Influencing factors of hospitalization costs for patients with subsequent osteoporotic hip fractures: multilevel model analysis by fracture episode.

BACKGROUND: Osteoporotic hip fractures often require immediate inpatient care and impose a substantial financial burden. Previous studies mainly focused on overall costs for patients with refractures, without disaggregating hospitalization costs by episode, leaving the determinants of first- and second-admission costs unclear in this subgroup. This study aimed to examine factors associated with hospitalization costs for each episode and to identify factors characterizing cost differences between the two hip-fracture hospitalizations in patients with subsequent osteoporotic hip fractures. METHODS: Data were obtained from the Shanghai Health Statistics Center, covering all medical institutions in Shanghai, China, from January 1, 2013, to December 31, 2022. Multilevel models were used to identify factors related to hospitalization costs for the first and second fractures. Hospitalization variations and cost differences between the two fracture episodes were examined, and multivariable quantile regression was used to explore factors associated with cost differences. RESULTS: Of 43,963 patients hospitalized for their first osteoporotic hip fracture, 1,152 experienced a subsequent hip fracture and required rehospitalization during the study period. At the first fracture, most patients were female (78.73%) and aged ≥80 years (57.47%). Grade-A tertiary hospitals represented 27.59% (n = 24) and 24.04% (n = 25) of treating hospitals for the first and second fractures, respectively, but admitted 55.64% (n = 641) and 51.39% (n = 592) of patients. For both episodes, fourth-degree surgery had the highest median costs ($9,787.15 and $9,726.55, respectively), whereas unoperated patients had the lowest ($1,147.75 and $1,133.36). Costs rose substantially with longer length of stay (LOS), reaching $11,197.72 and $11,729.95 for stays of ≥31 days. Null multilevel models of hospitalization costs revealed ICCs of 0.333 and 0.338 for the first and second fractures, respectively, both indicating significant hospital-level variance (all P < 0.001). Random-intercept models showed that grade-A tertiary hospitals showed higher adjusted costs than grade-B secondary hospitals for both fractures, and that higher surgery levels and longer LOS were positively associated with costs across both episodes (all P < 0.001). For the second fracture, an interval of 1 to 1.5 years between hospitalizations was associated with significantly lower costs compared with an interval of ≤1 year (exp(β) = 0.899; 95%CI, 0.824 to 0.980; P < 0.05). Quantile regression showed that a higher hospital level and longer LOS during the second hospitalization were associated with increased relative costs across quantiles (all P < 0.05). Conversely, a higher surgery level in the second hospitalization was associated with reduced relative costs at the 25th and 50th percentiles (all P < 0.05), indicating heterogeneous associations across different quantiles of the cost distribution. CONCLUSIONS: Among patients with subsequent osteoporotic hip fractures, grade-A tertiary hospitals were associated with higher adjusted costs than grade-B secondary hospitals for both fracture episodes, and inpatient costs increased progressively with higher surgery levels and longer LOS. Between-episode variations in hospital level, surgery level, and LOS exhibited heterogeneous effects on cost differences across quantiles. These findings highlight the need for episode-specific cost-control strategies that also account for variations between the two hospitalizations.