PDF(Pubmed)
Abstract：
BACKGROUND: Despite the significance of demand forecasting accuracy for the registered nurse (RN) workforce, few studies have evaluated past forecasts.
OBJECTIVE: This paper examined the ex post accuracy of past forecasting studies focusing on RN demand and explored its determinants on the accuracy of demand forecasts.
METHODS: Data were collected by systematically reviewing national reports or articles on RN demand forecasts. The mean absolute percentage error (MAPE) was measured for forecasting error by comparing the forecast with the actual demand (employed RNs). Nonparametric tests, the Mann‒Whitney test, and the Kruskal‒Wallis test were used to analyze the differences in the MAPE according to the variables, which are methodological and researcher factors.
RESULTS: A total of 105 forecast horizons and 196 forecasts were analyzed. The average MAPE of the total forecast horizon was 34.8%. Among the methodological factors, the most common determinant affecting forecast accuracy was the RN productivity assumption. The longer the length of the forecast horizon was, the greater the MAPE was. The longer the length of the data period was, the greater the MAPE was. Moreover, there was no significant difference among the researchers\' factors.
CONCLUSIONS: To improve demand forecast accuracy, future studies need to accurately measure RN workload and productivity in a manner consistent with the real world.
OBJECTIVE: This paper examined the ex post accuracy of past forecasting studies focusing on RN demand and explored its determinants on the accuracy of demand forecasts.
METHODS: Data were collected by systematically reviewing national reports or articles on RN demand forecasts. The mean absolute percentage error (MAPE) was measured for forecasting error by comparing the forecast with the actual demand (employed RNs). Nonparametric tests, the Mann‒Whitney test, and the Kruskal‒Wallis test were used to analyze the differences in the MAPE according to the variables, which are methodological and researcher factors.
RESULTS: A total of 105 forecast horizons and 196 forecasts were analyzed. The average MAPE of the total forecast horizon was 34.8%. Among the methodological factors, the most common determinant affecting forecast accuracy was the RN productivity assumption. The longer the length of the forecast horizon was, the greater the MAPE was. The longer the length of the data period was, the greater the MAPE was. Moreover, there was no significant difference among the researchers\' factors.
CONCLUSIONS: To improve demand forecast accuracy, future studies need to accurately measure RN workload and productivity in a manner consistent with the real world.
摘要：
背景：尽管需求预测准确性对注册护士（RN）劳动力具有重要意义，很少有研究评估过去的预测。
目的：本文研究了过去以RN需求为重点的预测研究的事后准确性，并探讨了其对需求预测准确性的决定因素。
方法：通过系统地审查国家报告或关于RN需求预测的文章来收集数据。通过将预测与实际需求（采用的RN）进行比较，可以测量预测误差的平均绝对百分比误差（MAPE）。非参数检验,Mann-Whitney测试,并使用Kruskal-Wallis检验根据变量分析MAPE的差异，这是方法论和研究者的因素。
结果：共分析了105个预测范围和196个预测。总预测范围的平均MAPE为34.8％。在方法论因素中，影响预测准确性的最常见的决定因素是RN生产率假设。预测范围越长，地图越大。数据周期的长度越长，地图越大。此外，研究人员的因素之间没有显着差异。
结论：为了提高需求预测的准确性，未来的研究需要以与现实世界一致的方式准确地测量RN工作量和生产率。
目的：本文研究了过去以RN需求为重点的预测研究的事后准确性，并探讨了其对需求预测准确性的决定因素。
方法：通过系统地审查国家报告或关于RN需求预测的文章来收集数据。通过将预测与实际需求（采用的RN）进行比较，可以测量预测误差的平均绝对百分比误差（MAPE）。非参数检验,Mann-Whitney测试,并使用Kruskal-Wallis检验根据变量分析MAPE的差异，这是方法论和研究者的因素。
结果：共分析了105个预测范围和196个预测。总预测范围的平均MAPE为34.8％。在方法论因素中，影响预测准确性的最常见的决定因素是RN生产率假设。预测范围越长，地图越大。数据周期的长度越长，地图越大。此外，研究人员的因素之间没有显着差异。
结论：为了提高需求预测的准确性，未来的研究需要以与现实世界一致的方式准确地测量RN工作量和生产率。
