2026-06-09 京都大学
操作変数法による解析の結果、入院が受診後30日および90日時点の死亡率を改善する、もしくは悪化させるという明確な証拠は得られなかった。一方、入院によって医療費の増加が認められた。エラーバーは95%信頼区間を示す。作成：池洲諒
＜関連情報＞
- https://www.kyoto-u.ac.jp/ja/research-news/2026-06-09
- https://www.acpjournals.org/doi/10.7326/ANNALS-25-03725
認知症患者における入院が医療成果および医療費に及ぼす影響の推定：準実験的研究 Estimating the Effect of Hospital Admission on Health Care Outcomes and Spending Among Persons With Dementia: A Quasi-experimental Study
Ryo Ikesu, MD, PhD, Elizabeth Rose Mayeda, PhD, MPH, Kathleen McGarry, PhD, Roch A. Nianogo, MD, PhD, Christina M. Ramirez, PhD, and Yusuke Tsugawa, MD, PhD
Annals of Internal Medicine Published:9 June 2026
DOI：10.7326/ANNALS-25-03725
Abstract
Background:
Because transitions in care settings can be confusing for persons with dementia (PWD), hospital admissions can negatively affect their physical and cognitive function. However, the effect of hospital admissions on patient outcomes and health care spending remains largely unknown.
Objective:
To estimate the effects of hospital admissions on health outcomes and health care spending among PWD.
Design:
Quasi-experimental instrumental variable method, using emergency physicians’ admission propensity as the instrument.
Setting:
United States.
Patients:
Medicare fee-for-service beneficiaries aged 66 years or older with dementia who visited an emergency department (ED) in 2017 to 2019.
Measurements:
Death, inpatient days (excluding the index hospital admission), and health care spending (excluding the index ED visit and hospital admission) within 30 and 90 days of ED visits.
Results:
Among 872 085 ED visits (62.9% women; mean age, 83.1 years) included in the analysis, 482 208 (55.3%) resulted in hospital admission. There was no evidence that hospital admission was associated with 30-day mortality rate (adjusted risk difference, −2.6 percentage points [pp] [95% CI, −5.2 to 0.1 pp]) or inpatient days (adjusted difference, 0.1 days [CI, −0.2 to 0.5 days]). Hospital admission was associated with higher 30-day health care spending (adjusted difference, $2547 [CI, $1390 to $3703]). Patterns were similar for 90-day outcomes.
Limitation:
Residual confounding.
Conclusion:
Among Medicare beneficiaries with dementia who visited an ED, there was no evidence that hospital admission was associated with mortality rates. Under conventional statistical criteria, an effect of hospital admissions between a 5.2-pp decrease and a 0.1-pp increase in 30-day mortality rates was highly compatible with the data. On the contrary, hospital admission was associated with higher health care spending.
Primary Funding Source:
None.
