Hospital at Home Outcomes and Cost Analysis from Propensity-Matched Study

Michael Shaw's research dataset contains results from a retrospective cohort study of 2,972 Hospital at Home episodes, with 1,488 episodes matched 1:1 to inpatient controls. It evaluates clinical outcomes, safety, and cost-effectiveness of a UK-based service from December 2021 to May 2024. The analysis includes primary outcomes like length of stay and bed-day costs, plus secondary metrics for readmission, mortality, and patient experience.

Use Cases

• Model the relationship between care setting (HAH vs inpatient) and length_of_stay using propensity score matching on patient characteristics.
• Predict 30_day_readmission rates based on care pathway (e.g., heart function, airway disease) and derived bed_day_savings.
• Analyze the impact of HAH on 90_day_mortality odds ratios using the matched control cohort data.
• Estimate total_bed_day_costs and net savings from 13,119 saved bed days over a 33-month period.

Strengths

• Includes 2,972 HAH episodes with 1,488 propensity-matched pairs for robust comparison.
• Reports statistically significant results (p < 1e-29 for length of stay) with confidence intervals for key outcomes.
• Covers multiple clinical pathways (heart function, airway disease, acute respiratory infection) and 33 months of service data.

Limitations

• Dataset is small (106.9 KB), likely containing summary results rather than patient-level records, limiting granular analysis.
• Geographic bias is present as data originates from a single HAH service in England.
• Temporal coverage ends in May 2024, and the data reflects a specific service model that may not generalize.

Provenance

Source

Michael Shaw via figshare.

Collection Method

Retrospective cohort study using propensity score matching and regression adjustment on hospital administrative data.

Time Range

December 2021 to May 2024.

Freshness

Last updated in April 2026, with underlying data through May 2024.

Geography

England, United Kingdom.