PRM189 - SUBJECTIVE MEASURE MAY BE MORE SENSITIVE IN IDENTIFYING PROGNOSTIC FACTORS OF FUNCTIONAL OUTCOME AMONG OLDER ADULTS WITH SURGICALLY-TREATED HIP FRACTURE – FINDINGS FROM A SYSTEMATIC REVIEW

Abstract

Prognostic studies identify predictors of disease outcomes to advise targeted interventions. This study examined the range of functional outcome measures used in prognostic studies among older adults with surgically-treated hip fracture and whether the choice of outcome measures (subjective vs objective) affect the findings of the prognostic factors. We performed a systematic search of four databases (MEDLINE, Embase, CINAHL, PsycINFO) for longitudinal studies examining prognostic factors of functional outcome at ≥1 month among older adults ≥50 years old with surgically-treated hip fracture. Two reviewers independently screened papers for inclusion and assessed their quality using the Quality in Prognosis Studies (QUIPS) Tool. We extracted the study characteristics, outcome measures, prognostic factors and p-values for each prognostic factor. We categorized the p-values into two (p≤0.05 or p>0.05). To examine whether the choice of outcome measures was associated with achieving statistical significance (p≤0.05) in examining prognostic factors, we performed logistic mixed model on R 3.4.4 (lme4 package) using study identifier as a random effect, adjusting for potential confounders. From 98 papers that met our inclusion criteria, we identified 144 outcome measures. Subjective measures (patient- or proxy-reported outcomes e.g. SF-36) dominated the literature (123/144), followed by objective measures e.g. walking time / speed (19/144) and composites of subjective and objective measures (2/144). We excluded the composite measures from analysis. Adjusting for type of prognostic factors, publication year, study quality, study design and sample sizes, subjective measures were associated with higher odds (OR 3.44; 95%CI 1.18-10.03) of achieving statistical significance in examining prognostic factors. Our findings suggest that subjective measures may be more sensitive in identifying prognostic factors compared to objective measures of functional outcome among older adults with surgically-treated hip fracture. This has implications on the choice of outcome measures and interpretation of findings in future prognostic studies.