International health spending forecasts: concepts and evaluation

Area of Science:

• Health economics
• Econometrics
• Forecasting methodologies

Background:

• National health care expenditures are influenced by historical organizational and financial decisions.
• These decisions are shaped by varying macroeconomic conditions over time.
• Current health care spending is a function of Gross Domestic Product (GDP) growth and inflation rates.

Purpose of the Study:

• To compare the accuracy of causal econometric models with time-series methods for forecasting health care expenditures.
• To evaluate the effectiveness of different forecasting approaches using data from OECD countries.
• To determine if combining forecasts enhances predictive accuracy.

Main Methods:

• Utilized causal econometric models to forecast health care expenditures.
• Compared econometric models against exponential smoothing, moving average, and ARIMA methods.
• Employed data from 19 OECD countries (1965-1979) for model calibration and ex ante forecasts (1980-1987).

Main Results:

• The aggregate econometric model across 19 OECD countries yielded the greatest reduction in Mean Absolute Error (MAE).
• Single-country econometric models, exponential smoothing, and international averaging also demonstrated effectiveness.
• A combination of all four forecasting methods reduced MAE by 25% compared to a constant growth projection.

Conclusions:

• Econometric models provide a robust framework for forecasting national health care expenditures.
• Combining diverse forecasting techniques offers superior accuracy over individual methods.
• Accurate forecasting of health care spending is crucial for effective health system management and financial planning.