Associations of hypothetical early intensive in-hospital rehabilitation with activities of daily living after hip fracture surgery in patients with and without dementia: emulating a randomized controlled trial using medical claims data

Takaaki Ikeda,Upul Cooray,Ryutaro Matsugaki,Yuta Suzuki,Michiaki Takagi,Keiji Muramatsu,Kiyohide Fushimi,Masayasu Murakami,Ken Osaka,Shinya Matsuda

DOI: https://doi.org/10.1016/j.jclinepi.2024.111550

2024-10-05

Abstract:Objectives: To investigate the impact of early intensive in-hospital rehabilitation, initiated within 2 days of surgery and lasting up to 7 days, on the recovery of activities of daily living in patients with and without dementia. Study design and setting: Medical claims data from 925 hospitals in Japan were analyzed. We enrolled patients aged ≥50 years who underwent hip fracture surgery within 2 days of admission between April 1, 2018, and December 31, 2019. Low- (20 minutes per day starting on day 2), highest- (60 minutes per day starting on day 1), and gradually increasing (20 minutes on day 1, 40 minutes on days 2-4, and 60 minutes per day thereafter) intensity regimens were used as exposures. The outcomes were Barthel Index (BI) scores at 14 and 30 days postoperatively. For per-protocol analysis, a target trial emulation framework with the sequential doubly robust estimator was used. Results: Among patients without dementia (N = 11,461), no significant differences in BI scores were observed at 14 days postoperatively across regimens. At 30 days postoperatively, significant differences in BI scores were noted between highest- and low-intensity regimens and between gradually increasing intensity and low-intensity regimens, with additive BI scores of 15.2 (95% CI, 10.7-19.7) and 14.7 (95% CI, 9.2-20.2), respectively. In patients with dementia (N = 14,302), significant differences in BI scores were noted at 14 days postoperatively between highest- and low-intensity regimens and between gradually increasing intensity and low-intensity regimens, with additive BI scores of 8.7 (95% CI, 5.2-12.2) and 10.7 (95% CI, 5.8-15.6), respectively. At 30 days postoperatively, a significant difference in BI scores was observed between gradually increasing intensity and low-intensity regimens, with an additive BI score of 17.9 (95% CI, 11.3-24.5). Conclusion: Early intensive in-hospital rehabilitation is highly relevant and beneficial for dementia patients.