Clinical Effectiveness of the Queen Square Intensive Comprehensive Aphasia Service for Patients With Poststroke Aphasia

Supplemental Digital Content is available in the text.

A phasia (an acquired disorder of language) has the greatest negative impact on quality of life of any medical condition, 1 but how best to treat it?
Systematic reviews of aphasia therapy support the hypothesis that the greater the number of therapy hours, the larger the gains in language recovery in the chronic stage poststroke. 2,3 People with aphasia (PWA) clearly benefit if they have on average over 100 hours of therapy. This target seems to be out-of-reach for most community-based health care services, including the UK NHS where a PWA might reasonably expect ≈8 sessions in total. 4 An efficient way to deliver the high doses required for chronic PWA is via an Intensive Comprehensive Aphasia Program (ICAP). Several have been trialled, mostly in the context of a research programme 5 and never before in the NHS. ICAPs provide high-intensity therapy at a minimum of 3 hours per day over at least 2 weeks; use a range of formats and treatment approaches including group and individual therapy; target impairment, activity and participation levels of language and communication functioning in line with the World Health Organization International Classification of Functioning framework; include education support for the individual and for families; and have a defined start and end date, with a set cohort of individuals entering and leaving the program together. Our ICAP contained all of these elements and is the first to have neuropsychological input for both assessment of PWA and therapy for both PWA and their families. Cohorts of 3 to 4 PWA attended the ICAP as day cases (nonresidential). Attendance was required for 7 hours per day, 5 days per week, for 3 weeks, much like the intensive upper-limb program that also runs at the National Hospital for Neurology and Neurosurgery. 6 The treating team of clinicians comprised speech and language therapists, speech and language therapist assistants, neuropsychologists, and a neurologist. See Materials in the Data Supplement for the full Template for Intervention Description and Replication (TIDieR) tool. This reporting guideline is designed to improve the documentation and implementation of clinical interventions. Post 3-week intervention, some patients were referred on to appropriate community services but did not receive any more therapy from the ICAP team.
The service was funded for 2 years, and we just completed the first one ( Here we report data from the 2 main outcome measures collected. (1) A language impairment-based measure, the Comprehensive Aphasia Test 7 from which we generated a score for each of the 4 main domains: spoken picture description (range, 9-87), written picture description (9-66), comprehension of spoken language (0-64), and comprehension of written language (0-62). (2) A functional communication measure, the Communicative Effectiveness Index (CETI), which is scored by the PWA's carer/relative/friend. 8 We also collected data on PWA's individual goals, as well as mood and quality of life outcomes for PWA and their carers. These data will be reported in subsequent articles. The data that support the findings of this study are available from the corresponding author upon reasonable request.
SPSS software was used for all analyses. Statistical significance was set at P<0.05 for all planned tests with a Bonferroni correction applied to all post hoc tests (P<0.0125). Greenhouse-Geisser estimates were used where sphericity was violated. We tested the following 3 hypotheses: Does the ICAP Significantly Improve Language Ability in PWA Across the 4 Language Domains?
A 2-way repeated measures multivariate ANOVA was conducted. Language domain (multiple dependent variable) had 4 levels (speaking, writing, auditory comprehension, and reading) and time had 3 (baseline, 3-week [immediately post-ICAP] and 12-week post-ICAP). We tested for an interaction between factors and explored these with post hoc tests. We also tested for significant changes between 3-week and 12-week scores. Because repeated measures tests reject any subjects with even a single data point missing, the analysis was performed on data from 36/46 patients. 20/552 data points were missing (3.6%). Two PWA were uncontactable at 12 weeks (COVID-19), and the majority of the remaining missing data was due to PWA not wishing to attempt written picture description.

Do Common Demographic Variables Affect the ICAP Language Outcomes?
A repeated measures multivariate ANCOVA was conducted to determine the impact of 3 demographic variables (age, months poststroke, and gender converted to categorical variables) on the main analysis outlined above.
Does the ICAP Significantly Improve Language Function for PWA?
The CETI was collected at 2 time points, baseline and at 12-week post-ICAP, so a paired t test was employed. Not all PWA had a carer/relative/friend, so this analysis was performed on data from 36/46 patients.

Effect Sizes
We quantified change using a standardized measure (Cohen's d for repeated measures data 9 ) and an unstandardized measure. The latter is a calculation of how much of the gap between the PWA's baseline score and the normal performance threshold has been closed: the back 2 normal percentage. 10 For the Comprehensive Aphasia Test, this threshold is the 95% cutoff derived from nonaphasic stroke patients; for the CETI it is a score of 88.

Hypothesis 1 (Domain-by-Time Interaction on the Comprehensive Aphasia Test Data)
There was an interaction between language domain and therapy time F(2.9,100.3)=12.7, P<0.0005, with monotonic increases in scores over time, with one exception ( Figure 1A).

Hypothesis 3 (Does the CETI Change Between Baseline and 12 Weeks?)
A paired t test demonstrated a significant effect of therapy, t(35)=5.4, P<0.0005, with the 12-week CETI scores higher by an average of 12.4 points compared with Baseline ( Figure 1B).

Effect Sizes
Both unstandardized and standardized scores demonstrated large effect sizes (>0.8 Cohen's d) for language production domains and the CETI, and medium effect sizes (>0.5 Cohen's d) for language perception domains. Speech production improved more than the other language domains (Figure 2).

DISCUSSION
PWA caused by stroke and who participated in the ICAP made, on average, large gains on both impairment-based and functional measures of language. Gains were sustained and in some cases continued to improve at followup. Given the therapeutic effects of the Queen Square ICAP, how does it compare with the other ICAPs? Of the 6 published ICAPs measuring change in language ability, 5 have shown significant improvements, with 2 showing large effect sizes similar to this study. 11,12 No other published ICAP studies have investigated therapy effects across language domains. Our finding of significantly greater effects on speech production may reflect that this is generally the domain that patients, especially in the chronic stage poststroke, care about and thus wish to work on most. Importantly, the presence of the therapyby-domain interaction strongly mitigates against either bias or test-retest explanations of our result. The significant improvements in 2 of the 4 domains at 12 weeks post-ICAP suggests that language gains were not only consolidated but also continued to improve. A similar carry-over effect was seen in the data from the Queen Square upper-limb programme 6 and one other ICAP. 13 The CETI scores support this with clinically significant changes (>11.4 8 ) in functional communication abilities similar to those reported in one of the other ICAP studies. 11 Related to this longer-term boost in function is the inclusion of neuropsychological input in our ICAP. Systemic therapy in this context has been shown to work 14 and may well be a key factor in the long-term effectiveness of our ICAP; as was the interdisciplinary focus on developing self-management skills. Future analysis of PWA and carer mood and quality of life outcomes collected as part of our service will investigate this further.

CONCLUSIONS
Given the effectiveness of this and other ICAPs, why are they not more widely available or considered as part of standard clinical practice? They certainly are resource and staff intensive, and cost is an issue. The Queen Square ICAP has running costs of ≈£300 000 pa which works out at £5500 per patient. This is much more than is currently spent on PWA in the United Kingdom, but it is only about the same as the cost of a hip replacement on the NHS. Rather than the current pragmatic state of service provision for PWA where, at least in the United Kingdom, woefully insufficient resources are shared out equally or in the United States, where resources are concentrated almost exclusively in the acute stage poststroke, should not the available evidence for clinical efficacy dictate how services should be set-up to benefit patients? Given that stroke prevalence is increasing and the compelling evidence we provide here on the effectiveness of an ICAP for chronic PWA, surely now is the time to provide these high-quality services and change lives.