Risk factors for hospital re-presentation among older adults following fragility fractures: a systematic review and meta-analysis

BACKGROUND Older adults hospitalized with fragility fractures are at high risk of negative events that can culminate in re-presentations to hospital emergency departments or readmissions to hospital. This systematic review aimed to identify patient, clinical, or hospital-related factors that are identifiable at the index admission and that may be associated with re-presentations to hospital emergency departments or hospital readmissions in older adults following fragility fractures. METHODS Four electronic databases (PubMed, CINAHL, Embase, and Scopus) were searched. A suite of search terms identified peer-reviewed English-language articles that examined potential correlates of hospital re-presentation in older adults (mean age ≥ 65 years) who were discharged from hospital following treatment for fragility fractures. A three-stage screening process (titles, abstracts, full text) was conducted by two researchers independently. Participant characteristics, study design, potential correlates examined, analyses, and findings were extracted for studies included in the review. Quality and risk of bias were assessed with the Effective Public Health Practice Project Quality Assessment Tool. The strength of evidence was incorporated into a best evidence synthesis, and meta-analysis was conducted where effect pooling was possible. RESULTS Eleven of 35 eligible studies were categorized as high quality studies. These studies reported that age, higher Cumulative Illness Rating scores, American Society of Anesthesiologists scores > 3, longer length of stay, male sex, cardiovascular disease, low post-operative hemoglobin, kidney disease, dementia and cancer were factors identified at the index admission that were predictive of subsequent re-presentation to hospital. Age was the only predictor for which pooling of effects across studies was possible: pooling was conducted for re-presentation ≤ 30 days (pooled OR, 1.27; 95 % CI, 1.14-1.43) and > 30 days (pooled OR, 1.23; 95 % CI, 1.01-1.50). CONCLUSIONS The best-evidence synthesis, in addition to the meta-analysis, identified a range of factors that may have utility in guiding clinical practice and policy guidelines for targeted interventions to reduce the need for re-presentation to hospital among this frail clinical population. The paucity of studies investigating re-presentations to hospital emergency departments without admission was an important gap in the literature identified in this review. Key limitations were exclusion of non-English language studies and grey literature. SYSTEMATIC REVIEW REGISTRATION PROSPERO CRD42015019379 .