Benchmarks of Significant Change After Aphasia Rehabilitation.

Most included within-group studies were lower-quality designs: 73% were level III, 26% were IIB, and 1% were IIA. Publication bias was detected for the within-group WAB-AQ and BNT analyses, although the adjusted point estimates shifted by less than 1 point. Between-group evidence was limited, with only 8 studies overall and only 5 studies for the BNT comparison. The analysis could not compare acute and chronic recovery stages because nearly all within-group studies were in the chronic phase. Only the WAB-AQ, CETI and BNT were analysed; less-used outcome measures, including contextual factors and quality of life, were excluded to keep enough studies for meta-analysis. Subgroup analyses were not possible for every measure or design because some groups had fewer than 5 studies. The WAB-AQ summary change was close to its stated SEM of 5, and the authors note measurement error varies by aphasia severity, so the benchmark may not fit mild or severe cases.

No author conflicts statement is included in the supplied text. The paper is tagged as supported by NIH extramural research.

Do not read the 5.03-point WAB-AQ, 10.37-point CETI and 3.30-point BNT changes as the minimum gain needed for one patient to improve. They are summary changes across studies, many without a control group, and publication bias was detected for two outcome measures.