TEMPLATE
Please note: all instructions are written in Italic, square brackets [ ] mark the places where you have to fill-in personalised information for the study
Retrospective mortality survey in the MSF catchment area in [insert in detail where the survey takes place, i.e. region, district/province, country]
Study protocol
[insert date of writing / DD.MM.20XX]
[insert version, i.e. first version or revision after comments from the co-investigators and/or ethical review board]
[insert name of principal investigator, degree i.e. PhD/MSc]
The template of the retrospective mortality survey protocol was prepared by Sibylle Gerstl (epidemiologist) with input from Ruby Siddiqui, Jane Greig, Philipp du Cros and Bhargavi Rao (Manson Unit, MSF-OCA) and Annick Lenglet (Public Health Department, MSF-OCA). This work was informed by reference to several retrospective mortality survey protocols from MSF and Epicentre.
First version [insert date of writing first version / DD.MM.20YY]
Mark every revision (i.e. after comments from the co-investigators and/or ethical review board, such as:
1. Revision [insert date of writing first revision / DD.MM.20YY]
2. Revision [insert date of writing second revision / DD.MM.20YY]
Study design Retrospective mortality survey
Study period [insert month(s) of estimated field part of the study]
Study site Randomly selected villages within catchment area of
MSF in [insert in detail where the survey takes place
i.e. region, district/province, country]
Principal investigator [insert name of principal investigator (insert qualification,
such as epidemiologist) for MSF-[OCX]
Email: [insert email]
Co-investigators Everybody who is involved in carrying out the study
needs to be listed here with name and affiliation
(i.e. John Doe (health advisor), MSF-XX, city, country)
Usually co-investigators are:
(1) the health advisor at MSF headquarters,
(2) the medical coordinator of the mission where the
morality study will be carried out, and
(3) a contact in the Ministry of Health of the country
where the study will be carried out
Data collection and analysis by MSF-[OCX]
Protocol and study design Everybody who is participating in writing the protocol
needs to be listed here with its name and affiliation
(i.e. John Doe (health advisor), MSF-XX, city, country)
This is always the principal investigator and usually one
to two more persons who contributed substantially to the writing of the protocol.
Collaborating institutions Very often the Ministry of Health of the country where the study takes place is a collaborating institution.
Further collaborating institutions could be universities and academic institutions of the country where the study takes place.
CONTENTS
(To update this table please go to the references tab in the WORD (2007) menu and select “update table”)
List of abbreviations 4
1. Introduction 5
1.1. Context 5
1.2. MSF presence in the country 5
1.3. Background - Justification for the study 6
2. OBJECTIVES 7
2.1. Primary objectives 7
2.2. Secondary objectives 7
3. Study design 8
4. Study AREA and period 8
5. Study population 8
5.1. Inclusion and exclusion criteria 9
6. Definitions 9
6.1. Household definitions 9
6.2. Recall period for reported deaths 10
7. SAMPLE SIZE and sampling 11
7.1. Sample size 11
7.2. Sampling 13
8. Data collection 15
9. Data entry and AnalysIS 15
10. Ethical issues 17
10.1. [Insert written or verbal] consent form 18
10.2. Risks and benefits of the study and contingency plans 18
11. Collaboration 19
12. Implementation of the study in the field 19
12.1. Selection and tasks of the study teams 19
12.2. Supervision 20
12.3. Suggested MSF support in the field 20
12.4. Training of the study team and pre-testing of the questionnaires 20
12.5. Timeframe in the field 21
13. LOGISTIC 22
13.1. Supplies needed 22
13.2. Transport needed 22
List of abbreviations
Insert all abbreviations used in the protocol
Generally used abbreviations are:
CMR Crude Mortality Rate
95% CI 95% confidence interval
IPTp Intermittent preventive treatment in pregnancy
MoH Ministry of Health
MSF Médecins sans Frontières
MSF-[OCX] Médecins sans Frontières – Operational Centre [insert MSF section]
U5MR Under 5 mortality rate (mortality rate in children under 5 years of age)
WHO World Health Organization
1. Introduction
1.1. Context
This chapter usually should include:
§ The rank where the country is rated in the human development index [http://hdr.undp.org/en/statistics] [insert year of the statistic] [insert (accessed month dd, 20XX)]
§ Special characteristics of the country relevant for the context of the study; such as ongoing crisis and conflicts, availability of natural resources (e.g. oil, gold, diamonds), information on infrastructure (education, roads, health system), key players present in the country and region
§ Population figures of the country
§ Demographic data such as life expectancy at birth stratified by sex, maternal mortality ratio, under-five mortality rate [http://www.globalhealthfacts.org/data/country/, accessed month dd, 20XX]
Finalise this chapter with a map of the country. Well arranged maps can be found at: http://www.un.org/Depts/Cartographic/english/htmain.htm
Figure 1 [insert map of the country]
1.2. MSF presence in the country
This chapter usually should include:
§ The name of the region where MSF is situated in the country
§ The year and month when MSF started its activities
§ An overview of the different MSF programmes/projects in the region always related to the planned mortality study
§ Statistics of the most recent “Monthly Medical Report” of the MSF project, such as
o Top diseases for admission to OPD and IPD usually stratified by persons aged 5 years and older and children under 5 years of age
o Inpatient mortality data (please note that you cannot directly draw conclusions from the inpatient mortality to the population mortality in the community, as there are different contributory factors to these rates such as quality of healthcare; nevertheless inpatient mortality is interesting to know in the planning of a community mortality survey)
Finalise this chapter with a map of the catchment area of the MSF project.
Normally the catchment area is equivalent to the study area chosen for the planned mortality study.
Figure 2 [insert map of MSF catchment area]
1.3. Background - Justification for the study
This chapter usually should include:
§ Crude mortality rates (CMR) and under-five mortality rates (U5MR) of all mortality studies carried out in the same region in the past, including their references/sources
§ CMR and U5MR of all mortality studies recently (over the last 5 years) carried out in the country, including their references/sources
§ A discussion of the results of these mortality studies and if they have exceeded emergency thresholds; reference rates for CMR and U5MR are listed in the table below
Reference rates for CMR and U5MR / CMR / U5MRIndustrialised countries / 0.3 / -
Developing countries / 0.5 / 1
Stable refugee camps / < 1 / < 2
Severe situation – emergency threshold / >= 1 / >= 2
Very severe situation – critical threshold / >= 2 / >= 4
Table 1: Reference rates for CMR and U5MR. Source: MSF (Médecins Sans Frontières): Rapid health assessment of refugee or displaced populations. MSF-OCF, Paris, France, 2006; UNHCR (United Nations High Commissioner for Refugees): Handbook for Emergencies. 2nd edition. Geneva, Switzerland, 2002.
Moreover, this chapter should provide a problem statement with justification for the need to carry out a mortality study. Justifications could be as follows:
§ Currently there is no regular assessment of community health indicators in the MSF catchment area, such as mortality.
§ It is not clear whether the [insert name of the project] project is seeing and responding sufficiently to the key health issues in the area.
§ There is a lack of data in our area of work, which we want to address with the present study.
Finally, an explanation of the expected use of the results of the study should be given.
Explanations could be as follows:
§ The planned survey will therefore give results on the scale of the health status in the population.
§ It will provide results on retrospective mortality and the likely symptoms / causes of deaths.
§ It will further provide results on [insert results according to supplementary chosen specific objectives (refer to chapter 2.2.)]
§ This information will help MSF to better target its medical programmes and advocacy strategies to the major causes of death.
2. OBJECTIVES
2.1. Primary objectives
§ To estimate retrospectively the crude mortality rate for the total population and for children under five years of age in the MSF catchment area
2.2. Secondary objectives
§ To describe the population surveyed by sex and age
§ To measure crude mortality rate for the total population and for children under five years of age
Depending on the recall period CMR and U5MR could be further stratified, for example by rainy and dry season, harvest and hunger gap periods, measles season, meningitis season.
§ To obtain an indication of the major causes of death, as well as the age and sex distribution of the deceased
To benefit from the study as much as possible it is recommended to add (only) a few specific objectives (and thus a few questions in the questionnaire) that go beyond the mortality part of the survey. The selection of questions should be discussed with headquarters and the field in order to tailor them to the needs of the project.
Common additional topics include proportion of malaria prevention during pregnancies, measles vaccination coverage, coverage and utilization of bednets and health seeking behaviour. Modules on nutrition, violence, health seeking behaviour and malaria with a list of possible questions are provided in the annex part of this document.
The following specific objectives could be added:
§ To determine the proportion of pregnant women (in the 2nd and 3rd trimester, age 15-49 years), in the study population who received intermittent preventive treatment (IPTp) for malaria
§ To determine the proportion of children (6-59 months) in the study population who are vaccinated against measles
§ To determine the proportion of persons (stratified by children under 5 years of age and pregnant women) in a household sleeping under long-lasting insecticide-treated bedNets (LLITNs)
§ To determine the coverage of LLITNs in the households
A list of possible questions on health seeking behaviour and malaria are provided in the annex part of this document. Please do not overload the mortality questionnaire with too many questions. Around 20 minutes is an acceptable duration for interviews.
§ To facilitate improvement of the MSF project activities in the catchment area
3. Study design
Retrospective mortality survey using a [select from below]
a) two-stage cluster sampling methodology as an adaptation of the standardized method recommended by the World Health Organization (WHO)[1] [this methodology is described in Chapter 7.2]
b) simple random sampling, a method by which households are selected by chance (‘out of a hat’) [e.g. used for an area where a list of every household or individual is available or for small target populations in small geographic areas – here exhaustive sampling can also be used but this is rare]
c) systematic random sampling, a method by which households are selected by chance but using a sampling interval and random start point [e.g. when there is a list of households or where the population is geographically concentrated and all the dwellings are arranged in a regular geometric pattern]
4. Study area and period
The study area will be the entire catchment area of the MSF-[OCX] [insert name of project in the country] project.
Please indicate if some areas of the MSF catchment area will be excluded and give the reason for exclusion, such as ‘inaccessible due to security problems’.
You can further insert here a quick overview of the typical climate of the region including annual mean temperatures. If it is relevant for the results of the mortality study, please indicate as well the months of the rainy and the dry season(s).
The recall period will include [include months or seasons of the mortality study] (see also chapter 6.2.)
5. Study population
The study population will consist of all people living in the villages, which are situated in the study area.
Information on current population estimates is based on [insert source, usually the last national population census] data and an estimated annual growth rate of [insert percentage of growth rate for the country].
Usually current population estimates are of limited availability. Very often the last population census on national level was carried out years before the time of the planned study. Therefore the population figures obtained may not entirely reflect the current population in the catchment area - a limitation of the study that should be mentioned in the discussion chapter of the study report.
Always try to get the most recent population figures of the study area and add the annual growth rate for the country. If no annual population growth rate is available for the country, use standard annual growth rates for regions and continents, which can be found at: http://www.unfpa.org/worldwide/africa.html (accessed April 3, 2012). The standard annual population growth rate for Sub-Saharan Africa is 2.2%.
You need to have population figures for all villages in the study area because cluster[2] allocation will be done by systematic sampling with the probability of allocation proportional to the respective population size of each village (see also chapter 7.2.). If there are no population figures available on village level, you need to generate them, ideally by counting the population in the study area. However, mainly estimates of village population figures will be carried out with the help of persons that know the study area very well (i.e. outreach supervisors).
The total population was [insert population figures of the latest available census or any other population figure source] in [insert year of latest available census] [insert reference in brackets]. With an estimated population growth of [insert available growth rate], the current study population is estimated at around [insert current population figures].
If there are no up-to-date data available, the standard population estimations for children under 5 years of age are 20% of the total population, and for pregnant women 10%.
5.1. Inclusion and exclusion criteria
A person will be included in the study if s/he satisfies all of the following criteria: