Background: Published guidelines for defining the “minimally conscious state” (MCS) included behaviours that characterise emergence, specifically “reliable and consistent” functional interactive communication (accurate yes/no responding) and functional use of objects. Guidelines were developed by consensus because of the lack of empirical data to guide definitions. Criticism emerged that individuals with severely impaired cognition would have difficulty achieving the requisite threshold of accuracy and consistency proposed to demonstrate emergence from MCS.
Objective: To determine the utility of the operational threshold for emergence from post-traumatic MCS, by evaluating a measure of yes/no accuracy (Cognitive Test for Delirium, auditory processing subtest (CTD-AP).)
Methods: Prospective, consecutive cohort of responsive patients recovering from traumatic brain injury (TBI), including a subset meeting criteria for MCS at neurorehabilitation admission who improved and were able to undergo the study protocol. Participants were evaluated at least weekly, and given the CTD-AP to assess yes/no responding.
Results: Of the 1434 observations from 336 participants, 767 observations yielded inaccurate yes/no responses. 75 participants (22%) never attained consistently accurate yes/no responses at any time during their hospitalisation. Generalised estimating equations analysis revealed that confused participants were more likely to respond inaccurately to yes/no questions. Further, the subset of individuals who were in MCS on rehabilitation admission and improved, were also more likely to respond inaccurately to yes/no questions.
Conclusions: Consistent yes/no accuracy is uncommon among responsive patients in early recovery from TBI. These results suggest that the operational threshold for yes/no response accuracy as a diagnostic criterion for emergence from MCS should be revisited.
Statistics from Altmetric.com
In 2002, the Aspen Neurobehavioural Workgroup proposed the term “minimally conscious state” (MCS) to define a state of recovery or decline in which persons demonstrate behavioural evidence of self or environmental awareness but do so inconsistently.1 Unlike unconscious patients in a coma or vegetative state, persons in MCS evidence non-reflexive behaviours such as following simple commands, gestural or verbal yes/no responses, intelligible verbalisation or purposeful movements, and demonstrate these on a reproducible or sustained basis. Two behaviours were proposed for emergence from MCS: (1) functional interactive communication and (2) functional use of two different objects.1 To enhance reliability, the workgroup proposed that functional communication be operationally defined as accurate yes/no responses to six of six situational orientation questions on two consecutive evaluations. Lacking sufficient data for an empirically based definition, these guidelines were developed by consensus.1
Controversy followed publication of the diagnostic criteria for MCS, particularly regarding the operational definition for MCS emergence. Critics asserted that conscious individuals with severely impaired cognition would have difficulty achieving the requisite threshold of accuracy and consistency proposed to demonstrate emergence from MCS.2 3 To determine the utility of the operational threshold for emergence from post-traumatic MCS, we prospectively evaluated a measure of yes/no reliability among a large, consecutive cohort of patients recovering from severe traumatic brain injury (TBI), including a subset that met workgroup criteria for vegetative state/MCS at neurorehabilitation admission. The purpose of the study was to evaluate the prevalence of failures in yes/no reliability among TBI groups with more normal conscious behaviour, including non-confused patients, that would provide empirical evidence that the operational definition for MCS emergence may be too difficult for some patients.
The study population consisted of all TBI Model Systems participants admitted to a freestanding neurorehabilitation hospital from January 1999 to February 2006 (see Gordon et al for TBI Model System criteria4). Of 376 qualified participants, those remaining unable to communicate via verbalisation/non-verbal gestures throughout hospitalisations (n = 31), pre-existing neurological disorders (n = 2), inability to speak English (n = 1) and abbreviated rehabilitation length of stay precluding clinical examination (n = 6) were excluded. Of the remaining 336 patients, 324 were responsive on rehabilitation admission and 12 were minimally responsive but improved and were able to participate in the study protocol. The study sample was primarily male (n = 231; 69%) with a median age of 29 years (range 16–87) and most (n = 189, 56%) had high school or greater education. Severe TBI, as classified by emergency department Glasgow Coma Scale score, was present in most of the sample (n = 191, 57%), with fewer moderate (n = 80, 24%) and mild (n = 63, 19%; two missing, 0.01%) injuries. Quartiles for post-injury duration at rehabilitation admission were 13, 20 and 30 days, respectively. The primary mode of injury was motor vehicle related (n = 251, 75%).
The Cognitive Test for Delirium5 (CTD) was developed to evaluate cognitive aspects of acute confusion. It is sensitive and specific to acute confusion in many study samples, including those composed exclusively of patients with TBI.5 6 Designed for use in acutely ill patients, the test permits use of non-verbal responses in the form of pointing, nodding the head or raising the hand. It consists of five subtests in the areas of orientation, attention span, memory, language comprehension, conceptual reasoning and vigilance. The comprehension section includes the auditory processing subtest (CTD-AP) which involves administration of four simple questions that require a yes or no response/gesture to determine auditory processing accuracy (eg, “Will a stone float on water?”). Scores range from 0 to 4 for the CTD-AP subtest. Two alternate forms of the CTD were used.
Individuals admitted in a minimally responsive state were evaluated using the Coma Recovery Scale.7 Our practice was to transition patients from assessment with the Coma Recovery Scale8 to the CTD protocol once they were able to use either verbal or non-verbal communication to respond to testing stimuli on the CTD. At least weekly, neuropsychologists rated responsive participants and determined whether the diagnostic criteria for delirium of the Diagnostic and Statistical Manual of Mental Disorders-fourth edition were met.9 Delirium diagnostic criteria (DDC) include: (A) disturbance of consciousness with attentional impairment, (B) change in cognition or development of a perceptual disturbance not attributable to dementia, (C) disturbance develops over a short time and fluctuates during the course of the day and (D) disturbance is caused by a general medical condition. Research assistants alternated administration of the CTD forms A and B during each week of hospitalisation and within 24 h of the neuropsychologist rating. Demographic and injury information (Glasgow Coma Scale scores at admission to emergency department) were obtained from medical records and interview with participants/family members.
During the study period, participants received multiple evaluations depending on their length of stay (range 1–22). Study analyses were limited to 1434 complete paired observations. The CTD data were missing for 29 observations and the DDC was missing for 10. A total of 324 participants were responsive on rehabilitation admission (non-MCS); 741 (56%) observations were completed when participants were confused (DDC+) and 597 (45%) when persons were no longer confused (DDC−). Twelve of the study participants were admitted in a vegetative or minimally conscious state but improved and were able to participate in the CTD protocol (emerged from MCS). The group of 12 patients who emerged from MCS after rehabilitation hospital admission yielded 96 study observations, of which 71 (74%) were obtained when participants met DDC and the remaining 25 (26%) when persons no longer met DDC. The four yes and no questions comprising the auditory processing subtest were answered correctly in 667 observations (46.5%) for both the emerged from MCS and non-MCS groups. Of the 336 responsive participants, 75 (22.3%) never attained consistently accurate yes/no responses at any time during their rehabilitation hospital stay. Group performance and comparisons are summarised in table 1.
A generalised estimating equations (GEE) method was used to analyse yes/no accuracy as a dichotomous dependent variable using a binomial distribution, logit link and an exchangeable correlation structure. The GEE takes into account the repeated measures design of the study and expected correlation for within subject observations over time.8 A GEE method was used with admission responsiveness status and DDC as independent variables. A total of 1434 observations were analysed. No interaction existed for DDC status across the non-MCS and emerged MCS groups, and the term was removed (p = 0.45). There were significant main effects for admission responsiveness status (non-MCS and emerged from MCS) and DDC status. Individuals who were DDC+ had 3.18 higher odds of being inaccurate than individuals who were DDC− (95% CI 2.42 to 4.17; p<0.0001). Individuals who emerged from MCS during rehabilitation had a 2.73 higher odds of being inaccurate than individuals who were never in MCS throughout rehabilitation (95% CI 1.47 to 5.06; p = 0.001).
The results of this study show that accurate yes/no responding to the CTD-AP items is challenging for responsive patients in early recovery from significant TBI. Regardless of delirium status, over half of the sample observations yielded inconsistent accuracy to the yes and no questions. Of the 622 observations obtained from non-confused participants, only 393 (59%) indicated consistent yes/no accuracy, indicating that CTD-AP items were challenging even for these patients. Of the 812 observations obtained from responsive although confused participants, only 274 (34%) indicated consistent yes/no accuracy. Accuracy was poorer for those who had recently emerged from a MCS compared with participants who were responsive at admission to inpatient rehabilitation.
A number of factors could explain inaccurate yes/no responses in our participants. The attentional demands of these items are greater than the demands of tasks such as following a simple command or functional use of common objects such as a comb or toothbrush. Patients in early recovery from TBI are likely to be distractible, potentially compromising the accuracy or consistency of responses. Confused patients in early recovery from TBI evidence fluctuating performance, thus decreasing accuracy on tasks requiring consistent responding.10 11 Also, the CTD-AP questions lack personal relevance to the participants which may have further diminished attention/motivation to give correct responses. Finally, these items may require more complex language processing and reasoning. Residual impairments in these abilities may have contributed to the high rates of failure.
While the observed frequency of yes/no response inaccuracy is striking, especially among non-confused patients, the implications of our findings for diagnostic criteria for emergence from MCS are unclear. The situational orientation questions (eg, Am I pointing to the ceiling?) recommended for determining accurate yes/no responding from patients in MCS seem less complex than the CTD-AP items. Situational orientation questions may place fewer demands on language processing and reasoning abilities. They may also place fewer demands on attention as the stimulus for these items is concrete and immediately present in the patient’s environment compared with the conceptual issues queried with the CTD-AP items. Nonetheless, the CTD-AP items are relatively simple and it is yet to be empirically determined that the type of yes/no questions influence accuracy when recovering from MCS.
The extremely high rates of failure observed for our patients in early recovery from TBI suggest that the issue of accuracy of yes/no responding as a diagnostic criterion for emergence from MCS should be revisited. While these results may imply that the present criterion for MCS emergence is too difficult, our study methods were not designed to provide empirical data that support a better alternative. Individuals with profound brain injuries often have multiple impairments that influence their ability to respond consistently and accurately to various behavioural tasks, including answering yes/no questions. Future research should investigate yes/no accuracy to alternative types of questions as well as other neurobehavioural responses that may reliably differentiate minimally responsive persons from responsive but cognitively impaired/confused persons.
Funding: This study was supported by The TBI Model System of Mississippi Grant, funded by the National Institute on Disability and Rehabilitation Research Grant #H133A020514.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.