Survey Manual for Supervisors

Retrospective Mortality Survey in UNITA Resettlement Areas
Bie, Cuando-Cubango, Huila, Malanje Provinces
Angola, August 2002

Context, study justification

With the resumption of the war in late 1998, UNITA fighters, their families, the populations under their control and the populations in the grey zones under FAA control progressively fell beyond the reach of humanitarian assistance. This was the result of insecurity, lack of will on both sides in the conflict, and an UN-imposed ban on relations with UNITA.

With the collapse of UNITA early 2002 this population re-emerged from the mata. MSF observed malnutrition and mortality rates amongst the highest seen in Africa.

This survey tries to trace back mortality to the pre-ceasefire period, when both populations were inaccessible to any kind of external humanitarian assistance. MSF feels that it is important to measure the mortality in the isolated war-affected areas during this time and also measure the real consequences of decisions taken by the international community, aid actors and parties to the conflict.

Objectives

To obtain mortality data of the population in resettlement centres (formerly "quartering and family reception areas") for the period July 2001�August 2002. Thus study population is composed of UNITA members and their families.

The specific objectives are:

To describe the demographic characteristics of the sampled population by age, sex and other key characteristics
To determine crude mortality rate and mortality rates for children less than five years of age for the period July 2001�August 2002
To identify the major causes of death
To assess the consequences of the mortality on the family structure.

Methods

Study population

The study population includes the residents of all nine resettlement camps in the provinces of Bie, Cuando-Cubango, Huila and Malanje. With a registered population of 149,106, the population in these camps comprises of 38% of the total registered population in the 35 resettlement camps country-wide.

Sampling

The survey will be conducted as a three-stage cluster sample with the household as the basic sampling unit. Thirty households in 30 clusters will be selected to achieve a sample size of 900 households �approximately 4500 persons. This sample size is required to estimate mortality with a sufficient precision.

At the first stage, thirty clusters have already been randomly allocated to the nine camps. Allocation was done by systematic sampling with probability of selection proportional to population size (PPS). Camp population figures were taken from the official registration list of all resettlement camps compiled by OCHA (August 2002).

Camp	Pop	Cum Pop.	# Clusters	Cum # clusters
Bie Ndele	8983	8983	2	2
Bie Ngamba I	8524	17507	1	3
Bie Ngamba II	4805	22312	1	4
CC Capembe	42249	64561	9	13
CC Soba Matias	11098	75659	2	15
CC Tchimongua	28094	103753	6	21
H N Galangue I, II, III	30122	133875	6	27
M Damba	8397	142272	2	29
M Ngangasol	6834	149106	1	30

At the second stage, the number of clusters allocated to each camp will be randomly distributed to individual camp sectors (again by systematic PPS sampling). This step will need to be carried out in the field, as a list of sectors and their respective population sizes is only available locally. This second stage will be carried out by the epidemiologist responsible for each of the provinces during the survey.
At the third stage, the standard WHO/EPI methodology will be used to select 30 households in a cluster within a sector:

Identify the geographical centre of the sector which has been selected (market place, church, health centre, etc).
In the centre, spin a bottle (throw a pen) to determine a random direction.
Following this direction, draw an imaginary line from the centre of the sector to its edge. Count all houses along this imaginary line.
Chose a random number between "1" and the total number of houses counted (use random number table provided).
Go back to the house which corresponds to the random number. This house is the first household to be interviewed.
When having completed the interview for the first household, proceed to the household physically closest to the first household. This is the next household to be interviewed.
In case there are two households "closest", choose the household on your right hand side.
Repeat the procedure until 30 households have been interviewed.
If no-one is present at the household, find out where the household members are. If they are likely to return during the day, re-visit the household later, otherwise replace household with the next household.

Data collection

Data will be collected by standardised, piloted questionnaire. The questionnaire will elicit information on demographic characteristics of the household and mortality throughout the period July 2001�August 2002. Questionnaires will be administered at the individual household level by a team of trained interviewers. Each team will consist of one national staff able to communicate in the local language and in Portuguese, and one expatriate able to speak Portuguese.

Guidelines for completing questionnaire

Introduce yourself to head of household and explain purpose of survey.
Ensure that people understand that this survey is not a registration process and entirely anonymous. There will be no incentives given for participating.
Interviews should be conducted with the head of household, or another adult member of the family in his/her absence.
Obtain oral consent from the respondent: he/she should be told that the survey is voluntary, and that he/she can stop answering questions any time he/she wishes
The definition of family for the purpose of this survey is a group of persons sleeping in the same home and sharing the same meals; visitors excluded. Make sure that the interviewee understands what family means in this particular context.

Coding of variables in questionnaire

Sex: Male=M, Female=F.
Age: Record age in full years. For children under 1 year of age, record age in months (eg, write "5 months")
Present now: Tick box for all family members who live in this household at the time of the survey.
Present SE (solar eclipse): Tick box for all family members who lived in this household/family at the time of the solar eclipse in July 2001.
Status: Classify all person on the list who are not "present now" as either

Died (X): Person used to live in this family at the time of the eclipse but died since.
Absent (A): Person used to live in this family at the time of the eclipse but left since. It is know that person is alive and where person lives.
Disappeared (D): Person used to live in this family at the time of the eclipse but disappeared since. It is not known if person still alive or dead, or where person lives, if still alive.

Date: This column collects all dates for

Died
Absent
Disappeared

Record dates as day/month/year. If only the month and the year are known, record as 15/month/year. Exception: If recording "15/month/year" would classify an event wrongly as occurring before or after arrival in the area, change the date to 01/month/year or 30/month year. For example, arrival in camp was on 08 April 2002, death occurred in "April 2002" but before arrival in the camp: In this case, record 1 April 2002.

Make use of a local calendar if needed. This calendar needs to be completed for each site after discussion with the camp coordinator: include, e.g., food riot in camp, big campaign in x, etc).

Calendar time	Local calendar	Comments
21 July 2001	Solar eclipse
03 August 2001	Birth date of Dr Jonas Savimbi
11 September	Terrorist attack in USA
mid-October 2001	Beginning of rainy season
11 November 2001	Angola Independence day
25 December 2001	Christmas
22 February 2002	Death of Jonas Savimbi	dependent on when news were first heard
31 March 2002	Easter
4 April 2002	Ceasefire	dependent on when news were first heard

Cause: Self-reported cause of death, classify as

1 Fever/Malaria

2 Diarrhoea

3 ARI

4 Measles

5 Malnutrition

6 Violence/War

7 Others

Order and structure of questions

1. How many family members live in your household today, including yourself?
Quanto sois na tua familia, incluindo a tua pessoa hoje em casa?

For all family members living in this household now, list sex (M/F), age in full years, and tick column "present now".

2. When did you arrive in this Area? (dd/mm/yyyy)
Quando e que tu chegaste nesta area?

Complete date of arrival for each name on the list.

Explain that we go now back in time to 21 July 2001 (solar eclipse). Ask in which municipality/province the family lived at the time of the solar eclipse and record at top of questionnaire.

3. At the time of the solar eclipse in July 2001, who among the family members living in this household now lived already in your household?

Havia um eclipse solar em Julho de 2001 ; quais dos membros familiares vive aqui.e que estava dentro de casa durante o eclipse solar ?

Tick column "present SE" for all members living in the household at the time of the eclipse and still present now.

4. At the time of the solar eclipse in July 2001, what other family members were living in your household?

Existe outros que estavao em casa durante o eclipse, e que nao vivem a dentro de casa agora ?

List other family members (sex, age) and tick "present SE".

5. What happened to those family members who lived in your household at the time of the solar eclipse but are now no longer living in this household?

O que aconteceu com os membros familiares, que estavao dentro de casa durante o tempo do eclipse, mais agora nao se encontrao em casa ?

Classify all members who are no longer living in the household as X, D, A in column "Status" and when they died, disappeared or left in column "Date". For all those who died, give cause of death in column "Cause".

As a check, ask directly if any family members died, left or disappeared between time of solar eclipse and now.

6. For family members who were not living in the household during the time of the solar eclipse but are present now: When did the family member join the household?

Ha membros familiares que nao se encontravao em casa durante o eclipse, mais estao presente agora . Quando e que eles vierao em casa ?