The impact of group therapy on word retrieval in people with chronic aphasia

1 Introduction

Anomia is the most common impairment associated with aphasia (Goodglass & Wingfield, 1997). Difficulty retrieving appropriate words presents a significant impact on an individual’s communication and often leads to conversation breakdown (Lesser & Algar, 1995). Treatment has been found to be effective for many individuals with anomia using a variety of different treatment approaches (e.g. Best, Herbert, Hickin, Osborne, & Howard, 2002; Boyle, 2004; Davis & Pring, 1991; Howard, Patterson, Franklin, Orchard-Lisle, & Morton, 1985; Nickels & Best, 1996; Renvall, Laine, Laakso, & Martin, 2003). However, structured one-on-one treatment models have limitations – crucially, it has been suggested that they may not lead to generalisation of skills to situations outside the therapeutic environment (Palchalska, 1991). The overriding goal of any treatment approach should be the improvement of communication in real-world situations, thus facilitating generalisation of skills beyond the clinical setting must be a key objective of intervention.

One-on-one face-to-face treatment is time intensive and therefore a costly treatment delivery model. Consequently, some authors have examined whether treatment delivered as a home programme could also be effective: It is clear that it can. For example, Nickels and Best (1996) showed clear gains from treatment delivered as worksheets completed independently at home. Similarly, Mason et al. (2011) showed that a simple word retrieval treatment (Repetition In the Presence of a Picture, RIPP) could not only produce gains in word retrieval using personally-chosen items (Renvall, Nickels, & Davidson, 2013a) but that for one participant these gains extended to connected speech.

Providing treatment in a group provides another potentially more cost effective option compared to one-on-one treatment. Importantly group therapy has been suggested to provide a number of additional specific benefits that are not always met by individual treatment. For example, some authors have suggested that generalisation may be supported by the adoption of a group treatment approach (Elman & Bernstein-Ellis, 1999; Kearns & Elman, 2001; Palchalska, 1991). Group treatment provides an increased range of conversational partners and conversational topics (Ross, Winslow, Marchant, & Brumfitt, 2006), increased friendships and support (Brown, Worrall, Davidson, & Howe, 2011). Groups have also been reported to provide a range of psychosocial benefits such as reduced depression (Brumfitt & Sheeran, 1997), improved social participation and reduced feelings of social isolation (Vickers, 2010), and increased confidence (Ross et al., 2006).

Here we focus on the evidence for improvement in the language impairment as a result of group treatment. Indeed, previous studies have argued that group treatment can be effective in the treatment of aphasia using a number of treatment approaches (e.g., Antonucci, 2009; Aten, Caligiuri, & Holland, 1982; Elman & Bernstein-Ellis, 1999; Bollinger, Musson, & Holland, 1993; Brumfitt & Sheeran, 1997; Marshall, 1993; van der Gaag et al., 2005). Brady, Kelly, Godwin, and Enderby (2012) in their systematic meta-analysis concluded that group treatment was as effective as individual treatment for both auditory comprehension and expressive language functional communication. Nevertheless, in their review of group therapy literature, Lanyon, Rose, and Worrall (2013) concluded that there was still a lack of clear empirical evidence into the effectiveness of group treatment in improving communication and for the aspects of treatment that made for an effective group.

In particular, methodologically sound studies investigating the specific effects of group treatment on word retrieval are limited, despite the importance of word retrieval for communication. A notable exception is a series of studies examining constraint-induced aphasia therapy (CIAT; e.g. Meinzer, Djundja, Barthel, Elbert, & Rockstroth, 2005; Meinzer et al., 2004; Pulvermuller et al., 2001) where there is clear evidence of improved word retrieval. However, CIAT is rather different to the ‘traditional’ aphasia therapy group in its intensity, highly structured and restricted task base (usually a card exchange game), and focus on oral language alone (without the use of multi-modal strategies). In our experience, typical ‘traditional’ (impairment-focused) aphasia groups, are usually held less frequently (often weekly), are often centred around topics of general interest (e.g., travel, cooking, gardening) and/or current affairs, include group conversation/discussion combined with a range of activities to promote language use and language recovery (e.g. Bollinger et al., 1993). This is supported by Rose and Attard’s (2015) recent survey of Australian community and outpatient aphasia groups. They found that when asked about group activities, all of the 156 respondents reported using group discussion and a sizable proportion (51%) reported the inclusion of specific language activities. The groups were most commonly run weekly (49.3%), and the most common location was in a hospital or rehabilitation centre (44%).

Eales and Pring (1998) report a study which examined the outcome of a twice-weekly conversation group without structured activities. They examined the effects on word retrieval of a group intervention which focused on conversations around ‘shopping’ or ‘holidays’. For example, shopping sessions discussed going shopping, which shops to visit, where to get different items, which items might be needed and making shopping lists. However, there was little clear evidence of improved naming as a result of the group sessions (unlike a preceding phase of individual treatment), due to steady improvement in naming throughout the study. Drummond and Simmons (1995) also found no specific improvement in lexical retrieval following a discussion-based group treatment. Several other studies, however, have argued that group discussion can improve language impairment, although many of these are hard to interpret due to design limitations. For example, Wertz et al. (1981) found participants had significantly improved language following group treatment focusing on discussion of news topics. However, the study was designed to compare the effectiveness of individual and group treatment approaches and could not exclude the possibility that the improvement in the (less effective) group treatment was a result of spontaneous recovery. Indeed, this was particularly likely as their participants were very soon post-onset of their aphasia (4 weeks).

As noted above, in our experience, aphasia groups typically include specific language-based activities, in addition to group discussion. However, there is a relative scarcity of evidence for the impact of this model of group therapy. Bollinger et al. (1993) used two group therapy approaches “structured television viewing treatment” and “contemporary group treatment”. While “structured television viewing” involved joint viewing of a television program and discussion, “contemporary group treatment” was closer to our experience of typical aphasia groups, consisting of group discussion and a range of language activities such as repetition and word association. While the two group types were not evaluated separately, Bollinger et al. found clear treatment-related improvements at the group level on the Porch Index of Communicative Ability (PICA; Porch, 1981) and the Communicative Abilities of Daily Living (CADL) test (Holland, 1980).

Given the prevalence of anomia, the importance of word retrieval for communication, and the need to facilitate generalisation, it would seem appropriate to address whether a traditional group treatment approach combining topic-focused conversation and discussion with language facilitation tasks achieves the goal of improving word retrieval both in naming and in connected speech. This seems particularly important in light of the evidence for the psychosocial benefits of aphasia groups and the potential cost-savings. Therefore, in this study, we examined the effects of group treatment on word retrieval with three participants with mild-moderate aphasia and anomia. All three participants presented with communication skills that were functional for daily tasks, but yet faced difficulties with more unstructured interactions such as conversation with peers. In planning the groups, we replicated a real-life clinical setting and a ‘typical’ aphasia group protocol as closely as possible in terms of timing, length and size of the group and the tasks, structure and stimuli used. The groups had a focus on current affairs and topics of general interest as this was felt to be appropriate for the communication needs of the participants. This choice is also consistent with the findings of Davidson, Worrall and Hickson (2003) who identified news and current affairs and expressing opinions on social issues amongst areas that were limited in the communication of people with aphasia when compared with healthy controls (see Renvall, Nickels, & Davidson, 2013b, for discussion). Improving communication on these topics may therefore help reduce the social isolation that is commonly experienced by people with aphasia (Parr, 2007): improving social interaction has been noted a key area of concern often identified by people with aphasia and their families (Brown et al., 2011).

In sum, our objective was test whether a ‘traditional’ group treatment approach for people with chronic aphasia would improve word retrieval when measured at both a single-word level and in connected speech. In addition, we examined whether supplementing the group treatment with a home treatment programme for vocabulary relevant to the topic (RIPP, Mason et al., 2011) would increase (or facilitate) the benefit for word retrieval. Hence, half of the topics discussed also received both the group and the RIPP home programme while half received only the group programme.

2 Method

3 Results

The three participants attended all group treatment sessions in the programme, except JDC who missed one session due to family commitments. Two of the three participants did not complete the home programme to protocol: RYT was unable to complete one week of the “health” topic home programme due to computer problems; RAA did not complete one week of the home programme for the “pollution” topic due to ill health.

4 Discussion

In this study we carried out a group treatment programme with three people with chronic aphasia. We used what we believe is a common ‘traditional’ model for aphasia language groups. Given the prevalence of anomia, and its impact on communication, we focused on the effects of the group therapy programme, and the same group programme supplemented by a word retrieval home programme, on word retrieval.

Evidence for treatment-related improvement on word retrieval in either picture naming or connected speech (semi-structured interviews) was equivocal for both the two groups of treatment topics (topics treated with the group programme and topics treated with the group programme with an additional RIPP home programme). However, for naming, although there was no clear differentiation in rate of improvement during the treatment phase compared to the no treatment phase (except for a result close to significance for RAA in the group+home programme condition), all of the participants showed a pattern of improved naming over the course of the study for treated topics that was not found for the untreated topics. A difference in the timeline of pretreatment baselines for the second control topic, with a longer gap between baselines than the treated topics, may have contributed at this first time point. However, the timing for both subsequent assessments was identical for the untreated topics and treated topics, and yet gains were found for treated and not untreated topics. Moreover, when we directly compared the gains, the gains on treated topics were significantly greater than those of untreated topics for two of the three participants, RYT and JDC, and trending towards significance for the third, RAA. It therefore seems likely that the continued gains that the participants made for the treated topics were as a result of intervention.

When the group was supplemented by a home programme, it was surprising that greater gains were not found, given that the same approach (RIPP), implemented in the same manner, was found to be effective for two of the three participants in our earlier study (Mason et al., 2011). We discuss the possible reasons for this below. However, there were numerically larger (and more strongly significant) gains over the course of the treatment for naming of topics which received an additional home programme compared to those that received the group alone, and the difference between the two treatment conditions was significant for one participant (JDC) and close to significance for a second (RAA). Hence, at least for these participants, it appears that supplementing the group with the home programme led to greater gains, than the group alone.

All the participants provided feedback that they felt the group was useful and that they discussed the topics with family members at home after the group. However, as the focus of this study was on word retrieval, this was not quantified. Moreover, we cannot be sure whether the willingness of the participants to discuss the topics in conversation outside the group was facilitated by improved word retrieval or was due to another factor such as increased confidence.

We suggest that several factors may have contributed to the limited improvement measured: i) the nature of the treatment, ii) the nature of the targets and iii) difficulty in measurement. We will discuss each in turn.

First, it is possible that the treatment itself was not optimal in terms of either the tasks used or the dosage of the treatment. For example, it is possible that a single treatment session addressing each topic may be insufficient and a programme which allows for increased exposures to topics and related words would be more likely to produce measureable change. Other authors have achieved this through combining group treatment with other intervention such as individual therapy or homework (e.g Bollinger et al., 1993), increased intensity (e.g. CIAT; Meinzer et al., 2004, 2005) or a longer duration of therapy programme (e.g. Aten et al., 1982; van der Gaag et al., 2004).

Second, the target words used in the picture naming task could be considered ‘difficult’. The target words had a number of complexities which would have impacted on the participants’ naming. In selecting vocabulary that would be important when discussing these current affairs topics, we included words of different word classes including verbs (e.g., “escape”), common nouns (e.g., “hospital”), proper nouns (e.g., “Iraq”) and adjectives (e.g., “silent”). The words also covered both concrete (“library”) and abstract concepts (“fear”; In Appendix A we provide a list of the full set of targeted words for each topic). Renvall et al. (2013b) note that a variety of word classes is needed for full participation in conversation and suggest that abstract words are particularly needed in the expression of thoughts, opinions and emotions. However, abstract and low frequency words are difficult for people with aphasia to name (e.g. Nickels & Howard, 1995). Moreover, while frequency information was not collected for the targets, many were specialized to the topic (e.g. “people-smugglers”, “petrol sniffing”) and it is not unreasonable to suppose that they would have low frequency. All of these factors added a level of difficulty to word retrieval. Mason et al. (2011) noted that the properties of the stimuli used in their home programme (long, low frequency words) most likely contributed to relatively smaller gains in naming compared to a clinician directed programme. However, the stimuli in the Mason et al. home programme, while functionally relevant, and relatively long and low frequency were, nonetheless, all concrete nouns (e.g. pottery tools, car parts, road names). Consequently, it seems likely that the difficulty of our target items may have led to smaller gains, and perhaps a larger ‘dose’ of treatment was required before clear improvement was seen as a result of the home programme (each word was only repeated a maximum of 10 times over the two weeks). However, given that, in general, compliance is less good with a home programme (where other life events conspire against even the most eager participant), increasing dose while making the task feasible may not be a straightforward issue. Nevertheless, we only required 10 repetitions in the RIPP therapy, and no opportunity was given for greater practice. When words are seen as personally and/or functionally relevant, at least some people with aphasia will meet the challenge given the opportunity and encouragement, even with low frequency, abstract stimuli.

In addition, the nature of the targets presented obvious difficulties in terms of measuring word retrieval in isolation. One of the biggest challenges for measuring gains in word retrieval for abstract stimuli is how to elicit the target names. We chose to use picture naming but appropriate stimuli were often difficult to obtain and even with the addition of verbal cues/definitions, some items may have remained unclear. There is no easy answer to this problem, but it is one that is important for researchers to address in order to be able to accurately measure gains as a result of treatment. Indeed, there is the question as to whether the use of a task assessing word retrieval in isolation is the most appropriate to obtain a meaningful picture of gains in lexical retrieval following the discussion-based group treatment approach. While the home programme directly targeted the words in a task similar to that used in assessment, this was not true for the group-treatment. Indeed, only around 50% of the target words were targeted directly in group therapy (in a written sentence completion task), with the others were targeted indirectly by way of the group discussion. While a link has been observed between picture naming and word retrieval in conversation (Best, Herbert, Hickin, Osborne, & Howard, 2002), Webster, Whitworth and Morris (2015) raise the point that the exact nature of the relationship between lexical retrieval at the single word level and in connected speech still requires clarification. Thus, it may be that a picture naming task may not be the best means to measure improvements in word retrieval when treatment uses indirect approaches such as group discussion. However, in this study, we also considered retrieval of the target words in a connected speech task - a semi-structured interview on the topics under discussion. This was chosen in order to simulate a plausible communicative situation where individuals are asked about their views on topical issues. In order to keep the format as consistent as possible, the same clinician conducted the interview on each occasion. Unfortunately, this may have been one of the factors that led the participants’ conversations to be variable, given the natural tendency not to discuss details that have previously been shared (for discussion see also Croot et al., 2014): Even across baseline sessions the number of target words successfully produced varied. This outcome reflects the known difficulty of examining word retrieval in conversation: Discourse is variable (e.g. Boles, 1998; Perkins, Crisp, & Walshaw, 1999; Webster, Franklin, & Howard, 2007) and subject to numerous outside influences. There is no easy answer, but there is clear need for further investigation into best practice for assessment of changes in word retrieval in connected speech particularly when the treatment involves discussion-focused group treatment approaches. This echoes the findings of Kearns and Elman (2001) and Aten (1991), who identify measuring outcomes as a major difficulty of group therapy.

5 Conclusions

In this study, we replicated group language therapy for aphasia as it is often conducted clinically. Group therapy has been suggested to have numerous advantages (e.g. improved generalisation, support from other people with aphasia). It also offers people with aphasia the opportunity to share opinions and engage in a range of communication topics with a range of communicative partners. Our participants showed a pattern of improved picture naming over the course of the study, yet there was no statistical evidence that could unambiguously attribute this to the effects of treatment. Nevertheless, in contrast to words in topics addressed in the group sessions, words from untreated topics showed neither a treatment effect nor a pattern of improved naming across the course of the study.

The steady improvement for the treated words and not the untreated words suggests that there may be potential for traditional group-based therapy to be effective in improving word retrieval. Moreover, the fact that the group with the home programme resulted in greater improvement than the group therapy alone, for at least some of the participants, suggests that supplementing group therapy with more targeted word retrieval treatment may be beneficial. Indeed, we would urge that our results should not be seen as a case against group treatment for anomia, potential problems in measurement may have contributed to our results.

For both the clinician and the researcher, the study raises considerations for the evaluation of treatment impact for group-based treatments. We attempted to test the outcomes of a group treatment approach for anomia using (verbally cued) picture naming, and a less constrained task: semi-structured interviews. Each approach presented with difficulties. In the case of picture naming, difficulties arose in the choice of appropriate target words and stimuli. A full range of word types is important for communicative function, however, as we found, can present difficulties for testing with a picture naming task. In order to obtain meaningful data from a picture naming task, we suggest that the clinician needs to consider how stimuli could be presented to minimise ambiguity. Issues to consider include: Are the in the words easily picturable? Could additional stimuli help clarify the target e.g. verbal cues, audio stimuli? For the less structured task of semi-structured interviews, we found no meaningful improvement when looking at target word productions and this perhaps reflects the overall complexity of testing in the context of conversation. Further studies into assessment of word retrieval in conversation in a way that is accessible for clinicians are greatly needed. We also noted anecdotal evidence that the participants were discussing treatment topics in conversation with family members outside of the clinical environment. We suggest that such findings indicate that an assessment communication confidence may be helpful as an assessment of treatment impact.

Nevertheless, our data should be taken as a cautionary note. If our results are found to be replicable, it is clear that we cannot assume that traditional group treatment approaches improve the language impairment in aphasia, no matter what other benefits they may confer for people with aphasia. Factors that need to be considered include: Are the kinds of tasks appropriate? Perhaps discussion-focused tasks while engaging, do not provide sufficient semantic and phonological focus to produce improved word retrieval. Is a greater amount of treatment required to improve word retrieval? A single session per topic may simply be insufficient. This is likely to be particularly the case when vocabulary is abstract and low frequency and even when supplemented by a home programme. We would hope that further research can provide the answers to these questions and provide further guidance regarding how group treatment can use engaging, ecologically valid materials and tasks and provide not only psychosocial benefits for people with aphasia but also improved languageskills.

Conflict of interest

The authors declare that there are no conflicts of interest.