As conflict continues over expanding medical school admissions, the Physician Workforce Supply and Demand Estimation Committee released item-by-item rebuttal materials against the Korean Medical Association Organization's claim that the physician demand forecast was distorted. As the government moved to increase medical school quotas based on the committee's results and the medical community pushed back, calling the figures an inflated "demand," the committee issued a direct explanation.
On the 13th, the committee said in an explanatory document that "this forecast is the best possible result derived from currently available data and mutually acceptable assumptions," and "the figures were not calculated on the premise of any specific policy conclusion."
The Korean Medical Association Organization has argued that demand was overestimated by including health care utilization data from 2020–2024, when the COVID-19 pandemic overlapped with government–medical community conflict. In response, the committee said it used all data from the period without arbitrarily excluding it. It explained that removing data from a specific period excessively erases recent changes in health care utilization, resulting instead in a steeper increase in medical demand.
It also rebutted criticism of the ARIMA (autoregressive integrated moving average) model used to predict health care utilization. The Korean Medical Association Organization pointed out that ARIMA mechanically extends a past upward trend into the future, but the committee said, "information from recent observations is directly reflected, and the influence of past information decreases over time," and "it is not a method that fixes and stretches a surge pattern from a particular period as is." It added that ARIMA is a time-series forecasting technique widely used in various fields, including health care.
Extending the analysis period back to 2000 also drew controversy. The Korean Medical Association Organization argued that the medical environment in the early 2000s differs from today and that only recent data should be used. However, the committee explained that in time-series analysis, uncertainty increases sharply if the forecast horizon becomes excessively long relative to the sample length. Using only recent data carries a high risk that special situations like COVID-19 will distort long-term trends.
It also addressed the method of estimating physician workload using medical billing data. The Korean Medical Association Organization has argued that medical bills, which include high-priced tests and equipment expense, cannot accurately reflect physician workload. The committee said it agrees that full-time equivalent (FTE) is ideal, but there are no official statistics or administrative data to calculate it consistently at the national level. Applying FTE with limited data could increase uncertainty in the forecast, so at this time it used medical bills as a proxy indicator with the highest comparability and objectivity.
Regarding the claim that it underestimated productivity gains for physicians from AI adoption, it countered that "overstating the impact is even less realistic." The committee said it judged that time saved by AI would not immediately translate into the same level of increased patient care and applied a composite scenario that considered both productivity gains and shorter working hours. It cited areas where physicians' judgment and communication are essential—such as patient counseling and explanation and multidisciplinary consultations—as difficult to replace.
There was also criticism of selective application because scenarios were applied only to some demand estimates. The committee explained that the cohort component method is a static model that assumes the base-year level of health care utilization remains unchanged, so applying additional scenarios is not methodologically appropriate by design. In contrast, time-series models like ARIMA are dynamic models in which future values are generated based on past trends and volatility structures, making it reasonable to incorporate scenarios separately.
The committee stressed that this supply and demand forecast was produced by an expert committee formed under the Framework Act on Health and Medical Services after 12 meetings. It said all minutes and materials have been made public and that advancing the forecasting methodology and building data have been set as ongoing tasks to be pursued in the regular five-year forecasting cycle.
Committee Chair Kim Tae-hyeon said, "Mid- to long-term physician supply and demand forecasting is inherently subject to uncertainty," and "this is the best possible result that can be derived under realistic constraints."