some Professor Bryan Jennett cases of repeated minor head injury in boxers (Department of Neurosurgery, Institute of (Roberts 1969), but remains a questionable Neurological Sciences, Glasgow G51 4TF, entity after a single severe injury (Corsellis & and the University of Glasgow) Brierley 1959). Fifteen patients in this series had progressive intellectual deterioration which could Predicting Outcome after Head Injury have been due to normal ageing superimposed on [Abstract] traumatic defect. Three others demented after improvement and could have had Alzheimer's We have heard from other speakers what can disease unrelated to the injury. In 3 paretic happen after head injury, but the doctor dealing limbs became more spastic after many years. Five with the head injured patient wants to know more, all with complicated injuries, developed what will happen. He must decide whether or not what appeared to be a communicating hydro- to admit the mild case to hospital and whether or cephalus after a delay of some years, and in 4 of not to persist with all-out intensive care for a severe injury. Deprived of reliable predictive data these after a further head injury due to falls. In almost 80 % of the consecutive series deaths doctors admit thousands of mild injuries briefly due to head injury took place within six months. to hospital every year, although only a tiny proIn those unconscious for longer than a month portion develop complications, whilst an undue who died of their injury there were some pro- proportion of effort on severe injuries is liable longed survivals, most of these mute, akinetic to be expended on those who either do not and decerebrate. It is to be hoped that earlier survive or are left very severely disabled. If we and more accurate prediction in the future may could recognize soon after injury which mildly injured patients were at risk from complications,, limit the number of these. and which severely injured patients had a reasonAcknowledgments: I am indebted to Mr Walpole able chance of recovery, we could concentrate Lewin, whose punchcard records and continuing our maximum effort on those patients. Reliable interest in head injury made this study possible; predictions would also make it possible to assess to Mrs D M Weir, Psychiatric Social Worker, for the efficacy of newly proposed methods of her invaluable assistance; and to Miss M Har- management. Prognosis is a probability statement which greaves for tracing many of the patients. The survey was financed by a research grant from the assumes a logical relationship between outcome and certain items of antecedent data. It requires Department of Health aqd Social Security. rigorous definition of outcome, and then identiREFERENCES fication of which items of data have predictive Breasted J H (1930) The Edwin Smith Surgical power and what is their relative influence alone Papyrus. University of Chicago Press; p 203 and in combination. The simplest kind of preCorsellis J A N & Brierley J B (1959) Journal of Mental Science 105, 714 diction is of a specific complication, such as Fahy T J, Irving M H & Millac P (1967) Lancet ii, 475 epilepsy or intracranial hmematoma, because Gurdjian E S & Webster J E (1958) Head Injuries, Mechanisms, Diagnosis and Management. Little, Brown, Boston; p 62 there are only two outcomes - either the compliHooper R S, McGregor J M & Nathan P W cation appears or it does not. The probability (1945) Journal of Mental Science 91, 458 that traumatic epilepsy will develop can now be Jennett B & Plum F (1972) Lancet i, 734 Jennett W B (1962) Epilepsy after Blunt calculated accurately on the basis of very few Head Injuries. Heinemann, London items of clinical data, and patients in the low Kremer M (1944) Proceedings of the Royal Society of Medicine 37, 564 risk category can be reassured and the others Logue V, Durward M, Pratt R T C, Piercy M & Nixon W L B given anticonvulsants (Jennett 1975). It is hoped (1968) British Journal ofPsychiatry 114, 137 Miller H & Stern G (1965) Lancet i, 225 that the risk of acute intracranial hzematoma may Ommaya A K & Gennarelli T A likewise prove predictable. Preliminary studies (1974) Brain 97, 633 Overgaard J, Christensen S, Haase J, Hern 0, Havid-Hansen 0, indicate that the risk in a fully conscious adult Land A, Pederson K K & Tweed W A (1973) Lancet ii, 631 without a skull fracture, even if he has been Pratt R T C & McKenzie W (1958) Lancet ii, 347 briefly unconscious, is very small indeed (GalRoberts A H (1963) British Journal ofPsychiatry 110, 191 braith 1973). Under the present system of large (1969) Brain Damage in Boxers. A Study of Prevalence of scale admission of mild head injuries to hospital, Traumatic Encephalopathy among Ex-professional Boxers. Pitman, London hmmatomas are frequently overlooked for many Rowbotham G F (1949) Journal of Mental Science 14, 336 hours. This may derive from an inability to mainRussell W R (1966) Proceedings of the Royal Society of Medicine 59,266 tain a high state of vigilance when the complicaRussell W R & Smith A (1961) Archives of Neurology 5, 4 tion rate is so low. More selective admission Smith E might not only reduce demand on acute surgical (1974) Journal ofNeurology, Neurosurgery and Psychiatry 37, 719 Storey P B (1967) British Medical Journal iii, 261 beds but might result in more effective manageTomlinson B E ment of those who are taken in. (1970) Journal oJ Clinical Pathology 23, Suppl 4, 154
