Computerized health information and the demand for medical care.
Computerized health information and the demand for medical care.
复制标题
DOI：
10.1046/j.1524-4733.2003.00155.x
复制
发表时间：
2003
期刊：
影响因子：
--
通讯作者：
T. Wagner;H. Jimison
中科院分区：
文献类型：
--
作者：
T. Wagner;H. Jimison
消费者健康信息，曾经是书籍和小册子的领域，已经变得越来越数字化，并在互联网上提供。本研究评估使用电脑化健康资讯对消费者医疗需求的影响。方法以过去一年就诊次数为因变量。关键的自变量是使用计算机化的健康信息，这被视为内源性的。我们使用普通最小二乘法、工具变量、固定效应和固定效应工具变量模型测试了使用计算机化健康信息对医生就诊的影响。工具性变量包括参与保健社区项目，这是一项全社区的保健信息干预措施;计算机拥有情况;以及因特网接入情况。在健康社区项目前后，随机向三个城市的家庭邮寄了调查问卷。总共收集了5909份调查，回复率为54%。结果在双变量和多变量分析中，使用计算机化的健康信息与自我报告的进入护理或访问次数无关。工具变量模型也没有发现差异，除了两阶段条件logit模型的进入护理的概率显着更大（P <0.05）。结论：虽然为人们提供健康信息是直观的吸引力，我们发现很少有证据表明使用计算机的健康信息和自我报告的医疗访问之间的关联在过去的一年。这项研究使用整体自我报告的利用率作为因变量，需要更多的研究来确定健康信息是否会以其他重要方式影响健康生产函数，例如护理的位置，获得护理的时间或治疗的强度。
OBJECTIVE Consumer health information, once the domain of books and booklets, has become increasingly digitized and available on the Internet. This study assessed the effect of using computerized health information on consumers' demand for medical care. METHODS The dependent variable was self-reported number of visits to the doctor in the past year. The key independent variable was the use of computerized health information, which was treated as endogenous. We tested the effect of using computerized health information on physician visits using ordinary least squares, instrumental variables, fixed effects, and fixed-effects instrumental variables models. The instrumental variables included exposure to the Healthwise Communities Project, a community-wide health information intervention; computer ownership; and Internet access. Random households in three cities were mailed questionnaires before and after the Healthwise Communities Project. In total, 5909 surveys were collected for a response rate of 54%. RESULTS In both the bivariate and the multivariate analyses, the use of computerized health information was not associated with self-reported entry into care or number of visits. The instrumental variables models also found no differences, with the exception that the probability of entering care was significantly greater with the two-stage conditional logit model (P <.05). CONCLUSIONS Although providing people with health information is intuitively appealing, we found little evidence of an association between using a computer for health information and self-reported medical visits in the past year. This study used overall self-reported utilizations as the dependent variable, and more research is needed to determine whether health information affects the health production function in other important ways, such as the location of care, the timing of getting care, or the intensity of treatment.