National Health Survey 1998, Interim Demographic and Health Survey - Cambodia

Abstract
---------------------------

The primary objective of the Cambodia National Health Survey is to provide the Ministry of Health with reliable, population-based, nationally representative data or infant/child mortality, fertility, and related health service indicators.

A secondary objective was to provide the ADB-financed Basic Helath Services Project (BHSP) and the World Bank finaced Cambodia Disease Control and Health Development Project (CDCP) with baseline information about their respective Project areas, against which project impact could later be assessed.

Geographic coverage
---------------------------

National coverage

Analysis unit
---------------------------

Household
Women age 15-49
Children under age 5

Kind of data
---------------------------

Sample survey data [ssd]

Sampling procedure
---------------------------

Sample Design and Selection
The NHS sample was designed to provide estimates of kwy health indicators including infant/ child mortality rates and fertility rates for the country as a whole, for urban and rural residence, and for the two project catchment areas (the Basic Health Services Project and the Cambodia Disease Control and Health Development Project). In addition, the design allows for estimates of most key variables (but not for the vaccination coverage of children, fertility rates, or mortality rates) for 14 Provinces. In the other Provinces, the sample size is not sufficiently large to allow for province-level estimates. In order to provide sufficient cases to meet the survey objectives, the number of households selected in the NHS sample from each Province was disproportional to the size of the population in the Province. The above arrangements imply stratification into 40 strata, with 40 different sampling fractions. These strata are 20 Provinces, each divided into an urban and a rural sector. As a result, the NHS sample is self-weighting within strata; weights are only necessary when making estimates across more than one stratum.

For a more complete description of the NHS sample design, see Appendix A of the survey final report.

Mode of data collection
---------------------------

Face-to-face [f2f]

Research instrument
---------------------------

The NHS involved two types of questionnaires: a household questionnaire and an individual questionnaire. The household questionnaire was administered to all selected households; the individual questionnaire was administered to all women aged 15-49 identified in the household questionnaire as either usual residents of the household or visitors who stayed there on the night before the day of interview. These questionnaires were developed to measure the desired indicators identified by the MOH and Technical Steering Committee. Wording and structure of the questionnaires, where applicable, was based on the model survey instruments Macro International has used in similar surveys worldwide.

The household questionnaire consisted of three parts: 1) a household schedule giving demographic details of all usual household members and overnight visitors; 2) a series of questions relating to the utilization of health services for any household members who had been ill or injured in the past 30 days; and 3) questions about wall and roof materials of the home and household possessions, which in turn were used to compose a measure of overall household socio-economic status.

The individual questionnaire administered to women aged 15-49 gathered detailed information about the woman's reproductive history, and maternal and child health related knowledge and practices. Questions specific to child health practices were limited to children born after January 1993. (i.e., children under age 5)

The questionnaire was developed in English, translated into Khmer, then back translated and corrected. Following this, a three day pretest covering 100 households was conducted in Phnom Penh and rural Kandal Province by twenty interviewers after initial two week training. The questionnaires were finalized following the pretest.

Cleaning operations
---------------------------

Data Processing was conducted by NIPH with technical assistance form Macro International. The NIPH central office collected questionnaires form supervisors as soon as a cluster was completed. Office editors reviewed questionnaires for consistency and completeness. The data from the questionnaires were then entered and edited on microcomputers using the Integrated System for Survey Analysis (ISSA), a software package developed especially for such surveys by Macro International. During the machine entry, all questionnaires were reentered for verification. Entry and editing of data began one week after the fieldwork started and was completed by the beginning of August 1998.
To provide feedback for the field teams, quality tables were produced every two weeks during the fieldwork. These tables were designed to identify major systematic errors in data collection (e.g. age displacement). The fieldwork coordinators reviewed these tables and, if they found a problem, notified and advised all teams of the steps to be taken to avoid this problem in the future.

Response rate
---------------------------

A total of 7,654 women were identified as eligible to be interviewed. Questionnaires were completed for 7,630 of those women, a response rate of 99.7 precent. There is a little difference between the household and individual response rates in urban and rural areas. The same is true for the two project areas.

Sampling error estimates
---------------------------

The estimate from a sample survey is affected by two types of errors: 1) nonsampling errors, and 2) sampling errors. Nonsampling errors are the results of mistake made in implementing data collection and data processing, such as failure to locate and interview the correct household, misunderstanding of the questions on the part of either the interviewer or the respondent, and data entry errors. Although numerous efforts were made during the implementation of the National Health Survey (NHS) to minimize this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling error, on the other hand, can be evaluated statistically. The sample of respondents selected in the NHS is only one of many samples that could have been selected from the same population, using the same design and expected size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability between all possible samples. Although the degree of variability is nor known exactly, it can be estimated from the survey results.

A sampling error is usually measured in terms of the standard error for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistics will fall within a range of plus or minus two times the standard error of that statistic in 95 percent of all possible sample of identical size and design.

If the sample of respondents had been selected as a simple random sample, it would have been possible to use straightforward formulas for calculating sampling errors. However, the NHS sample is the result of a multi-stage stratified design and consequently, it was necessary to use more complex formulae. The computer software used to calculate sampling errors for the NHS is the ISSA Sampling Error Module. This module used the Taylor linearization method of variance estimation for survey estimates that are means of proportions. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.

For details of sampling error estimations information see Appendix B of the final survey report.

Data appraisal
---------------------------

Data Quality Tables
- Household age distribution
- Births by calendar year
- Reporting of age at death in days
- Reporting of age at death in months

Note: See detailed tables in APPENDIX C of the survey report.

摘要
---------------------------
柬埔寨国家健康调查的主要目标是为卫生部提供可靠、基于人口、全国代表性的数据，以评估婴儿/儿童死亡率、生育率及相关健康服务指标。

次要目标是为亚洲开发银行资助的基本卫生服务项目（BHSP）和世界银行资助的柬埔寨疾病控制和卫生发展项目（CDCP）提供各自项目区域的基线信息，以便日后评估项目影响。

地理覆盖范围
---------------------------
全国覆盖

分析单元
---------------------------
家庭
15-49岁女性
5岁以下儿童

数据类型
---------------------------
样本调查数据 [ssd]

抽样程序
---------------------------
样本设计和选择

国家健康调查（NHS）的样本设计旨在提供包括婴儿/儿童死亡率、生育率等关键健康指标的估计值，这些指标针对整个国家、城市和农村居住区，以及两个项目受益区域（基本卫生服务项目和柬埔寨疾病控制和卫生发展项目）。此外，该设计允许估计14个省份的大部分关键变量（但不包括儿童的疫苗接种覆盖率、生育率或死亡率）。在其他省份，样本量不足以进行省级估计。为了满足调查目标，每个省份在NHS样本中选择的户数与该省人口规模不成比例。上述安排意味着分为40个层次，每个层次有40个不同的抽样比例。这些层次是20个省份，每个省份分为城市和农村两个部分。因此，NHS样本在层次内自我加权；只有在跨越多个层次进行估计时才需要权重。

有关NHS样本设计的更完整描述，请参阅调查最终报告的附录A。

数据收集方式
---------------------------
面对面 [f2f]

研究工具
---------------------------
NHS涉及两种类型的问卷：家庭问卷和个人问卷。家庭问卷被发放给所有选定的家庭；个人问卷被发放给在家庭问卷中被识别为该家庭常住居民或前一天晚上在该处过夜的访客的所有15-49岁女性。这些问卷的开发是为了测量由卫生部和技术指导委员会确定的目标指标。在适用的情况下，问卷的文字和结构基于宏国际（Macro International）在全球类似调查中使用的模型调查工具。

家庭问卷包括三个部分：1）家庭日程，提供所有常住家庭成员和过夜访客的人口统计细节；2）一系列问题，涉及过去30天内任何生病或受伤的家庭成员的健康服务利用情况；3）有关家庭墙壁和屋顶材料以及家庭财产的问题，这些信息随后被用来编制总体家庭社会经济状况的衡量指标。

对15-49岁女性进行个人问卷调查，收集有关女性生殖历史、以及母亲和儿童健康相关知识和实践方面的详细信息。针对儿童健康实践的具体问题仅限于1993年1月以后出生的儿童（即5岁以下儿童）。

问卷最初用英语编写，翻译成高棉语，然后进行回译和校正。随后，在金边和卡达省农村地区由20名调查员进行了为期三天的预测试，预测试覆盖了100个家庭。在最初的两周培训之后，问卷最终确定。

数据清洗操作
---------------------------
数据处理由NIPH进行，并得到了宏国际的技术协助。NIPH中央办公室在完成一个簇后立即从监督员那里收集问卷。办公室编辑人员审查问卷以确保一致性和完整性。然后，使用宏国际专门为此类调查开发的集成调查分析系统（ISSA）软件包在微计算机上录入和编辑问卷数据。在机器录入过程中，所有问卷都被重新录入以进行验证。数据录入和编辑始于实地工作开始后的一周，并于1998年8月初完成。

为了为实地团队提供反馈，在实地工作期间每两周产生一次质量表。这些表格旨在识别数据收集中的主要系统性错误（例如年龄错位）。实地工作协调员审查这些表格，并在发现问题时通知并建议所有团队采取避免未来出现此类问题的措施。

回应率
---------------------------
共有7,654名女性被确定为有资格接受访谈。完成了7,630名女性的问卷，回应率为99.7%。城市和农村地区的家庭和个人回应率之间略有差异。对于两个项目区域也是如此。

抽样误差估计
---------------------------
样本调查的估计值受到两种类型误差的影响：1）非抽样误差，2）抽样误差。非抽样误差是实施数据收集和数据处理中出现的错误的结果，例如未能找到并访谈正确的家庭、访谈员或受访者对问题的误解，以及数据录入错误。尽管在实施国家健康调查（NHS）期间做出了大量努力以最大限度地减少此类错误，但非抽样误差是不可避免的，并且难以进行统计评估。

另一方面，抽样误差可以通过统计方法进行评估。NHS中选定的受访者样本只是从同一人口中可能选出的许多样本之一，使用相同的结构和预期规模。这些样本中的每一个都会产生与实际选定样本结果略有不同的结果。抽样误差是衡量所有可能样本之间差异的指标。虽然差异程度并不完全清楚，但它可以从调查结果中估计。

抽样误差通常以特定统计量（平均值、百分比等）的标准误差来衡量，这是方差的平方根。可以合理地假设标准误差将落在某个范围内。例如，对于从样本调查中计算出的任何给定的统计量，该统计量的值将在95%的所有可能样本中落在该统计量的标准误差的两倍范围内。

如果受访者样本被选为简单随机样本，则可以使用简单的公式来计算抽样误差。然而，NHS样本是多层次分层设计的结果，因此有必要使用更复杂的公式。用于计算NHS抽样误差的计算机软件是ISSA抽样误差模块。该模块使用泰勒线性化方法进行方差估计，用于调查估计值是比例的平均值。对于生育率和死亡率等更复杂的统计数据，使用Jackknife重复复制方法进行方差估计。

有关抽样误差估计的详细信息，请参阅最终调查报告的附录B。

数据评估
---------------------------
数据质量表
- 家庭年龄分布
- 按日历年出生
- 死亡年龄按天数报告
- 死亡年龄按月份报告

注意：请参阅调查报告附录C中的详细表格。