High-performance work systems (HPWS) are intended to benefit healthcare employees and service performance, yet paired evidence of such gains is limited. This systematic review distinguished distal employee or retention indicators from proximal attitudes and behaviours and classified strict mutual gains, pathway evidence, one-sided findings, and conditional or null outcomes. Scopus, PubMed/MEDLINE and ProQuest were searched from inception through 19 September 2026, with reporting informed by PRISMA 2020 and SWiM. Methodological appraisal was reported using MMAT 2018, with narrative sensitivity comparisons based on study design and outcome source. Among 476 record occurrences, 218 unique records were screened, 54 full texts assessed and 51 studies included. Three (5.9%) met the conservative paired-endpoint definition, 17 (33.3%) provided pathway evidence, 15 (29.4%) reported performance or service gains only, ten (19.6%) reported employee or retention gains only, five (9.8%) yielded conflicting or conditional results, and one (2.0%) reported a temporally stable null relationship. Wheeler et al. (2012) was classified as employee/retention evidence because its workaround measure captured employee behaviour rather than a distinct service outcome. Forty-nine studies (96.1%) were non-temporal, and 39 relied predominantly on self-report. An exploratory MMAT-based subgroup contained one strict case among ten originally labelled as having no critical concern, but those appraisal assignments still require source-level verification. The eligible literature does not establish general causal mutual gains, while the outcome rule limits conclusions about employee harm. Future studies should pair distal employee and patient or service outcomes in prospective, multi-source designs.