THE SCALES FOR ESTIMATION OF INJURY SEVERITY AND PREDICTION OF OUTCOMES OF INJURIES
Irkutsk Scientific Center of Surgery and Traumatology,
Irkutsk City Clinical Hospital No.3,
Irkutsk, Russia
The list of abbreviations:
MFS-Triage – the scale of the military field surgery chair – triage
MFS-I (GSW) – the scale of the military field surgery chair – injury (gunshot wound)
MFS-I (NBW) – the scale of the military field surgery chair – injury (non-ballistic wound)
MFS-I (MI) – the scale of the military field surgery chair – injury (mechanical injury)
MFS-SE – the scale of the military field surgery chair – state upon entry
MFS-SA – the scale of the military field surgery chair – state after admission
DAI – diffuse axonal injury
IATBI – Index of associated traumatic brain injury
ICD – International classification of diseases
MSCT – multispiral computed tomography
ARD – anesthesiology and reanimation department
ATBI – associated traumatic brain injury
STBI – severe traumatic brain injury
SATBI – severe associated traumatic brain injury
GCS – Glasgow Coma Scale
AI – Anatomic index
AIS – Abbreviated Injury Scale
APACHE – Acute Physiology and Chronic Health Evaluation
APS – Anatomic Profile Score
ASCOT – A Severity Characterization of Trauma
CHOP-Index – estimation of changes in some key values of blood and hemodynamics analysis: C – serum creatinine, H – hematocrit, O – plasma osmolarity, P – systolic arterial pressure
CRAMS – a modification of Trauma Index. It estimates 5 signs: Ñ – blood circulation; R – breathing; À – abdominal and thoracic injury; Ì – motion responses; S – voice reaction
CRIS – Comprehensive Research Injury Scale
ICD -–international classification of disease
ICISS – International Classification of Disease Based Injury Severity Score
ICU – intensive care unit
ISS – Injury Severity Score
LODS – Logistic organ dysfunction score
MaxAIS – Maximal Abbreviated Injury Scale
MODS – Multiple Organ Dysfunction Score
MPM – Mortality Probability Models
NISS – New Injury Severity Score
PIM – pediatric index of mortality
PELOD – pediatric logistic organ dysfunction scales
POSSUM – physiological and operative score for enumeration of mortality and morbidity
PRISM – pediatric risk of mortality score
P-MODS – Pediatric Multiple Organ Dysfunction Score
PTS – Polytrauma-Schlüssels
RTS – Revised Trauma Score
SAPS – Simplified Acute Physiology Score
SNAP – Schedule for Non-Adaptive and Adaptive Personality
SOFA – Sequential Organ Failure Assessment
TRIOS – 3 days recalibrated ICU outcome score
TRISS – Trauma and Injury Severity Score
VSS – Vital Signs Score
24-hour ICU score – 24-hour scale for the intensive care unit
Development and implementation of various digital scales in different fields of human activity and particularly in medicine is an old practice. The objective of such work is objectivization and standardization of the approaches to solution of specific tasks and problems, i.e. digital conversion of information and minimizing the human factor influencing on outcomes of work.
The examples of such approaches are several dozens of point scales (for the last 45 years) for objective estimation of severity of injuries in the human’s body. Many scales were developed for achievement of specific and sometimes narrowed objectives beginning from estimation of severity of individual injuries at different stages of medical aid to optimization of statistical analysis of results of treatment and selection of optimal curative techniques [3]. Currently, the actual issue is the appropriateness of usage of the estimation scales for arrangement of assistance for patients with severe associated traumatic brain injury (SATBI) at the hospital stage. Which scales could and should be used in conditions of the hospital? Is it correctly to use efforts and time in relation to such scales at the moment of salvation of the life of the patient?
The modern principles of arrangement of medical aid for patients with SATBI in the trauma center suggest transferring the patient from the anti-shock unit (Fig. 1) of the reception ward through the radial diagnostics unit (usually multispiral computed tomography, MSCT). If necessary, the patient is transferred through the anti-shock surgical room to the intensive care unit (ICU) (Fig. 2) [30].
Figure 1
The trauma room of the admission unit of the level 1 trauma center (St.Joseph’s Hospital and Medical Center, Phoenix, USA)
Figure 2
The traffic route of the patient with the severe associated injury
Multiple modern methods of instrumental and laboratory diagnostics are used. The data can be used for various scales. Time deficiency during mathematical calculations, continuous improvement and changes in medical technologies and different levels of equipping in trauma centers challenge the possibility of development of the uniform standard for objective estimation of injury severity in conditions of trauma centers.
There are 4 main tasks which could be solved by various authors after development of a new diagnostic scale: 1 – classification of injuries, 2 – triage, 3 – objective estimation of injury severity, 4 – the most correct prediction of injury outcome [3]. These tasks are correlated: the more severe injury, the more rapid assistance is required, and the worse predicted outcome is anticipated [9, 10]. It is evident that the route of hospital transfer of the patient with severe trauma can change the composition of the medical team [11], suspicion in identification of severe injuries and the volume of diagnostic examinations. The paramount importance relates to the predictive value of the scale, because concise possibility of injury outcome can influence on calculations of treatment costs and can assist in retrospective estimation of quality of medical aid.
The objective of the present study is to conduct the analysis of multiple comparative studies of the diagnostic scales, to estimate specificity and sensitivity in terms of prediction of injury outcome, with selection of the most appropriate methods for estimation of injury severity and prediction of its outcome in condition of the trauma centers in our country.
MATERIALS AND METHODS
The analysis included 66 articles from the domestic and foreign literature for the last 40-50 years. PubMed search system and the key words were used.
RESULTS AND DISCUSSION
There are more than 50 scales for estimation of injury severity (Fig. 3). Some authors [2, 3, 7, 12, 15] indicate 3 main approaches for estimation of injuries: 1) the quantitative approach – calculation and estimation of injury severity in individual regions of the body; 2) the qualitative approach – estimation of severity of general condition of the patient with emphasis on the vital functions and the physiological parameters; 3) the combined approach – the combination of two previous variants.
Figure 3
The main estimation scales of injury severity (described in the text)
The scales can be general (for estimation of all systems of the body) and specific (for isolated injuries or diseases in individual parts of the body), as well as the scales for thoracic surgery, cardiosurgery etc. (for example, Thoracoscore, Euroscore and others) [16]. Some scales are based on the real examination of injuries, for example, with AIS. Other scales are the derivatives from the estimation scales, with prediction of outcomes on the basis of retrospective analysis of treatment outcomes of numerous patients in concordance with the latest international classification of diseases (ICD). The examples are ICISS (Osler T. et al., 1996) – the modified ISS with prediction of probability of lethal outcomes with empirical technique. Any variant of an injury is coded in concordance with ICD – 8, 9, 10 (ICD – 8, 12, 13).
Each code relates to the calculated mortality on the basis of the analysis of the data of treatment outcomes, for example, ICISS-9 – 300,000 trauma patients for 5 years [49]. The scales AI and CRIS, which are derivatives of ICD, may be called as empirical ones. They require periodical revision, because ICD and disability after the same injuries change constantly. Also there are other scales, for example, for ICU (LODS, TRIOS, POSSUM and others) [38, 50, 58, 60], the pediatric scales (SNAP, PRISM, P-MODS, PELOD, PIM) [18, 26, 27, 54], which are based on estimation of the physiological parameters of the body with consideration of the age and periods of development of the child.
The abundance of the scales caused the need for realization of scientific works for comparison of the predictive efficiency and choice of the most effective scales that are shown in multiple foreign and in some domestic studies [1, 3, 31, 45, 51-53, 56, 62, 63, 66]. However, the authors often compared the scales, which were different in terms of tasks and methods of estimation. In most studies the efficiency of the tests was estimated with the objective parameters (receiver operating characteristic – ROC), which were diagramed as the curves of sensitivity and specificity of the examined test (ROC-curve), sometimes with changing the square under the curves [1, 41]. Sensitivity of the test is the reliability of identifications of a sign (expressed as percentage) among patients with real availability of this sign. Specificity is the reliability of identification of absence of a sign (expressed as percentage) among patients without that sign. The higher percent of sensitivity and specificity gives more reliability of the test.
The attempt of comparison of the predictive value of 4 scales (ISS, TRISS, NISS and ICISS) by means of the metaanalysis of 64 articles (MEDLINE and Embase, 1990 – 2008) with construction of 157 ROC-curves was ineffective because of unhomogenous values in these scales [61]. The authors of the study made a non-precise conclusion that NISS was more specific in prediction of mortality of blunt injuries in comparison with ISS. The comparison of NISS and TRISS did not find any advantages. ICISS was considered as less stable and with lower predictive efficiency than AIS. It is associated with presence of the high amount of various parameters requiring computed analysis.
Another similar study showed the comparison of efficiency of predicted mortality of 9 scales: 4 algorithms on the basis of AIS (ISS, NISS, APS (Anatomic Profile Score), maxAIS), 4 ICD-derivatives of these scales, and the empirical scale ICISS-9 (the derivative of ISS on the basis of ICD-9) [45, 49]. The predictive efficiency of the scales was estimated with construction of ROC-curves and computed statistical preparation (Hosmer-Lemeshow) [42]. Any significant differences in predictive value of the scales were not found. ICISS was more sensitive and specific, whereas ICD-derived scale APS demonstrated better predictive values according to Hosmer-Lemeshow. ICD-derived AIS demonstrated lower sensitivity and moderate specificity of prediction. According to the opinion by the authors such complex scales as ICISS and APS have better classification capacity than maxAIS and ISS and are recommended for registration of injuries. ISS and maxAIS showed moderate predictive efficiency, but good convenience for usage at the patient’s bed.
The advantage of ICISS against ISS in terms of prediction of lethal outcomes in ROC-analysis of the scales has been shown [49]. The error in predicted survival was 7.67 % for ISS and 5.95 % for ICISS.
In the article “The end of ISS and TRISS: ICISS is better than ISS and TRISS for prediction of survival rate, treatment costs and duration of hospital treatment” the authors published the results of the retrospective analysis of the cases of the trauma patients in the federal hospital US database (estimation of ICISS) and the American state trauma register (estimation of ISS and TRISS) – total of 7,705 cases [51]. As for ICISS, the special attention was given to additional estimation of neurological disorders. As result, ICISS exceled ISS and TRISS at its possibilities for prediction of mortality, hospital stay and treatment costs. Estimation of neurological disorders significantly improved its predictive capacity. TRISS (and its derivative ASCOT) has interest from scientific point of view. However because of its complexity, TRISS is inappropriate for real time practice with triage and prediction of injury outcome. Also its inappropriateness is associated with reliability for retrospective analysis of quality of arrangement of medical aid for trauma patients. From one side, use of mathematical formulae and complex multiform coefficients is explained by the attempt by the authors to decrease the error in estimation of severity and prediction, but, from other side, the calculations were based on the scales AIS, ISS, RTS and TS which include subjective estimation. The objective data (for example, results of laboratory examination) is not considered in this scale.
The comparison of APACHE II, TRISS and 24-hour ICU was made on the basis of the retrospective analysis of 1,000 patients in the ICU (364 patients with injuries). The authors of 24-hour ICU showed higher predictive value of their scale in comparison with TRISS and also criticized APACHE II, because, according to their observations, it significantly increased the risk of lethal outcome in relation to low points and underestimated the risk of death in high points [62].
Another statistical study included the retrospective examination of predictive value of APACHE II and SAPS II for 672 neurosurgical patients with subarachnoidal hemorrhage and severe TBI. The examination was conducted with multivariant mathematical analysis within the first 24 hours after admission to ICU [53]. The predicted mortality for APACHE II and SAPS II was 37.7 % and 38.4 % correspondingly, but the real value was 24.8 %. The metaanalysis of the scientific articles about efficiency of APACHE III, SAPS II and MPM II did not find any predictive advantages of any scale [42]. Another comparison of predictive efficiency of APACHE, SAPS, APACHE II, SAPS II and MPM on the basis of the regression analysis of 630 patients in ICU [55] showed that prediction of outcomes of patient’s condition could be realized with only 5 values: presence of multiple organ insufficiency, level of blood glucose, serum calcium, prothrombin ratio, plasma osmolarity.
Higher efficiency of APACHE II and MPM II0 was found after comparison of mortality prediction of APACHE II, SAPS II, MPM II0 and MPM II24 after the analysis of results of treatment of 969 surgical and non-surgical patients with use of ROC-curves [65]. MPM II24 was more sensitive, SAPS II – less sensitive and specific. The authors recommend to predict mortality with use of MPM II24 because of its simplicity and high sensitivity.
The comparison of efficiency of SOFA and MODS in the prospective cohort study of 209 patients with severe TBI (with ROC-curves) showed the advantage of SOFA for prediction of mortality and unfavorable neurological outcomes [66].
One cannot observe any wide-spread use of the whole range of the diagnostic predictive values in Russia. As result, there are few scientific works with comparison of such scales. Being the representatives of different medical specialties, neurosurgeons, traumatologists, general surgeons, anesthesiologists and other physicians use some scales for treatment of one and the same patient. Gabdulkhakov et al. [1] conducted a study of predictive significance of several, mainly, anesthetic and diagnostic, scales (SAPS II, ASCOT, 24-hour ICU Trauma Score, MPM II72). ROC-analysis was conducted. According to the author’s opinion, the above mentioned anatomic and structural systems for estimation of injury severity show the low sensitivity for prediction of lethal outcomes. The offered systems for injury severity estimation are ineffective for prediction of probability of the outcome in each specific case. The scales only allow stratification of patients in dependence on the risk of lethal outcomes. SAPS II shows the highest sensitivity and accurateness in prediction of lethal outcomes on the first day. SAPS II, ASCOT, 24-hour ICU Trauma Score, MPM II72 allow corrective estimation of predicted outcomes, and, as result, optimizing medical tactics. Nevertheless, the indicated scales are periodically used by various specialists in various medical facilities. The uniform approach is absent. The most common scales in Russian practice are Glasgow Coma Scale (GCS), the scale by Nazarenko-Tsibin, MFS scales by Gumanenko [2, 7, 57].
Previously we discussed some difficulties of implementation of GCS into daily practice in our country [8]. In the foreign countries GCS is perceived controversially. The retrospective analysis of efficiency of GCS in prediction of disease outcomes among 1,390 patients (GCS = 8-14) found the dependence on the total amount of points and combination of results of tests: the patients with the same total amount of points demonstrated some significant differences in the mortality [59].
The blank technique by Nazarenko-Tsibin is being implemented into the activities of the trauma centers in our country [7]. It is convenient for usage, especially if the forms are completed with the medical history of the patient, but it has some disadvantages for patients with TBI. The scale does not consider an important sign – disordered consciousness. As result, the technique by Tsibin-Nazarenko is offered to be used in combination with GCS [7]. However the scales become oppositely directed: the highest point of Tsibin-Nazarenko scale indicates the worst patient’s condition, and conversely for GCS. The above mentioned variants of injuries to the anatomic regions included shockogenic types of injuries, but without consideration of severe cerebral injuries (diffuse axonal injury, brain compression with intracerebral hematoma and others), which influence on injury outcome. Also spinal cord injuries are not considered. They often lead to arterial hypotensia and bradycardia after spinal shock.
The scales of military field surgery present the certain interest, mainly, for field conditions. Several scales are offered for the victim (the patient) at different stages of evacuation and treatment: MFS-Triage, MFS-SE, MFS-SA and one of the scales MFS-I (MI) [2]. From one side, the different tasks (and also estimation of trauma, prediction, classification) are solved by the different scales, but from other side, the scales often use some similar signs, which are estimated with points by different means (for example, response to pain, systolic arterial pressure, patterns of external breathing (the table 1)), and it leads to some misunderstanding.
| Table 1 | |||||||||||||||||||
| Differences in point estimation of vital values in the scales of Military Field Surgery-Sorting (MFS-Sort), Military Field Surgery-Admission Status (MFS-AS) and Military Field Surgery-Hospital Status (MFS-HS) | |||||||||||||||||||
MFS-Triage and MFS-SE do not estimate the level of consciousness. The high amount of tables and complex calculations of triage and estimating scales inevitably lead to diagnostic and predictive errors. Moreover, some signs are subjective, for example, guide values of blood loss in ml (5 degrees of severity), intestinal noise (3 degrees of severity), the patterns of external respiration (MFS-SE) and others. Real time use of such scales is difficult without computing technics or special tables and trained employees (specialists in statistics) who make only this type of work. Time costs for use of such scales by surgeons and anesthesiologists are unreasonable because of absence of real medical aid during triage of patients in case of massive admission.
A simple and convenient scale is necessary for estimation of injury severity at the stage of the reception ward (triage). The higher simplicity and comfort is better choice despite of possible errors, because any complicacy leads to refusal from a scale and to even higher bias in empirical estimation of severity of patients’ condition without any algorithms. Because of time deficiency the relevance is given to the scale, which deals with estimation of condition severity and does not require complex calculations (also with computing techniques).
| Table 2 | |||||||||
| Calculation of level of consciousness by means of summarizing the results of testing of three clinical signs | |||||||||
| Table 3 | ||||||||||||
| The relationship between six-point scale of consciousness with the common classifiction of degrees of disorders of consciousness | ||||||||||||
From this point of view, the index of associated traumatic brain injury (IATBI) is the most simple and comfortable tool, which was offered in 2007 [4, 8] for estimation of patient’s condition severity and prediction of injury outcomes on the basis of the analysis of two key signs upon admission. The scoring system for estimation of disordered consciousness was developed (the tables 2, 3).
Also the estimation according to the degree of arterial hypotension in TBI was developed and tested (the table 4).
| Table 4 | |||
| Point estimation of arterial hypotension | |||
The feature of IATBI is not summing the score points, but their deduction (Fig. 4).
The method for calculation of the index of associated traumatic brain injury (IATBI)
We have been using IATBT in our hospital (the level 2 trauma center) for 10 years. Our center realizes urgent 24 hour aid for trauma patients. The developed and implemented scale has successfully proved its significance from the perspective of its simplicity and accuracy of predicted outcomes. In the retrospective analysis of outcomes of associated traumatic brain injury the diagnosis of associated injury should correspond to the definition from 3rd All-Soviet Union Convention of Traumatologist-Orthopedists (1975). The direct relationship has been found between the value of the index and the probability of a lethal outcome (the table 5).
| Table 5 | |||||||||
| Mortality among patients with associated TBI in concordance with the index of associated TBI | |||||||||
The mortality was 100 % among the patients with associated traumatic brain injury who were admitted in 30-40 minutes after injury with IATBI < 0. This fact is extremely important from the perspective of triage in high rate of hospital admission.
CONCLUSION
The conducted analysis of the literature showed that standardization of diagnostic procedures and implementation of scoring scales require proper investigation of multiple materials with consideration of stages of medical aid, readiness of medical employees and logistic provision in medical facilities. In the day and night clinic, when the group of severe patients is already identified, and diagnostic equipment and enough amount of time are available, some complex scoring scales can be useful (the scale by Tsibin-Nazarenko, anesthesiology scales). With use of computers, such scales consider multiple objective findings (results of diagnostic tests, laboratory examinations etc.), taking into account the equipment status in the hospital. But one has to agree that all complex things do not take the roots in clinical practice [5]. The key factor of urgent aid is time. Therefore, we recommend the index of associated traumatic brain injury for use at the prehospital stage or in the admission ward, especially for patients with disordered consciousness. The index allows rapid prediction of injury outcomes without administration of computing techniques.