Data for: Comparative evaluation of interrupted time series analytical methods for healthcare quality improvement research: a Monte Carlo simulation study

This dataset supports a Monte Carlo simulation study comparing four interrupted time series (ITS) analytical methods—ARIMAX, GAM, Prophet, and LSTM—for healthcare quality improvement research. The simulation uses a full 2⁹ factorial design with 512 scenarios and 500 replicates per scenario (256,000 simulated time series in total). The nine factors are time series length, AR(1) coefficient, level- and slope-change intervention effects, trend type, seasonality, noise standard deviation, external shock, and number of covariates. Each replicate is stored as a CSV file under `scenario_<id>/rep_<j>.csv`; `scenarios_key.csv` maps each scenario ID to its factor levels. The complete analysis code, environment specifications, and step-by-step reproducibility guide are available at https://github.com/SeungmanKim/ITS. Full methodological details are described in the accompanying manuscript, currently under peer review at BMC Medical Research Methodology. License: Creative Commons Attribution 4.0 International (CC BY 4.0).本数据集支持一项蒙特卡洛（Monte Carlo）模拟研究，该研究针对医疗质量改进研究场景，对比了四种间断时间序列（Interrupted Time Series, ITS）分析方法：ARIMAX、GAM、Prophet及LSTM。 本次模拟采用完整的2⁹析因设计，共设置512种情景，每种情景开展500次重复实验，总计生成256,000条模拟时间序列。本次研究的九个影响因子分别为：时间序列长度、AR(1)系数、水平与斜率变化干预效应、趋势类型、季节性特征、噪声标准差、外部冲击及协变量数量。 每一次重复实验的结果均以CSV文件格式存储于`scenario_<id>/rep_<j>.csv`路径下；`scenarios_key.csv`文件可将各情景ID映射至其对应的因子水平配置。 完整的分析代码、环境配置说明与分步可复现指南可通过https://github.com/SeungmanKim/ITS获取。完整的方法学细节详见随附手稿，该手稿目前正在《BMC Medical Research Methodology》期刊接受同行评审。 许可协议：知识共享署名4.0国际许可（Creative Commons Attribution 4.0 International, CC BY 4.0）。