The 2006-07 Sri Lanka Demographic and Health Survey (SLDHS) is the fourth in a series of DHS surveys to be held in Sri Lanka-the first three having been implemented in 1987, 1993, and 2000. Teams visited 2,106 sample points across Sri Lanka and collected data from a nationally representative sample of almost 20,000 households and over 14,700 women age 15-49.
A nationally representative sample of 21,600 housing units was selected for the survey and 19,872 households were enumerated to give district level estimates (excluding Northern Province). Detailed information was collected from all ever-married women aged 15-49 years and about their children below five years at the time of the survey. Within the households interviewed, a total of 15,068 eligible women were identified, of whom 14,692 were successfully interviewed.
The Department of Census and Statistics (DCS) carried out the 2006-07 SLDHS for the Health Sector Development Project (HSDP) of the Ministry of Healthcare and Nutrition, a project funded by the World Bank. The objective of the survey is to provide data needed to monitor and evaluate the impact of population, health, and nutrition programmes implemented by different government agencies. Additionally, it also aims to measure the impact of interventions made under the HSDP towards improving the quality and efficiency of health care services as a whole.
All 25 districts of Sri Lanka were included at the design stage. The final sample has only 20 districts, however, after dropping the 5 districts of the Northern Province (Jaffna, Kilinochchi, Mannar, Vavuniya, and Mullativu), due to the security situation there.
The objective of this report is to publish the final findings of the 2006-07 SLDHS. This final report provides information mainly on background characteristics of respondents, fertility, reproductive health and maternal care, child health, nutrition, women's empowerment, and awareness of HIV/AIDS and prevention. It is expected that the content of this report will satisfy the urgent needs of users of this information.
Survey results indicate that there has been a slight upturn in the total fertility rate since the 2000 SLDHS. The total fertility rate for Sri Lanka is 2.3, meaning that, if current age-specific fertility rates were to remain unchanged in the future, a woman in Sri Lanka would have an average of 2.3 children by the end of her childbearing period. This is somewhat higher than the total fertility rate of 1.9 measured in the 2000 SLDHS.
Fertility is only slightly lower in urban areas than in rural areas (2.2 and 2.3 children per woman, respectively); however, it is higher in the estate areas (2.5 children per woman). Interpretation of variations in fertility by administrative districts is limited by the small samples in some districts. Nevertheless, results indicate that Galle and Puttalam districts have fertility rates of 2.1 or below, which is at what is known as “replacement level” fertility, i.e., the level that is necessary to maintain population size over time. Differences in fertility by level of women's education and a measure of relative wealth status are minimal.
According to the survey findings, knowledge of any method of family planning is almost universal in Sri Lanka and there are almost no differences between ever-married and currently married women. Over 90 percent of currently married women have heard about pills, injectables, female sterilization, and the IUD. Eight out of ten respondents know about some traditional method of delaying or avoiding pregnancies.
Although the proportion of currently married women who have heard of at least one method of family planning has been high for some time, knowledge of some specific methods has increased recently. Since 1993, knowledge of implants has increased five-fold-from about 10 percent in 1993 to over 50 percent in 2006-07. Awareness about pill, IUD, injectables, implants, and withdrawal has also increased. On the other hand, awareness of male sterilization has dropped by 14 percentage points.
The study of infant and child mortality is critical for assessment of population and health policies and programmes. Infant and child mortality rates are also regarded as indices reflecting the degree of poverty and deprivation of a population. Survey data show that for the most recent five-year period before the survey, the infant mortality rate is 15 deaths per 1,000 live births and under-five mortality is 21 deaths per 1,000 live births. Thus, one in every 48 Sri Lankan children dies before reaching age five. The neonatal mortality rate is 11 deaths per 1,000 live births and the postneonatal mortality rate is 5 deaths per 1,000 live births. The child mortality rate is 5 deaths per 1,000 children surviving to age one year.
The survey shows that virtually all mothers (99 percent) in Sri Lanka receive antenatal care from a health professional (doctor specialist, doctor, or midwife). The proportion receiving care from a skilled provider is remarkably uniform across all categories for age, residence, district, woman's education, and household wealth quintile. Even in the estate sector, antenatal care usage is at the same high level. Although doctors are the most frequently seen provider (96 percent), women also go to public health midwives often for prenatal care (44 percent).
BREASTFEEDING AND NUTRITION
Poor nutritional status is one of the most important health and welfare problems facing Sri Lanka today and particularly affects women and children. The survey data show that 17 percent of children under five are stunted or short for their age, while 15 percent of children under five are wasted or too thin for their height. Overall, 21 percent of children are underweight, which may reflect stunting, wasting, or both. As for women, at the national level, 16 percent of women are considered to be thin (with a body mass index < 18.5); however, only 6 percent of women are considered to be moderately or severely thin.
Poor breastfeeding and infant feeding practices can have adverse consequences for the health and nutritional status of children. Fortunately, breastfeeding in Sri Lanka is universal and generally of fairly long duration; 97 percent of newborns are breastfed within one day after delivery and 76 percent of infants under 6 months are exclusively breastfed, lower than the recommended 100 percent exclusive breastfeeding for children under 6 months. The median duration of any breastfeeding is 33 months in Sri Lanka and the median duration of exclusive breastfeeding is 5 months.
The HIV/AIDS pandemic is a serious health concern in the world today because of its high case fatality rate and the lack of a cure. Awareness of AIDS is almost universal among Sri Lankan adults, with 92 percent of ever-married women saying that they have heard about AIDS. Nevertheless, only 22 percent of ever-married women are classified as having “comprehensive knowledge” about AIDS, i.e., knowing that consistent use of condoms and having just one faithful partner can reduce the chance of getting infected, knowing that a healthy-looking person can be infected, and knowing that AIDS cannot be transmitted by sharing food or by mosquito bites. Such a low level of knowledge about AIDS implies that a concerted effort is needed to address misconceptions about HIV transmission. Programs might be focused in the estate sector and especially in Batticaloa, Ampara, and Nuwara Eliya districts where comprehensive knowledge is lowest.
Moreover, a composite indicator on stigma towards HIV-infected people shows that only 8 percent of ever-married women expressed accepting attitudes toward persons living with HIV/AIDS. Overall, only about one- half of ever-married women age 15-49 years know where to get an HIV test.
WOMEN'S EMPOWERMENT AND DEMOGRAPHIC AND HEALTH OUTCOMES
The 2006-07 SLDHS collected data on women's empowerment, their participation in decisionmaking, and attitudes towards wife beating. Survey results show that more than 90 percent of currently married women, either alone or jointly with their husband, make decisions on how their income is used. However, husbands' control over women's earnings is higher among women with no education (15 percent) than among women with higher education (4 percent).
In Sri Lanka, the husband is usually the main source of household income; two-thirds of women earn less than their husband. Although the majority of women earn less than their husband, almost half have autonomy in decisions about how to spend their earnings.
The survey also collected information on who decides how the husband's cash earnings are spent. The majority of couples (60 percent) make joint decisions on how the husband's cash income is used. More than 1 in 5 women (23 percent) reported that they decide how their husband's earnings are used; another 16 percent of the women reported that their husband mainly decides how his earnings are spent.
Kind of data
Sample survey data
A nationally representative sample of 21,600 housing units was selected for the survey and 19,872 households were enumerated to give district level estimates (excluding Northern Province).
Unit of analysis
- Women age 15-49
- Children under five
In principle, the sample was designed to cover private households in the areas sampled. The population residing in institutions and institutional households was excluded. For the detailed individual interview, the eligibility criteria wereall ever-married women aged 15-49 years who slept in the household the previous night and about their children below five years at the time of the survey.
Producers and sponsors
Department of Census and Statistics (DCS)
Health Sector Development Project (HSDP)
Ministry of Healthcare and Nutrition
Macro International Inc.
The World Bank
United Nations Children's Fund
Provider of technical material
The 2006-07 SLDHS sample was designed to produce key indicators for the country as a whole, and for sectors (urban, rural, and estate) and districts. All 25 districts of Sri Lanka were included at the design stage. The final sample has only 20 districts, however, after dropping the 5 districts of the Northern Province (Jaffna, Kilinochchi, Mannar, Vavuniya, and Mullativu), due to the security situation there. The districts in the other 8 provinces are given below:
1. Western Province: Colombo, Gampaha, Kalutara
2. Southern Province: Galle, Matara, Hambantota
3. Sabaragamuwa Province: Ratnapura, Kegalle
4. Uva Province: Badulla, Monaragala
5. Central Province: Kandy, Matale, Nuwara Eliya
6. Eastern Province: Ampara, Batticaloa, Trincomalee
7. North Central Province: Anuradhapura, Polonnaruwa
8. North Western Province: Puttalam, Kurunegala
The SLDHS used a stratified two-stage cluster sample design. The sample was spread geographically more or less proportionally to the population. The first stage involved selecting 2,500 enumeration areas (clusters) from the list of about 100,000 enumeration areas (EAs) formed in the 2001 Population Census. The objectives of the sampling design were to provide reasonably accurate estimates at three levels-national, sector (urban, rural, estate), and district-and to provide estimates for tsunami-affected areas as well. In order to provide reliable estimates for these levels, some districts were oversampled. Therefore, the final sample is not self-weighting, and weighting factors were used to make the sample be proportional to the population. Weighted data are used throughout the report, unless otherwise noted.
An enumeration area (EA) is a subdivision of a Grama Niladari area, which consists of about 80 housing units in urban areas and about 65 units in rural or estate areas. The criterion used in creating an EA is that one enumerator can visit all the units in the area within six hours to take a count of all the units and the people residing therein. A household list of each EA (including information on housing units) along with a map is available at DCS to be used as a frame for selection of samples for surveys.
The second stage of selection involved the systematic sampling of 10 households listed in each enumeration area. Thus, 2,500 clusters were selected initially: 469 urban, 1,831 rural and 200 estate. In the end, information was collected from 2,106 clusters. The remaining 394 clusters were not enumerated (340 clusters from Northern Province due to unsettled conditions prevailing in the province at the time of the survey, and 54 clusters from other areas). The reasons for elimination were: temporary tsunami camps (5 clusters), unstable security conditions (45), and landslides that prevented access to the clusters (4).
For the final sample (which excluded the Northern Province), 21,060 housing units were selected, and 19,862 households were interviewed. All ever-married women age 15-49 years living in these households were eligible to be interviewed. Eligible respondents were either usual residents of the households or visitors present in the household on the night before the interview date.
There are certain limitations in comparing the findings of this survey with those from the 2000 SLDHS. The earlier survey did not include the Eastern Province, whereas the 2006-07 SLDHS did include it. However, neither the 2000 DHS nor the 2006-07 DHS covered the Northern Province. Thus when comparisons are made to the 2000 SLDHS, the data for the current survey are shown with the Eastern Province also excluded. As a result, the statistics for 2006-07 in such tables may differ slightly from tables presenting the 2006-07 SLDHS data only.
A total of 21,600 housing units were selected for the sample, from which 21,357 households were located, and 20,317 were occupied at the time of the survey. Of those existing households, 19,862 were successfully interviewed, yielding a household response rate of 98 percent. The household response rate is slightly higher in the rural sector than in the urban and estate sectors.
Within the households interviewed, a total of 15,068 eligible women (ever-married women age 15-49) were identified, of whom 14,692 were successfully interviewed, yielding a response rate of 98 percent. The eligible women's response rate is also slightly higher in rural areas than urban and estate areas.
The principal reason for non-response among eligible women was the failure to meet them at home despite repeated visits to the households. There were very few partially completed cases, and refusals were very minimal.
Dates of collection
Mode of data collection
Data collection supervision
Senior staff of DCS were appointed as district coordinators. The overall responsibility of the coordinators was to ensure the smooth implementation and good quality of fieldwork. Their tasks included checking completed questionnaires for quality and managing logistics, such as transport and accommodations. A survey expert from Macro also made two visits in 2006 (in August-September and late October-early November) to observe field work and review completed questionnaires and give feedback to interviewers.
A new feature was added to this round of the SLDHS to improve the quality of fieldwork. A special set of tables was generated, which were used to measure the quality of fieldwork. If any deviations from the expected targets were found, the teams were informed and instructed on remedial actions.
The 2006-07 SLDHS questionnaire, which was used to collect information from households and eligible women through personal interviews, contained the topics mentioned below. An effort was made to incorporate globally recommended standard questions as much as possible. Model questionnaires developed by MEASURE DHS were used with some modifications to match the local situation. Additional questions were also included to satisfy the needs of the health sector and also to provide data for the compilation of UNICEF's World Fit for Children and Millennium Development Goals (MDG) indicators.
The questionnaire had two main sections, namely, a household section and a section on women and children.
a) The first section served two purposes. One was to list all the usual members and visitors in the selected household together with some basic information about them-such as age, education, marital status, nature of residence, and relationship to the head of household. This information was used to identify eligible women and children for the main interview and also to provide denominators for analysis of some of the household characteristics. The second purpose was to collect information on characteristics of the household's dwelling. Information of this type included source of drinking water, toilet facilities, construction materials of the home, land tenure, garbage disposal, ownership of livestock and various durable goods, use of iodized salt, and use of mosquito nets (including treated nets). There were also questions about several non-communicable diseases, adequacy of basic requirements for school-going children, and information about orphanhood. Children under five and women 15-49 were eligible to be weighed and measured and have their haemoglobin level measured.
b) The second section, the questionnaire for ever-married women, covered the following topics:
- Background characteristics (education, marital status, media exposure, etc.)
- Marriage and sexual activity
- Reproductive history
- Knowledge and use of family planning methods
- Antenatal, delivery, and postnatal care
- Child immunization and health
- Child and women's nutrition
- Sexual activity
- Fertility preferences
- Woman's work and husband's background characteristics
- Awareness about AIDS and other sexually transmitted infections (STIs)
- Use of drugs, tobacco, and alcohol by household members
- Other health issues
Respondents were also asked an extensive series of questions about their children below 5 years at the time of the survey. Topics covered were vaccinations, childhood illnesses, nutrition status, and breastfeeding. In addition, a calendar of events was used to record information related to respondent's marriage, pregnancies and births, and contraceptive use in a specially designed chart for a five-year period prior to the survey.
The complete questionnaire was pre-tested by a team of experienced staff to test the feasibility, sequence, skipping, and timing before it was finalized. The questionnaire was prepared in Sinhala and Tamil; an English version was used occasionally for a few interviews.
Department of Census and Statistics
Processing SLDHS data began a few weeks after the fieldwork commenced. Completed questionnaires were returned periodically from the field to the Data Processing Division in Colombo, where they were coded manually by staff specially trained for this task. Processing the data concurrently with fieldwork was a distinct advantage for data quality, since the survey unit was able to advise field teams of errors detected during data entry.
A group of experienced data entry persons at DCS did data entry. Two consultants from Macro assisted the Data Processing Division and DHS Unit staff by giving necessary guidance for manual coding and editing, data entry, verification, online editing, and machine editing. Data were entered using the CSPro computer package. All data were entered twice (100 percent verification). Two DCS staff members who were specially trained for this purpose did machine editing. Data entry was completed in December 2007. DCS completed machine editing of the data file in mid-January 2008. A Macro Data Processing Specialist visited Colombo in February 2008 to work with DCS staff to produce the final data file and run preliminary tables.
The estimates from a sample survey are affected by two types of errors: nonsampling errors and sampling errors. Nonsampling errors are the results of mistakes 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 2006-07 Sri Lanka Demographic and Health Survey (SLDHS) to minimize this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.
Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the 2006-07 SLDHS 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 not 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 used to calculate confidence intervals within which the true value for the population can reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistic will fall within a range of plus or minus two times the standard error of that statistic in 95 percent of all possible samples 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 2006-07 SLDHS 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 2006-07 SLDHS is a Macro SAS procedure. This module used the Taylor linearization method of variance estimation for survey estimates that are means or proportions. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.
The Jackknife repeated replication method derives estimates of complex rates from each of several replications of the parent sample, and calculates standard errors for these estimates using simple formulae. Each replication considers all but one cluster in the calculation of the estimates. Pseudo-independent replications are thus created. In the 2006-07 SLDHS, there were 2,106 non-empty clusters. Hence, 2,106 replications were created.
In addition to the standard error, the procedure computes the design effect (DEFT) for each estimate, which is defined as the ratio between the standard error using the given sample design and the standard error that would result if a simple random sample had been used. A DEFT value of 1.0 indicates that the sample design is as efficient as a simple random sample, while a value greater than 1.0 indicates the increase in the sampling error due to the use of a more complex and less statistically efficient design. The procedure also computes the relative error and confidence limits for the estimates.
Sampling errors for the 2006-07 SLDHS are calculated for selected variables considered to be of primary interest for the women's survey and for the men's surveys, respectively. The results are presented in an appendix of the Final Report for the country as a whole, for urban, rural and estate areas, and for each of the 20 districts of the country. For each variable, the type of statistic (mean, proportion, or rate) and the base population are given in Table B.1 of the Final Report. Tables B.2 to B.25 present the value of the statistic (R), its standard error (SE), the number of unweighted (N-UNWE) and weighted (N-WEIG) cases, the design effect (DEFT), the relative standard error (SE/R), and the 95 percent confidence limits (R±2SE), for each variable. The DEFT is considered undefined when the standard error considering simple random sample is zero (when the estimate is close to 0 or 1). In the case of the total fertility rate, the number of unweighted cases is not relevant, as there is no known unweighted value for woman-years of exposure to child-bearing.
The confidence interval (e.g., as calculated for children ever born) can be interpreted as follows: the overall average from the national sample is 1.434 and its standard error is 0.026. Therefore, to obtain the 95 percent confidence limits, one adds and subtracts twice the standard error to the sample estimate, i.e., 1.434±2×0.026. There is a high probability (95 percent) that the true average number of children ever born to all women is between 1.381 and 1.487.
For the total sample, the value of the DEFT, averaged over all variables, is 1.104. This means that, due to multi-stage clustering of the sample, the average standard error is increased by a factor of 1.104 over that in an equivalent simple random sample.
Department of Census and Statistics (DCS) and Ministry of Healthcare and Nutrition (MOH). 2009. Sri Lanka Demographic and Health Survey 2006-07. Colombo, Sri Lanka: DCS and MOH.
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.