Secondary Logo

Journal Logo

Original Articles

Development of a Measure of Nociception for Patients With Severe Brain Injury

Whyte, John MD, PhD*; Poulsen, Ingrid RN, PhD†,‡; Ni, Pengsheng MD, MPH§; Eskildsen, Marianne SHCA; Guldager, Rikke RN, MSc, PhD†,∥

Author Information
The Clinical Journal of Pain: April 2020 - Volume 36 - Issue 4 - p 281-288
doi: 10.1097/AJP.0000000000000811
  • Open


Traumatic brain injury (TBI) is the largest cause of traumatic death and disability worldwide,1 and pain is a consequence for many patients. TBI severity is typically graded on the basis of initial depth of unconsciousness, using the Glasgow Coma Scale (GCS),2 or the duration in the disturbance of consciousness and residual confusion, referred to as posttraumatic amnesia.3 Patients who survive severe TBI and the first phase of coma will often go through the following phases in their recovery process: the vegetative state/unresponsive wakefulness syndrome (VS/UWS), wherein the patient shows preserved arousal but no signs of consciousness; the minimally conscious state (MCS), wherein the patient shows preserved arousal and reproducible but fluctuating behavioral signs of consciousness, followed by a period of confusion and impairment in day-to-day memory.4 In combination, these states are termed disorders of consciousness (DOC).4 Some who fully regain consciousness have residual communication impairments. Patients with DOCs or severe communication impairments are unable to accurately self-report pain or distress.

Pain has both objective and subjective elements5 that are important to evaluate and manage in the clinical care of individuals with TBI.6 Injuries and tissue pathology produce nociceptive stimuli that receive restricted neural processing and lead to autonomic and reflexive responses.7 Further processing by the full “pain matrix,” including limbic circuitry, leads to the subjective experience of pain, and goal-directed behaviors to end it.8 Undertreatment of subjective pain, when present, may limit the patient’s participation in rehabilitation, and lead to chronic physiological and psychological conditions.9,10 However, nociceptive reactions may be important indicators of pathology, irrespective of subjective pain. TBI is often accompanied by painful comorbidities,11–13 at a time when subjective awareness or reporting of pain may be impaired by DOC or communication.

The goal of assessment determines whether the measurement approach should focus on indicators of nociception alone, or on indicators of subjective pain and distress as well. Considerable research has been published addressing the question of whether patients in the chronic VS/UWS can experience pain.7 Indeed, this is an important question for ethical and medical-legal reasons. However, in the acute phase after injury, the priority is to recognize indicators of nociception regardless of the presence of subjective distress, to facilitate diagnosis and treatment of these conditions. Moreover, accurate diagnosis of a patient’s state of consciousness is, itself, a complex and error-prone task, particularly in the context of rapidly evolving changes in consciousness.14 Thus, an instrument that detects nociceptive stimuli similarly in conscious and unconscious patients would avoid this confound.

During the recovery of consciousness, some patients with TBI develop agitation, which may include behaviors such as moaning, grimacing, or rocking.15 These behaviors may be interpreted as signs of pain, leading to overtreatment with opioids and other sedating agents, which also limit participation in rehabilitation.16,17 Ideally, assessment of nociception or pain should avoid confounding with agitation. In summary, accurate assessment of nociception in persons with TBI is important to optimize alertness, and aid in diagnosing potentially serious conditions or injuries. Where subjective pain is present, its detection and management can alleviate suffering and distress.

Standard pain assessment uses self-report measures that are not feasible in patients with alteration of consciousness or impaired communication.18–20 Observational rating scales have been developed for other populations with limited abiity to communicate.21,22 However, the behavioral manifestations of pain/nociception in individuals with severe brain injury may differ from those with advanced dementia, sedated patients, or normal neonates, requiring a tailored approach.

The Nociception Coma Scale (NCS) and subsequent Nociception Coma Scale-Revised (NCS-R),23,24 were developed specifically for severe brain injury. The NCS-R consists of 3 subscales: Motor response, Verbal response, and Facial expression. Each of the subscales is scored from 0 to 3 points, wherein 3 points equals the greatest possible sign of pain for each individual item. A cutoff value at 4 points is set to indicate pain.24 However, these measures contain a mixture of items that assess nociception regardless of consciousness, and those that require conscious reactions to subjective pain. For example, the highest score on the Motor Response and Verbal Response subscales require “localization to painful stimulation” (eg, the patient’s hand attempts to remove the noxious stimulus), and “intelligible verbalization,” respectively, both of which require consciousness. Patients in the VS/UWS may alter their posture or groan in response to noxious stimuli, but they do not exhibit the above cortically mediated behaviors. Moreover, as discussed by Vink et al,25,26 “vocalization” and “groaning” are problematic items in patients who are intubated or have tracheostomies. A similar concern is raised for patients with paresis and aphasia concerning the subscores in the Motor and Verbal Response categories. Furthermore, in a qualitative study of nurses’ experience with the NCS-R in a clinical rehabilitation practice, they argued that the scale was unable to assess patients in the state of VS/UWS fairly, as they very rarely would reach the set cutoff level of 4 points as a sign of pain due to a general low functional level.27

Thus, the maximal score that a patient in VS/UWS can obtain is lower than for those in an MCS, regardless of the intensity of nociception. Indeed, NCS-R scores are correlated with those on the Coma Recovery Scale-Revised (CRS-R),28 confirming that NCS-R scores differ systematically by the level of consciousness.29 As a result, a patient’s current state of consciousness must be known, to interpret the NCS-R score.

The Pain Assessment Scale (PAS), also developed for brain injury, includes behavioral indicators overlapping with the NCS-R, along with a set of autonomic indicators.30 The preliminary validation study on PAS showed very good interrater agreement on 13 and good agreement on 8 of 27 items. Nevertheless, some of the items were dichotomous and others impractical to score. A significant change after repositioning of patients provided preliminary validation.30 Overall, it appeared that many of the PAS items held the potential for inclusion in a new, more comprehensively developed and validated scale for determining pain.

The aim of the present study was to develop a scale that can be used to measure the intensity of nociception in patients with TBI who are unable to provide subjective pain reports. In developing such a scale, we intended to have its score be unrelated to the level of consciousness (ie, to be measuring the intensity of nociception, rather than subjective pain), and to the level of behavioral agitation as well. We aimed to validate this measure as an indicator of nociception, to demonstrate that scores are not confounded with the level of consciousness or behavioral agitation, and to assess the amount of time required to perform an accurate assessment.



Participants were enrolled between 2013 and 2017 from MossRehab Hospital, Elkins Park, PA, and the Department of Neurorehabilitation, Traumatic Brain Injury, Rigshospitalet, Denmark. We included patients with moderate to severe TBI >17 years of age and with a FIM score <5 ( on Comprehension or Expression, indicating unreliable verbal pain reporting ability.31 We excluded patients without a consenting Legally Authorized Representative (LAR), and those whose attending physician or LAR was unwilling to alter non-opioid analgesic orders for the 48-hour rating period.

Development of the Brain Injury Nociception Assessment Measure (BINAM)

We reviewed observational pain measures developed for other populations,9,10,20–23 the NCS-R29 and the PAS,30 for items that appeared relevant and feasible for assessment in severe TBI, and which were unlikely to be confounded with the level of consciousness or agitation. We began with a set of 15 items (Table 1), which did not appear related to the level of consciousness, could be assessed quickly, and might be expected to change promptly in response to analgesia or new nociceptive stimuli.

Candidate Items for the Brain Injury Nociception Assessment Measure


In this study, severity sufficient for inclusion was defined in terms of the inability to reliably communicate in receptive or expressive terms. The FIM is an 18-item ordinal rating scale of functional independence,32 from which the comprehension and expression items were used to identify a sample unable to provide accurate self-reports of pain. Within this range, the degree of impairment in consciousness was assessed using the CRS-R. The CRS-R is designed to measure the level of consciousness and is validated for use in clinical and research settings. It consists of 23 items, grouped into 6 subscales. The test items are operationally defined, as are the patient responses that are scored. The CRS-R can distinguish between VS/UWS MCS, and higher states of consciousness, and was used to characterize the study participants’ level of consciousness at the time of examination.28

The Agitated Behavior Scale (ABS) is designed to assess agitation in patients who have sustained a TBI.15 The ABS is an observational rating scale that ranges from 14 to 56 points, with 3 subscales assessing disinhibition, aggression, and lability. A score >21 indicates clinically significant agitation. The ABS has been recommended as the best validated instrument for such assessments.33,34 The ABS15,33,34 was used to characterize the study participants’ level of agitation at the time of participation.

The study was approved by the Ethics Committee at the Danish site (H-1-2012-149), and by the local Institutional Review Board at the US site. In Denmark, the study also was approved by the Data Protection Agency with Journal no: 2007-58-0015. Written informed consent was provided by the LAR of each participant.


For ethical reasons, it was felt that patients who were already on opioid analgesics could not have those medications withdrawn for study reasons. Therefore, we hoped to validate the BINAM’s sensitivity to the level of analgesia with the more modest manipulation of acetaminophen. Regardless of what other analgesic medications a participant was on, if she/he was not on acetaminophen, this was added for one of the 2 study days. Participants who were on standing doses of acetaminophen (with or without other analgesics) had the acetaminophen withdrawn for 1 of the study days. Regardless, participants were randomly assigned, in a 1:1 ratio, to receive acetaminophen 1 g every 6 hours during the first or second assessment day, beginning the previous evening. CRS-R assessment was performed <24 hours before the first BINAM assessment, and <24 hours after the last. ABS was rated by the patient’s primary nurse during the day and evening shift on each of the 2 assessment days.

A trained nurse or research assistant completed the BINAM in 15-minute observation blocks on each of 2 days in 4 settings chosen for potential differences in discomfort/nociception: (1) in bed shortly after repositioning; (2) in a chair or wheelchair shortly after repositioning; (3) in bed or wheelchair after prolonged sitting/lying; and (4) during physical therapy involving gross motor activities (eg, standing, transferring, range of motion exercises). Before data collection began, an appropriate activity was selected for each participant’s physical therapy data collection sessions, and standardized across the 2 days. An activity was chosen that was clinically appropriate and within the participant’s capability, involving weight-bearing or range of motion/stretch, but not involving significant aerobic activity. During each 15-minute rating session, the data collector observed the patient for about a minute and then recorded the observations for about a minute, making a complete set of ratings 7 times/session on a scannable rating sheet. The order of conditions was held constant for each participant across the 2 days, but it varied from participant to participant on the basis of their normal activity schedule. To ensure that the 2 sites carried out the assessments in a reliable way, in-person visits, frequent telephone meetings, and close monitoring by the Principal Investigators were conducted.

Data Analysis

The protocol called for ratings of lateralized items to be on a consistent side of the body. However, due to a procedural error, nearly half of the participants had a mixture of sides used for data collection. Day-to-day correlations were compared between pairs of ratings collected on the same side versus opposite sides (data not shown). Inconsistency had little effect on ratings of blood pressure, heart rate, or pupil size, but, for the degree of eye-opening (arousal), correlations were substantially lower for ratings performed on different sides, than for those performed on the same side. This was addressed during subsequent analysis.

Nine participants had an alteration in the assigned acetaminophen schedule. In these cases, individual sessions were classified as being on or off acetaminophen accordingly. When it was unclear how well medicated the participant was at the time of data collection, drug condition was set to missing.

We applied Rasch Analysis, using the Partial Credit Model (PCM) to calibrate the response data and assessed the item fit by examining mean squares (MNSQ). Power calculations suggested that an analyzable sample of 158 participants would provide stable item calibration within +/−0.5 logits with 99% confidence. Items with MNSQ<0.7 or >1.3 were removed from further analysis. The joint maximum likelihood estimation method implemented in the WINSTEPS program was used. Valid joint maximum likelihood estimation results can be obtained even for participants with some missing data, on the basis of missing-at-random assumption. The final scores were converted to a 100-point scale with larger scores indicating greater nociception intensity.

We examined the model fit by fitting 2 restricted item response theory (IRT) models, one examining the invariance of item parameters of repeated observations of the same item by fitting the PCM (restricted PCM) with the equivalent item parameters of the repeated observations for the same item, and the other examining the assumption of independence of sessions within a participant by fitting a multilevel PCM model (which takes into account the correlation between sessions within a participant) based on the restricted PCM. We used information criteria (Akaike information criterion [AIC], Bayesian information criterion [BIC]) to assess whether the more restricted models improved model fit (lower AIC and BIC values). In multilevel PCM, each participant has 2 scores—at the participant and at the session level; the sum of those scores was compared to the person scores generated from the restricted PCM. We also used analysis of variance to examine the effect of drug condition (acetaminophen or no acetaminophen) and activity (the 4 types of sessions) using the different models. We used the marginal maximum likelihood estimation method to fit all 3 models. The person scores were generated from the Mplus program using Expected A Posteriori (EAP) method.

In IRT, the item response given by a participant is only determined by the level of the latent construct(s) being measured and not by other demographic variables. We examined violations of this assumption (differential item functioning—DIF), for each item by sex and by age (participants older and younger than the median of 55), by testing the target item, anchoring all other items. For example, when we examined “heart rate” for male and female participants, we set the parameters to be equal for male and female participants on all other items except “heart rate,” and then conducted the Wald χ2 test to examine the difference of item parameters by sex. These analyses were carried out separately for each item. We calculated the weighted absolute difference between item characteristic curves (wABC) to assess the impact of DIF. wABC values exceeding 0.12, 0.18, 0.24, 0.3 for 2-, 3-, 4-, and 5-category items, respectively, were considered to indicate that DIF had a large impact. We compared the test characteristic curves across groups to examine the impact of DIF at the summed score level.

The proportion of study participants who were in VS/UWS (~14%), or who were rated as clinically agitated (~17%) at the time of data collection, was too small to allow this method of assessment of DIF. Consequently, we analyzed the data with the full sample and then after removing patients in VS/UWS or those who were agitated, respectively, and calculated the correlation of the item parameter estimates based on the full sample and subsamples. We also calculated the correlation between CRS-R scores and BINAM scores as continuous variables, and tested for any impact of analgesic medication on ABS scores to further explore the validity of BINAM scores as indicative of nociception rather than agitation, using the Wilcoxon Signed-Rank test.

We examined the psychometric properties of the 10-item scale after removal of 5 misfitting items, based on the restricted PCM model. We generated the item map and the distribution of the person score standard error. We assessed whether the item map covered the whole range of the scale and examined the number of patients with SEs above the threshold of score reliability of 0.9 (the threshold of standard error with score reliability >0.9 was calculated as sqrt[{1−0.9}×score variance]).

We also examined the impact of decreasing the number of repeated observations or of setting some of the items to missing on the reliability of the person score estimates. Using the cases with data for all items and all 7 observations, we examined the correlations between the person scores generated from the complete data set versus those generated from fewer (4, 5, 6) observations. Similarly, correlations were calculated after removing “pupil,” or “BP_sys,” or “BP_dys,” or “heart rate” or all 3 cardiovascular items, as these were the most frequently missing. The person scores were estimated on the basis of weighted likelihood estimations, implemented in SAS.


One hundred seventy-six participants were enrolled in the study (83 from MossRehab; 93 from Rigshospitalet), but BINAM data were not collected on 19 participants due to the emergence of higher-level communication or intercurrent illness. One hundred fifty-seven participants contributed data to the analysis (71 from MossRehab; 86 from Rigshospitalet) (Table 2).

Characteristics of the Participants

Item Calibration and Fit

One participant missed a complete day of assessment (4 sessions), and 9 participants missed 1 to 2 sessions, for logistical reasons, resulting in the exclusion of 15 data collection sessions in total. Data from certain additional sessions were excluded due to ambiguity in whether the participant was medicated with acetaminophen at the time of the session (27 sessions) or failure to record the eye used for ratings (10 sessions). The final data set represented 1204 sessions from 157 individuals (96% of intended sessions). Higher scores for all items were assigned to categories predicted to reflect greater nociception intensity. Original rating categories were merged because of a sample size <10 in some of the categories, and then further merged for 9 items because of violation of monotonicity on initial calibration. This resulted in 2 to 5 rating categories for all items (Table 1).

Five items (Respiratory Rate, Goosebumps, Skin Wetness, Skin Temperature, Tears) with MNSQ values exceeding acceptable threshold values were removed from the item bank. The fit indices of the remaining 10 items were acceptable. Multilevel PCM had the lowest AIC and BIC values among the 3 models assessed (PCM, restricted PCM, and multilevel PCM) (Table 3). The correlation between the person score estimates from the restricted PCM and multilevel PCM was 0.9982. We compared the relative validity of Multilevel PCM with the other 2 models in detecting an impact of acetaminophen or physical therapy versus resting conditions (Table 4). The PCM had similar validity as the restricted PCM for known groups, and was better able to discriminate experimental conditions than the multilevel PCM. Adding the fact that restricted PCM is easily understood, we used the restricted PCM in subsequent analyses.

Fit Comparison Among Partial Credit Model (PCM), Restricted PCM, and Multilevel PCM Models
Brain Injury Nociception Assessment Measure Scores by Activity and Drug Condition (Mean [SD]) and RV of the Different Models

Figure 1 shows the item map for the remaining 10 items. Note that a change in arousal, as indicated by the degree of eye-opening, is the most sensitive item for detecting mild nociception, whereas skin color tends to change only with the most intense nociception.

Item Map of the final 10 items from the restricted Partial Credit Model. The x-axis reflects the nociceptive dimension with greater nociception intensity to the right. The items are arranged from most sensitive to nociception (arousal, at the bottom) to least sensitive (change in skin color, at the top). Different shades of gray reflect different item scores.

As is typical, the reliability of the person score estimate was greatest in the midrange of the scale and decreased with more extreme scores (Fig. 2). However, reliability was above 0.9 for approximately a 4-logit range of the scale.

Relationship between SEs and the person scores based on the restricted Partial Credit Model. The dotted line represents the SE with score reliability=0.9. The dots below the line have a person score reliability >0.9; the dots above the line have a person score reliability <0.9.

Initial Validation

We predicted that BINAM scores would be modulated by activity, with the highest scores recorded during physical therapy, followed by the prolonged position condition, with the 2 assessments conducted shortly after repositioning at rest being least painful. We also predicted that scores would be lower with acetaminophen than without it. Indeed, a 2-way analysis of variance revealed main effects of analgesia (F=8.86, 1 df, P=0.003) and activity (F=44.52, 3 df, P<0.0001) on BINAM scores, with no significant interaction (F=0.68, 3 df, P=0.56). Tukey post hoc pairwise comparison of scores for individual activity conditions revealed that scores were higher in physical therapy than in each of the other conditions, and that scores were higher after prolonged sitting or lying than after recent repositioning in bed (Table 5). Of interest, ABS scores did not differ on the basis of acetaminophen administration (on acetaminophen: mean=17.97, median=15, range=14 to 45; off acetaminophen: mean=17.92, median=15, range=14 to 47; S=−68, P=0.74).

Effect of Activity and Analgesia on Brain Injury Nociception Assessment Measure Scores


No items showed DIF with respect to sex. One item (vocalization) demonstrated DIF between older and younger individuals, but the impact involving this single item was negligible on the overall summed score (maximal difference<0.13 SDs of the summed score across the range of the scale). The effect of rating arousal from a consistent versus inconsistent eye was small.

With respect to the 2 subgroups of interest (VS/UWS; agitation), the item parameters were correlated>0.99 between the full sample and the subsample after exclusion of either group (Fig. 3), suggesting that level of consciousness and level of agitation had negligible effects on the scores. The fact that acetaminophen modulated BINAM scores but not ABS scores provides further evidence of the independence of measurement of nociception even in agitated patients. The average correlation between the CRS-R and BINAM scores across conditions was small but significant (Spearman ρ=0.18, P=0.0269).

Impact of unconsciousness and agitation on parameter estimates. The correlation between item parameters calculated on the whole sample (on the x-axis) and those calculated after excluding patients with agitation (filled circles) or unconsciousness (open circles), respectively (on the y-axis). VS indicates vegetative state.

Impact of Observation Time and Missing Data

We assessed whether 15 minutes of observation (7 rating episodes) are required to arrive at reliable scores, and the impact of missing data for specific items. The correlations between the person scores computed with partial versus complete data are shown in Table 6. There was little impact on reliability, with as few as 4 observations, compared with 7 (person score correlations >0.98). Heart rate and systolic and diastolic blood pressure were the most frequently missing items, due to the failure of the automatic blood pressure cuff to read during episodes of arm movement or behavioral agitation. Pupil size could sometimes not be rated because of the patient’s position. With completely missing data for Pupil, DiasBP, SysBP, or HR, the person score estimates still correlated highly (>0.9) with scores derived from the full data set. When data were missing for all 3 of the cardiovascular items, person score correlations with the true score dropped to ∼0.7.

Impact of Reduced Observation Time and Missing Items on Person Scores


This study is the first to validate a nociception scale consisting of both behavioral and physiological items designed for noncommunicative patients with severe TBI, using IRT methods. IRT allows interval measurement of constructs such as the intensity of nociception using a set of dichotomous or ordinally scored items.35 Our data suggest that the final BINAM, composed of 10 behavioral and physiological items, is capable of producing reliable person scores in ∼10 minutes of observation, and that the score is sensitive to activities likely to exacerbate nociception, and to medications likely to ameliorate nociception. Moreover, these results provide preliminary evidence that the BINAM can assess nociception similarly in patients in VS/UWS or MCS, and in patients with or without agitated behavior, thus disentangling assessment of nociception from that of consciousness or agitation.

Several study limitations should be considered when interpreting these results. The cardiovascular items collected with an automated cuff were the most likely items to be missing, followed by pupil size. This suggests that the scale will be most difficult to apply to uncooperative or highly mobile patients. However, the scale was relatively robust to missing data for certain items. In addition, one might be concerned that certain neurological injuries would invalidate some of the assessment items. For example, a patient with an efferent pupillary defect might not display larger pupils, or a severely paralyzed patient might not engage in posturing, in the presence of nociception. This concern was blunted by the rating instructions, which encourage clinicians to use less impaired body parts for the rating, or to consider the patient’s baseline range of motion before assessing posturing. In the end, however, the performance of the scale suggests that these are not overwhelming confounds. The fact that Rasch analysis could arrive at a stable item hierarchy implies that the primary variation in the rating items was due to nociception, not due to idiosyncratic patient factors.

Validating a measure of nociception in a noncommunicative population is challenging, without an objective standard of nociception intensity or a subjective self-report to compare with. Validation studies support the use of the NCS-R, but recent research suggests that the previously recommended fixed cutoff score indicating nociception is not reliable and that one should focus on the detection of within-subject change.26,29 However, although within-subject change can help in managing known sources of nociception, it cannot identify which patients are experiencing nociceptive stimuli. In addition, as noted previously, NCS-R and CRS-R scores are substantially correlated, complicating the use of the NCS-R to detect pathology and nociception in patients with unclear or rapidly evolving levels of consciousness. Although BINAM and CRS-R scores are also significantly correlated, the magnitude of the correlation is smaller than that between CRS-R and the NCS-R reported by Chatelle et al,29 suggesting less confounding with state of consciousness. Moreover, as the current study involved naturally occurring injuries rather than experimentally induced nociception, there is the possibility that the actual burden of nociceptive stimuli was somewhat higher in the more conscious patients.

With respect to use in patients with agitated behavior, complete independence of BINAM and ABS scores may not be realistic. Although agitation is believed to be an independent construct, nociceptive stimuli may increase agitation in susceptible individuals. However, the fact that acetaminophen was not associated with a change in ABS scores suggests that the BINAM is more targeted to measuring nociception itself, and can still be used to assess nociception in agitated patients.

The response of the BINAM to activity provides partial construct validation. However, there is the possibility that the cardiovascular items of the BINAM responded to postural and aerobic stimuli in physical therapy and to nociception as well. Prolonged positioning in bed or chair was the second most painful activity, but it was only significantly more painful than repositioning in bed, but not repositioning in the chair. As we did not record whether the prolonged position was sitting or supine, we cannot rule out that this difference was also partly postural in nature. However, the significant effect of acetaminophen, which has no known cardiovascular effects, demonstrates that the BINAM score must be driven substantially by nociception. The IRT fit statistics support the instrument’s reliability in discriminating pain levels among individuals, not merely within-subject changes in nociception input. Efforts are underway to further validate the BINAM as a measure of nociception/pain by examining its relationship to the objective burden of injury.

In conclusion, this study provides evidence that the 10-item BINAM, rated over a 10-minute observation period, can reliably assess the severity of nociceptive factors in individuals with severe TBI, and that the resulting score is largely independent of the level of consciousness or agitation. Further research to validate the BINAM, and to compare its accuracy and feasibility in clinical practice with the NCS-R, is warranted.


The authors thank the patients and relatives for their participation in the project. They are also grateful to Neurological Consultant Lars Westergaard, Kongens Lyngby, Denmark; Pia Brix Køge, RN, Denmark; Sylvia Andersen, RN, Copenhagen, Denmark; Devon Kratchman, BS, Research Assistant, Moss Rehabilitation Research Institute, Albert Einstein Healthcare Network, Elkins Park, PA (currently Research Coordinator at PolicyLab and Center for Pediatric Clinical Effectiveness at Children’s Hospital of Philadelphia, Philadelphia, PA); Stephen Faha, MSA, MSE, Database Coordinator, Moss Rehabilitation Research Institute, Albert Einstein Healthcare Network, Elkins Park, PA.


1. Dewan MC, Rattani A, Gupta S, et al. Estimating the global incidence of traumatic brain injury. J Neurosurg. 2018;1:1–18.
2. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. Lancet. 1974;2:81–84.
3. Levin HS, O’Donnell VM, Grossman RG. The Galveston Orientation and Amnesia Test: a practical scale to assess cognition after head injury. J Nerv Ment Dis. 1979;167:675–684.
4. Giacino JT, Katz DI, Whyte J. Neurorehabilitation in disorders of consciousness. Semin Neurol. 2013;33:142–146.
5. Vallath N, Salins N, Kumar M. Unpleasant subjective emotional experiencing of pain. Indian J Palliat Care. 2013;19:12–19.
6. Morone NE, Weiner DK. Pain as the fifth vital sign: exposing the vital need for pain education. Clin Ther. 2013;35:1728–1732.
7. Pistoia F, Sacco S, Sara M, et al. The perception of pain and its management in disorders of consciousness. Curr Pain Headache Rep. 2013;17:374.
8. Bastuji H, Frot M, Perchet C, et al. Pain networks from the inside: spatiotemporal analysis of brain responses leading from nociception to conscious perception. Hum Brain Mapp. 2016;37:4301–4315.
9. Li D, Puntillo K, Miaskowski C. A review of objective pain measures for use with critical care adult patients unable to self-report. J Pain. 2008;9:2–10.
10. Herr K, Bjoro K, Decker S. Tools for assessment of pain in nonverbal older adults with dementia: a state-of-the-science review. J Pain Symptom Manage. 2006;31:170–192.
11. Whyte J, Ponsford J, Watanabe T, et alFrontera WR, Delisa JD, Gans BM, Walsh NA, Robinson L. Traumatic brain injury. Delisa’s Physical Medicine and Rehabilitation: Principles and Practice, 5th ed. Philadelphia, PA: Lippincott Williams & Wilkins Health; 2010:575–623.
12. Hoffman JM, Lucas S, Dikmen S, et al. Natural history of headache after traumatic brain injury. J Neurotrauma. 2011;28:1719–1725.
13. Ofek H, Defrin R. The characteristics of chronic central pain after traumatic brain injury. Pain. 2007;131:330–340.
14. Day KV, DiNapoli MV, Whyte J. Detecting early recovery of consciousness: a comparison of methods. Neuropsychol Rehabil. 2018;28:1233–1241.
15. Corrigan JD. Development of a scale for assessment of agitation following traumatic brain injury. J Clin Exp Neuropsychol. 1989;11:261–277.
16. Bogner J, Barrett RS, Hammond FM, et al. Predictors of agitated behavior during inpatient rehabilitation for traumatic brain injury. Arch Phys Med Rehabil. 2015;96:S274–S281.
17. Kadyan V, Mysiw WJ, Bogner JA, et al. Gender differences in agitation after traumatic brain injury. Am J Phys Med Rehabil. 2004;83:747–752.
18. Katz DI, Polyak M, Coughlan D, et al. Natural history of recovery from brain injury after prolonged disorders of consciousness: outcome of patients admitted to inpatient rehabilitation with 1-4 year follow-up. Prog Brain Res. 2009;177:73–88.
19. Laureys S, Celesia GG, Cohadon F, et al. Unresponsive wakefulness syndrome: a new name for the vegetative state or apallic syndrome. BMC Med. 2010;8:68.
20. Zwakhalen SM, Hamers JP, Abu-Saad HH, et al. Pain in elderly people with severe dementia: a systematic review of behavioural pain assessment tools. BMC Geriatr. 2006;6:3.
21. Hummel P, van Dijk M. Pain assessment: current status and challenges. Semin Fetal Neonatal Med. 2006;11:237–245.
22. AGS Panel on Persistent Pain in Older Persons. The management of persistent pain in older persons. J Am Geriatr Soc. 2002;5:S205–S224.
23. Schnakers C, Chatelle C, Majerus S, et al. Assessment and detection of pain in noncommunicative severely brain-injured patients. Expert Rev Neurother. 2010;10:1725–1731.
24. Chatelle C, Majerus S, Whyte J, et al. A sensitive scale to assess nociceptive pain in patients with disorders of consciousness. J Neurol Neurosurg Psychiatry. 2012;83:1233–1237.
25. Vink P, Eskes AM, Lindeboom R, et al. Nurses assessing pain with the Nociception Coma Scale: interrater reliability and validity. Pain Manag Nurs. 2014;15:881–887.
26. Vink P, Lucas C, Maaskant JM, et al. Clinimetric properties of the Nociception Coma Scale (-Revised): a systematic review. Eur J Pain. 2017;21:1463–1474.
27. Poulsen I, Balle M, Givard KL. Nociception Coma Scale-Revised: Nurses’ Experience in Clinical Practice. Pain Manag Nurs. 2019;20:592–598.
28. Giacino JT, Kalmar K, Whyte J. The JFK Coma Recovery Scale-Revised: measurement characteristics and diagnostic utility. Arch Phys Med Rehabil. 2004;85:2020–2029.
29. Chatelle C, Hauger SL, Martial C, et al. Assessment of nociception and pain in participants in an unresponsive or minimally conscious state after acquired brain injury: the relation between the Coma Recovery Scale-Revised and the Nociception Coma Scale-Revised. Arch Phys Med Rehabil. 2018;99:1755–1762.
30. Poulsen I, Brix P, Andersen S, et al. Pain assessment scale for patients with disorders of consciousness: a preliminary validation study. J Neurosci Nurs. 2016;48:124–131.
31. Granger CV. The emerging science of functional assessment: our tool for outcomes analysis. Arch Phys Med Rehabil. 1998;79:235–240.
32. Wilde EA, Whiteneck GG, Bogner J, et al. Recommendations for the use of common outcome measures in traumatic brain injury. Arch Phys Med Rehabil. 2010;91:1650–1660.
33. Lombard LA, Zafonte RD. Agitation after traumatic brain injury: considerations and treatment options. Am J Phys Med Rehabil. 2005;84:797–812.
34. Sandel ME, Mysiw WJ. The agitated brain injured patient. Part 1: definitions, differential diagnosis, and assessment. Arch Phys Med Rehabil. 1996;77:617–623.
35. Kreiner SCKvon Davier MC. Validity and objectivity in health related summated scales: analysis by graphical loglinear Rasch models. Multivariate and Mixture Distribution Rasch models: Extensions and Applications. New York, NY: Springer; 2007:398.

brain injuries; consciousness disorders; nociception; pain; pain measurement

Copyright © 2020 The Author(s). Published by Wolters Kluwer Health, Inc.