Zhu Jun, Davis James, Taira Deborah A, Yamashita Marisa
Hawaii Medical Service Association, Care Management, PO Box 860, Honolulu, HI 96808-0860,
Prev Chronic Dis. 2006 Apr;3(2):A56. Epub 2006 Mar 15.
Preventive screening is widely recognized as a key component of cost-effective, high-quality health care. Even so, national screening for cancer, diabetes, and cholesterol falls far short of U.S. Preventive Services Task Force recommendations. Although evidence has shown that reminder programs improve preventive screening rates, this study is one of the first to examine the characteristics of health plan members who respond to screening reminders.
The study sample included active members of a large health plan in Hawaii who were identified by an algorithm as not having received one or more recommended screenings based on age and sex criteria (2000-2003) for breast cancer (n = 44,331), cervical cancer (n = 73,875), colon cancer (n = 131,860), diabetes (n = 86,216), and cholesterol (n = 54,843). Statistical analyses were conducted using Cox proportional hazard and logistic regression models. In the proportional hazard models, reminder letters were treated as time-varying exposures. Hazard ratios, or rate ratios, were used to examine the relationship between health plan member and physician characteristics and the likelihood of responding to the reminders. The effects of additional or multiple reminders among health plan members receiving more than one reminder were examined in multivariable regression models.
The impact of health plan member characteristics and number of office visits on the response to reminders varied among the five health-screening types. Health plan members responded better to reminders for diabetes screening than for colon cancer screening. Members sent their second annual reminders were less likely to obtain screening than members sent their first reminder. Members receiving their third (or more) annual reminder were especially recalcitrant.
Our findings suggest that the response to reminders differs according to patient characteristics. In particular, targeted interventions may be needed to encourage screening for younger and healthier members whose response rate to reminders was low. Further research is needed to determine how health plans can best reach members who do not respond to patient reminders.
预防性筛查被广泛认为是具有成本效益的高质量医疗保健的关键组成部分。即便如此，美国针对癌症、糖尿病和胆固醇的全国性筛查仍远远未达到美国预防服务工作组的建议。尽管有证据表明提醒计划可提高预防性筛查率，但本研究是首批考察对筛查提醒做出回应的健康计划成员特征的研究之一。
研究样本包括夏威夷一个大型健康计划的活跃成员，这些成员通过一种算法被确定为基于年龄和性别标准（2000 - 2003年）未接受一项或多项推荐筛查，涉及乳腺癌（n = 44,331）、宫颈癌（n = 73,875）、结肠癌（n = 131,860）、糖尿病（n = 86,216）和胆固醇（n = 54,843）。使用Cox比例风险模型和逻辑回归模型进行统计分析。在比例风险模型中，提醒信被视为随时间变化的暴露因素。风险比或率比用于考察健康计划成员和医生特征与对提醒做出回应的可能性之间的关系。在多变量回归模型中考察了接受不止一次提醒的健康计划成员中额外或多次提醒的效果。
健康计划成员特征和门诊就诊次数对提醒回应的影响在五种健康筛查类型中各不相同。健康计划成员对糖尿病筛查提醒的回应比对结肠癌筛查提醒的回应更好。收到第二次年度提醒的成员比收到第一次提醒的成员进行筛查的可能性更小。收到第三次（或更多次）年度提醒的成员尤其顽固。
我们的研究结果表明，对提醒的回应因患者特征而异。特别是，可能需要有针对性的干预措施来鼓励对提醒回应率较低的年轻且健康的成员进行筛查。需要进一步研究以确定健康计划如何能最好地联系到那些对患者提醒不做出回应的成员。
