Objective: The Glasgow Coma Scale (GCS) is widely used to objectively describe the extent of patients’ impaired consciousness. However, there are known variations in scoring GCS both in adults and children which may impact patient management. The aim of this audit was to assess the application of GCS by medical and nursing staff in pediatric medical patients. Subject and Methods: An online questionnaire was distributed amongst doctors and nurses working in the Department of Child and Adolescent Health at Mater Dei Hospital in Malta. The participants assigned GCS for 8 case scenarios involving children of different ages with varying levels of consciousness. Results were analyzed by calculating percentage agreement and by Cronbach’s alpha. Results: Sixty-six participants were studied, with a response rate of 52%. Performance was poor overall, with Cronbach alpha 0.53. Correlation was better at the upper and lower ends of the scale and the worst performance was for verbal response. Only respondents with 5–10 years of experience achieved acceptable consistency in the application of the GCS (Cronbach alpha 0.78). Conclusion: There is considerable variation in application of GCS in pediatric patients, highlighting the need for education and training to improve consistency for this commonly used neurological assessment tool.

Highlights of the Study

  • There are known variations of scoring the Glasgow Coma Scale in both adults and children, and these may impact management.

  • There is considerable variation in the application of Glasgow Coma Scale in pediatric patients, highlighting the need for education and training to improve consistency for this commonly used neurological assessment tool.

The Glasgow Coma Scale (GCS) was designed in 1974 with the aim of classifying traumatic brain injury and was the first grading score to offer an objective assessment for patients’ level of consciousness [1]. The scale is based on three independent behavioral aspects: best eye response (E), best motor response (M), and best verbal response (V), with the GCS score (also summarized as GCS) representing the total of all 3 components.

The GCS has become one of the most widely used and validated assessment tools to objectively describe the extent of impaired consciousness in acute medical and trauma patients [2]. The standard GCS is not appropriate for use in preverbal children in view of the need for further interaction. This led to the introduction of the pediatric GCS (pGCS), which accounts for developmental differences in younger children up to 5 years of age [2, 3] and has been validated for use in children under 2 years of age [4]. The GCS also has prognostic significance [4], with the eye component being the least predictive of mortality, except when verbal response cannot be assessed [5]. There are known variations in scoring of GCS, both in adults and in children [6] which may impact patient management. The aim of this study was to assess the application of GCS by medical and nursing staff for pediatric medical patients outside the intensive care setting.

An online questionnaire was distributed in February 2023 to all doctors (pediatric trainees at all levels of training and pediatric specialists) and all nurses (including charge nurses) working in the pediatric medical wards and pediatric emergency within the Department of Child and Adolescent Health at Mater Dei Hospital in Malta. Two reminders were then sent at weekly intervals to encourage participation. Authorization for this study was obtained from the Chair for the Department of Child and Adolescent Health and from the Nursing and Midwifery Director. Ethical approval was not required as no patient contact was made.

The questionnaire consisted of 8 case scenarios for children between ages 1–15 years presenting with varying levels of consciousness secondary to different underlying diagnoses commonly encountered in the pediatric medical wards and pediatric emergency department. Half of the scenarios involved children under 5 years, to represent the use of the pGCS. Participants were asked to score the individual components of the GCS for each scenario and to calculate the total GCS. The full questionnaire is attached as an online supplementary Appendix 1 (for all online suppl. material, see https://doi.org/10.1159/000534797). The questionnaire was not timed. No preparatory course on GCS was offered, but a copy of a standard pediatric and adult GCS was distributed with the questionnaire, as shown in Table 1.

Table 1.

Glasgow Coma Scale

Score Adult Pediatric 
 Eyes 
  4 Open Open 
  3 To command To shout 
  2 To pain To pain 
  1 No response No response 
 Motor 
  6 Obeys commands Normal spontaneous movements 
  5 Localizes pain Localizes pain 
  4 Withdraws from pain Withdraws from pain 
  3 Abnormal flexion Abnormal flexion 
  2 Abnormal extension Abnormal extension 
  1 No response No response 
 Voice 
  5 Converses Alert, babbles, coos, words, or sentences to usual ability 
  4 Confused Less than usual ability, cries but consolable 
  3 Inappropriate words Cries to pain or inconsolable 
  2 Incomprehensible sounds Moans to pain 
  1 No response No response 
Score Adult Pediatric 
 Eyes 
  4 Open Open 
  3 To command To shout 
  2 To pain To pain 
  1 No response No response 
 Motor 
  6 Obeys commands Normal spontaneous movements 
  5 Localizes pain Localizes pain 
  4 Withdraws from pain Withdraws from pain 
  3 Abnormal flexion Abnormal flexion 
  2 Abnormal extension Abnormal extension 
  1 No response No response 
 Voice 
  5 Converses Alert, babbles, coos, words, or sentences to usual ability 
  4 Confused Less than usual ability, cries but consolable 
  3 Inappropriate words Cries to pain or inconsolable 
  2 Incomprehensible sounds Moans to pain 
  1 No response No response 

The respective GCS component is scored as non-testable (NT) if an external factor is interfering with a patient’s response – e.g., eye swelling (E), intubation (V), paralysis (M).

Data collection was anonymized. Results for individual components of the GCS were analyzed by calculating the percentage agreement and by Cronbach’s alpha to measure the internal consistency of the questionnaire and assess for variability. Cronbach’s alpha was generated using the Real Statistics Resource Pack software (Release 7.6) [7]. The resulting range of variation for the total Glasgow Coma Score was also determined.

The questionnaire was distributed to 92 doctors and 36 nurses in February 2023, with a response rate of 52% (66/128). Of the responders, 53 (80%) were medical doctors and 13 (20%) were nurses. The average time to complete the questionnaire was 15 min. Work experience was mixed, with 27 responders (41%) having more than 10 years of experience, 25 responders (38%) with under 5 years of experience and 14 (21%) having between 5 and 10 years of experience. All 66 participants completed the questionnaire in full.

Table 2 shows the percentage of correct replies for a range of scores across the different components of the GCS for each of the 8 scenarios. The answers are divided based on the age of the child in question, with a cut off age of 5 years.

Table 2.

Overall correct answers for individual components of GCS

Glasgow Coma ScaleCorrect reply (%) (corresponding question number in brackets)
patient age <5 yearspatient age >5 years
E1 91 (Q29)  
E2 94 (Q1)  
E3 89 (Q25) 91 (Q9) 
E4 100 (Q5) 97 (Q13) 
E NT (conscious pt)  86 (Q17) 
E NT (unconscious pt)  36 (Q21) 
M1  100 (Q22) 
M3 86 (Q30)  
M4 82 (Q2) 47 (Q18) 
M5 41 (Q26) 58 (Q10) 
M6 100 (Q6) 56 (Q14) 
V1  100 (Q23) 
V2 89 (Q31)  
V3 77 (Q3) 26 (Q11) 
V4 83 (Q27) 83 (Q19) 
V5 100 (Q7) 100 (Q15) 
Glasgow Coma ScaleCorrect reply (%) (corresponding question number in brackets)
patient age <5 yearspatient age >5 years
E1 91 (Q29)  
E2 94 (Q1)  
E3 89 (Q25) 91 (Q9) 
E4 100 (Q5) 97 (Q13) 
E NT (conscious pt)  86 (Q17) 
E NT (unconscious pt)  36 (Q21) 
M1  100 (Q22) 
M3 86 (Q30)  
M4 82 (Q2) 47 (Q18) 
M5 41 (Q26) 58 (Q10) 
M6 100 (Q6) 56 (Q14) 
V1  100 (Q23) 
V2 89 (Q31)  
V3 77 (Q3) 26 (Q11) 
V4 83 (Q27) 83 (Q19) 
V5 100 (Q7) 100 (Q15) 

NT, non-testable; pt, patient.

In general, performance was better at the upper and lower ends of the scale. The eye scale was determined most accurately overall. There was one exception in the case of an unconscious patient with swollen eyes (Q21), with the majority of respondents incorrectly scoring the eye component as 1.

Respondents also found difficulty in discerning between motor scores 4 and 5 (Q2, 10, 18, 26) and to recognize that the higher score should be applied for asymmetric motor response (56% correct – Q14). With regard to vocal response, the major difficulty was differentiating between V3 and V4 (Q11).

Cronbach’s alpha for different patient and respondent categories is shown in Table 3. Table 4 shows interpretation for the coefficient. There was poor scoring for all GCS fields across all patient age groups. Only respondents with 5–10 years of experience achieved acceptable consistency in application of the GCS. In addition, it was noted that nurses performed better when compared to doctors with a Cronbach alpha of 0.62 and 0.3 overall, respectively.

Table 3.

Variance in calculation of GCS

CategoryGCSCronbach’s alpha
All All 0.53 
Work experience <5 years All 0.46 
Work experience 5–10 years All 0.78 
Work experience >10 years All 0.17 
All patients Eye opening 0.49 
Motor 0.40 
Verbal 0.02 
Patient age <5 years Eye opening 0.47 
Motor 0.32 
Verbal 0.12 
Patient age >5 years Eye opening 0.22 
Motor 0.22 
Verbal 0.20 
CategoryGCSCronbach’s alpha
All All 0.53 
Work experience <5 years All 0.46 
Work experience 5–10 years All 0.78 
Work experience >10 years All 0.17 
All patients Eye opening 0.49 
Motor 0.40 
Verbal 0.02 
Patient age <5 years Eye opening 0.47 
Motor 0.32 
Verbal 0.12 
Patient age >5 years Eye opening 0.22 
Motor 0.22 
Verbal 0.20 
Table 4.

Interpretation of Cronbach’s alpha

Cronbach’s alpha
>0.9 Excellent 
>0.8 Good 
>0.7 Acceptable 
>0.6 Questionable 
>0.5 Poor 
<0.5 Unacceptable 
Cronbach’s alpha
>0.9 Excellent 
>0.8 Good 
>0.7 Acceptable 
>0.6 Questionable 
>0.5 Poor 
<0.5 Unacceptable 

These variations in interpretation for the individual components of the GCS led to marked inaccuracies for the total GCS, in the range of −6 to +6 of the actual score, as shown in Table 5. For the questions involving non-testable components, scores which included a reference to the missing eye component were also counted as correct. Of note, most participants were not aware that the total GCS cannot be calculated if there are non-testable components.

Table 5.

Proportion of answers showing correct total GCS

GCS score% correct GCS scoreGCS below correct scoreGCS above correct scoreQuestion number
range% repliesrange% replies
76 4–5 7–12 15 Q34 
70 6–8* 10–13 21 Q6 
11 29 10 18 13–15 53 Q14 
12 41 10–11 48 13–15 11 Q30 
15 55 9–14* 44  Q18 
15 100   Q10 
NT 11 NA  NA  Q22 
NT NA  NA  Q26 
GCS score% correct GCS scoreGCS below correct scoreGCS above correct scoreQuestion number
range% repliesrange% replies
76 4–5 7–12 15 Q34 
70 6–8* 10–13 21 Q6 
11 29 10 18 13–15 53 Q14 
12 41 10–11 48 13–15 11 Q30 
15 55 9–14* 44  Q18 
15 100   Q10 
NT 11 NA  NA  Q22 
NT NA  NA  Q26 

*One answer non-testable, not included in the table.

Our findings show poor performance overall when assessing GCS in pediatric patients, with overall Cronbach alpha value of 0.53, which is lower than the mean noted in other studies [6]. This could be partly due to inexperience in health care professionals with <5 years of experience and lack of direct use of GCS in the line of work for the more experienced health care workers. This discrepancy in performance between medical and nursing staff may be accounted for by more frequent use of GCS by the nursing staff in general. However, the low response rate from the nursing staff (36%) when compared to medical staff (58%) is a limitation when comparing inter-rater reliability between different professions. The worst performance overall was for the verbal response, and there was more variation in the middle-range scores for motor and verbal responses, in keeping with other studies [6]. These discrepancies may have an impact on patient management, for example, in decisions regarding imaging in head injuries [8] and may affect prognosis, especially in the setting of traumatic brain injury [4].

A systematic review assessing reliability of GCS scoring system had shown consistency ≥0.6 in 85% of assessments, with better correlation for GCS components (when compared to total GCS) [6]. Unfortunately, most of the included studies (43/52) were of poor quality. A subsequent systematic review [9] then identified three main factors influencing reliability of GCS: (1) training and education in the application of the GCS, which were associated with enhanced reliability in a majority of the studies; (2) type of (painful) stimuli used; (3) level of consciousness, with the least reliable GCS assessments being obtained in the middle-range GCS scores. Conflicting evidence was found for the assessors’ years of experience.

Reilly et al. [10] first proposed the pGCS version in 1988 and it is intended for children below 5 years of age. A valid assessment tool requires appropriate inter-rater reliability to be accurate. A systematic review by Borgialli et al. [4] found that pGCS compared well to the standard GCS when it comes to diagnosing clinically important traumatic brain injury, with the verbal component having the best predictive performance in all age groups, mainly involving children with minor head trauma. However, there was less inter-rater agreement for the pGCS with better correlation in all fields for adult GCS, with the motor score showing the best inter-user correlation (0.80). Our study showed different findings, with the eye response achieving best consistency, albeit still poor (0.49). A more recent study involving nurses in a pediatric ICU setting and showed decreased inter-rater reliability for intermediate-range pGCS responses, with the best inter-rater correlation being noted for verbal response [11].

Alternative tools to the pGCS such as the Full Outline of UnResponsiveness (FOUR) score have been developed; it assesses four components: eye and motor response, respiration, and brainstem reflexes and therefore eliminates the verbal component, which may be difficult to assess especially in preverbal children. The score was initially developed for the adult population but has been validated for use in the pediatric population, with excellent inter-rater correlation (weighted kappa coefficient 0.94) [12]. Its performance in predicting outcome is similar to the pGCS [12, 13], but it is not as widely used as the GCS.

Although applying the pGCS to a non-verbal child may be more challenging [9], our study noted better performance for GCS scoring in children <5 years of age in comparison to adolescents. This could be explained by the fact that a majority of respondents to the questionnaire had >5 years of experience in pediatric care.

To our knowledge, this is the first study to assess interpretation of GCS by use of an online questionnaire which facilitated simultaneous identical assessment for medical and nursing staff. There is no validated questionnaire available, but our questionnaire was based on commonly encountered pediatric clinical scenarios and was tested by the study authors prior to distribution to optimize consistency.

Unfortunately, we found significant variations in interpretation of the GCS in pediatric patients. This may have implications for patient management and prognosis and may confound GCS monitoring for patients with altered level of consciousness.

There is considerable variation in application of the GCS in pediatric patients, highlighting the need for education and training to improve consistency for this commonly used neurological assessment tool.

We acknowledge all the staff who took the time to complete this questionnaire.

This study did not require ethical approval as no patient contact was carried out.

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

Maria Caruana: data collection, design of the study, interpretation of the results, and review of the paper. Sophie Noelle Hackenbruch: data collection, write-up, and review of the paper. Victor Grech: statistical interpretation of the results and write-up and review of the paper. Ruth Farrugia: study design, statistical analysis, and review of the paper.

Data can be accessed upon request from authors.

1.
Teasdale G, Jennett B. Assessment of coma and impaired consciousness: a practical scale. Lancet. 1974;2(7872):81–4.
2.
Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G. The Glasgow Coma Scale at 40 years: standing the test of time. Lancet Neurol. 2014;13(8):844–54.
3.
Royal College of Pediatrics and Child Health. Management of children and young people with an acute decrease in conscious level: a nationally developed evidence-based guideline for practitioners; 2015. Available from: https://www.rcpch.ac.uk/resources/management-children-young-people-acute-decrease-conscious-level-clinical-guideline [accessed 13 April 2023].
4.
Borgialli DA, Mahajan P, Hoyle JDJr, Powell EC, Nadel FM, Tunik MG, et al. Performance of the pediatric Glasgow Coma Scale Score in the evaluation of children with blunt head trauma. Acad Emerg Med. 2016;23(8):878–84.
5.
Moore L, Lavoie A, Camden S, Le Sage N, Sampalis JS, Bergeron E, et al. Statistical validation of the Glasgow coma score. J Trauma. 2006;60(6):1238–43; discussion 1243-4.
6.
Reith FC, Van den Brande R, Synnot A, Gruen R, Maas AI. The reliability of the Glasgow Coma Scale: a systematic review. Intensive Care Med. 2016;42(1):3–15.
7.
Zaiontz C. Real statistics using Excel; 2020. Available from: www.real-statistics.com [accessed 13 April 2023].
8.
Nuttall AG, Paton KM, Kemp AM. To what extent are GCS and AVPU equivalent to each other when assessing the level of consciousness of children with head injury? A cross-sectional study of UK hospital admissions. BMJ Open. 2018;8(11):e023216.
9.
Reith FCM, Synnot A, van den Brande R, Gruen RL, Maas AIR. Factors influencing the reliability of the Glasgow coma scale: a systematic review. Neurosurgery. 2017;80(6):829–39.
10.
Reilly PL, Simpson DA, Sprod R, Thomas L. Assessing the conscious level in infants and young children: a paediatric version of the Glasgow Coma Scale. Childs Nerv Syst. 1988;4(1):30–3.
11.
Kirschen MP, Snyder M, Smith K, Lourie K, Agarwal K, DiDonato P, et al. Inter-rater reliability between critical care nurses performing a pediatric modification to the Glasgow coma scale. Pediatr Crit Care Med. 2019;20(7):660–6.
12.
Cohen J. Interrater reliability and predictive validity of the FOUR score coma scale in a pediatric population. J Neurosci Nurs. 2009;41(5):261–7; quiz 268-9.
13.
Czaikowski BL, Liang H, Stewart CT. A pediatric FOUR score coma scale: interrater reliability and predictive validity. J Neurosci Nurs. 2014;46(2):79–87.