Three statistical models for estimating length of stay

Area of Science:

• Biostatistics
• Health Services Research

Background:

• Estimating length of stay is crucial for healthcare resource allocation.
• Existing methods may rely on distributional assumptions that limit applicability.

Purpose of the Study:

• To derive probability density functions for institutional length of stay data.
• To calculate unbiased estimates of expected length of stay using these functions.
• To compare methods, including a distribution-independent approach.

Main Methods:

• Derivation of probability density functions for three data collection methods.
• Calculation of unbiased estimates of expected length of stay.
• Application of a distribution-independent method to real-world healthcare data.

Main Results:

• Identified probability density functions for different length of stay data collection methods.
• Developed unbiased estimation techniques for expected length of stay.
• Demonstrated the utility of a distribution-independent method.

Conclusions:

• The distribution-independent method offers a robust approach to estimating length of stay.
• Accurate length of stay estimation is vital for healthcare management.
• Findings are applicable to skilled nursing and intermediate care facilities.