Rehabilitation Interventions in Head and Neck Cancer: A Review

Abstract

<h3>Objective(s)</h3> To investigate the evidence and quality of rehabilitation interventions on functional outcomes in head and neck cancer (HNC). <h3>Data Sources</h3> PubMed, EMBASE, and CINAHL electronic databases were searched for "head and neck neoplasms" and "rehabilitation" limited to English language and human studies from 5/1/2017 - 7/22/2020. Additional interventional studies from 1/1/1990 - 4/30/2017 were sourced from our previously published scoping review (Parke et al. 2019). <h3>Study Selection</h3> Included studies were interventional trials with majority HNC population and rehabilitation-specific interventions and outcomes. Exclusion criteria included surgical, dental, economic, or smoking interventions. Studies were screened for inclusion by two independent authors. Sixty-nine of 1176 studies retrieved met inclusion criteria. <h3>Data Extraction</h3> Pairs of authors independently extracted data and evaluated study quality with the PEDro tool. <h3>Data Synthesis</h3> Studies were organized by intervention category, population, setting, timing, and study quality. The largest intervention categories included jaw exercise (13 studies), swallow exercise (11), and whole body exercise (13). Most interventions occurred in the outpatient setting for oropharyngeal cancer. Timing was predominantly post-treatment in the restorative Dietz category (51). Median study sample size was 50 (range 15 to 399). Median study quality was 5 with no studies of excellent quality. Major limitations included lack of therapist blinding (97%), lack of concealed allocation (64%), lack of intention to treat analysis (59%), and high participant drop out (58%). <h3>Conclusions</h3> Most interventional trials in HNC rehabilitation are focused on restorative exercise for the jaw, swallow, or the whole body. Most studies have substantial methodological weaknesses. Higher quality interventional studies in HNC rehabilitation are needed. <h3>Author(s) Disclosures</h3> None.