HOUSEHOLD SURVEY MANUAL:
DIARRHOEA AND
ACUTE RESPIRATORY INFECTIONS
Division for the Control of Diarrhoea1 and Acute Respiratory Disease World Health Organization
Geneva, 1994
WHO/CDR/94.8
THE HOUSEHOLD SURVEY MANUAL This manual was prepared by the World Health Organization's Division for the Control of Diarrhoea1 and Acute Respiratory Disease through a contract with ACT International, Atlanta, Georgia, USA.
HOUSEHOLD SURVEY MANUAL TABLE OF CONTENTS
Section-Page INTRODUCTION
. . .
A. . .
Key WHOIUNICEF Programme Indicators Measured A-2
Flowchart . . . A 4
...
1.0 PLAN THE SURVEY B
. . . .
1.1 Select Survey Coordinator B-l
1.2 Decide on Rates to Measure, Sections of Survey
. . . .
to Include, and Geographic Scope of Survey B-l
. . .
Rate List: CDDIARI Household Survey B 4
. . . .
1.3 Decide on the Time Period to be Covered by the Survey B-19
. . .
1.4 Determine Sample Size B-20
. . .
Denominator Sample Sizes for an Initial CDDIARI Survey B-24 Worksheet for Determining Sample Size for an Initial CDDIARI Household Survey . B-25
. . . .
1.5 Prepare Budget B-29
...
2.0 PREPARE TO CONDUCT THE SURVEY C
2.1 Adapt and Translate Questionnaire and Other Survey Forms as Necessary . . . C-l . . . .
2.2 Pretest Survey Forms and Revise them as Necessary C-l l
. . . .
2.3 Select Communities where Clusters will be Located C-14
. . . .
Sampling Process C-16
2.4 Prepare Information and Guidelines Needed for Selection of
Households in Clusters . . . C-20 . . .
Process for Selecting Households in Clusters C-21
2.5 Select Survey Personnel . . . C-23
2.6 Copy Questionnaire
.
.. .
. ..
. .. .
..
.. .
. ..
. ..
. . ..
. . ..
.. .
.. . .
..
. . . C-27 2.7 Make Other Administrative Arrangements.
. .. .
.. .
.. .
. . ..
.. .
. ..
. .. .
..
C-32 2.8 Train Survey Personnel .. .
.. . .
.. .
.. .
.. .
. .. .
. ..
. . ..
.. . .
. ..
.. .
..
C-33 3.0 CONDUCTISUPERVISE THE SURVEY. . . . . . . . . . . . . . . . .
3.1 Select Households in Clusters .
. .
. .. .
. ..
. .. . .
.. . .
.. .
.. .
.. .
3.2 Conduct Interviews . .
. .
. .. .
.. . .
. ..
. .. .
. .. .
.. .
. .. . .
. ..
. .. .
.. .
. 3.3 Investigate ORS Access.
. ..
. . .. .
.. . .
.. .
. .. .
. .. . . .
.. .
.. . .
.. .
.. .
3.4 Monitor Performance of Survey Personnel .
. . .
.. .
.. . . .
.. .
. .. .
. .. .
.. .
. 3.5 CompileData . . . .3.6 Provide Feedback and Solve Problems
.
.. . .
.. . .
. .. . .
.. .
. . ..
. .. .
..
Examples of Problems, Causes, and Remedial Actions .
. .
. . . ..
. . .. .
. . . 4.0ANALYSESURVEYRESULTS...
E4.1 Calculate Rates .
. .
.. . .
. .. .
. .. . .
. .. . .
.. . .
. .. .
. . .. . .
.. .
. ..
. ..
E-l 4.2 Calculate Limits of Precision .. . .
. . ..
. . .. . .
.. . .
. . .. . .
.. .
. . .. .
.. .
. E-l lWorksheet for Calculating Limits of Precision . . . .
.
. . . .. .
. . .. .
. . ..
. .. .
E-13 4.3 Interpret Survey Results. . . .
. .. . .
.. . . .
.. . .
.. . . .
. . .. .
. . .. . .
.. .
. . E-15 4.4 Report Survey Results . .. . .
. . .. . .
. . .. .
..
.. .
. . .. . .
. .. . .
. .. .
. . ..
E-17 ANNEXESCDDIARI Core Household Survey Questionnaire
.
. ..
.. .
. . . ..
..
. . ..
. . .. .
. F Question-by-Question Explanation of Core Questionnaire .. . .
. . .. .
. . . .. . .
. .. .
. G ORS Access Investigation and Question-by-Question Explanation. . .
. .. . .
. . .. . .
..
H Forms Used in Compiling and Analysing Core Survey Data. . .
. .. .
. .. . .
. .. . .
..
I Green Drug Option (Cough) . . ..
. .. .
. . .. . .
. . . .. .
. ..
. . . .. .
. .. .
J. . .
Yellow Drug Option (ANA) K
. . .
Pink Drug Option (Diarrhoea) L
. . .
Breastfeeding Option M
Ageoption
. . .
NGenderoption
. . . .
P. . . .
Instructions for the Drug Analyst Q
. . . .
Alphabetical Drug List 4-16
Converting 2-Week Diarrhoea Prevalence to 2-Week Incidence and
. . . .
Estimating Annual Incidence Rate from 2-Week Incidence R
. . . .
Random Selection S
. . . .
Calculating Sample Size for Follow-up Surveys T
. . .
Calculator Programmes for Limits of Precision U
Coordinator's Checklist and Time Chart, Sample Supervisors' Checklist,
. . .
Sample Surveyors' Checklist V
. . . .
Training Surveyors and Supervisors W
. . .
Notes and Exercises, Core Survey W-6
. . .
Tips on Good Interviewing Technique W-10
. . .
Sample Training Agenda, Core Survey W-67
. . .
Notes and Exercises, Survey with Options W-73
. . .
Sample Training Agenda, Survey with Options W-109
Directory of Software to Support Survey and
. . .
Instructions on Using Lotus Spreadsheets X
Glossary
. . .
YBlank CDD/ARI Household Survey Questionnaire
Spreadsheets
National programmes to control diarrhoea1 disease and acute respiratory infections (CDD and ARI programmes) need information on the population's knowledge about the diseases,
careseeking practices, and treatment practices in order to effectively plan, manage, and evaluate activities. Some of this information can be obtained through reports from health facilities. However, cases seen in health facilities represent only a fraction of actual cases occurring in the community. Many cases may be treated at home or get well without
treatment. Some cases will get worse and die at home without coming to the attention of the health system. Therefore, it is useful to visit households to ask about illnesses, types of care sought, and treatment given. A methodology for a sample survey of households is presented in this manual.
This household survey can be used in the following ways:
- to establish a baseline against which changes in the population's knowledge, careseeking practices, and treatment practices can later be measured,
-
to measure progress and identify problems in existing CDD and ARI programmes by measuring key WHOIUNICEF programme indicators outlined in the following table, and by measuring other indicators and rates summarized in Section B.For the household survey to contribute positively to your national programme, it is essential that you plan the survey with a clear idea of what rates you need in order to evaluate
programme activities, identify problems, and make decisions about subsequent programme activities.
Carefully conducted periodic surveys of this type will provide reliable information on long- term trends to determine the impact and future direction of a CDD or ARI programme.
Follow-up surveys should be done when sufficient activities have taken place to expect detectable changes in knowledge, careseeking practices, or treatment practices.
The following indicators are among the key child survival programme indicators agreed upon by WHO and UNICEF from the UNICEF Child Survival Monitoring Report. Additional important indicators measured by this survey are presented in Section B of this manual. See the Evaluation modules of the CDD and ARI Programme Management courses for further information on indicators useful to these programmes.
Diarrhoea
Acute Respiratory Infection ( A W
Caretaker knowledge of home case management Use of ORT (increased fluids) plus continued feeding Access to oral rehydration salts (ORS)
Caretaker knowledge of when to seek care for ARI Careseeking from appropriate providers for ARIs
needing assessment (i.e., having fast or difficult breathing)
W
HYC
ONDUCTC
OMBINEDS
URVEYS OFDIARRHOEAL
ANDA
CUTER
ESPIRATORYD
ISEASES?
CDD and ARI programmes both have key programme indicators which require a household survey for measurement. For example, CDD programmes must survey households in order to measure caretakers' knowledge of home therapy for diarrhoea. ARI programmes must survey households in order to measure caretakers' knowledge of when to seek care for A N . Both CDD and A N programmes are primarily concerned with the same target population, that is, children less than 5 years of age.
Often CDD and ARI programmes are within the same division in the Ministry of Health and share personnel and other resources. Using a combined survey can help conserve these resources. Surveys often mean that 20
-
30 health workers must be absent from their usual clinical jobs for several weeks to participate as surveyors. Combining the surveys means these absences will occur less often.This manual provides a combined survey instrument which measures rates important to both CDD and ARI programmes.
PURPOSE OF THIS MANUAL
This manual is written for coordinators of CDDIARI household surveys. By "coordinator,"
we mean the person responsible for planning the survey, selecting and training survey personnel, overseeing the survey, and analyzing the results. If these functions are divided among a number of people, then each person will need to study the relevant sections of the manual.
In some countries the CDD or ARI programme manager or the maternal and child health evaluation officer may serve as the survey coordinator. In any case, the relevant programme managers should be responsible for the initial decisions about the rates to measure in the survey, the sections of the survey to include, and the timing and geographic scope of the survey.
Here are some examples of how this manual can help the survey coordinator. This manual:
-
describes all the steps involved in planning and conducting a household survey. These steps are shown on the flowchart on the next page. Sections B, C, D, and E of the manual correspond to the major steps 1.0, 2.0, 3.0 and 4.0 on the flowchart.- provides a core survey instrument for measuring rates most important to ARI and CDD programmes.
-
describes options for expanding the core survey instrument to obtain breastfeeding rates, drug use rates, and other rates which may be needed by some programmes.- provides worksheets for determining sample size.
- provides an extensive section on training of surveyors and supervisors, including agendas, exercises, training notes, and pages which can be photocopied and given to survey supervisors and surveyors to study during their training.
-
includes forms for compiling data, calculating rates and limits of precision, and analysing survey results.-
provides a directory of supporting software available from the Division for the Control of Diarrhoea1 and Acute Respiratory Disease (CDR) at WHO.- includes a checklist and time chart of tasks to be done by the survey coordinator, as well as examples of checklists for surveyors and supervisors.
Before beginning study of the manual, the survey coordinator may wish to refer to the Coordinator's Checklist and Time Chart in Section V for an overview of the survey process and the time required for planning and preparation.
SECTION B:
PLAN THE SURVEY
Flowchart: Conducting CDDIARI Household Surveys
1.2 1.3
Decide on Decide 1.4 1.5
On --)Determine + prepare
survey the time sample budget.
coordinator. Seaions period size. 8-29
6-1 of survey to be 8-20
to include, covered
and by the
geographic survey scope of 8-79 SUNey.
6-1
SECTION C:
2.0 PREPARE TO CONDUCT THE
Adapt and Pretest Select translate survey communities
+ questionnaire + forms and +where -
and other revise clusters survey forms them as will be as necessary necessary, located.
C- 1 C-1 l C-14
Page references are in italics below each step.
2 4 Prepare information and +guidelines
needed for selection of households m clusters.
C-20
2.5 Select Survey
9 personnel.
C-23
2.7
+Make administrative arrangements other
33
C-32
SECTION D:
2.8 L
Train survey 'personnel.
C-33
3.0 CONDUCT1 SUPERVISE THE SURVEY
SECTION E:
Select Conduct
L
4.1 4.2 4.3 4.4households interviews. Calmlate Cakulate Interpret
-
Reportin clusters. limits of survey survey
D 1 E-l precision. results. results
Compile E-1 1 E-15 E-17
data.
Investigate 3 0 R S
access.
D 2
3.4 3.6
+Monitor Provide performance - j
::Fck
Of survey
personnel. solve D 2 problems.
D 1 4
SECTION B
1.0 PLAN THE SURVEY
The survey coordinator is responsible for planning the survey, selecting and training survey personnel, overseeing the survey, and analysing the results. If any of these functions are delegated to other people, the survey coordinator is responsible for preparing those people to do their jobs and supervising their work. A checklist of tasks that the coordinator must do or oversee, and a timetable for those tasks, is given in Section V.
The survey coordinator should be very familiar with the national CDD andlor ARI programmes and be in a position to commit programme resources. If possible, he or she should have experience in conducting sample household surveys such as previous diarrhoea household surveys. If possible, the survey coordinator should be familiar with the use of a computer and be able to enter data on a Lotus spreadsheet; if not, tasks requiring these skills may be delegated to another person.
D
ECIDE ONR
ATES TO MEASURE,S
ECTIONS OFS
URVEY TOINCLUDE,
ANDG
EOGRAPHICS
COPE OFS
URVEY Rates to MeasureThe core questionnaire provided in Section F yields rates which will be useful to most CDD and ARI programmes. This core questionnaire includes four sections
corresponding to the different topics being investigated. Each section is a different colour:
White - questions asked at all households with children under age 5 about
symptoms experienced in the past 2 weeks, caretaker knowledge of when to seek care, and knowledge of how to treat diarrhoea at home
Green - questions about careseeking asked for children who had illness with cough in the last 2 weeks
Yellow - questions about careseeking asked for children with ARIs needing assessment (hereafter called ANAs*) in the last 2 weeks
Pink - questions asked for children with diarrhoea in the last 2 weeks about careseeking for diarrhoea, treatment of diarrhoea, use and preparation of ORS and recommended home fluids, and sources of ORS
Carefully study the core questionnaire and the core rate list provided in this section of the manual (pages B-4 through B-10). In most countries it will be preferable to include all core sections of the questionnaire. This will maintain the integrity of the questionnaire and yield the most data for the expense. Also, it does not take significantly longer to conduct the survey using all sections than it would with some sections omitted.
Optional Sections of the Survey
In addition to the core questionnaire, there are a number of optional questionnaire sections which countries may include if needed. These options, provided in Sections J-P, include:
- Drug Use options for - cough
- ANA, and - diarrhoea
- Breastfeeding option
- Age option
- Measles option
- Gender option
The rates measured by these optional sections are listed on pages B-l1 to B-18. In deciding whether to include options, remember to avoid overburdening the surveyors, respondents, and analysts with too many optional sections. Most programmes do need to measure all of the optional rates. If they did measure all of the optional rates, most programmes would not
m
the information. This is wasteful, since incorporating options requires additional preparation, longer time interviewing in the home, and extra work in the tabulation, analysis, and interpretation of rates. Only incorporate an optional section if your programme really needs and will use the resulting rates. For many ARI and CDD programmes, only the core questionnaire is needed.*ANA is an acronym for an Acute Respiratory Infection (ARI) Needing Assessment. In this survey children are classified as ANAs based on the caretaker's report of fast or difficult breathing, which are common symptoms of pneumonia. ANAs need to be taken to a health worker who can examine them and determine if they need treatment for pneumonia. Some ANAs will be found to have pneumonia, and some will be found to have a simple cough or cold. It is not possible, by caretaker recall of an episode of disease, to determine if a child had pneumonia, but it is possible to determine if the child had an ARI episode with fast or difficult breathing, an ANA. The classification ANA includes almost all children who had pneumonia, plus some who did not. The rates for ANAs should be interpreted with this in mind.
Many optional rates are interesting but not
useful
for national programme evaluation and planning. Use only the optional sections that your programme really needs. Think through how you will use the information before collecting it.If more than one programme is sponsoring the household survey, the programme managers need to sit together and decide on a manageable number of questions to include.
Occasionally a programme may wish to add questions or do extra analysis using existing questions. It is important to plan these additions well in advance so that they are included in pretesting and so that the forms for data compilation and analysis can be adapted. This adaptation is a demanding task.
Geographic Scope of Survey
The survey method described in this manual is designed to estimate rates for the whole area from which the sample is drawn. It will not provide estimates for subsections of that area.
If you wish to obtain national estimates of rates, the survey sample should be drawn from the population of the whole country. In many countries, however, a national survey will be impractical or unnecessary. The alternative is to select a region or regions of the country in which to conduct the survey. A region may be selected because:
- it is considered more or less typical of the country as a whole,
- it is where CDD or ARI programme impact is expected to be greatest, or - it is a priority region for the country.
In countries where CDD or ARI programme impact is expected to vary greatly from region to region, it may be best to do separate surveys in a number of regions so that differences will be clear in the survey results.
In deciding whether to conduct a national survey, a regional survey, or several regional surveys, you should consider first the CDD and ARI programmes' needs and then the economic and logistical constraints. Remember that it will take more resources to train and supervise survey personnel who are widespread geographically.
Remember:
Training and supervision of survey personnel are critical.
Ensure that the scope of the survey is manageable.
RATE
LIST: CDDIARI HOUSEHOLD SURVEYRates obtained through the core sections of this survey are based on information from the 2-week period prior to the survey. Mothers or other caretakers of children less than 5 years old are asked questions about the occurrence, careseeking for, and treatment of diarrhoea and ARI in "the past 2 weeks." The reason for this is that family members are more likely to remember episodes of illness if asked about a short, recent time period.
You will notice that there are more CDD rates measured in this survey than ARI rates.
The reason is that more of the CDD programme indicators must be measured in the home, because appropriate home treatment of diarrhoea can contribute significantly to reduction of diarrhoea1 mortality. For ARI, on the other hand, the most important indicators measured in the home are caretaker knowledge of when to seek care, and careseeking practices when a child develops signs of possible pneumonia (i.e., fast or difficult breathing). A child with an AFU with fast or difficult breathing is considered to have an ANA a R I needing assessment).
There are no indicators for home treatment of ARI, as treatment of ANAs must be sought from health care providers outside the home.
Rates marked by a
*
are key WHOKJNICEF programme indicators. Formulae for all core A N and CDD rates are given in Section I on the form titled "Summary of Survey Results."Core ARI Rates (Obtained from Core White, Green, and Yellow Pages)
*l. Caretaker knowledge of when to seek care for ARI
-
This is the proportion of caretakers1 who know when to seek care from a health worker for a child with cough.In order to count as having correct knowledge, caretakers must mention at least one of the following signs: fast breathing, difficult breathing, or a local term for fast or difficult breathing or pneumonia. These are the key signs to recognize to assure antibiotic treatment early in the course of pneumonia.
2. Caretaker knowledge of a specific breathing-related reason for careseeking
-
Three rates may be calculated: the proportions of caretakers who specifically mention fast breathing, difficult breathing, or a local term as a reason for careseeking for cough. Comparing these three rates can help programmes judge how well their health education messages are conveying each of these key reasons for careseeking.
3. Caretakers who consider fever a reason for careseeking
-
This is the proportion of caretakers who mention fever as a reason for careseeking for cough. This rate will help the programme understand other symptoms that caretakers focus on when deciding to seek care.'caretakers are the adults who are responsible for children in the household. In most cases, the caretaker interviewed is the mother.
Core ARI Rates, continued
. . .
Careseeking for ARI needing assessment (ANA) outside the home - This is the proportion of ANAs for whom care was sought outside the home. ANAs are cases of acute respiratory infection needing assessment, i.e., having fast or difficult breathing.
This rate measures whether caretakers, after recognizing that a child had fast or difficult breathing, sought care outside the home. This rate includes all careseeking for ANAs outside the home, regardless of the quality of the provider from whom care was sought. This rate does not include careseeking for a child who had just a cough or cold without fast or difficult breathing, or careseeking for an ANA before fast or difficult breathing developed.
Careseeking for ARI needing assessment (ANA) from appropriate providers - This is the proportion of ANAs for whom care was sought from appro~riate providers. This is the most important ARI indicator measured by the survey, as it suggests how likely ANAs are to have received appropriate care. Appropriate providers are those who have been trained in standard ARI case management and supplied with appropriate antibiotics, or other providers expected to deliver good case management. This rate does not include careseeking for a child who had just a cough or cold without fast or difficult breathing, or careseeking for an ANA before fast or difficult breathing developed.
Careseeking for ARI needing assessment (ANA) from appropriate providers
first
after development of fast or difficult breathing
-
This is the proportion of ANAs for whom care was soughtfirst
from an appropriate provider after fastldifficult breathing developed. The emphasis here is on the word "first." This rate can help programmes estimate whether careseeking from an appropriate provider is fast enough. It is important to seek care for possible pneumonia from an appropriate provider quickly. Visiting other providers first can cause significant delay.Careseeking for ARI needing assessment (ANA) from a specific type of provider
-
This is the proportion of ANAs who sought care from a specific type of provider.
Rates for 11 different types of providers may be calculated. Comparing these rates will inform programmes about which providers are most frequently visited for ANA.
This may help guide decisions concerning training and other activities to promote improved ARI case management. These rates for ANA may be compared to similar rates for cough (see 9 below).
Careseeking for cough outside the home - This is the proportion of children with simple cough for whom care was sought outside the home. If the programme is doing a good job teaching mothers the signs of ANA (fast or difficult breathing), this rate should be lower than rate 4. In other words, parents should be more likely to seek care outside the home for ANA than for simple cough.
Careseeking for cough from a specific type of provider
-
This is the proportion of children with simple cough for whom care was sought from a specific provider.Rates for 11 different types of providers may be calculated, compared, and used as a rationale for where to focus efforts for improving case management.
Core CDD Rates (Obtained from Core White and Pink Pages)
2-week diarrhoea prevalence - This is the proportion of children under age 5 who had diarrhoea at any time in the 2-week period prior to the survey. This prevalence rate may be roughly converted to a 2-week diarrhoea incidence rate (proportion of
children with diarrhoea starting in the 2-week period prior to the survey) and to an annual incidence rate as described in Section R. Changes in incidence rates can show the impact of interventions intended to prevent diarrhoea.
Caretaker knowledge about increasing fluids - This is the proportion of all caretakers of children under age 5 who know the first rule of home case management for
diarrhoea, which is to increase fluids.
Caretaker knowledge about continuing food - This is the proportion of all caretakers of children under age 5 who know the second rule of home case management for diarrhoea, which is to continue giving food. Caretakers who say that a child with diarrrhoea should be given about the same amount or more food than usual are counted as having correct knowledge.
Caretaker knowledge about careseeking for diarrhoea - This is the proportion of caretakers of children under age 5 who know the third rule of home case
management, which is to take the child to a health worker if he or she develops any one of seven signs (many watery stools, repeated vomiting, marked thirst, not eating or drinking well, fever, blood in the stool, not getting betterlgetting sickerlvery sick).
Caretakers who can remember and name 2 or more of these 7 signs are counted in this rate as having correct knowledge about careseeking.
Comparing rates 2, 3, and 4 will tell a programme manager whether certain rules of home case management are being communicated more effectively than others.
Caretaker knowledge of 3 rules of home case management - This is the proportion of caretakers who know all 3 rules of home case management described above. This rate is a measure of the programme's overall effectiveness in communicating messages about home case management to mothers.
Careseeking for diarrhoea outside the home - This is the proportion of diarrhoea cases for whom care was sought outside the home. This rate includes careseeking from any type of provider outside the home.
Careseeking for diarrhoea from a specific type of provider - This is the proportion of children with diarrhoea who sought care from a specific type of provider. Rates for 11 different types of providers may be calculated. Comparing these rates will inform programmes about which providers are most frequently visited. This may help guide decisions concerning training and other activities to promote improved diarrhoea case management.
Core CDD Rates, continued
.. .
Proportion of diarrhoea cases that are dysentery
-
This is the proportion of diarrhoea cases that had blood in the stool. Since antibiotics are needed for dysentery, it is important to know what proportion of cases are dysentery. This will help aprogramme project antibiotic needs. It will also help interpret drug use rates, if such rates are obtained. (See rates for the Pink Drug Option for diarrhoea.)
Proportion of diarrhoea that is diarrhoea needing assessment (DNA) - This is the proportion of diarrhoea cases reported to have had any of these signs: many watery stools, repeated vomiting, marked thirst, not eatingldrinking well, fever, blood in the stool, or not getting betterlgetting sickerlvery sick. Children who have diarrhoea with any of these signs should be taken to a health worker or health facility for assessment.
Careseeking for diarrhoea needing assessment (DNA) outside the home - This is the proportion of DNAs for whom care was sought outside the home. CDD programmes encourage parents to seek care for DNAs. The rate includes careseeking from any type of provider outside the home.
ORT use (increased fluid intake) - This is the proportion of diarrhoea cases in the last 2 weeks who actually consumed more fluid during their diarrhoea than they usually consume.
Continued feeding - This is the proportion of diarrhoea cases in the last 2 weeks who actually consumed about the same amount, or more food during their diarrhoea than they usually consume. Note: "about the same amount" includes "slightly less than the usual amount. "
ORT (increased fluid intake) phi continued feeding - This is the proportion of diarrhoea cases in the last 2 weeks who received both increased fluid and continued feeding. Increased fluid means more fluid consumed during diarrhoea than is usually consumed; continued feeding means about the same or more food consumed during diarrhoea than is usually consumed. This rate is a measure of whether the
programme's messages about good case management are being put into practice.
Children offered food 4 or more times daily during diarrhoea - This is the proportion of diarrhoea cases in the last 2 weeks who were offered food 4 or more times daily during diarrhoea. This rate is another measure of whether caretakers are continuing feeding during diarrhoea. The focus here is on how many times food was offered, whether or not it was consumed.
Core CDD Rates, continued
...
ORS use
-
This is the proportion of diarrhoea cases in the last 2 weeks who received ORS solution. Since ORS solution is not needed for every case of diarrhoea, it is hard to say what this rate should be. If it goes up over time, it may mean that ORS is more accessible and acceptable than before (a good thing). If the rate goes down, it may mean that more cases are being effectively treated with home fluids so that ORS is not needed as often (also a good thing). This rate will need to be considered in light of other rates and the likely effects of programme activities that haveoccurred.
16. RHF use - This is the proportion of diarrhoea cases in the last 2 weeks who received one or more recommended home fluids (RHFs). These are fluids recommended by the CDD programme as being especially good for children with diarrhoea (e.g., specific soups, rice water, yoghurt drinks). This is a rate which should go up as messages about RHFs are communicated to the public.
17. ORS andtor RHF use - This is the proportion of diarrhoea cases in the last 2 weeks who received ORS andtor an RHF. This rate should increase over time as the
programme communicates its messages and trains and supplies health workers. (In the earlier version of the CDD Household Survey, this rate was referred to as ORT use, i.e., use of oral rehydration therapy.)
18. ORS use among those who sought care outside the home
-
Of those diarrhoea cases who sought care outside the home, this is the proportion who also received ORS.Most countries have a policy of giving ORS whenever care is sought for diarrhoea.
This rate measures the extent to which children who sought care received ORS.
19. Source of advice to use ORS
-
Of the diarrhoea cases who received ORS, this is the proportion who received advice to use ORS from a specific source. Rates for 12 possible sources can be calculated. Knowing which sources are advising ORS and which are not will help the CDD programme to focus its training and ORSdistribution efforts.
Dispensing source of ORS
-
Of the diarrhoea cases who received ORS, this is the proportion who obtainedlpurchased the ORS from a specific source. Rates for 12 possible sources can be calculated. If the programme feels that it would be more useful to know the dispensing source of ORS (i.e., where the ORS was obtained or purchased) rather than the source of advice to use ORS, they may alter the wording of the questionnaire to obtain these rates instead.Core CDD Rates, continued
...
Correct ORS preparation
-
Of caretakers who have given ORS solution to a child with diarrhoea in the past 2 weeks, this is the proportion who prepare the solution correctly (i.e., using the whole packet and completely dissolving the contents in an appropriate amount of water). To measure this rate, each caretaker who gave ORS in to a child with diarrhoea in the past 2 weeks is asked to prepare ORS solution from a packet like she used.Caretakers who completely dissolved ORS - This is the proportion of caretakers who completely dissolved ORS when preparing ORS solution. When compared with the proportions obtained in 22 and 23 below, this will help programmes know which steps in mixing ORS are easy for mothers and which are difficult. This will help programmes improve their efforts to teach mothers.
Caretakers who used entire ORS packet
-
This is the proportion of caretakers who used the entire packet when preparing ORS solution.Caretakers who used an appropriate amount of water in mixing ORS
-
This is the proportion of caretakers who used the correct amount of water for the packet size when preparing ORS solution.Correct recommended home fluid (RHF) preparation
-
Of caretakers who gave a recommended home fluid (RHF) to a child with diarrhoea in the past 2 weeks, this is the proportion who can prepare the RHF correctly. This rate is obtained only if the programme has an RHF which requires special preparation. If this rate is needed, then the survey can include questions or ask for a demonstration to discern whether caretakers can prepare the fluid correctly. These questions will be designed by the programme and inserted on pink page 5 of the questionnaire.Access to ORS
-
This is the proportion of the population with a regular supply of ORS in their community. This rate is obtained by a separate investigation done by survey supervisors while the surveyors are interviewing at households. (See Section H for more information on measuring this rate.)Other CDD Rates Available with Extm Analysis of the Core Questionnaire Results
Although the next three rates can be obtained from the core questionnaire, they will require some extra preparation in order to summarize the necessary data and calculate the rates, as they are not now included in the data summary forms or the Lotus spreadsheets used in analysis. The formulae for these rates are given in Section I. If your programme needs these rates, it will be necessary to:
- revise the core questionnaire to get cumulative totals for questions W6, P10, and P19, respectively; and
- revise or create new data summary forms to obtain the necessary numerator and denominator data.
26. Caretaker knowledge of specific signs to seek care for dianhoea - This is the proportion of caretakers of children under age 5 who know a specific sign of diarrhoea needing assessment (DNA): many watery stools, repeated vomiting, marked thirst, not eatingldrinking well, fever, blood in stool, or not getting better/
getting sickerlvery sick. Separate rates could be obtained for each sign to show which signs of DNA caretakers most often knew.
27. Cases given specific fluids other than ORS - This is the proportion of diarrhoea cases in the last 2 weeks given a specific fluid other than ORS. If desired, totals and rates for any fluids given during diarrhoea could be obtained. These rates might suggest another possibility for a recommended home fluid, or they might indicate a fluid which caretakers should be taught nnt to give during diarrhoea.
28. Average amount of water used with ORS packets of a specific size - For each packet size used in the survey, this is the average amount of water used by caretakers in preparing ORS solution. This average would tell programmes how far off caretakers are with their measuring and whether they tend to use too much water or too little. If averages are calculated for several different packet sizes, a comparison might reveal which amounts caretakers are best able to measure with vessels in their homes.
RATES MEASURED BY OPTIONAL SECTIONS OF THE QUESTIONNAIRE Green Drug Option
-
CoughThese rates describe use of drugs for simple cough or cold (no fast or difficult breathing).
Antibiotics are not needed for simple cough or cold, and ARI programmes hope to reduce inappropriate use of antibiotics for these symptoms. ARI programmes also hope to reduce use of harmful cough and cold remedies, for example, those containing codeine, alcohol, high doses of antihistamines, isoniazid (INH) , or carcinogens. These rates may help programmes identify problems with antibiotic use for simple cough/cold, or use of harmful cougWcold remedies, and focus training or other efforts on providers who are giving inappropriate advice or dispensing inappropriate treatment.
Choosing this drug option, or the drug options for ANA or diarrhoea, requires substantial preparatory work in the classification of drugs commonly used in the area. In many settings, there is already ample evidence that inappropriate antibiotic use is common for children with cough, and no specific harmful remedies have been identified to be targeted by the
programme. In such settings, you will need to decide whether it is worth the effort to document high rates of inappropriate antibiotic use, or to determine the nature and extent of harmful drug use.
Formulae for the drug rates for cough are found on page 5-7. All of these rates are for drug use for simple cough in the 2 weeks prior to the survey.
1. Antibiotic use for cough
-
This is the proportion of children with a simple cough or cold (no fast or difficult breathing) who were inappropriately given an antibiotic.Appropriate antibiotic use for ear infection, skin problem, etc. is excluded from the numerator.
2. Harmful drug use for cough - This is the proportion of children with simple cough or cold (no fast or difficult breathing) who were given a harmful drug. Before the survey, drugs commonly used for ARI in the area are categorized as harmful or harmless (as described in Section Q, "Instructions for the Drug Analyst"). Any injection given for a simple cough or cold is also considered potentially harmful.
3. Source of advice to use antibiotic for cough
-
This is the proportion of antibiotics given for cough which were advised or prescribed by a specific provider. Rates can be calculated for 12 possible providers. Knowing the most frequent sources of inappropriate advice may help the ARI programme decide where to focus efforts to reduce antibiotic use for cough.or
-
Dispensing source of antibiotic for cough - This is the proportion of antibiotics given for cough which were dispensed by a specific provider. If the programme feels that it would be more useful to know the dispensing source of the antibiotic (i.e., where it was obtained or purchased), rather than the source of the advice or prescription, the questionnaire may be altered to obtain these rates instead of the above.
m e n (Cough) Drug Option Rates, continued
4. S o m e of advice to use hannful drug for cough
-
This is the proportion of harmful drugs given for cough which were advised or prescribed by a specific provider.Rates for 12 possible providers may be calculated, compared, and used as a rationale for where to focus programme efforts.
Qt
Dispensing s o m e of hannful drug for cough - This is the proportion of harmful drugs given for cough which were dispensed by a specific provider. If the
programme feels that it would be more useful to know the dispensing source of the harmful drug, they may alter the wording of the questionnaire to obtain these rates instead of the above.
5. Drugs given for cough that could not be identified
-
This is the proportion of all drugs given for cough that could not be identified. This rate is important to help interpret rates 1 and 2 (antibiotic and harmful drug use rates). In rates 1 and 2 unidentified drugs are not counted in the numerator. If a high proportion of drugs cannot be identified, the rates of inappropriate antibiotic use and harmful drug use for cough could be underestimated.Yellow Drug Option
-
ANA (ARI Needing Assessment)Antibiotic use may be appropriate for ANA (ARI needing assessment because of fast or difficult breathing), as some of these children will be found on examination to have
pneumonia. However, there are some drugs that are harmful for children with ANA as well as for children with cough, such as drugs containing codeine, alcohol, high doses of
antihistamines, isoniazid (INH), or carcinogens. The ARI programme may make efforts to reduce use of specific harmful drugs.
Formulae for the drug rates for ANA are found on page K-7. All of these rates are for drug use for ANA in the 2 weeks prior to the survey.
1. Harmful drug use for ARI needing assessment (ANA)
-
This is the proportion of children with ANA who were given a harmful drug. Before the survey, drugs commonly used for ARI in the area are categorized as harmful or harmless (as described in Section Q, "Instructions for the Drug Analyst").2. Source of advice to use harmful drug for ARI needing assessment (ANA)
-
Thisis the proportion of harmful drugs given for ANA which were advised or prescribed by a specific provider. Rates for 12 possible sources of advice or prescriptions may be calculated and compared. Knowing the most frequent sources of harmful advice may help the ARI programme decide where to focus its efforts to reduce harmful drug use.
or
-
Dispensing source of harmful drug for ANA
-
This is the proportion of harmful drugs given for ANA which were dispensed by a specific provider. If theprogramme feels that it would be more useful to know the dispensing source of the harmful drug, the questionnaire may be altered to obtain rates for the dispensing source instead of the source of advicelprescription.
3. Drugs given for ANA that could not be identified
-
This is the proportion of all drugs given for ANA that could not be identified. This proportion is important to help interpret rate 1 (harmful drug use for ANA). In rate 1 unidentified drugs are not counted in the numerator. If a high proportion of drugs cannot be identified, the rate of harmful drug use for ANA could be underestimated.Pink Drug Option
-
DiarrhoeaMost drug use for diarrhoea is inappropriate. Antibiotics should only be used for dysentery and cholera. Other drugs, such as antidiarrhoeals, do not help and may actually harm the child. The diarrhoea drug rates are intended to help programmes identify drug use problems and focus their efforts to solve them.
Formulae for diarrhoea drug rates are found on page L-8. All of these rates are for drug use for diarrhoea in the 2 weeks prior to the survey.
Drug use for diarrhoea- This is the proportion of children with diarrhoea who received any type of drug (pill, syrup, capsule, or injection) for the diarrhoea. Herb or plant remedies are not counted as drugs, nor is ORS counted as a drug in this context. Since most drug use for diarrhoea is inappropriate, this rate should be low or should go down over time as the programme delivers its messages.
Source of advice to use drugs for diarrhoea
-
This is the proportion of drugs given to children with diarrhoea which were advised or prescribed by a specific provider.Rates for 12 possible sources of advice or prescriptions may be calculated and
compared. Knowing which sources are advising drugs will help the CDD programme to focus its efforts to reduce inappropriate drug use.
or
-
Dispensing source of drugs for diarrhoea - This is the proportion of drugs given to children with diarrhoea which were dispensed by a specific provider. If the
programme feels that it would be more useful to know the dispensing source of the drug (i.e., where it was obtained or purchased), rather than the source of the advice or prescription, the questionnaire may be altered to obtain these rates instead of the above.
Antibiotic use for diarrhoea- This is the proportion of diarrhoea cases who received antibiotics for their diarrhoea. Antibiotics given for ear infection, skin problem, or pneumonia are excluded from the numerator. Of the remaining antibiotic use, some may be appropriate (i.e., for cholera and dysentery) and some not appropriate.
Dysentery cases given an appropriate antibiotic - This is the proportion of diarrhoea cases with blood in the stool who were treated with an appropriate antibiotic.
Antidiarrhoeal use - This is the proportion of diarrhoea cases who received
antidiarrhoeals. Since antidiarrhoeals are never needed and may be very harmful, it is hoped that this rate will go down over time as the programme implements activities to reduce inappropriate drug use.
Antiprotozoal use - This is the proportion of diarrhoea cases who received
antiprotozoals (other than antimalarials). Antiprotozoals are frequently misused in the treatment of childhood diarrhoea.
Pink (Diarrhoea) Drug Option Rates, continued
7. Drugs given for diarrhoea that could not be identified - This is the proportion of drugs given for diarrhoea that could not be identified. This proportion is important to help interpret rates 3 - 6. If there is a high proportion of unidentified drugs, then rates 3 - 6 could be underestimated.
8. Use of 1,2,3, or 4+ drugs
-
Four rates may be calculated: the proportion ofdiarrhoea cases who received 1, 2, 3, or 4 or more drugs. These rates will help the programme determine how many drugs are typically being given to a child who receives drugs for diarrhoea. If multiple drugs are typically being given, this is an even more serious problem than single drug use. These rates will also give the programme an idea whether families are spending money inappropriately on drugs for diarrhoea.
Breastfeeding Option
Many epidemiological studies have demonstrated the benefits of breastfeeding in reducing diarrhoea and pneumonia in infants. Breastfeeding indicators can be used to assess
breastfeeding practices and evaluate the progress of promotional programmes. All of the indicators below are measured using a 24-hour recall period; in other words, caretakers are asked what their children drank or ate "since this time yesterday." Only caretakers of children less than 2 years old (0-23 months) are asked questions about breastfeeding.
Rates marked by a
*
are key WHOIUNICEF breastfeeding indicators. Further information on these and other breastfeeding indicators can be found in the document titled "Indicators for Assessing Breastfeeding Practices in Households" WHO/CDD/SER/91.14 revised.Formulae for breastfeeding rates are found on page M-8.
1. Any breastfeeding in children less than 4 months old+- This is the proportion of children in this age group who receive any breastmilk at all, whether or not they are given other fluidslfoods as well.
*2. Exclusive breastfeeding in children less than 4 months old
-
This is the proportion of children less than 4 months old who are exclusively breastfed. Exclusivebreastfeeding means that the child receives only breastmilk and no additional fluidslfoods (with the exception of vitamins, mineral supplements, or medicines, if needed). CDD programmes recommend that all children be exclusively breastfed until 4-6 months of age.
*3. Predominant breastfeeding in children less than 4 months old
-
This is the proportion of children under 4 months old whose main source of nourishment is breastmilk, but who also receive other fluids. The fluids allowed are water,sweetened or flavoured water, fruit juices, tea, bottled soft drinks, or ORS solution.
No infant formula, animal milk, or food-based fluids are allowed if the infant is to be considered predominantly breastfed. Thus, the child is not counted as predominantly breastfed even if he or she receives only one infant formula feed during the day. As the proportion of infants exclusively breastfed may be quite low, this measurement of predominant breastfeeding is needed to give a more complete picture of breastfeeding practices.
"4. Timely complementary feeding in children 6-9 months old + - This is the proportion of children 6-9 months old who are receiving breastmilk and complementary foods.
*5. Bottle-feeding rate in children less than 12 months old' - This is the proportion of infants less than 12 months old who are receiving any food or drink from a bottle with a nipple or teat.
+ In all breastfeeding rates, less than 4 months (0-3 months) equals 0-120 days; 6-9 months equals 182-303 days; and less than 12 months (0-11 months) equals 0-365 days.
Breastfeeding Rates, continued
*6. Continued breastfeeding in children 12
-
15 months old-
This is the proportion of children 12 - 15 months old who are still breastfed to any extent. Some programmes promote breastfeeding for "one year or longer" while some programmes promote breastfeeding for "at least 2 years." This rate gives an indication of breastfeeding beyond 1 year.*7. Continued breastfeeding in children 20
-
23 months old-
This is the proportion of children 20-
23 months old who are still breastfed to any extent. For programmes promoting breastfeeding for "at least 2 years," this rate gives an indication of breastfeeding in children almost 2 years old.Age Option
The Age Option adds a question to the survey to determine each child's age. This gives programmes the potential to recalculate all rates obtained in the survey by age. However, this additional analysis would require a great deal of work and would probably not be very useful. The main reason that the Age Option is offered is to determine if careseeking rates for young infants indicate a problem.
In many cultures, young infants 0-1 month old(i.e, less than 2 months old) are less likely to be taken outside the home for care, even when they have signs suggesting a serious illness such as ANA. 30% of ARI deaths occur in this age group, so it is useful to know how careseeking in this age group compares to careseeking for ANA in older children. The age groups of interest are 0-1 months, 2-1 1 months, and 1-4 years. By comparing careseeking for these age groups, the programme can see if the careseeking rates for young infants indicates a problem requiring special programme efforts.
Formulae for rates obtained from the Age Option are given on page N-8.
1. Appropriate careseeking for ANAs by age group (3 rates) - This is the proportion of ANAs in each age group who were taken to an appropriate health care provider (i.e., one who is trained and supplied with appropriate antibiotics). Rates may be obtained for the three age groups mentioned above.
(Note: the Age Option is not needed if the Breastfeeding Option is also being used, as the Breastfeeding Option includes a question about the child's age.)
Measles Option
In some cultures, infants and young children with measles are kept at home and not taken for care, even when they have signs suggesting a serious illness such as an ANA (ARI needing assessment). Comparing careseeking rates for ANA for children with and without measles in the past month could alert a programme to this problem, which could be
addressed through special health education messages, particularly during measles epidemics, or through other interventions to improve careseeking in children with measles.
Only use the Measles Option if measles cases have been occurring recently in the
communities surveyed, and if there is a reason to suspect that careseeking may be poor for children with measles.
Formulae for rates obtained from the Measles Option are given on page 0-6.
1. Appropriate careseeking for ANA with recent measles - Of children with ANA who also had recent measles, this is the proportion for whom care was sought from appropriate providers (trained and supplied with antibiotics).
2. Appropriate careseeking for ANA without recent measles
-
Of children with ANA without recent measles, this is the proportion for whom care was sought fromappropriate providers (trained and supplied with antibiotics). This rate should be compared to rate 1 to see if recent measles affects careseeking for ANAs.
Gender Option
The gender of the child may be recorded if the programme is interested in comparing careseeking rates or other rates for girls and boys. In some cultures girls are taken for health care less frequently than boys, and their illnesses may be treated differently at home.
If this type of problem is found, it may be addressed through special health education messages. The rates for comparison between boys and girls can be defined, tabulated, and analysed in the same way as other rates provided in this manual.
1.3 D
ECIDE ON THE TIME PERIOD TO BEC
OVERED BY THES
URVEYIn most countries diarrhoea prevalence is greater during some times of the year than others. This means that during some seasons it will be easier to find the required number of diarrhoea cases than others. Coughs are likely to be common all year, but pneumonia (and ANAs) may occur more often in particular seasons. The timing of the survey should thus be decided based on diarrhoea and pneumonia seasonality.
If possible, the survey should be conducted at a time when the pneumonia and diarrhoea seasons overlap. If the pneumonia and diarrhoea seasons do not overlap, the survey should be conducted during the high season of the disease with lower expected prevalence.
Of course, if there is no seasonality to one of the diseases, either diarrhoea or pneumonia, the survey should be conducted during the high season of disease which does occur in seasonal patterns. If neither disease has seasonality, it does not matter when the survey is conducted.
Follow-up surveys should always be carried out at the same period of the year as the initial survey.
What Can We Conclude From a Sample Survey?
The only way to measure a rate with total accuracy is to survey the entire population.
Since this is impractical, you can survey a sample of the population to obtain an estimate of the rate. This estimate will not be an exact measurement, but a range in which we can be reasonably confident the actual rate falls. In this manual the upper and lower limits of this range will be called the "limits of precision." For example, an estimate of the ORS use rate might be 0.30 with limits of precision of plus or minus (+) 0.10. This would mean that we can be reasonably confident that the actual rate is between 0.20 and 0.40.
But what do we mean by "reasonably confident"? In this manual, statistical theory has been applied with a view to ensuring that there will be only a 5 % chance that the actual rate in this population is outside of the range determined by the survey. In other words, we can be 95 % confident that the rate in the population lies within the determined range.
When comparing two estimates of a particular rate, from initial and follow-up surveys, a similar principle is applied. If analysis of the survey results determines that the
difference between the two estimates is "statistically significant," this means that there is only a 5% chance that the difference is due to chance variation associated with the
sampling. In other words, we can be 95 % confident that the difference detected between the two surveys is real.
If, on the other hand, analysis of the survey results shows that the difference between the two rates is not statistically significant, this does not necessarily mean that no change has occurred. It simply means that there is a chance greater than 5 % that the difference is due to chance variation associated with the sampling. Thus, we are less than 95%
confident that the difference between the two rates is real.
Calculating Sample Size
To determine the sample size required to estimate a particular rate, you need to know:
- a rough estimate of the rate, and
- the change you want to measure in the rate at a subsequent follow-up survey, or the level of precision required.
Measuring less frequent events will require larger sample sizes. For example, it takes a larger sample sue to measure careseeking for ANAs than careseeking for cough because ANAs occur less frequently. Also, the greater the precision required, the larger the sample s u e will need to be. Thus, detecting a small change in a rate between two surveys may require a large sample size.
The next pages describe how to determine the required sample size for an i&id survey and provide an example of the process. Section T describes how to calculate the required sample size for a follow-up survey. If you are interested in measuring progress since a past survey, you should use Section T instead of the following pages.
DETERMINING SAMPLE SIZE FOR AN INITIAL CDDIARI SURVEY*
In this survey you will seek a number of different rates (for example, an ORS use rate, a correct ORS preparation rate, careseeking rates for cough, ANA, and diarrhoea). To determine the sample size for such a survey, you will need to determine which of the rates requires the largest sample size. To do this, you will need to consider the following for selected rates:
- a rough estimate of the current rate,
-
the change you want to measure in the rate at a subsequent follow-up survey, or the precision required in this survey's findings, and- the appropriate denominator for the rate.
We have selected certain rates for a worksheet which will yield the highest sample size needed within reason. (The worksheet considers the most important rates, but does not consider rates from the optional sections which would require an extremely large sample size, such as the rate of dysentery cases given an appropriate antibiotic.) An example of a completed worksheet is on page B-27; fold out that page now and study the example as you read the following instructions. After studying the example and instructions, determine the sample size for your own survey using the blank worksheet on page B-25.
Instructions
At the top of the worksheet, enter an estimate of 2-week diarrhoea prevalence during the time of year that the survey will be conducted. (This is the proportion of children less than 5 years old expected to have diarrhoea during a 2-week period at this time of year.) Base this estimate on data from previous surveys, hospital data, ORS distribution records, etc.
Also enter an estimate of 2-week ANA prevalence during the time that the survey will be conducted. This should be somewhat higher than the expected prevalence of pneumonia seen in hospitals and health centres. In many countries the 2-week prevalence of ANA will be about 5 % during the peak pneumonia season.
The steps below correspond to column headings on the worksheet.
1. For each rate listed in the left column, make a rough estimate of the current rate.
(This may involve estimating based on previous surveys, hospital data, health centre records, ORS or antibiotic distribution records, etc.)
2. Decide what change you would be interested in detecting in a future follow-up survey. (This may be smaller than the change predicted in your programme
subtargets.) Express the change as an increase or decrease in percentage points. (For example, if you wanted to detect an increase in a rate from 30% to 50%, that would be an increase of 20 percentage points.)
* If you are interested in comparing results of this survey with a past survey, you may wish to calculate sample size as for a follow-up survey, as described in Section T.
3. Determine what the limits of precision would have to be to detect the change specified in step 2. To do this, divide the change in percentage points by 2.
4. Decide whether the limits of precision found in step 3 above are satisfactory: Do you need a more precise estimate of the current rate for planning or other purposes?
For example, if the expected current rate were 20%, then
+
15 percentage points (that is, 5%-
35%) might be an unsatisfactory range of precision. On the other hand, if the expected current rate were 50% and the limits of precision were+
10percentage points (that is, 40% - 60%), that would usually be satisfactory.
5. If the limits found in step 3 are satisfactory, go on to step 6. If they are too wide, choose narrower limits; usually
f
5% or+
10% will be satisfactory.6. Specify the appropriate denominator for the rate. For example, in an ORS use rate the denominator is children under age 5 who had diarrhoea in the past 2 weeks. In the rate of caretaker knowledge of when to seek care for ARI, the denominator is all caretakers of children under age 5 (since all caretakers should know when to seek care, even if their child has not recently had an ARI). The appropriate denominators for the rates on the worksheet are specified for you.
7. Determine the denominator sample size required to measure the rate with specified limits of precision. To do this, refer to the table on page B-24. Find the row for the estimated current rate and the column for the limits of precision. The intersection of this row and column will tell you the required denominator sample size.
8. Determine the factor needed to convert the denominator sample size to the sample size of children under age 5.
For example, for the ORS use rate on page B-27, the denominator sample size is 576 children under age 5 with diarrhoea in the past 2 weeks. You need to figure out how many children under age 5 you would have to include in your sample in order to find 576 who have had diarrhoea. Since the estimated 2-week diarrhoea prevalence rate is 20% (recorded at the top of the worksheet), the conversion factor is 5 (100120); that is, you would need about 5 children under age 5 to find one with diarrhoea in the past 2 weeks.
For the correct ORS preparation rate, you must calculate the conversion factor for a denominator of caretakers who have given ORS to children with diarrhoea in the past 2 weeks. First you need an estimate of 2-week diarrhoea prevalence (recorded at the top of the worksheet) and the ORS use rate (recorded in the first column). In the example, the estimated 2-week prevalence is 20%, so you would need 5 children (100J20) to find one with diarrhoea in the past 2 weeks. The estimated ORS use rate islO%, so you would need to find10 children who have had diarrhoea (100110) to find one whose caretaker had given ORS. Thus the factor to convert the denominator sample size to the sample size of children under age 5 would be 50 (5 X 10).
