eVIN assessment

Electronic Vaccine Intelligence Network (eVIN) assessment was conducted in 12 states (Assam, Chhattisgarh, Gujarat, Jharkhand, Manipur, Nagaland, Odisha, Bihar Himachal Pradesh, Madhya Pradesh, Rajasthan and Uttar Pradesh) where eVIN was launched initially. A pre-post comparison was used on key performance indicators for programmatic assessment. As obtaining data for one-year prior to implementation of eVIN was challenging, therefore, six months period was chosen as the reference period for pre-eVIN phase. The pre-eVIN reference period was of six months for all CCPs, however it varied for different states and districts due to different timeframe of eVIN rollout. The specific duration of pre-eVIN phase are as mentioned: most of the CCPs in a district were covered either in between the period of April 2015 to September 2015 (13 districts), April 2016 to September 2016 (12 districts), October 2015-March 2016 (11 districts) and October 2011- March 2012 (1 district). The post eVIN reference period was from October 2017 to March 2018 for the assessment. Minimum number of CCPs required for the study was calculated to be 502 considering 43% of PHCs reported instances of stock-out, 10% non-response rate and 1.2 design effect. It was further increased to 617 CCPs in order to draw valid conclusions at the state levels. Selection of CCPs in the eVIN states were done using two-stage sampling design. In the first stage, districts were selected followed by selection of CCPs in the second stage. In each eVIN state, number of sampled districts was decided based on Probability Proportion to Size of CCPs. In total 37 districts were selected using systematic random sampling technique after arranging the districts in ascending order based on the proportionate share of cold chain point in the total cold chain point in state. Further, CCPs were randomly selected in each of the selected district. A detailed methodology is available in the larger study document. Quantitative data was obtained using structured questionnaire from CCPs pertaining to stock management, temperature monitoring, cold chain equipment, and documentation aspects of vaccine supply chain. The primary data for pre-eVIN phase was done from stock registers, vaccine distribution registers, temperature log books and other important registers. Completeness and accuracy were analyzed in the assessment. Completeness was seen of Indent form, vaccine stock register, and temperature log book. Accuracy was assessed through stock register and eVIN record, eVIN record and physical count. Computer Assisted Personal Interviewing (CAPI) technique was employed using tablets/mobiles for real-time data collection and data entry.电子疫苗智能网络（Electronic Vaccine Intelligence Network, eVIN）评估在首批部署该系统的12个邦（阿萨姆邦、恰蒂斯加尔邦、古吉拉特邦、贾坎德邦、曼尼普尔邦、那加兰邦、奥里萨邦、比哈尔邦、喜马偕尔邦、中央邦、拉贾斯坦邦与北方邦）开展。本研究采用前后对照设计，以核心绩效指标为评估维度实施项目评估。由于获取eVIN正式部署前1年的相关数据存在较大难度，故选取6个月作为eVIN实施前阶段的参考周期。尽管所有冷链点（Cold Chain Points, CCPs）的eVIN前参考周期均设定为6个月，但由于各邦及地区的eVIN部署进度存在差异，该周期在不同区域有所不同。各地区多数冷链点的覆盖周期具体分为四类：2015年4月至2015年9月（13个地区）、2016年4月至2016年9月（12个地区）、2015年10月至2016年3月（11个地区）以及2011年10月至2012年3月（1个地区）。本次评估的eVIN实施后参考周期为2017年10月至2018年3月。研究初始测算所需最小冷链点样本量为502个，该计算基于43%的基层医疗保健中心（Primary Health Centers, PHCs）存在疫苗缺货情况、10%的无应答率以及1.2的设计效应。为确保能在邦级层面得出有效研究结论，该样本量进一步调整至617个冷链点。eVIN覆盖邦内的冷链点采用两阶段抽样法选取：第一阶段抽取地区，第二阶段抽取冷链点。每个eVIN覆盖邦的抽样地区数量依据冷链点占该邦总冷链点的比例规模（Probability Proportion to Size, PPS）确定。本研究先按各地区冷链点占所在邦总冷链点的比例将地区升序排列，再采用系统随机抽样法共抽取37个地区。随后在每个抽中的地区内随机选取冷链点。详细研究方法可查阅完整研究文档。定量数据通过结构化问卷收集，调研内容涵盖冷链点的库存管理、温度监控、冷链设备及疫苗供应链文件记录等维度。eVIN实施前阶段的基础数据来源于库存登记簿、疫苗分发登记簿、温度日志本及其他相关登记资料。本次评估对数据的完整性与准确性进行了分析：完整性分析涉及订货单、疫苗库存登记簿及温度日志本；准确性则通过库存登记簿与eVIN系统记录、eVIN系统记录与实物盘点的比对进行评估。本研究采用平板电脑/移动设备开展个人计算机辅助访谈（Computer Assisted Personal Interviewing, CAPI），实现实时数据采集与录入。