METHODOLOGIC ISSUES RELATED TO RANDOM DIGIT DIALING

提出了两个项目来研究两种方法 与使用Waksberg随机程序有关的问题 数字拨号在基于人群的病例对照研究。 的 具体目标是：1）确定和最佳时间，当以前 拨打不合格的电话号码（定义为非 工作，业务，错误的县，快忙碌，慢忙碌，没有 联系电话号码可以返回到采样池 以及2）确定地理聚类是否发生超过 当一个以上的控制是偶然的， 根据主采样单元（PSU）选择。 的总目标 这两个项目是：制定指导方针和建议，1） 恢复不符合资格的电话号码， 采样池，以及2）可能选择一个以上 每个PSU的控制。 在项目1中， 将确定可追溯到1979年的六项病例对照研究。 不符合资格的电话号码的百分比（约9，061 已转换为住宅号码的电话号码， 通过呼叫号码和/或检查其状态来确定 针对电话公司的ScanTel查号台 数据库。 将在所有住宅区进行住户普查。 号码 这些数据将用于1）估计 按年龄和性别选择的概率，如果 被列为不符合资格的数字被转换为住宅 在不同时间点的数字和2）提供 关于将数字返回到 取样池 在项目2中，大约1，029个 将从先前的三项病例对照研究中选择对照， 映射以确定是否使用不同的 Waksberg方法的变体（例如，只有主号码， 主号码和次号码）群集 期待偶然。 这两个项目的结果可能会有一个 对研究设计和研究费用的重大影响-两个方面 这将在所提出的建议中加以处理。

Two projects are proposed to investigate two methodological issues related to the use of Waksberg's procedures for random digit dialing in population-based case-control studies. The specific aims are: 1) to identify and optimal time when previously dialed ineligible telephone telephone numbers (defined as non- working, business, wrong county, fast busy, slow busy, and no contact telephone numbers can be returned to the sampling pool and 2) to determine if geographical clustering occurs more than would be expected by chance when more than one control is selected per Primary Sampling Unit (PSU). The general aims of both projects are: to develop guidelines and recommendations, 1) for the restoration of ineligible telephone numbers to the sampling pool, and 2) for the possible selection of more than one control per PSU. In Project 1, ineligible telephone numbers from six case-control studies dating back to 1979 will be identified. The percentage of ineligible telephone numbers (about 9,061 numbers) that have converted to residential numbers will be determined by calling the number and/or checking its status against the telephone company's ScanTel Directory Assistance Data Base. A household census will be taken at all residential numbers. These data will be used 1) to estimate changes in the probability of selection by age and sex that would occur if the numbers classified as ineligible were converted to residential numbers at various points in time and 2) to provide recommendations about the optimal time to return numbers to the sampling pool. In Project 2, the addresses of approximately 1,029 controls selected from three previous case-control studies will be mapped to determine if controls selected using different variations of the Waksberg method (e.g., primary numbers only, primary and secondary numbers) cluster more than would be expected by chance. The findings from both projects could have a significant impact on study design and study costs -- two areas that will be addressed in the recommendations that are made.