Predicting recovery in patients suffering from traumatic brain injury by using admission variables and physiological data: a comparison between decision tree analysis and logistic regression

Restricted access

Object. Decision tree analysis highlights patient subgroups and critical values in variables assessed. Importantly, the results are visually informative and often present clear clinical interpretation about risk factors faced by patients in these subgroups. The aim of this prospective study was to compare results of logistic regression with those of decision tree analysis of an observational, head-injury data set, including a wide range of secondary insults and 12-month outcomes.

Methods. One hundred twenty-four adult head-injured patients were studied during their stay in an intensive care unit by using a computerized data collection system. Verified values falling outside threshold limits were analyzed according to insult grade and duration with the aid of logistic regression. A decision tree was automatically produced from root node to target classes (Glasgow Outcome Scale [GOS] score).

Among 69 patients, in whom eight insult categories could be assessed, outcome at 12 months was analyzed using logistic regression to determine the relative influence of patient age, admission Glasgow Coma Scale score, Injury Severity Score (ISS), pupillary response on admission, and insult duration. The most significant predictors of mortality in this patient set were duration of hypotensive, pyrexic, and hypoxemic insults. When good and poor outcomes were compared, hypotensive insults and pupillary response on admission were significant.

Using decision tree analysis, the authors found that hypotension and low cerebral perfusion pressure (CPP) are the best predictors of death, with a 9.2% improvement in predictive accuracy (PA) over that obtained by simply predicting the largest outcome category as the outcome for each patient. Hypotension was a significant predictor of poor outcome (GOS Score 1–3). Low CPP, patient age, hypocarbia, and pupillary response were also good predictors of outcome (good/poor), with a 5.1% improvement in PA. In certain subgroups of patients pyrexia was a predictor of good outcome.

Conclusions. Decision tree analysis confirmed some of the results of logistic regression and challenged others. This investigation shows that there is knowledge to be gained from analyzing observational data with the aid of decision tree analysis.

Article Information

Contributor Notes

Address reprint requests to: Peter J. D. Andrews, M.D., Department of Anaesthetics, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, United Kingdom. email: Pandrews@ed.ac.uk.
Headings
References
  • 1.

    Alvarez MNava JMRue Met al: Mortality prediction in head trauma patients: performance of Glasgow Coma Score and general severity systems. Crit Care Med 26:1421481998Alvarez M Nava JM Rue M et al: Mortality prediction in head trauma patients: performance of Glasgow Coma Score and general severity systems. Crit Care Med 26:142–148 1998

    • Search Google Scholar
    • Export Citation
  • 2.

    Andrews PJ: Head injury: complications and management. Curr Opin Anaesthesiol 11:4734771998Andrews PJ: Head injury: complications and management. Curr Opin Anaesthesiol 11:473–477 1998

    • Search Google Scholar
    • Export Citation
  • 3.

    Andrews PJDearden NMMiller JD: Jugular bulb cannulation: description of a cannulation technique and validation of a new continuous monitor. Br J Anaesth 67:5535581991Andrews PJ Dearden NM Miller JD: Jugular bulb cannulation: description of a cannulation technique and validation of a new continuous monitor. Br J Anaesth 67:553–558 1991

    • Search Google Scholar
    • Export Citation
  • 4.

    Andrews PJPiper IRDearden NMet al: Secondary insults during intrahospital transport of head-injured patients. Lancet 335:3273301990Andrews PJ Piper IR Dearden NM et al: Secondary insults during intrahospital transport of head-injured patients. Lancet 335:327–330 1990

    • Search Google Scholar
    • Export Citation
  • 5.

    Baker SPO'Neill BHaddon W Jret al: The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 14:1871961974Baker SP O'Neill B Haddon W Jr et al: The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 14:187–196 1974

    • Search Google Scholar
    • Export Citation
  • 6.

    Bekar AIpekoglu ZTureyen Ket al: Secondary insults during intrahospital transport of neurosurgical intensive care patients. Neurosurg Rev 21:981011998Bekar A Ipekoglu Z Tureyen K et al: Secondary insults during intrahospital transport of neurosurgical intensive care patients. Neurosurg Rev 21:98–101 1998

    • Search Google Scholar
    • Export Citation
  • 7.

    Block EF: Diagnostic modalities in acute trauma. New Horiz 7:10251999Block EF: Diagnostic modalities in acute trauma. New Horiz 7:10–25 1999

    • Search Google Scholar
    • Export Citation
  • 8.

    Blumer AC: ID3 Decision tree learning algorithm. Electrical Engineering and Computer Science at Tufts University. (http://www.eecs.tufts.edu/g/150ML/slides/id3.html) [Accessed 5 June 2002]Blumer AC: ID3 Decision tree learning algorithm. Electrical Engineering and Computer Science at Tufts University. (http://www.eecs.tufts.edu/g/150ML/slides/id3.html) [Accessed 5 June 2002]

    • Search Google Scholar
    • Export Citation
  • 9.

    Chesnut RM: The management of severe traumatic brain injury. Emerg Med Clin North Am 15:5816041997Chesnut RM: The management of severe traumatic brain injury. Emerg Med Clin North Am 15:581–604 1997

    • Search Google Scholar
    • Export Citation
  • 10.

    Chesnut RM: Medical management of severe head injury: present and future. New Horiz 3:5815931995Chesnut RM: Medical management of severe head injury: present and future. New Horiz 3:581–593 1995

    • Search Google Scholar
    • Export Citation
  • 11.

    Chesnut RM: Secondary brain insults after head injury: clinical perspectives. New Horiz 3:3663751995Chesnut RM: Secondary brain insults after head injury: clinical perspectives. New Horiz 3:366–375 1995

    • Search Google Scholar
    • Export Citation
  • 12.

    Choi SCMuizelaar JPBarnes TYet al: Prediction tree for severely head-injured patients. J Neurosurg 75:2512551991Choi SC Muizelaar JP Barnes TY et al: Prediction tree for severely head-injured patients. J Neurosurg 75:251–255 1991

    • Search Google Scholar
    • Export Citation
  • 13.

    Davies GDeakin CWilson A: The effect of a rigid collar on intracranial pressure. Injury 27:6476491996Davies G Deakin C Wilson A: The effect of a rigid collar on intracranial pressure. Injury 27:647–649 1996

    • Search Google Scholar
    • Export Citation
  • 14.

    Dearden NM: Mechanisms and prevention of secondary brain damage during intensive care. Clin Neuropathol 17:2212281998Dearden NM: Mechanisms and prevention of secondary brain damage during intensive care. Clin Neuropathol 17:221–228 1998

    • Search Google Scholar
    • Export Citation
  • 15.

    Doberstein CEHovda DABecker DP: Clinical considerations in the reduction of secondary brain injury. Ann Emerg Med 22:9939971993Doberstein CE Hovda DA Becker DP: Clinical considerations in the reduction of secondary brain injury. Ann Emerg Med 22:993–997 1993

    • Search Google Scholar
    • Export Citation
  • 16.

    Dunn LT: Secondary insults during the interhospital transfer of head-injured patients: an audit of transfers in the Mersey Region. Injury 28:4274311997Dunn LT: Secondary insults during the interhospital transfer of head-injured patients: an audit of transfers in the Mersey Region. Injury 28:427–431 1997

    • Search Google Scholar
    • Export Citation
  • 17.

    Dzeroski SLavrac N: Rule induction and instance-based learning applied in medical diagnosis. Technol Health Care 4:2032211996Dzeroski S Lavrac N: Rule induction and instance-based learning applied in medical diagnosis. Technol Health Care 4:203–221 1996

    • Search Google Scholar
    • Export Citation
  • 18.

    Gennarelli TAChampion HRCopes WSet al: Comparison of mortality, morbidity, and severity of 59,713 head injured patients with 114,447 patients with extracranial injuries. J Trauma 37:9629681994Gennarelli TA Champion HR Copes WS et al: Comparison of mortality morbidity and severity of 59713 head injured patients with 114447 patients with extracranial injuries. J Trauma 37:962–968 1994

    • Search Google Scholar
    • Export Citation
  • 19.

    Graham DIAdams JHNicoll JAet al: The nature, distribution and causes of traumatic brain injury. Brain Pathol 5:3974061995Graham DI Adams JH Nicoll JA et al: The nature distribution and causes of traumatic brain injury. Brain Pathol 5:397–406 1995

    • Search Google Scholar
    • Export Citation
  • 20.

    Graham DIFord IAdams JHet al: Ischaemic brain damage is still common in fatal non-missile head injury. J Neurol Neurosurg Psychiatry 52:3463501989Graham DI Ford I Adams JH et al: Ischaemic brain damage is still common in fatal non-missile head injury. J Neurol Neurosurg Psychiatry 52:346–350 1989

    • Search Google Scholar
    • Export Citation
  • 21.

    Jennett BBond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:4804841975Jennett B Bond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:480–484 1975

    • Search Google Scholar
    • Export Citation
  • 22.

    Jones PAAndrews PJMidgley Set al: Measuring the burden of secondary insults in head-injured patients during intensive care. J Neurosurg Anesthesiol 6:4141994Jones PA Andrews PJ Midgley S et al: Measuring the burden of secondary insults in head-injured patients during intensive care. J Neurosurg Anesthesiol 6:4–14 1994

    • Search Google Scholar
    • Export Citation
  • 23.

    Juul NMorris GFMarshall SBet al: Intracranial hypertension and cerebral perfusion pressure: influence on neurological deterioration and outcome in severe head injury. The Executive Committee of the International Selfotel Trial. J Neurosurg 92:162000Juul N Morris GF Marshall SB et al: Intracranial hypertension and cerebral perfusion pressure: influence on neurological deterioration and outcome in severe head injury. The Executive Committee of the International Selfotel Trial. J Neurosurg 92:1–6 2000

    • Search Google Scholar
    • Export Citation
  • 24.

    Knaus WA: APACHE 1978–2001: the development of a quality assurance system based on prognosis: milestones and personal reflections. Arch Surg 137:37412002Knaus WA: APACHE 1978–2001: the development of a quality assurance system based on prognosis: milestones and personal reflections. Arch Surg 137:37–41 2002

    • Search Google Scholar
    • Export Citation
  • 25.

    Lavrac N: Selected techniques for data mining in medicine. Artif Intell Med 16:3231999Lavrac N: Selected techniques for data mining in medicine. Artif Intell Med 16:3–23 1999

    • Search Google Scholar
    • Export Citation
  • 26.

    Maas AIDearden MTeasdale GMet al: EBIC-guidelines for management of severe head injury in adults. European Brain Injury Consortium. Acta Neurochir 139:2862941997Maas AI Dearden M Teasdale GM et al: EBIC-guidelines for management of severe head injury in adults. European Brain Injury Consortium. Acta Neurochir 139:286–294 1997

    • Search Google Scholar
    • Export Citation
  • 27.

    McQuatt A: Using machine learning techniques to predict clinical outcome. Thesis. Aberdeen: University of Aberdeen1998McQuatt A: Using machine learning techniques to predict clinical outcome. Thesis. Aberdeen: University of Aberdeen 1998

    • Search Google Scholar
    • Export Citation
  • 28.

    Morganti-Kossmann MCHans VHLenzlinger PMet al: TGF-beta is elevated in the CSF of patients with severe traumatic brain injuries and parallels blood-brain barrier function. J Neurotrauma 16:6176281999Morganti-Kossmann MC Hans VH Lenzlinger PM et al: TGF-beta is elevated in the CSF of patients with severe traumatic brain injuries and parallels blood-brain barrier function. J Neurotrauma 16:617–628 1999

    • Search Google Scholar
    • Export Citation
  • 29.

    Pilah IAMladenic DLavrac Net al: Data analysis of patients with severe head injury in Lavrac NKeravnou ETZupan B (eds): Intelligent Data Analysis in Medicine and Pharmacology. Boston: Kluwer Academic1997Pilah IA Mladenic D Lavrac N et al: Data analysis of patients with severe head injury in Lavrac N Keravnou ET Zupan B (eds): Intelligent Data Analysis in Medicine and Pharmacology. Boston: Kluwer Academic 1997

    • Search Google Scholar
    • Export Citation
  • 30.

    Quinlan JR: Data mining tools C5.0 and See5. Rulequest Research. (http://www.rulequest.com/see5-info.html) [Accessed 5 June 2002]Quinlan JR: Data mining tools C5.0 and See5. Rulequest Research. (http://www.rulequest.com/see5-info.html) [Accessed 5 June 2002]

    • Search Google Scholar
    • Export Citation
  • 31.

    Quinlan JR: Induction of decision trees. Mach Learn 1:811061986Quinlan JR: Induction of decision trees. Mach Learn 1:81–106 1986

    • Search Google Scholar
    • Export Citation
  • 32.

    Quinlan JR: Simplifying decision trees. Int J Hum-Comput St 51:4975101999Quinlan JR: Simplifying decision trees. Int J Hum-Comput St 51:497–510 1999

    • Search Google Scholar
    • Export Citation
  • 33.

    Robertson CSGopinath SPGoodman JCet al: SjvO2 monitoring in head-injured patients. J Neurotrauma 12:8918961995Robertson CS Gopinath SP Goodman JC et al: SjvO2 monitoring in head-injured patients. J Neurotrauma 12:891–896 1995

    • Search Google Scholar
    • Export Citation
  • 34.

    Sarnaik APLieh-Lai MW: Transporting the neurologically compromised child. Pediatr Clin North Am 40:3373541993Sarnaik AP Lieh-Lai MW: Transporting the neurologically compromised child. Pediatr Clin North Am 40:337–354 1993

    • Search Google Scholar
    • Export Citation
  • 35.

    Signorini DFAndrews PJJones PAet al: Adding insult to injury: the prognostic value of early secondary insults for survival after traumatic brain injury. J Neurol Neurosurg Psychiatry 66:26311999Signorini DF Andrews PJ Jones PA et al: Adding insult to injury: the prognostic value of early secondary insults for survival after traumatic brain injury. J Neurol Neurosurg Psychiatry 66:26–31 1999

    • Search Google Scholar
    • Export Citation
  • 36.

    Verweij BHMuizelaar JP: Avoiding secondary brain injury after severe head trauma: monitoring and management. J Craniomaxillofac Trauma 2:8171996Verweij BH Muizelaar JP: Avoiding secondary brain injury after severe head trauma: monitoring and management. J Craniomaxillofac Trauma 2:8–17 1996

    • Search Google Scholar
    • Export Citation
  • 37.

    Wald SLShackford SRFenwick J: The effect of secondary insults on mortality and long-term disability after severe head injury in a rural region without a trauma system. J Trauma 34:3773821993Wald SL Shackford SR Fenwick J: The effect of secondary insults on mortality and long-term disability after severe head injury in a rural region without a trauma system. J Trauma 34:377–382 1993

    • Search Google Scholar
    • Export Citation
  • 38.

    Young JS: Cerebral perfusion pressure or intracranial pressure? J Neurosurg 92:1911922000 (Letter)Young JS: Cerebral perfusion pressure or intracranial pressure? J Neurosurg 92:191–192 2000 (Letter)

    • Search Google Scholar
    • Export Citation
TrendMD
Cited By
Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 475 455 38
Full Text Views 305 103 7
PDF Downloads 142 46 4
EPUB Downloads 0 0 0
PubMed
Google Scholar