Recovery from moderate aphasia in the first year poststroke: Effect of type of therapy☆☆☆★★★♢♢♢♦

Article Outline

• Abstract
• Methods
  • Subjects
  • Procedure
  • Evaluation
    • Reading and writing
    • Aphasia
    • Communicative abilities
  • Statistical analysis
• Results
• Discussion
• Conclusion
• Acknowledgements
• References
• Copyright

Abstract 

Carlomagno S, Pandolfi M, Labruna L, Colombo A, Razzano C. Recovery from moderate aphasia in the first year poststroke: effect of type of therapy. Arch Phys Med Rehabil 2001;82:1073-80. Objectives: (1) To determine whether 2 model-based remediation programs affect writing performance in unselected subjects with moderate aphasia and whether there is consequent improvement in everyday life, and (2) to interpret the potential changes observed by recourse to a theoretical model. Design: Consecutive sample, multiple baseline, within subject crossover study. Setting: Ambulatory care units. Participants: Eight subjects with moderate aphasia from 6 to 12 months postonset. Intervention: A standardized test for reading and writing skills was given at the beginning and the end of each therapy program and 1 month after therapy stopped. Main Outcome Measures: Functional outcome measures were the Communicative Abilities in Daily Living (CADL) test and subtests from standardized aphasia assessment. Results: After the 2 programs, there was improved writing performance, which was maintained after therapy stopped. Patterns of improvement corresponded to each of the 2 programs. Learning transfer was observed on the CADL test and functional writing, but gains on oral language were limited. Only 1 program was effective for 6 of the 8 patients. Conclusion: Specific rehabilitation programs aid recovery from aphasic symptoms from 6 to 12 months postonset. Individual response is linked to type of treatment. The interpretation is linked to a model-based description of aphasic symptoms and mechanisms of functional recovery. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation

Keywords:  Aphasia, Cerebrovascular accident, Italy, Neuropsychological tests, Reading, Rehabilitation, Writing

DURING THE FIRST FEW MONTHS after the onset of aphasia, language abilities improve in a number of patients, but the precise temporal limit of this process is not well established. As a result of some studies, aphasia treatment is believed to enhance recovery¹, ², ³, ⁴ and to extend its temporal limit beyond the sixth month postonset.¹, ², ³, ⁴, ⁵ However, aphasic subjects who received 2 hours of treatment for 24 weeks did not show greater improvement than subjects who received no treatment at all⁶ and, in another study,⁷ aphasics who were given 30 hours of language stimulation by professional speech therapists showed a recovery curve that overlapped with subjects who received only general support from volunteers.

It is noteworthy that studies¹, ², ³, ⁴ reporting therapy effects used longer (6mo-1yr) and more intensive (≥3hr/wk) treatment than studies⁶, ⁷ that failed to show significant treatment effects. These differences suggest that the amount of treatment may be a critical factor in obtaining greater language improvement than with spontaneous recovery alone. Moreover, during intensive and prolonged treatment, speech therapists may have more opportunities to design programs to match the pattern of language deficit in the individual patient. Effects of the type of treatment, however, have scarcely been addressed in the literature on aphasia rehabilitation and some group studies have usually been assumed as evidence against this view.

The Veterans Administration Cooperative Study on Aphasia,² for instance, compared individual language stimulation with group therapy that stimulated conversation. It was found that the 2 treatments had been equally effective on measures of informativeness and overall language score. Another study³ investigated the effect of 3 treatment modalities—the Language Oriented Treatment, individual language stimulation, and unstructured group communication. In this case, benefit from the 3 therapies with respect to untreated patients was found, but no differences could be shown among the 3 groups who had received treatment. It should be noted, that in the first study² only the individual language stimulation group improved on the subtest dealing with written language. This indicated that the type of treatment may affect the pattern of recovery from dysphasic symptoms. In the latter study,³ the outcome measure was a composite score reflecting oral and written language subtests and no analytical measures were used to explore whether different patterns of improvement could result from the 3 different therapeutic approaches.

This aspect was addressed in a further study comparing the effect of 2 different types of language therapy, the Systematic Therapy Program for Auditory Comprehension Disorders (STADCAP) and traditional language stimulation.⁸ To determine whether and how functional recovery had occurred, a multiple baseline assessment method was used and a control group was included. No significant differences among the 3 groups were observed at the posttest even on the auditory comprehension subtests, which used practice materials from STADCAP. This trial, however, included a considerable proportion of patients with aphasia lasting more than 12 months and was limited to 40 therapy sessions during a 5-month period. Furthermore, the STADCAP program used 28 treatment tasks covering a wide range of linguistic elements and constructions (from phonology to lexical semantics and morphosyntax). This suggests that the STADCAP was too extensive to be learned in such a treatment regimen, and the reduced amount of treatment did not allow even patients who received conventional language stimulation to obtain significant gains. Thus, no firm conclusion could be drawn about the relevance of the type of treatment on recovery from aphasia.

In recent years, other studies have examined the effects of the type of treatment on recovery from aphasia by investigating relationships between the pattern of language deficit(s) at the pretest and aspects of the therapeutic program. These studies involve the so-called model-based approaches to therapy, ie, the therapy program is directed at a specific “locus of impairment” within a model of linguistic processing.⁹, ¹⁰ Within this theoretic framework, the effects of the type of therapy have been shown by the correlation between therapeutic program and pattern of improvement. However, the effectiveness of these programs has usually been proven in single case studies that used patients with a long history of aphasia and sufficiently consistent pattern of errors to justify targeting the therapeutic program. Generalization of these results to everyday clinical practice raises a number of theoretical and practical issues. For instance, aphasic patients in the first months after onset are usually regarded as more variable in their responding than chronic patients.¹¹ Furthermore, the pattern of language deficits in the aphasic individuals who receive treatment is usually more complex than those described in patients included in single case studies.¹² Finally, although therapy programs for single case studies have been described as laborious and time-consuming, generalization of effect to everyday activities has not been demonstrated for most of them.

In our study, the issue about effect of type of treatment on recovery from aphasia is addressed by comparing 2 model-based treatment methods for writing disturbances in a small group of patients with moderate aphasia lasting from 6 to 12 months, ie, a postonset period during which language recovery can occur.¹, ², ³, ⁴, ⁵ Contrary to previous group studies,², ³, ⁸ our experiment benefited from both the multiple baseline design, which allowed us to analyze in-depth correspondence between pattern of recovery and therapeutic program, and the fact that each patient was exposed to multiple treatment programs to identify possible selective response.

More specifically, the study attempted to answer the following questions: (1) Can model-based therapeutic programs for chronic aphasics be generalized to treatment of unselected patients with less stable symptoms?; (2) Does retraining aphasic subjects in writing have a positive effect on functional recovery?; and (3) Is functional recovery linked to type of treatment?

Back to Article Outline

Methods 

Subjects 

From January 1989 to May 1996, 12 consecutive aphasic subjects completed 2 treatment programs for adult dysgraphia at the IRCCS Santa Lucia in Rome or in Napoli. Speech therapists under the supervision of the Neuropsychology Unit, Seconda Università degli Studi di Napoli, administered the programs. The 2 programs had been devised according to current models for describing writing disturbances in aphasic subjects and used in a pilot study to retrain 6 subjects¹³, ¹⁴ together with a communication-oriented therapy.¹⁵

Following this study, other patients were enrolled. All received standard language stimulation for at least 3 months at the speech therapy unit of the Policlinico Universitario Federico II in Napoli and had been cooperatively referred by their own therapists to the Neuropsychology Unit for assessment with the reading and writing subtests of the Batteria per l'Analisi dei Deficit Afasici¹⁶ (BADA). Nine consecutive patients, with at least 8 years of education, were examined at 6 to 9 months postonset. Their scores on the shortened version of the Token Test (range, 12-22; maximum score, 36) indicated moderate aphasia. For these patients, it was suggested that phonologic and lexical difficulties could be tackled by structured stimulation of written language abilities. Indeed, learning transfer between oral and written language modalities has been found to occur in the course of standard language stimulation treatment.¹⁷

Furthermore, the effect of communication-oriented treatment has been observed even in absence of language recovery,¹⁸ and the new treatment cycle included a number of therapy sessions dealing with functional communication.¹⁵ In 1 case, the consideration that relevant dysarthric difficulties could be compensated for by improved writing abilities also justified our proposal.¹⁹ The therapy cycle was first discussed for each patient with her/his therapist. Then, it was proposed to the patient and relatives, and informed consent was obtained from 7 of the 9 subjects. The other 2 families refused because of difficulties with the intensive treatment regimen.

We reviewed the clinical records of the 7 new patients who underwent therapy and of the 6 who had participated in the pilot study. Eight subjects were included in our study, using the following criteria: (1) time from onset of aphasic symptoms to the beginning of the first treatment program from 6 to 9 months (mean, 7.4mo); (2) aphasia resulting from a single left-hemisphere lesion; (3) aphasia of moderate severity as assessed by a standardized aphasia test²⁰; and (4) the patient had received the 2 treatment programs within an interval of 30 to 60 days (mean, 44.2d). Four subjects, previously included in the pilot study, were excluded because of long postonset; a fifth dropped out during the first therapy program. Clinical and demographic features of the 8 subjects in this study are summarized in table 1.

Table 1: Clinical and Demographic Features of the 8 Patients at the T1 Aphasia Assessment

Abbreviations: M, male; F, female; None, no deficit; Mild, mild deficit; Mod, moderate deficit; Sev, severe deficit.

Procedure 

The baseline scores for reading and writing abilities at time 0 (RWA-T0) were obtained while patients were still receiving conventional language stimulation. This continued for about 1 month. At the end of this stimulation period, patients received the Communicative Abilities in Daily Living²¹ (CADL) test, an aphasia standardized assessment,²⁰ and the RWA-T1 (time 1). Each patient was then assigned to 1 of the 2 treatments programs with the decision about which of the 2 treatments was the first being administered to each patient being left to his/her own therapist. On completion of the first treatment program, patients were given the RWA-T2, which was followed by 1 month with the modified Promoting Aphasics' Communicative Effectiveness (PACE) therapy.¹⁵ A new RWA (RWA-T3) was then given, after which the other treatment program for writing disturbances was given. At the end of this, patients received the RWA-T4, aphasia, and CADL tests. The RWA test was again administered 1 month after therapy stopped (RWA-T5).

Each of the 2 writing therapies consisted of 20 to 24 1-hour sessions in 5 to 6 weeks. The pilot study¹³, ¹⁴ provided empirical evidence that this number of sessions was sufficient, for both programs, to train patients to a predetermined level of performance on practice stimuli.

The first treatment—“lexical” therapy—used a written naming task structured as a crossword puzzle. This was intended to stimulate retrieval of whole word spelling by providing the patient with visuolexical cues (eg, the number of the letters composing the stimulus, presence of relevant orthographic group, the first letter, letters belonging to the stimulus in an anagram format). The cues were progressively faded out during the treatment until 80% correct responses on practice lists were obtained in exercises where the patient was provided only the number of letters composing the stimulus word. Each therapy session involved a sublist of 10 stimuli. Sublists were alternated across sessions to avoid practice. The entire treatment program used a list of stimuli (line-drawings of 40 objects and 40 actions) with names matched for length and orthographic complexity with those used in the assessment.

The second treatment—“nonlexical” therapy—involved exercises of repetition, phoneme discrimination, phonemic segmentation of a dictated stimulus, letter-to-sound matching, writing syllables, or pronounceable nonsense letter strings under dictation. Sublists of nonsense letter strings included 10 to 12 stimuli where one-to-one sound to letter correspondences apply. These stimuli were matched for length with those of the RWA. Single phoneme to multiple letter string correspondences, eg, /k/ → “ch”, were not addressed in the program. The 2 treatment programs have been described in detail,¹³, ¹⁴ as has the functional treatment program.¹⁵ Five patients included in this study first received the nonlexical therapy.

Evaluation 

Each of the 2 treatment programs was intended to stimulate only 1 of the 2 writing routines (lexical, nonlexical) used by unimpaired or brain-damaged subjects.²², ²³, ²⁴ Thus, the resulting improvement was expected to show a peculiar pattern. For instance, following the lexical therapy, patients were expected to show improved performance on written naming or word writing on dictation tasks, but not in writing dictated nonsense stimuli. On the contrary, after the nonlexical therapy, patients would show significant improvement when writing nonsense stimuli and words on dictation, but not in the written naming task.

Reading and writing 

According to the multiple baseline design, the RWAs used the BADA test,¹⁶ which includes different tasks to check the effectiveness of lexical and nonlexical routines for reading and writing. This approach corresponds to other cognitively oriented methods for aphasia assessment, eg, the Psycholinguistic Assessment of Language Processing in Aphasia (PALPA) for English-speaking subjects.²⁵

As far as this study is concerned, we focused on tasks requiring written naming of objects (40 items) and actions (40 items), as well as writing on dictation word (80 items) and on nonsense stimuli (80 items). The stimuli in the 3 lists were matched for length (4-7 letters), and orthographic complexity, and lexical stimuli were also matched for frequency. In each of the 3 lists, half of the stimuli may be written by applying one-to-one sound to letter correspondences, eg, /d/ → “d”; half of the stimuli contained at least 1 phoneme corresponding to a multiletter string, eg: /k/ → “ch.” This list composition allowed us to control a potential bias for analyzing relationships between patterns of improvement and the 2 therapy programs.¹³, ¹⁴

All the patients in the study performed better on the oral naming subtest than on written naming task of aphasia assessment. Thus, improved written naming could have been observed after the nonlexical therapy because of the improved phonology to orthography segmental conversion of phonological information. To overcome such a bias, when relationships between the type of treatment and pattern of improvement were analyzed, results on the written naming tasks only took into account the 40 stimuli containing 1 phoneme to multiletter string correspondences. These stimuli had not been addressed with the nonlexical therapy and could not be concerned with improved phonology to orthography nonlexical conversion. Conversely, with respect to the nonlexical therapy, the results from the nonsense stimuli writing task only took into account the 40 stimuli in which one-to-one sound to letter correspondences apply.

Aphasia 

The word reading subtest from the BADA¹⁶ provided a first measure of learning transfer. Another outcome measure was provided by standard aphasia test before and after the study period (T1 and T4). This test did not include an overall aphasia score.²⁰ Therefore, changes on standard language examination were analyzed on subtests from the battery regarding functional communication, ie, oral and written naming, oral and written language comprehension. Individual improvement on the aphasia assessment subtests was acknowledged after consensus between the first author (SC), the therapist involved in doing therapy, and the psychometrist involved in the assessment. Additional data were provided by another task included in the aphasia test which was also administered 3 months posttherapy. In this task, patients were requested to write a 1-page letter about a personal concern.¹, ⁴

Communicative abilities 

Finally, the CADL test²¹ was given at the beginning (T1) and at the end (T4) of the therapy period. In the interval between the 2 treatments, each patient received 18 to 24 sessions of communicative-oriented therapy. Because of concerns about bias, we focused on the results from the 21 CADL items that explore functional reading and writing (READ/WRITE subtest; maximum score = 42).²¹

Psychosocial adjustment was examined by unstructured interview with patients' relatives in the 3 to 6 months after treatment stopped. Testing was performed by psychometrists not involved in doing therapies.

Statistical analysis 

The first analysis concerned overall writing performance at the 6 RWAs. The number of the correct responses on the 3 writing tasks were summed and the resulting scores were entered a repeated-measures analysis of variance (ANOVA). The same procedure was applied to scores of the word reading task from the BADA.

The second analysis concerned relationships between type of treatment and pattern of improvement. To do so, scores from the 3 writing tasks were entered into a 3-way (type of treatment, writing task, pre-post) repeated-measures ANOVA. Scores on the 40 stimuli that could have been related to the treatment method were doubled in order to equate them to the 80 on the word writing task. Results from the RWA-T0 and the RWA-T5 were not included in this analysis.

Changes on the CADL's READ/WRITE subtest were analyzed by paired comparison as well as the results on aphasia assessment subtests. Significance was set at p less than .01.

Changes at the individual level in writing performance were analysed by a separate 2-factor Mantel-Haenszel chi-square test²⁶ on the results of the 4 × 8 pre-post RWAs. The first factor was the pre-post change in performance, the second the type of treatment. This method allowed us to recognize the interaction between the 2 factors and to avoid a large number of unconstrained chi-square tests. It is noteworthy that, in any event, type I error did not constitute a real problem because our aim was to show that, in some instances, each of the 2 therapy programs could be uneffective.

Statistical analyses were performed by using StatView 4.5^a running on Macintosh computer,^b with the exception of the Mantel-Haenszel chi-square test, which was performed with Epi Info 6^c from The Centers for Disease Control and Prevention.

Back to Article Outline

Results 

The overall writing performance on the 6 RWAs are reported in table 2. Significant improvement (p < .001) was observed during the study period (T0-T5). Posthoc comparisons (Sheffé's method) indicated stable writing performance before writing therapy started (T0-T1) and during the PACE treatment period (T2-T3), whereas changes in performance were linked to the 2 therapy programs (T1-T2 and T3-T4 comparisons, p < .01, respectively). Furthermore, learning remained stable 1 month after therapy stopped (T4-T5 comparison, nonsignificant).

Table 2: Overall Writing Score at the 6 Reading and Writing Assessments

Abbreviations: NL, nonlexical therapy; L, lexical therapy; Mean writing, mean performance at the BADA writing task (240 items); Mean reading, mean performance at the BADA reading task (188 items).

Improved performance on word reading aloud task from the BADA was also found (p < .001). In the last case, however, no firm relationships could be shown with the 2 treatment programs because, with the exception of the T0-T1 comparison, increase in score was observed across all testing sessions (table 2).

On the aphasia assessment, significant improvement was found only on written naming and reading subtests (p < .01); no changes were found on oral language tasks (table 3).

Table 3: Mean Score on the Aphasia Assessment Subtests and on the 21 READ/WRITE CADL Items Before (T1) and After (T4) the Entire Therapy Period

However, changes at the CADL evaluation (READ/WRITE subtest) indicated that some improvement had occurred in daily life activities (p < .01).

At the T1 assessment, none of the patients was able to write a letter, with the exception of patient 5, whose sample (54 words) included 17 mispellings (the target word being in most cases identifiable). At T4, 7 patients produced spontaneous writing samples that contained about 100 words. Patients 4 and 5 were able to do so with fewer than 5% mispelled words. This performance was maintained 3 months after therapy stopped. Patients 2, 6, 7, and 8 produced fewer than 10% mispelled words. However, comparable performance was maintained only by patient 6 (8%) and patient 2 (5%) at 3 months posttherapy but deteriorated for the other 2 patients. In the period from T4 to the 3-month posttherapy assessment, patient 3's performance improved from 22/98 to 5/89 mispelled words. Also at the T4 assessment, patient 1 was unable to accomplish the task. After producing 6 nonpronounceable letter strings and 5 1-letter substitution errors out of 22 orthographic units, he refused to continue. His performance was no better 3 months after therapy stopped. It is noteworthy that the overall writing score at the RWA-T5 was not suitable to predict performance on the writing a letter test at 3 months posttherapy. For instance, although the T5 writing overall score was 203/240 for patient 7 and 180/240 in patient 8, they remained poor writers on functional assessment, whereas good outcome was observed in patients 2 and 3, whose overall writing scores were 113 and 140/240, respectively at the RWA-T5. However, during the writing a letter test, the first 2 patients failed to advance beyond their phonetic (patient 7) and lexical-syntactic (patients 7, 8) difficulties and did not exhibit consistent writing behavior. The latter 2 correctly applied whole word form (patient 2) and segmental phonology to orthography conversion (patient 3) strategies on simple self-dictated sentences.

At the 3-month posttherapy control patients 2 and 3 had regained partial activity: patient 2 became a salesclerk and patient 3 became a nursery school attendant. The best outcome was observed for patient 4, who regained his prestroke position as a business manager. Patient 6 was a nun who 2 months after T5 assessment returned to her mission in Kenya. She sent us letters at regular intervals describing her reduced work schedule. For the other patients, no significant gains in psychosocial adjustment were noticed with the exception of patient 1, a retired employer, who became involved in parish social activities.

Figure 1 shows changes in performance on the 3 writing tasks as a function of the 2 treatment programs.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Mean number of correct responses out of 80 items at the pre-posttherapy assessment for each type of writing therapy program. (■—■ nonword writing, ●—● word writing, ○—○ written naming.)

ANOVA showed significant pre-post variation (p < .001), significant (p < .001) task effect, and significant (p < .038) high-order (pre-post × treatment × task) interaction. Interaction between pre-post variation and type of treatment was not significant. This indicated that the 2 programs had been equally effective. However, according to our prediction, post hoc comparisons indicated that, after the lexical treatment, significant changes (p < .05) were confined to written naming and word writing on dictation tasks. After the nonlexical treatment, writing word and nonsense stimuli improved (p < .05), whereas written naming did not.

Overall writing scores at the 4 pre-posttherapy assessments were then analyzed in each individual patient. In patients 7 and 8, both therapy programs had been effective. For the other 6 patients, highly significant (p < .005) interaction between type of treatment and pre-post evaluation was found. This indicated that 1 program had produced no effect, that is, the nonlexical for patients 1, 2, and 4, or the lexical for patients 3, 5, and 6, (table 2.). It is noteworthy that the pretherapy performance did not predict which of the 2 programs would be effective and, in 3 of these 6 patients, only the second program was effective. Furthermore, the ceiling effect could not be invoked for patients 4 and 5. After the nonlexical therapy, patient 4's performance in writing nonsense stimuli only increased from 24 to 35 correct responses while, after the lexical, patient 5's performance in the written naming task remained at 32. Finally, the comparison of data from , showed no trend for the fluency dimension or education.

Such a selective response corresponded to features of patient compliance during the development of each of the 2 treatment programs. For instance, patients 1, 2, and 4 were often observed to write in nonlinear order at the posttherapy RWA, which indicated using visuolexical cues.¹⁹ During the nonlexical treatment, these patients found exercises of phonemic segmentation and analysis very difficult to perform because of the deficit in phonological processing (patients 1, 2) or the articulatory difficulties (patient 4). Patient 3, on the other hand, was almost unable to use visuolexical cues during the lexical treatment. After the nonlexical treatment, she was observed to generalize phonology to orthography segmental conversion to 1 phoneme to multiletter string correspondences and to practice phonemic segmentation of difficult consonant clusters. Patients 5 and 8 used exercises of the lexical program more as an extension of the nonlexical treatment than training with a visuolexical strategy.

During the entire therapy period, patient 7 was more concerned with his dysarthric difficulties than with writing.

Finally, individual results at the pre-posttherapy testing for oral language skills were analyzed. Improved performance on oral confrontation naming was acknowledged for patients 2 and 4, whose writing performance had increased after the lexical therapy. These 2 patients were often observed to prime their oral response by writing down the first letter of the stimulus. Improved repetition skills in patients 3, 7, and 8 corresponded to better phonological processing after the nonlexical therapy. Increased repetition score, however, was also acknowledged in patient 4.

Back to Article Outline

Discussion 

The results from our pilot study¹³, ¹⁴ had demonstrated that treatment methods devised for chronic patients with peculiar dysgraphic symptoms may be generalized to patients with chronic aphasia. Our present results extend this possibility to 6 to 12 months postonset aphasics who are still considered candidates for conventional language stimulation.¹, ², ³, ⁴, ⁵ After the 2 writing therapy programs, significant gains in performance were observed, which did not occur after comparable periods of conventional language stimulation or modified PACE therapy. Subjects included in our study also showed increased functional communication after the intensive functional treatment between the 2 writing programs.¹⁵ Increased communication skills were evident on CADL scores but were not generalized to the standard aphasia assessment. They could be analyzed only by testing methods to assess the particular response pattern.¹⁵ These methods, however, changed during development of the present study and do not allow us to show results from this cohort in a comprehensive way.

The improved writing performance had clinical relevance because reduced disability was observed on the CADL's READ/WRITE subtest and, for 5 patients (2-6), on an explicit measure of written language competence.¹, ⁴ Some authors²⁷, ²⁸ have cautioned against the significance of reading and writing abilities in aphasic subjects with respect to their psychosocial adjustment. However, 4 patients regained employment that involved reading and writing activities and in a fifth better psychosocial adjustment was noticed.

For 6 of the 8 patients, 1 treatment program did not increase writing performance. The last finding was not linked to floor effect because only the second program was effective for 3 of them and the ceiling effect could not be invoked for the other 2. Furthermore, the effectiveness of each program could not be predicted by the severity of writing disturbances at the pretest or education. The former result could indicate that patients with more severe dysgraphia had not responded to the more difficult program; the latter, that demographic variables determined the individual response to therapy. But this was not the case. Our findings, instead, indicated that type of treatment plays an important role in the recovery from dysphasic symptoms at 6 to 12 months postonset.

To avoid practice effect, the tests for assessing functional writing and learning transfer to oral language or daily activities could not be administered at the end of each writing therapy program. Thus, a question may be raised as to whether functional recovery is linked to the particular therapeutic program or to the entire stimulation period. The last hypothesis cannot be ruled out by these results. Nevertheless, each of the 2 treatment programs produced a specific pattern of improvement and, at the individual level, learning transfer to oral language skills could be predicted by the therapeutic strategy shown to be effective. Behavioral correlates of this improvement have already been described at length. These findings indicated that the effective nonlexical therapy had provided some patients with writing strategy and improved phonological skills that they could implement in functional writing. While following the lexical program, other patients could better exploit lexical knowledge.

Because of selective learning transfer to oral language, our results differed from those in studies showing therapy effects generalized to all language modalities.¹, ², ³, ⁴ Some reasons may account for the difference. It could depend on selection bias, which occurred in our experiment. Here, the proposal of the therapy focusing on writing abilities could also be justified by the therapists' (covert) opinion of poor oral language recovery at the pretherapy evaluation and evidence of relevant dysgraphic disturbances. However, pretherapy aphasia assessment disclosed no important differences between our patients and those with mild to moderate aphasia from a large Italian cohort.²⁴ A second explanation concerns the outcome measures used in the different studies. Our criterion for acknowledging improvement on aphasia assessment subtests was probably too conservative. This possibility should be addressed by further studies.

Neverthless, a parsimonious explanation points to the structure of the therapy programs. During prolonged and intensive individual language stimulation treatment, with exercises tapping oral and written language, there is in fact a considerable possibility that a number of familiar stimuli are treated in all language modalities. This treatment paradigm is consistent with the final result of learning spread over all language tasks of common aphasia assessment methods, because such tasks usually include familiar stimuli. The multimodal treatment paradigm is also consistent with unchanged performance at more detailed reading and writing assessment (see the T0-T1 RWA comparison in our study). However, it is unlikely that therapy programs focusing on a particular language domain will result in wide generalization of therapy effects, although significant improvement may occur on more exacting language assessment tasks.

The 2 therapy programs used here focused on training patients in 2 writing procedures—say writing by phonology to orthography segmental conversion rules and writing by lexical routine. These routines are assumed to be necessary for writing in normal literate adults.²², ²³ Patients with dissociated performance, ie, resulting from a selective damage to 1 of the 2 routines, have been described in an Italian cohort of mild to moderate aphasic subjects²⁴ as well as in English-speaking²², ²³ aphasics. Damage to both components of the writing system can also occur²⁴ and the pretherapy overall score less than 100 in the BADA test likely indicated that this was the case in at least 5 patients from our cohort. Contrary to the English writing system, written Italian is considered to have shallow orthography, which permits writing at an acceptable level of accuracy even when using only 1 of the 2 routines.²⁴ That was assumed to be the goal for each of the 2 treatment programs where a series of successively more complex tasks led the patient to approach functional writing. The 2 programs were thus endowed with a rationale from neurolinguistic models of language processing, ie, restoring (or reorganizing) a damaged component of the writing system. Both programs also furnished a functional view of the outcome.

The cognitive rationale of the 2 writing therapy programs was confirmed by the correspondence between type of therapy and pattern of improvement on the group results. However, although all subjects did well on the trained stimuli, in 6 of the 8 patients, generalization of the treatment effects to untrained writing stimuli (and untrained language tasks) occurred for only 1 of the 2. This result indicates that learning across therapy steps is not the only factor accounting for functional recovery at the posttest.

These findings are in agreement with results from recent studies²⁹, ³⁰, ³¹ on microcomputer-assisted rehabilitation of written confrontation naming in aphasics. This therapy program was administered to 18 aphasic subjects and focused on written naming from the keyboard of 80 objects displayed with semantic or lexical cues on the screen. The group analysis at first showed significant learning and generalization of benefits to untrained items and to untrained oral naming.²⁹ However, subsequent multiple single-case analysis indicated various patterns of improvement and of generalization.³¹ Despite actual learning during treatment, for some patients there were no benefits on untreated items whereas generalization to untreated items and/or generalization to untrained oral modality occurred in others. These findings, together with those from this study, suggest that obtaining consistent functional recovery (eg, writing a letter) with focused therapeutic programs is constrained by interaction between the improved functional behavior and the entire pattern of language deficits the individual patient shows.

However, the lack of increased performance in writing nonsense stimuli after the nonlexical therapy for patients 1, 2, and 4, as well as the unimproved written naming for patients 3, 5, and 6 following the lexical therapy, also indicated that restoration of damaged phonology to orthography segmental conversion routine or damaged lexical writing routine had not occurred. Thus, even highly focused (and theoretically oriented) therapy programs do not guarantee success; thus other variables should be taken into account for predicting outcome at the individual level.

Among these, natural mechanisms underlying functional recovery (restitution of neural processing or developing compensatory strategies³²) likely play an important role. Restitution of neural processing, for instance, may be obtained by repeated exposure to a treatment task following the traditional stimulus-response paradigm. This approach may result in learning across therapy sessions, but generalization of treatment effect to untrained stimuli might not occur.

Compensatory strategies, on the other hand, might develop after targeting appropriate behaviors, in particular functional domains, and account for learning transfer to untrained stimuli and untrained language tasks. However, reorganizing cognitive processes to obtain functional behavior(s) is constrained by the relative integrity of cognitive processes to which the therapeutic strategy proposes to have recourse.³³ Thus, a detailed (perhaps cognitively oriented) evaluation of the pretherapy pattern of dysfunction represents a further crucial variable for predicting the effectiveness of the therapy program. It should be remembered, for instance, that, in the previously mentioned trial on written naming rehabilitation, in 2 patients' patterns of improvement could be predicted on the basis of their writing errors at the pretest.³⁰

With reference to the last point, this study cannot provide relevant insight. First, most of our patients showed important damage to both writing routines. Furthermore, they often showed additional language or memory deficits. This aspect kept us from placing them into the diagnostic categories of dysgraphic impairment, which are usually used by cognitively oriented therapists to study relationships between treatment program and pretherapy pattern of dysfunction.⁹, ¹⁰ It also suggests that actual model-based assessment methods, such as the BADA¹⁶ and, perhaps, PALPA²⁵ tests, are not suitable to predict effective therapy and outcome in the more impaired subjects.

Back to Article Outline

Conclusion 

Our observation, that a considerable proportion of patients with aphasia may not improve with even highly structured therapy programs, poses some problems about the best therapeutic option for treating moderate aphasic disturbances 6 to 12 months after onset. Because of repeated demonstration of significant effects by conventional language stimulation,¹, ², ³, ⁴, ⁵ one might still rely on the personal expertise of speech therapists in doing intensive language stimulation and disregard particular therapeutic programs in the current aphasia therapy literature. This usually occurs in clinical practice. Indeed, specific therapy programs such as Melodic Intonation Therapy,³⁴ Syntax Stimulation Therapy,³⁵ functional treatment,¹⁸ or PACE³⁶ have been devised for particular groups of patients. Most of them have been effective in pilot studies, but they are usually considered extrema ratio for chronic aphasic symptoms.

There are, however, 2 findings from our experiment that argue against the view of empirically driven language stimulation as the best option. First, on 6 occasions, the 2 therapy programs resulted in no significant gains, although during the treatment patients showed sufficient learning on practice items. This suggests that, during the development of the therapeutic project, the therapist's clinical impression may be at times misleading. Second, the choice about which of the 2 treatment programs had to be administered first was made for each patient by his/her own therapist. However, in 3 of the 8 subjects only the second therapy program was effective. This suggests that the choice of therapeutic strategy based on the therapist's clinical impression does not guarantee the better choice for a given patient. Even the fluent-nonfluent dimension, widely used in the clinical definition of aphasic symptoms, could not predict the effectiveness of the 2 therapy programs in individual patients or the choice of the first program. The last appeared to indicate that in 1 case (ie, patient 3) therapists assumed that severe deficit in phonologic processing had to be addressed by direct training while, in another case (ie, patient 5) it could be compensated by improved lexical processing.

We thus concur with the conclusions of other studies⁸, ¹¹ that important questions concerning the effects of therapy and choice of therapeutic method in the single patient can be solved only by a large number of studies where individual patients are trained by specific therapy programs. In these studies, testing methods should not only include an assessment about whether and how functional recovery has occurred, but also identify cognitive resources that are a prerequisite for developing the therapeutic project. Therapy programs should also be detailed with reference not only to the functional behavior for which the program is addressed but also to the cognitive component(s) on which the therapeutic strategy focuses and the neurologic rationale (restoration of neural processing or developing compensatory strategies) on which the therapeutic strategy depends. Finally, these studies should not be confined to patients with mild chronic aphasia; patients with less stable or more severe symptoms should also be included.

Back to Article Outline

Acknowledgements 

We are indebted to Serena Emanuelli, ST, and Paola Casadio, ST (IRCCS Santa Lucia, Roma); Felicita Romano, ST, and Maria Tedeschi, ST (Speech Therapy Unit, Policlinico Federico II, Napoli); and Francesca Faccioli, ST (Neuropsychology Laboratory, II Università di Napoli) for technical assistance.

Back to Article Outline

References 

1. Basso A, Capitani E, Vignolo LA. Influence of rehabilitation on language skills in aphasic patients: a controlled study. Arch Neurol. 1979;36:190–196
   • View In Article
   • MEDLINE
2. Wertz RT, Collins MJ, Weiss DG, Kurtze JF, Friden T, Brookshire RH, et al.  Veterans Administration cooperative study of aphasia: a comparison of individual and group treatment. J Speech Hear Res. 1981;24:580–594
   • View In Article
   • MEDLINE
3. Shewan CM, Kertesz A. Effects of speech and language treatment on recovery from aphasia. Brian Lang. 1984;23:272–299
   • View In Article
4. Mazzoni M, Vista M, Geri E, Avila L, Bianchi F, Moretti P. Comparison of language recovery in rehabilitated and matched, non-rehabilitated aphasic patients. Aphasiology. 1995;9:553–563
   • View In Article
5. Wertz RT, Weiss DG, Aten JL, Brookshire RH, Garcia-Bunuel L, Holland AL, et al.  Comparison of clinic, home, and deferred language treatment for aphasia. A Veterans Administration cooperative study. Arch Neurol. 1986;43:653–658
   • View In Article
   • MEDLINE
6. Lincoln NB, McGuirk E, Mulley GP, Lendrem W, Jones AC, Mitchell JR. Effectiveness of speech therapy for aphasic stroke patients. A randomised controlled trial. Lancet. 1984;1:1197–1200
   • View In Article
   • MEDLINE
7. David R, Enderby P, Bainton D. Treatment of acquired aphasia: speech therapist and volunteers compared. J Neurol Neurosurg Psychiatry. 1982;45:957–961
   • View In Article
   • MEDLINE
   • CrossRef
8. Prins RS, Schoonen R, Vermeulen J. Efficacy of two different types of speech therapy for aphasic stroke patients. Appl Psychol. 1989;10:85–123
   • View In Article
9. In:  Seron X,  Deloche G editor. Cognitive approaches in neuropsychological rehabilitation. Hillsdale (NJ): : Lawrence Erlbaum; 1989;
   • View In Article
10. In:  Riddoch MJ,  Humphreys GW editor. Cognitive neuropsychology and cognitive rehabilitation. Hove (UK): : Lawrence Erlbaum; 1994;
    • View In Article
11. Linebaugh CW, Baron CR, Corcoran KJ. Assessing treatment efficacy in acute aphasia: paradoxes, presumptions, problems and principles. Aphasiology. 1998;12:519–536
    • View In Article
12. Basso A. Spontaneous recovery and language rehabilitation. In:  Seron X,  Deloche G editor. Cognitive approches in neuropsychological rehabilitation. Hillsdale (NJ): : Lawrence Erlbaum; 1989;p. 17–38
    • View In Article
13. Carlomagno S, Colombo A, Casadio P, Emanuelli S, Razzano C. Cognitive approaches writing rehabilitation in aphasics: evaluation of two treatment strategies. Aphasiology. 1991;5:355–360
    • View In Article
14. Carlomagno S, Iavarone A, Colombo A. Cognitive approaches to writing rehabilitation: from single case to group studies. In:  Riddoch MJ,  Humphreys GW editor. Cognitive neuropsychology and cognitive rehabilitation. Hove (UK): : Lawrence Erlbaum; 1994;p. 485–502
    • View In Article
15. Carlomagno S, Losanno N, Emanuelli S, Casadio P. Expressive language recovery or improved communicative skills: effects of PACE therapy on aphasics' referential communication and story retelling. Aphasiology. 1991;5:419–424
    • View In Article
16. Miceli G, Burani C, Laudanna A. Batteria per l'analisi dei deficit afasici. Milano: : ASRN; 1991;
    • View In Article
17. Basso A, Capitani E, Zanobio ME. Pattern of recovery of oral and written expression and comprehension in aphasic patients. Behav Brain Res. 1982;6:115–128
    • View In Article
    • MEDLINE
    • CrossRef
18. Aten JL, Caligiuri MP, Holland AL. The efficacy of functional communication therapy for chronic aphasic patients. J Speech Hear Dis. 1982;47:93–96
    • View In Article
19. Hatfield FM, Weddell R. Re-training in writing in severe aphasia. In:  Hoops R,  Lebrun Y editor. Recovery in aphasics. Amsterdam: : Swets & Zeitlinger; 1976;p. 65–78
    • View In Article
20. Pizzamiglio L, Mammucari A. Evidence for sex differences in brain organization in recovery in aphasia. Brain Lang. 1985;25:213–233
    • View In Article
    • MEDLINE
    • CrossRef
21. Holland AL. Communicative abilities in daily living. Baltimore: : University Park Pr; 1980;
    • View In Article
22. Ellis AW. Reading, writing and dyslexia: a cognitive analysis. Hillsdale (NJ): : Lawrence Erlbaum; 1984;
    • View In Article
23. Roeltgen DP, Heilman KM. Lexical agraphia, further support for the two-strategy hypothesis of linguistic agraphia. Brain. 1984;107:811–827
    • View In Article
24. Luzzatti C, Laiacona M, Allamano N, De Tanti A, Inzaghi MG. Writing disorders in Italian aphasic patients: a multiple single-case study of disgraphia in a language with shallow orthography. Brain. 1998;121:1721–1734
    • View In Article
    • CrossRef
25. Kay J, Lesser R, Coltheart M. Psycholinguistic assessment of language processing in aphasia (PALPA). Hove (UK): : Lawrence Erlbaum; 1992;
    • View In Article
26. Norman GR, Streiner DL. Biostatistics: the bare essential. St Louis (MO): : Mosby-Year Book; 1994;
    • View In Article
27. Parr S. Everyday reading and writing practices normal adults: implications for aphasia assessment. Aphasiology. 1992;6:273–283
    • View In Article
28. Holland AL. Pragmatic aspects of intervention in aphasia. J Neurol. 1991;6:197–211
    • View In Article
29. Deloche G, Ferrand I, Metz-Lutz MN, Dordain M, Kremin H, Hannequin D, et al.  Confrontation naming rehabilitation in aphasics: a computerized written technique. Neuropsychol Rehabil. 1992;2:117–124
    • View In Article
    • CrossRef
30. Deloche G, Hannequin D, Dordain M, Metz-Lutz MN, Kremin H, Tessier C, et al.  Diversity of patterns of improvement in confrontation naming rehabilitation: some tentative hypotheses. J Commun Dis. 1997;30:11–22
    • View In Article
31. Deloche G, Dordain M, Kremins H. Rehabilitation of confrontation naming in aphasia: relation between oral and written modalities. Aphasiology. 1993;7:201–216
    • View In Article
32. Singer W. Recovery mechanisms in the mammalian brain. In:  Nicholls JG editors. Repair and regeneration of the nervous system. Berlin: : Springer Verlag; 1982;p. 96–104
    • View In Article
33. Seron X. Reeducation strategies in neuropsychology: cognitive and pragmatic approaches. In:  Rose FC editors. Advances in neurology: progress in aphasiology. Vol 42:New York: : Raven Pr; 1984;p. 317–325
    • View In Article
34. Sparks R, Helm N, Albert M. Aphasia rehabilitation resulting from melodic intonation therapy. Cortex. 1974;10:303–316
    • View In Article
    • MEDLINE
35. Helm-Estabrooks N, Fitzpatrick PM, Barresi B. Response of an agrammatic patient to a syntax stimulation program for aphasia. J Speech Hear Dis. 1981;46:422–427
    • View In Article
36. In:  Davis G,  Wilcox M editor. Adult aphasia rehabilitation: applied pragmatics. Windsor (UK): : NFER-Nelson; 1985;
    • View In Article