Statistical forecasting in a hospital clinical laboratory.

Area of Science:

• Laboratory operations management
• Healthcare analytics
• Time series forecasting

Background:

• Accurate forecasting of laboratory test volumes is crucial for effective financial planning and resource allocation.
• Existing forecasting methods may not adequately address the complexities of laboratory test data.

Purpose of the Study:

• To identify the optimal methodology for forecasting monthly laboratory test counts for the upcoming fiscal year.
• To support reimbursement, income, and laboratory operations management decisions.

Main Methods:

• Application of three distinct forecasting methodologies to historical monthly laboratory test count data.
• Utilizing Box-Jenkins Autoregressive Integrated Moving Average (ARIMA) models.
• Comparing forecasts generated from aggregated test packages versus individually forecasted inpatient and outpatient test counts.

Main Results:

• Box-Jenkins ARIMA models demonstrated superior performance across all tested scenarios.
• Forecasting individual test counts separately for inpatients and outpatients, then aggregating, significantly improved forecast accuracy.
• With two years of data, the annual forecast error achieved was approximately 4.5%.

Conclusions:

• Box-Jenkins ARIMA modeling is the recommended approach for forecasting laboratory test volumes.
• A segmented forecasting approach (inpatient/outpatient) enhances predictive accuracy for operational and financial planning.
• The established methodology provides reliable forecasts with a low annual error rate.