%0期刊文章%@ 1438-8871 %I Gunther Eysenbach %V 10% N 4% P e30% T对医疗保健提供者和管理人员的互联网调查的响应审计:决定反应率的影响%A Dobrow,Mark J %A Orchard,Margo C %A Golden,Brian %A Holowaty,Eric %A Paszat,Lawrence %A Brown, adalstein D %A Sullivan,Terrence + 620大学大道，癌症服务和政策研究单位，安大略省，多伦多，ON M5G 2L7，加拿大，+1 416 217 1380,mark.dobrow@utoronto.ca %K卫生保健调查%K互联网%K调查方法%D 2008 %7 16.10.2008 %9原始论文%J J医学互联网Res %G英文%X背景:与传统的调查方式相比，互联网调查方式往往是不利的，特别是在回复率方面。对Internet调查的响应可能受到所采用的分发选项和响应/收集功能以及自动(不在办公室的)回复、自动转发、服务器拒绝和组织或个人垃圾邮件过滤器的影响。然而，互联网调查也提供了无与伦比的机会来跟踪研究对象和检查许多影响回复率决定的因素。可用于互联网调查的跟踪数据提供了其他调查方式无法比拟的关于响应的重要组成部分的详细信息和即时反馈。本文介绍了对加拿大安大略省5000多名癌症护理提供者和管理人员的大型互联网调查的响应审计。目的:在互联网电子调查结果报告核对表的基础上，本文的主要目的是(a)评估一系列因素对决定互联网调查回复率的影响，以及(b)建议改进已公布的互联网调查方法描述的步骤。方法:我们审核调查响应数据，分析影响最终响应的分子和分母的因素。我们还进行了敏感性分析，以解释与某些因素对响应率的影响相关的固有不确定性。 Results: The survey was initially sent out to 5636 health care providers and administrators. The determination of the numerator was influenced by duplicate/unattached responses and response completeness. The numerator varied from a maximum of 2031 crude (unadjusted) responses to 1849 unique views, 1769 participants, and 1616 complete responses. The determination of the denominator was influenced by forwarding of the invitation email to unknown individuals, server rejections, automated replies, spam filters, and ‘opt out’ options. Based on these factors, the denominator varied from a minimum of 5106 to a maximum of 5922. Considering the different assumptions for the numerator and the denominator, the sensitivity analysis resulted in a 12.5% variation in the response rate (from minimum of 27.3% to maximum of 39.8%) with a best estimate of 32.8%. Conclusions: Depending on how the numerator and denominator are chosen, the resulting response rates can vary widely. The CHERRIES statement was an important advance in identifying key characteristics of Internet surveys that can influence response rates. This response audit suggests the need to further clarify some of these factors when reporting on Internet surveys for health care providers and administrators, particularly when using commercially available Internet survey packages for specified, rather than convenience, samples. %M 18926979 %R 10.2196/jmir.1090 %U //www.mybigtv.com/2008/4/e30/ %U https://doi.org/10.2196/jmir.1090 %U http://www.ncbi.nlm.nih.gov/pubmed/18926979