Patterns of early conversational recovery for people with traumatic brain injury and their communication partners

Abstract

Primary objective:

To investigate whether the degree of participation by people with severe Traumatic Brain Injury (TBI), and the degree of support by their communication partners (CPs) changes in conversation during subacute recovery.

Methods and procedures:

Seventeen pairs of participants with TBI and their CPs were video-recorded during a 10 min casual conversation at 3 and 6 months post-injury. Communication behaviors were rated using the adapted Measure of Participation in Conversation (MPC) and the adapted Measure of Support in Conversation (MSC) at both time points and compared.

Results:

Inferential analyses showed that there was no significant change in the degree of participation in conversation by participants with TBI and the degree of conversation support by their CPs from 3 to 6 months post. Comparison of qualitative field notes revealed that specific conversational behaviors changed over time, including better turn-taking and topic maintenance.

Conclusion:

Documenting early communication recovery is a complex and challenging endeavor. The lack of change in conversational effectiveness during the sub-acute period using global rating scales highlights the need for social communication tools that are sensitive to communication recovery following severe TBI.

Introduction

Traumatic Brain Injury (TBI) is the world’s leading cause of death and disability for young adult populations (1). Cognitive functions are commonly impaired after TBI (2) and due to the intrinsic relationship between cognition and language, a disruption in cognition is likely to be concomitant with social communication impairments (3). These communication deficits are characterized by the American Speech-Language-Hearing Association (ASHA) as ‘cognitive-communication disorders’ (2). Cognitive-communication disorders reduce an individual’s ability to engage in pre-morbid interests and daily activities at the same functional level (4) prior to their injury. Limitations in their ability to return to pre-morbid activities will, in turn, affect family interactions (3,5), academic performance (6), vocational success (7,8), social integration and quality of life (9). Therefore, it is essential to help individuals with severe TBI develop necessary communication strategies to raise their level of social activity and participation and hence productivity and life satisfaction.

Communication characteristics in individuals with severe TBI

There is much heterogeneity in the conversational abilities of individuals with severe TBI because diffuse brain damage can lead to different cognitive deficits in executive function, memory deficits and attention (10–12). Executive function impairment may result in disinhibition, poor planning, and word retrieval or verbal fluency difficulties (13,14), resulting in reduced utterance length, difficulty in initiating, sustaining or repairing conversations (15) and inappropriate comments or responses (16,17). Memory is another cognitive domain that can be impaired after severe TBI (18–21), reflected in conversational deficits such as poor topic maintenance (22,23) and inappropriate responses (3). Deficits in attention may lead to reduced speech and efficiency of language comprehension (24), resulting in frequent pauses (25) or requests for repetition (22). Hence, interactions with people with severe TBI can be perceived as ‘burdensome’ or ‘unrewarding’ (15,22,26) and will require greater dependence on CPs to maintain the conversation (15,27,28).

Cognitive-communication recovery in TBI

Recovery of cognitive-communication deficits is potentially linked to the recovery of a person’s cognitive abilities. However, few studies have investigated if cognitive-communication deficits associated with cognitive skills also recover during the sub-acute period. Cognitive-communication abilities in people with moderate-severe TBI recovers most rapidly between 3 and 6 months post-injury, but can also then decline in the longer term (29–31). Reduced ability to participate in conversation, places people with TBI at risk of poor long term quality-of-life outcomes (5,32). Snow and colleagues (1997) found that participants with TBI at 3–6 months post-injury had a greater frequency of cognitive-communication deficits such as false starts, hesitations, and poor turn taking (23) than the control group. A follow-up study found that after 2 years, one-third of the sample had deteriorated over time (22). At this sub-acute stage (3–−6 months post-injury), individuals are medically stable, but will need further rehabilitation to facilitate the transition to community settings (33). If clinicians are able to understand the early recovery of conversational behaviors of people with TBI and their CPs, they will be able to support and manage these behaviors during this crucial phase. Currently, we are unsure of the time point when maladaptive CP behaviors may commence, and seek to understand this further in this study.

Role of communication partners

Most literature on cognitive-communication recovery focuses on the person with TBI rather than examining the changes in CP support, or the whole interaction (22,23,34,35). However, previous studies that examined the discourse of the person with TBI showed that CPs play a vital role in effective conversations (27,36–39). Togher, Hand, and Code found that CPs of people with TBI displayed maladaptive behaviors, including frequent questioning, requests for clarification and slowed speech production. Such behaviors may disempower the person with TBI to contribute to conversations (27). Other studies showed that people with TBI provided more information with CPs who used support strategies such as asking elaborative questions and responding to requests for clarification (26,40). However, these studies tend to focus on analysing the discourse of the person with TBI but not their CPs. With an understanding that CPs do have an impact on the participation of the person with TBI, one study showed that with training, CPs are able to negotiate communication breakdowns to optimize the conversational skills of the person with TBI in the chronic phase (i.e. 18 months to 30 years post injury) (38,41). Overall, these findings demonstrate that CP contributions can limit or facilitate conversational effectiveness in the chronic phase (i.e. 18 months to 30 years post-injury).

As these studies tend to focus on training of CPs, they do not investigate the pattern of change of CP support during the early stages of recovery (3–6 months post-injury). This subacute period is a crucial time when people with TBI are transitioning from inpatient rehabilitation units back into their community (33). Effective interactions between the person with TBI and their significant CPs (e.g. spouses, paid caregivers) are essential to smoothen this transition. Understanding how CPs change in their conversational support is important because people with TBI may have limited awareness of their cognitive-communication deficits (42,43) during the early stages of recovery, making it challenging for clinicians to teach them conversational strategies. CPs can assist recovery during this time by implementing educational strategies and enhancing interactions with people with TBI (44). Due to the lack of literature investigating the sub-acute period, it is unclear at which point CPs may change their behaviors in response to the person with TBI and whether that change occurs in the more predominantly sub-acute inpatient setting at 3 months or at 6 months, when they have been discharged home. Looking at these patterns of recovery will provide clinicians with insight on which behaviors to target during therapy.

The adapted Measure of Participation in Conversation (MPC) and adapted Measure of Support in Conversation (MSC)

Togher and colleagues (36) adapted the Adapted Measure of Participation in Conversation (MPC) and Adapted Measure of Support in Conversation (MSC) from the aphasia literature (45) to assess conversational interaction and effectiveness between individuals with TBI and their CP.

The adapted MPC measures the ability of a person with TBI to participate in the social connection component (measured by the Interaction sub-scale) and the content component of a conversation (measured by the Transaction sub-scale). The adapted MSC measures the support provided by the CP to acknowledge the competence of the person with TBI (Acknowledging competence subscale) and their skill in facilitating the person with TBI to exchange content that helps reveal his/her competence (Revealing competence subscale) (36).

These outcomes measures have been used in a clinical trial, and have been successful in showing that training CPs improves everyday conversational behaviors (38). Moreover, it can be performed more quickly in a clinical setting than traditional in-depth conversational analysis. Additionally, no other outcome measures other than the MPC/MSC capture the conversational behaviors of CPs even though the literature states that CPs may demonstrate maladaptive behaviors that can impacts on a person with TBI’s participation in the conversation. With an understanding of both sides of the interaction, it would allow for consideration of whether the contributions of the CP is a barrier or facilitator of the conversation and at what time point this is critical.

Aim

There is an absence of understanding of cognitive-communication difficulties in the early stages of recovery and the pattern of conversational skills demonstrated by adults with TBI and their CPs 3–6 months post-injury. Therefore, this study utilized an observational design to examine the process of spontaneous recovery of conversational skills between 3 and 6 months, as well as identifying the presence and emergence of positive and negative communication behaviors by the CPs. No direct treatment or training was administered to the participants (including the CPs) as the primary aim of the study was to investigate behaviors presented in the sub-acute phase and determine whether there were any differences in the degree of support provided earlier in rehabilitation and at a later time period. The MPC/MSC scales allow comparison of communication behaviors over time for both interactants in order to determine whether there were any differences in the degree of support provided earlier in rehabilitation and at a later time period. An improvement in MPC/MSC scores may indicate that recovery of interactions occurred from 3 to 6 months post injury or worse scores may indicate that communication in later contexts may be more demanding and difficult for CPS. These observations and objective scores may be analyzed and help clinicians identify strengths and weaknesses in CP support, which can guide them on what to focus on during a CP training program and at which time period during their recovery that it should be best conducted.

Research questions

1. a) What is the degree of participation in conversations by people with severe TBI during early recovery? 1 b) Does this degree of participation changeover a period of 3 to 6 months post-injury?

2. a) What is the degree of conversational support provided by CPs to people with severe TBI during early recovery? 2 b) Does this degree of support changeover a period of 3 to 6 months post-injury?

We hypothesize that people with TBI will have some degree of difficulty participating in the conversation at 3 months post-injury, and that they may improve at 6 months post-injury (23,46). As for the CPs, we hypothesize that their degree of support will change from 3 to 6 months. These research questions focus on interrogating the process of change in conversational participation, interaction, and support during the sub-acute phase of recovery.

Method

This observational study is part of a larger, longitudinal communication study tracing communication recovery in people with severe TBI over a 2-year period post-injury. Participants completed comprehensive speech pathology and neuropsychology batteries at 3, 6, 9, 12 and 24 months post-injury. The primary task for this present study was a casual conversational discourse sample.

Participants

Participants with TBI and their CPs (n = 17 pairs) were recruited from three Brain Injury Units across metropolitan Sydney. Inclusion criteria for participants with TBI as part of this study were:

1. Aged between 16 and 65 years at the time of the injury.

2. Severe TBI defined as a score of eight or less on the Glasgow Coma Scale (GCS) and/or a period of Post Traumatic Amnesia (PTA) more than 24 h.

3. Reside in Sydney, Australia metropolitan area or within 3-h travelling distance.

4. Participants from diverse cultural and linguistic background were considered on an individual basis depending on English proficiency.(no interpreters available)

5. Have corrected hearing and vision that were within normal limits.

6. Able to identify a CP with whom they interacted with on a daily basis. This CP had to be available to accompany the person with TBI to both sessions at 3 and 6 months. The CP should not have sustained a brain injury or have a known psychiatric history.

7. Consent given to have their conversations video-taped.

Exclusion criteria were:

1. Consent unable to be obtained from a person with TBI or from significant other.

2. History of neurological illness, injury or significant medical history (i.e. developmental delay).

3. People with TBI who were more than 6 months post-injury.

4. Participants not available for follow up testing at 6 months.

Demographic and neuropsychological data of the participants with TBI and the CPs are presented in Table 1.

Table 1.

Basic demographic and neuropsychological data for participants with TBI (n = 17). Mean, standard deviation (SD) and range are presented for numerical data and frequency for categorical data.

	Participants with TBI Mean SD (Range/Frequency)	CPs Mean SD (Range/Frequency)
Gender (male: female, n)	15:2	16:1
Age (years)	34.06 SD 13.23	46.3 SD 9.52
	(17–55)	(27–66)
Primary language (English: other, n)	15:2	16:1
Education (years)	14.3 SD 2.82	13.9 SD 2.59
	(10–20)	(10–17)
Knew person with TBI prior to injury (yes:no, n)	NA*	17:0
Relationship to person with TBI (parent:partner: sibling)	NA	8:8:1
Duration of premorbid relationship (years)	NA	26.6 SD 13.8
Cause of TBI (MVA: fall: assault, n)	11:5:1	(15–58) NA
Length of PTA (days)	58.2 SD 33.1	NA
Diagnosis of Aphasia** (yes: no, n)	(6–122) 10:17	NA
Diagnosis of Dysarthria† (yes: no, n)	15:2	NA
Receiving speech pathology intervention at time of assessment (yes: no, n)	16:1	NA

^*NA = Not applicable

^**Based on Western Aphasia Battery-Revised Aphasia Quotient > 93.8

^†Based on a score of < 7 on any subtest on the Frenchay Dysarthria Assessment – 2

Outcome measures

Primary outcome measures

The first two research questions aimed to investigate the degree of participation by people with TBI and the degree of support by their CPs during conversation at 3 and 6 months post-injury, and its changes over time. To address these questions, the adapted MPC and the adapted MSC were used to measure the participants’ conversational skills at 3 and 6 months. These scales are a time-effective way to analyze natural conversation and provide useful information in setting intervention goals (47).

The MPC provides an index of participation in conversation by the person with TBI. The person with TBI is rated on two subscales:

1. Interaction – related to the ability to engage socially with a partner.

2. Transaction – related to the ability to exchange information, opinions, and feelings.

The MSC provides an index of the support provided by the CP for the person with TBI. The CP is rated on two subscales:

1. Acknowledging competence (AC) – related to the ability of the partner to use natural adult talk appropriate to the context and to show sensitivity to the person with TBI.

2. Revealing competence (RC) – related to the ability of the partner to use scaffolding to promote the potential participation of the person with TBI.

Both the MPC and MSC are 9-point Likert scales with a range of 0 to 4 with 0.5 intervals. A score of four represents a higher level of participation on the MPC, while a score of 4 represents a higher level of support on the MSC. Excellent inter- and intra-rater reliability for the adapted MPC and MSC scales have been reported (ICCs ranged from 0.84 to 0.97), and are comparable with those reported by Kagan and colleagues (45,48). The adapted MPC/MSC and scoring information are outlined in appendix 1.

Raters in this study also took notes on the participants’ behaviors to support their ratings. Although this is not part of the routine procedure for using the MPC/MSC scales, a previous study employing this method reported that qualitative note-taking supplemented the quantitative MPC/MSC data, and provided an avenue for in-depth discussion of changes in conversational behavior (47).

It should be noted that the MPC/MSC rating scales were not designed to distinguish conversational behaviors between people with TBI and healthy control participants. Instead, they were designed to provide a global rating of participation and support abilities during conversation, possibly shedding light on any CP maladaptive behaviors that have been documented in the literature. Therefore, the observational design was most suited to this aim and area of research.

Assessment procedure

A 10-min casual conversation sample between each of the 17 TBI-CP pairs was recorded using a video recorder (Canon Vixia HFS200) placed in the corner of the room. The conversations occurred in a quiet setting; 15 samples were recorded at the rehabilitation unit and two samples were recorded in the participants’ home. Participants were instructed: ‘I am going to leave the room for 10 min. I want you to have a chat while l am gone.’ Videos were edited to remove identifying information and content which would reveal the time of recording (3 or 6 months) to the blinded raters.

Rating of conversational samples

Two final year speech pathology students were trained in rating the adapted MPC and MSC scales. Both raters underwent a standardized training procedure and followed the same rating procedures to minimize personal interpretation (49). Training was conducted by an experienced speech pathologist and an author of the adapted MPC and MSC scales (EP) (36). Training took place over the course of 3 days (17 hours) and involved raters familiarising themselves with the scale descriptors, exemplar videos and rating practice with feedback provided by the trainer. After training was completed, 34 randomized conversational samples were rated using the MPC/MSC scales. Raters watched each video together, rated it separately, and took notes on the participants’ behaviors. Both raters’ observational notes were qualitatively analyzed for the reasons resulting in a change in conversational ratings. Inter-rater reliability was not calculated (36) as the scales were previously reported to be reliable. After rating the videos separately, the raters discussed observations before reaching a consensus on the final rating. For intra-rater reliability, one rater rated the 34 samples 2 weeks later with re-orientation to the scales. Data was entered into IBM Corp SPSS Statistics 22.0 (50) and reliability was analyzed using lntraclass Correlation Coefficients (lCCs). In accordance with the previous study (36), there was a moderate-strong positive correlation on intra-rater reliability on all measures (Interaction, r = 0.846; Transaction, r = 0.846; Acknowledging competence r = 0.867; Revealing competence, r = 0.696).

Data analysis

All data from the outcome measures were entered into IBM Corp SPSS Statistics 22.0 [IBM]. Descriptive statistics were performed for all key measures and the data was screened for normality using the Shapiro-Wilk tests of normality (51).

To answer both research questions (changes in conversational participation and support), the two-tailed Wilcoxon-signed rank test (alpha level of 0.05) was used to compare ratings obtained on the MPC/MSC at 3 and 6 months. Non-parametric statistics were chosen because the MPC/MSC are ordinal data and the homogeneity of variance assumption for parametric tests were not met. The MPC and MSC ratings were also sorted according to value, and the magnitude of change (i.e. most and least change). Descriptive analysis was performed to identify the frequency of each score at 3 months to determine the baseline function of the participants.

Results

Descriptive statistics and statistical comparisons between 3 and 6-month MPC/MSC scores for the included participants are presented in Table 2 to answer questions one and two (change in the degree of participation and support in conversation over time). No significant differences were observed between the 3 and 6-month scores for the participants with TBI and their CPs.

Table 2.

Adapted MPC/MSC scores for participants included in this study (n = 17): mean SD (range), with statistical comparisons between 3 and 6 months.

	Outcome measure	3 Months	6 Months	Difference between 3 and 6 monthsc	Statistic (p)
Participants with TBI (MPC) CPs (MSC)	Interactiona	2.38 SD 0.626 (1–3.5)	2.44 SD 0.705 (1–3.5)	0.059 SD 0.555 (−1–1)	z = −0.440 (p = 0.660)
	Transactiona	2.30 SD 0.561 (1–3)	2.41 SD 0.690 (1–3.5)	0.118 SD 0.516 (−1–1)	z = −0.921(p = 0.357)
	Acknowledging	2.38 SD 0.416 (1.5–3)	2.5 SD 0.707 (1–3.5)	0.118 SD 0.674 (−1.5–1)	z = −0.836 (p = 0.403)
	competenceb Revealing competenceb	2.36 SD 0.409 (1.5–3)	2.38 SD 0.714 (0.5–3.5)	0.453 SD 0.421 (0–1.8)	z = −1.18 (p = 0.238)

^aA score of 0 = No participation, 2 = Basic participation, 4 = Full participation

^bA score of 0 = Not supportive, 2 = Basic skill in support, 4 = Highly skilled support (41)

^cA change of 0.5 is considered a minimally clinical important difference

^dWilcoxon-signed rank test was performed.

The frequency of the participants achieving each score on the MPC and MSC are reflected in Figure 1a and b respectively. On the Interaction scale, 59% (n = 10) participants were able to mostly take responsibility for sharing the conversation while on the Transaction scale 52.9% (n = 9) had more than ‘some participation’ in the conversation. For CPs, 64.7% (n = 11) demonstrated more than ‘basic support skills’ on the AC scale and 70.6% (n = 12) had ‘basic support skills’ on the RC scale.

Figure 1.

(a) Frequency of participants with TBI (n = 17) who achieved a score on the MPC at 3 and 6 months. (b) Frequency of CPs (n = 17) who achieved a score on the MSC at 3 and 6 months.

According to Norman, Sloan, and Wynivich (52), improvement by more than one-half of the outcome score’s standard deviation is considered a minimally clinical important difference (MCID). In this case, one-half of the score’s standard deviation ranges from 0.205 to 0.357, and therefore a MCID in the present study would be 0.5 on the MPC/MSC. A descriptive analysis of the degree of change in scores from 3 to 6 months revealed that, in contrast to the outcomes of the inferential statistical comparison above, 82.3% (n = 14) participants made a MCID on the Interaction scale, 52.9% (n = 9) on the Transaction scale, 76.5% (n = 13) on the AC scale, and 58.9% (n = 10) on the RC scale.

The raters’ qualitative notes were collated and compared between both time points. Results indicated that conversational recovery was characterized by the cessation of some behaviors and emergence of new maladaptive behaviors at 6 months. For example, for one of the participant dyads (#57), the person with TBI scored 2.5 on the MPC at 3 months, and made no changeover time. According to the raters, this score was assigned because even though he displayed good conversational behaviors (e.g. appropriate eye contact, providing information in an organized way), he would occasionally interrupt his CP when she was talking. At 6 months the person with TBI had improved turn-taking skills but another deficit had emerged – the person with TBI was unaware of his CP’s frustration during conversation and did not use clarifying questions to alleviate the situation. A similar trend was observed for his CP. At 3 months, the CP initiated appropriate topics and acknowledged responses from the person with TBI. However, at 6 months, she started asking testing questions that ‘demanded’ rather than ‘invited’ information, behaviors not evident in the conversation at 3 months post-injury.

Discussion

Conversational skills are complex to evaluate given the multifaceted composition of sociolinguistics, psycholinguistics and social cognitive elements necessary for a rewarding interaction. For people with severe TBI, conversational effectiveness can be compromised due to a breakdown in cognitive-communication networks and changes in the responses of their CPs. Recovery of conversational skills during the first six months of recovery has received little attention. In this exploratory study, it was found that on the MPC/MSC scales, the degree of participation and support of the people with TBI and their CPs did not show a statistically significant change from 3 to 6 months. These findings do not support our preliminary hypothesis of a positive change in participants with TBI. This lack of change may be due to the following factors as explored below:

(a). Characteristics of the participants

Firstly, the characteristics of participants in this study may have accounted for the lack of change in MPC scores. Despite a diagnosis of aphasia (37%, n = 10) and dysarthria (88.2%, n = 15)), a majority of participants with TBI were already displaying some participation behaviors at baseline (score of 2), typical of interactive conversations (53). Thus, there is possibly little room for change. In addition, people with TBI who have 12–15years of education are five times more likely to recover compared to those with less than 12 years’ education (53). Therefore, participants in this study may have a large cognitive reserve as modulated by their high number of educational years pre-injury (82.3% had at least 12 years of education) (53).

(b). Nature and variability of recovery of cognitive-communication deficits

Recovery of cognitive communication deficits is not a stepwise logical re-acquisition of skills. The heterogeneity of cognitive-communication deficits may account for why recovery was observed with some skills whilst other problems emerged later in recovery (54). Insight into this variability was made possible by considering observations of interactional behaviors over time for participants. For example, for participant #57 (as discussed earlier), early maladaptive behaviors may have reduced at 6 months but other undesirable behaviors that emerged could have cancelled out that improvement on the global measure, resulting in no significant change.

Likewise, some CPs displayed one strategy at 3 months, but not at 6 months in response to a different presentation of deficits by the person with TBI. For example, the CP of participant #57 demonstrated active listening by repeating and acknowledging the person with TBI’s responses whenever he took an appropriate conversational turn. However, at 6 months, when the person with TBI displayed poorer awareness and understanding (which he did not at 3 months), she encouraged the use of cognitive supports (e.g. diary/phone) to aid his comprehension. This heterogeneity, including variable recovery of self-awareness and possible social cognition abilities, adds to the complexity in understanding these changes during the early recovery period (55,56).

(c). Influence of communication partners

People with TBI may exchange more information with familiar CPs (i.e. mother vs a paid caregiver) (27,37) and use more conversational strategies with a peer than other CPs (40). All the CPs in this study shared history with the person with severe TBI and although the injury may have had a considerable emotional impact on them, CPs were likely to be in the early adjustment phase where they may feel positive about the person with TBI’s communication abilities (57,58), Moreover, this study only investigates the sub-acute phase, when people with TBI are beginning to transit back home (33) and CPs have yet to become fully aware of the functional implications of cognitive-communication deficits (57,59–62). Overall, the TBI-CP relationship, the CP’s perception of communication abilities, and their reactions at different phases of recovery could impact on the findings of cognitive-communication recovery on the MPC/MSC.

(d). Nature of the outcome measures

The outcome measures used (adapted MPC/MSC) in this study have limitations which could have impacted the findings. Firstly, the adapted MPC/MSC have been used to describe communication abilities (37,47) and evaluate CP training programs (38,63), but not to examine the subacute period following severe TBI. The global MPC/MSC scales may not have been sufficiently sensitive to detect subtle changes in specific conversational behaviors (36,45). Nonetheless, another study that used a more specific linguistic approach with a procedural task also found similar results to the present study (i.e. no changes overall) (64). Another consideration is that without sufficient training, inexperienced clinicians may rate interactants at a seemingly basic level (score of 2), that does not reflect their conversational ability accurately. One drawback of training clinicians is that it was time-consuming (17 hours in this study) for them to develop an internal consistency with the scales, which may not be practical, clinically. However, another study showed that individual raters require different amounts of training regardless of experience, and student raters had similar scores to experienced clinicians (65). One strategy that inexperienced clinicians can employ is taking notes on behaviors that justify their ratings. The behaviors noted (whether positive or maladaptive) can also be used to set goals for CP training programs. Although the MPC/MSC scores may not reflect the early interactional change in a group study, it is recommended that clinicians still refer to the descriptors/anchors during the intervention to document progress and supplement this with other detailed analyses (45).

(e). Nature of the conversational task

The nature of the conversational task may have impacted changes on the MPC/MSC data. Communication abilities following severe TBI have been examined in genres such as narratives (35,66–69) and procedural discourse (34,64,68). During casual conversations, interactants are free to make conversational moves to negotiate their social identity and develop and maintain their interpersonal relationships. Such negotiation moves involve the use of executive functions, which are typically impaired in people with TBI (14). However, participants in this study had relatively unimpaired executive functioning, reflected in their baseline score of 2, which provided less capacity to show changeover time. Furthermore, the joint nature of casual conversations provides opportunities for CPs to facilitate the participation of the person with TBI (69), thereby potentially masking their cognitive-communication deficits. Nonetheless, other genres were elicited (e.g. personal recounts, picture description, etc.) as part of the larger study. Examining these genres may help determine which genre and analysis are the most suitable to diagnose and measure the change in this population.

The structure and predictability of a 10 min recorded interaction may provide an incentive for participants to use greater effort to engage in conversation (70), or employ support strategies more explicitly. Having a conversation in a quiet setting could increase participants’ language perception and processing thereby easing the burden to participate or provide support in conversation (71). Also, the heterogeneity of cognitive-communication deficits may not allow for the same behavior to occur frequently enough to be represented within 10-min samples (72). In summary, the nature of the conversational task may limit capacity to capture changes over time. This limitation implies that clinicians should attempt to obtain conversational samples across genres, settings, and CPs in order to gain a complete communication profile of the person with TBI.

(f). Interplay of other communication difficulties

The extent and impact of speech and language disorders following severe TBI are not well documented in literature. Speech and language disorders may limit the person with TBI’s ability to interact and convey content and reduce listener’s comfort during conversation (73) therefore limiting CP conversational support. Many of our participants presented with dysarthria (n = 15) and aphasia (n = 10) at the time of assessment and their speech and language difficulties could have contributed to the lack of interactional change during early recovery.

Summary of clinical and future research implications

Cognitive-communication behaviors are heterogeneous in nature and documenting recovery is challenging. Based on the findings of this study, some people with severe TBI have seemingly unimpaired conversational behaviors during the sub-acute phase. However, clinicians should monitor conversational behaviors and target any maladaptive behaviors early on as cognitive-communication deficits may become evident or worsen at a later stage of recovery. When evaluating conversational effectiveness, the CPs perceptions of the recovery process may influence early interactional behaviors and need to be taken into account during education and intervention. Clinicians are also advised to obtain conversational samples in various genres and supplement the MPC/MSC with other measures such as global impressions (26) and detailed conversational analyses (74). Observational scores obtained from the MPC/MSC may potentially help clinicians identify clients who are a priority for CP training i.e. those with a lower MPC/MSC scores and need more training. Given that conversational behaviors in the early stages may be subtle and complex to monitor, early intervention and carer education are essential to strengthen supportive facilitatory behaviors and prevent maladaptive behaviors in the sub-acute phase. Future research should work to investigate the recovery of conversational effectiveness over time in a variety of genres, settings, and types of CPs (e.g. familiar vs unfamiliar).

Limitations

There are several limitations associated with this exploratory study. Firstly, the relatively small sample size used means that the findings should be interpreted with caution. Secondly, there may be confounding variables beyond the researchers’ control, such as environmental, physical, and emotional factors that could have affected the participants’ behaviors during the assessment. Thirdly, the conversational behaviors were only rated on casual conversations, thus changes in other genres may not be reflected in this study’s findings. Lastly, using ratings scales may result in central tendency or extreme end-aversion bias, whereby raters favor the middle range of the scale (75,76) producing relatively unimpaired scores (i.e. 2/2) on the MPC/MSC.

Conclusion

On average, participants demonstrated no marked conversational difficulties in the early stages of recovery. However, observational data indicated that a small number of participants did display some interactional difficulties. In conclusion, documenting patterns of early cognitive-communication recovery is a complex task, laced with constraints surrounding the heterogeneity of communication behaviors, the influence of CPs and the paucity of suitable outcome measures.