THREE DECADES OF RESEARCH ON POPULATION AND HEALTH:

THE &STON EXPERIENCE IN RURAL SENEGAL: 1962-1991. .

(draft)

Michel Garenne

and

* Pierre Cantrelle

i Reviseci

September 15, 1991

Paper prepared for the IUSSP Seminar on Longitudinal Studies, Saly Portudal, 7-11 October 1991.

Correspondance :

Michel Garenne, Harvard Center for Population and Development Studies, 9 Bow Street, Cambridge, MA 02138 (USA).

?

l

I -

1

II

Plan

.

Background

Initial Impetus

The 1962-1966 project

P. Cantrelle's itinerary

Measles vaccines

The 1961 vital registration act

IIIEvolution of objectives

Phase 1 : Sine-Saloum, 1962-1966

Phase II : Ndemene-Ngayokheme, 1967-1982

Phase III : Niakhar, 1983-1991

IV Study population

v Methodological issues

Annual census

Continuous recording of events

Registration and guestionning method

Definition of residence

Mapping

Computerization

Discussion

VI Interventions

VIILegacy

1 EMXGROUND

When African countries became independent, in the late 1950's

and early 196O's, research on population was virtually non .

existent in tropical Africa. At this time, modern demographp was

emerging as a new science. Until then, demoqraphy focused mostly

cn population genetics and on mathematical modzls aiming at

dascribing population dynamics. Modem epidemiological and

statistical methods, now widely used throuqhout the world, were

iargely ignored in the field of population studies.

Anthropologists paid little attention to demoqraphic and health

processes. Their fscus was on local myths, religions and social

structure. Little attention was devoted to the cultural factors

of population and health in Tropical Africa. The concept of

multidiscipiinary research was not yet fashionable and most

researchers were working isolated in their own field.

Before 1954, most of the little available demographic data was

based on censuses and vital registration of European populations

living in Africa. There was only a handful of scattered bodies of

reliable data based on African populations: these were usually

localized and based on a small sample: for instance local vital

registration systems in cities and parish registers in rural

._. . ‘.

. ,’

l .

. c

areas. Other large scale demographic data were considered

unreliable: for instance the administrative enumerations

consistently showed a strong undercount of the total population,

especially of Young children.

The first systematic studies of African populations started

in the early 60s. A team of researchers working at Princeton

University published one of the first comprehensive account of

African demography in 1966 (Brass et al.). At about the same

time, a group of French demographers published a summary of their

experiences and findings on Tropical Africa, with emphasis on the

sample surveys conducted by INSEE since 1954 (Cantrelle et al.

1967). The first seminar on African demography was organized by

Franck Lorimer and held in Paris in 1959. It was followed by two

other seminars, in Ibadan in 1966 (Cadwell and Okonjo) and in

Nairobi in 1969 (Ominde and Ejiogu). The first IUSSP conference

on African Uemography was held in Accra, Ghana in 1971

(Cantrelle, 1974).

Likewise, public health research in Tropical Africa was

strongly biased towards the needs of adults and of expatriates.

Health research was restricted to tropical diseases, especially

those diseases whose gerns recently became identified or for

which a vaccine or a treatment recently became available.

4

According to Becker and Collignon (1989), most of the scientific

publications prior to 1960 dealt with the following diseases:

yellow-fever, malaria, trypanosomiasis, plague, tuberculosis,

schistosomiasis, filariasis, leptospirosis and dysenteries. Other

diseases, now recognized 'as major causes of death or major s sources of morbidity, were virtually ignored, such as watery

diarrhea, acute respiratoryinfections (ARI), measles, pertussis,

tetanus, meningitis and poliomyelitis. There was very little

research done on materna1 and Child health, on materna1

mortality, on health systems, on family planning, on impact of

vaccination and other topics which now are seen as the most

challenging health problems in the region.

II INITIAL IMPETUS:

1) The 1962-1966 proiect

Prospective community studies in rural Senegal were started

within the context of the reorganization of the country after it

declared independence on April 4, 1960. The first "Economie and

Social Plan, 1961-1964" invited to undertake a project designed

to improve knowledge of demographic rates. More specifically, the

plan estimated that it would be important to have a series of

demographic data based on vital registration. This series would

5

. I

. ,’

! 1 f

,

test the validity of the 1960 sample survey, the first national

demographic survey conducted in Senegal. In addition, a new

legislation for universal vital registration was passed in 1961.

There was a need to evaluate the completeness of vital

registration and to assess the functioning of the new vital

registration system.

For this project on demographic data collection, the UN-TAB

(the United Nations Technical Assistance Bureau, now UN-FPA)

created a position for one year, from July 1962 to June 1963.

Pierre Cantrelle was chosen as the project director, under the

supervision of Louis Verriere who was a statistician from INSEE

(the French Institut National de la Statistique et des Etudes

Economiques). Cantrelle met Louis Verriere for the first time in

September 1961 in New-York at the IUSSP conference. After the

conference, they went to visit F. Lorimer, A. Coale and W. Brass

at Princeton university and discussed the projectwith them. They

agreed that the priority was to conduct a feasibility study of

recording births, deaths and marriages for at least three years.

The UN-TAR position, initially created for one year, was extendeà

for one more year, until June 1964. In 1964, ORSTOM, a French

research organisation founded in 1943 (l'office de la Recherche

Scientifique et Technique Outre-Mer), created a position for a

demographer. Cantrelle was recruited and the project was later

managed and supporred by ORSTOM.

6

The field work of the original project was supported by the

FAC (the French Fond d'Aide et de Coopération). A budget was set

up for one year and extended for two more years (1963-1965). The

Population Council.participated in the financing of the data

analysis. Later, the project was primarily supported by, ORSTDM

and the marginal cost of specific projects was supported by other

institutions.

P. Cantrelle's itinerary

Pierre Cantrelle was a physician who was first recruited in

1954 at the IFAN (1' Institut Fondamental d'Afrique Noire) in

Dakar, for a position in physical anthropology. He soon became

interested in public health issues and denography. He

participated in the 1954 National Denographic Survey in Guinée,

the first demographic sample survey conducted in tropical Africa.

He conducted a micro-demographic and anthropological study in a

village in Fouta Jallon (Dantari). In 1957-1958, P. Cantrelle

conducted the demographic, consumption and clinical part of a

major multidisciplinary study in the Senegal River Valley

(Boutillier et al., 1962) called the MI%ES (Mission Socio-

Economique du Fleuve Senegal). During this study, the importance

of measles as a cause of death in the African population first

became documented. The MISOES study also highlighted specific

7

. -;-~‘._., l l., ’ .

. .

. I’

features of the level and age pattern of mortality in infancy and

childhood and the prevalence of malnutrition in the population.

The survey also made an early attempt to use non-medical

personnel for community based studies of morbidity and mortality

and to have a local vital registration system organized by

village volunteers. Cantrelle also participated in other pioneer

demographic sample surveys conducted by INSEE in Tropical Africa:

Upper-Volta (now Burkina-Faso) in 1960, Dahomey (now Benin) in

1961. During the Dahomey sunrey, an attempt to organise vital

registration was also conducted.

Measles vaccines

The early 1960's witnessed the production of the first measles

vaccines, after the isolation of the measles virus by Enders in

1954. In 1963, two Edmonston-B vaccines became available, one

inactivated and one live vaccine. The live vaccine was tested in

a rural community in 1963 in Tattaguine and in Niakhar, Senegal,

a study in which Cantrelle played a major role (Rey et a1.1964).

The vaccine was found to produce many adverse reactions. Later,

a further attenuated vaccine was studied in 1966 in three other

rural communities: Khombole, Niakhar and Paos Roto. The new

vaccine was found safe and efficient.

.L I . . , c_._ _ _- . .~---.--- . . ..-^._. -.. ,.-_, _ -. _

The 1961 vital reuistration act

The newly independent nation of Senegal felt that monitoring

the dynamics of the national population was a priority. In 1960,

vital registration had a low coverage nationwide, with marked

differences that could be explained by its complex history.

Initially, vital registration was compulsory for French citizens

only. In 1916, the registration of births, deaths and marriages

became compulso,ry for the residents of the four communes (Dakar,

Saint-Louis, Rufisgue and Gorée) who becaine de facto French

citizens. In 1933, vital registration was extended to certain

categories of the African population: military personnel,

government employees and other tax payers. In 1950, vital

registration was extended to a11 those residing within 10

kilometers of a vital registration tenter. The vital registration

act of June 23, 1961 made the registration of births and deaths

compulsory for a11 Senegalese citizens.

It is within this context of research on vital registration,

population dynamics and measles prevention that the prospective

community studios began in rural Senegal.

9

c

III EVOLUTION OF OEUECTIVES

The studies had three main phases: 1962-1966, 1966-1983, 1983-

1991. These are summarized in the following table:

Table 1 : ORSTOM Prospective Community Studies in Senegal

Fatick Nioro du Rip (North West) (South East)

.Phase 1 : Sine-Saloum : December 1962 - February 1966

a11 Niakhar arrondissement 1/2 Paos-Xoto arrondissement 65 villages 135 villages 35,187 people in 1566 18,988 people in 1966

Phase II : Ndemene-Ngayokhene : December 1962 - February 1983

a11 secco Ngayokheme a11 secco Ndemene 8 villages 30 villages about 5,000 people about 6,000 people subsample of previous area subsample of previous area

Phase III : Niakhar : March 1983 - ongoing

CR Ngayokhene + 1/2 CR Diarere 30 villages

(tenninated)

about 25,000 people extension of previous area 18 villages fron phase 1 8 villages from phase II

NB: in phase 1, birth cohorts were followed up until April 1968.

10

Phase 1 : Sine-Saloum. 1963-1966

Phase 1 Etudies focused on demographic data collection. The

main question the project sought to address was: could reliable

demographic data be collected in rural areas of tropical Africa?

It was decided to collect precise prospective vital data in a

small sample, rather than conducting large scale retrospective

surveys of doubtful guality. In addition, data on measles

morbidity and mortality were systematically recorded. An

evaluation of the cost of data collection was built into the

project. The study was directed by Cantrelle and had two

Senegalese supervisors: KM Mamadou Diagne and Boubacar Fall. A

Sereer sop -iologist was recruited to help with the definitions of

age, marriages and deaths, but he stayed only for a few months

and left for persona1 reasons. Field workers were chosen among

the staff of the statistics division (Direction de la

Statistique).

Phase II : Ndomene-Nuavokheme, 1967-1982

In 1966, when the project was taken over by ORSTOX and

external financing ceased, Cantrelle urged the Senegalese

authorities and the Institute to continue a long tenn demographic

observation in a sample of villages. The idea was to gather long

term trends in mortality, fertility and nuptiality, to do more

11

?

. I’

studies of social structure, to monitor the future of the recent

Vital Registration Act and to evaluate the effects of the 1966

crisis in rural areas which was due to low yield of the two main

crops: millet and peanuts. The budget was roughly limited to the

standard annual allocation of researchers at ORSTOM. Therefore,

two small subsamples of about 5,000 people each were selected.

The longitudinal observation of 8 villages in the North-Western

area and of the 30 villages in the South-Eastern area lasted

until March 1983. Since it included villages from the previous

project, the phase II data include the 20 years of observation

between December 1962 and January 1983.

The field work was conducted by a team of 3 to 4 professional

field workers who were tenured employees of ORSTOM. Al1 of them

are still working for the Institute. The work was conducted under

the supervision of Cantrelle, who left Senegal in 1969. Many

Young demographers participated in the study and were trained to

do field work during this period: Bernard Lacombe, Jacques

Vaugelade, Francis Gendreau, Benoit Ferry, Dominique

Waltisperger, Christine Guiton, Gilles Pison and Michel Garenne.

Several'of them undertook similar studies later in various parts

of Africa. A comprehensive assessment of the demographic

situation was done in 1981, with funding from the Ford

foundation.

12

Phase III : Niakhar. 1983-1991

When Michel Garenne arrived in Senegal, in December 1982, he

gathered a multidisciplinary'research team to do more studies

based on micro-demographic data. The focus was -on- the

interactions between demography, epidemiolog-y and anthropology to

study mortality determinants. The first major study that helped

to reshape the project was on the relationship between

nutritional status and mortality. A study of causes of death

estimated by verbal autopsies was built into the project. This

study first determined the sample sire (about 5,000 children age

O-4 were needed) and the choice of the study area: Niakhar. The

second study area (Ndemene) was dropped primarily because of the

difficulty of working in several different laquages. The Niakhar

area was predominantly Sereer whereas the Paos-Koto area, which

was an area of recent settlers, included five significant ethnie

groups and laquages (Wolof, Poular, Mandikas, Sereer, Tukuler)

plus other minorities. A last census was taken in Janua-ry and

February of 1983: the Ndemene area was closed afterwards and the

Ngayokheme area was included in the new Niakhar area which

encompassed 22 new villages, many of them belonging to the phase

1 Niakhar area.

Michel Garenne came from the field of mathematics and

statistics. He had a MA training in demography in Paris

13

?

University and earned a P~D in demography at the University of

Pennsylvania, Philadelphia. His main interestwas the integration

of epidemiologic and anthropologie information to promote better

understanding of demographic phenomena. He gathered a

mutidisciplinary team of nutritionial epidemiologists and

anthropologists (Bernard Maire, Olivier Fontaine, Jean Pierre

Beau, Andre Briend, Ehady Dieng, René Collignon, Charles Becker).

The team started a study on the relationship between nutritional

status and mortality, funded by the European Economie Community

(EEC), a study on verbal autopsies, a study of measles

transmission, and a study on materna1 mortality, funded by Fanily

Health International (FHI). Later, specific morbidity and

mortality studies on diarrhea, choiera and malaria were also

conducted. At the same time, in 1983, Garenne founded the

Research Unit "Population et Santé" at ORSTOM, which was first

chaired by Cantrelle and which gathered people interested in

multidisciplinary research in the field of population and health:

demographers, nutritionists, sociologists and geographers. A new

team of 12 field workers was recruited and trained to meet the

needs of the new area.

Afterthe 1983-1986 projects were completed, the research unit

felt the need to participate more actively in the worldwide

effort on Child survival and to contribute directly to the

control of mortality in the population. In 1986, a new Expanded

14

Program for Immunisation (EPI) was started in Senegal. The

research unit decided to contribute to it directly by doing a

study of high titer measles vaccines and their potential for

reducing early mortality from measles and to evaluate the impact

of pertussis and tetanus immunisation. A separate study of risk

factors of neonatal tetanus was also conducted during the same

period by Odile Leroy. The study of high titer neasles vaccines

became a randomized vaccine tria1 monitored by the Task Force for

Child Survival, with funds from the Rockefeller Foundation, the

World Health Organization (WHO), the World Bank, the United

Nations Program for Development (UNDP) and the United Nations

Children's Fund (UNICEF).

During the 1983-1989 period, many Young researchers, African,

European and American, were welcomed in the team and trained for

Y! or PhD dissertations. An interesting dynamic was created and

became very favorable for multidisciplinary research. The field

station attracted other researchers from different backgrounds,

in particular economists, agriculture specialists and ecologists

who studied other aspects of the society, notably a group

gathered by André Lericollais and Pierre Killeville, who spent

several years in Niakhar (Lericollais, 1991).

15

P

. ,’

I

IV STUDY POPULATION

Phase 1

Phase 1 studies were conducted in two separate areas of Sine

Saloum: Niakhar and Paos-Koto (see maps 1 and 2). Sine Saloum was

at that time the most populated region of Senegal, located in the

heart of the peanut growing region. Initially the target was to

follow up 2 areas totalling about 70,000 persons. This target was

.later reduced to 50,000 persons because of budget constraints and

of the unexpected length of time which it took to complete the

first census. Initially a sample of villages in the Sine-Saloum

region was thought to be preferable. But contiguity helped reduce

the costs of transportation and was found to be a necessary

condition for gaining knowledge of the population, and for in-

depth research. However, two separate areas were chosen for

studying two different demographic situations. The Niakhar area

was located in the Sine, a dry part of the orchard Savannah, an

area of high population pressure and of net ont-migration. The

Paos-Koto area was located in the Saloum, an area of heavier

rainfail, with a lot of available land and strong in-migration

flows. The Niakhar area was ethnically, economically and

culturally homogeneous, with 96% of the population belonging to

the Sereer group whereas the Paos-Koto area was ethnically

heterogenous, and was more of a settlement of pioneers Vho were

16

on the average richer than there counterparts in the Sine.

The limits of the study area were defined by administrative

boundaries. The Sine area included the whole arrondissement of

Niakhar. The Saloum part included half of the Paos-Rot0

arrondissement. The clioice of the arrondissements of Niakhar and

Paos-Koto from among the other arrondissements of Sine-Saloum

were made with in consult with the local authorities (préfet).

Phase II

The villages of phase II areas were chosen as a subsample of

phase I villages. The administrative unit was the "secco", a

administrative unit organized around the collection of peanuts,

a practice which was nationalized at that time. The sample sise

was fixed by budget constraints to about 5,000 people. The secco

of Ngayokheme was selected because it was included by the

Senegalese Governement as a pilot area for the future rural

communities (Communautés Rurales). The choice of Ndemene was made

because it was the only secco of approximately 5,000 persons.

Phase III

The phase III area was chosen as an extension of the phase II

area of Ngayokheme because the homogeneity of the population was

17

.

, I’

. ’

considered a priority for in-depth multidisciplinary research.

Rural conununities (CR) were created in 1974 and the Ngayokheme CR .

included 9 new villages in addition to the 8 villages included in

the former Secco. However, the 17 villages of the Ngayokheme CR

did not account for enough children for the project on

malnutrition. There were two possibilities of extension: towards

the East (Patar) and towards the West (Diohine). The West

extension was chosen mainly because of the local private

dispensary of Diohine which provided the opportunity to compare

two local systems of health tare. In addition, Diohine had

already been studied by the team when evaluating a food

supplementation program (PPNS).

18

‘-

,

. ..-

v xBTB0wLoG1CAL IssuEs

Earlier exoeriences

When the first study began in 1962, there was very little

experience in longitudinal demographic data collection in rural

areas of less developed countries. Only two published studies

were available to P. Cantrelle when.he started: the Yangtse River

Valley study in China (Chiao et al. 1938) and the Guanabara study

in Brazil (1964). The Khanna study had just finished and had not

yet been published, the Keneba study and the Pakistan Grow-th

Experiment were underway.

At that time, demographic data collection was hampered by the

idea that recall biases were dominant. It was felt that people in

tropical Africa were unable to properly report births and deaths

when asked more than a few months after the event occurred.

Nowadays, a better relationship exists between field workers and

the population, people understand better what is obtained from *

demographic surveys and are less likely to lie systemâtically,

methods of posing questions have improved, such that we now

realize that this was mostly a misconception. It is now common to

have good retrospective sunreys on events going back 20 to 30

years before interview.

19

Annual census

The phase 1 studies began with a vital registration based on

an annual census and continuous recording of vital events in the

villages. For the annual census, the list of compounds was

updated with the head of the village and completed in the field.

Compounds are locally defined with the word "mbind". Each

compound is visited and each individual resident at the previous

census is called by name, which provides information on deaths

and outmigrants; new resident are detected by recording births,

incoming wives and other immigrants. Changes in marital statu.5

and in other variables of interest are also recorded during the

roll call. The annual census was maintained from 1964 to 1987 as

the sole means of recording demographic events with a few

exceptions of years without a census (1967, 1075, 1976, 1979) and

of years of recording demographic events during morbidity and

nutrition surveys in 1983-1984.

Continuous recordins of events

During tha first 20 years, the continuous recording of events

by field workers was used in 1963 only, and was conducted every

3 months, that is 3 visits in addition to the annual census. TO

reduce the cost of going to each household too often, the

recording was based on the local vital registration, on

20

independent records made by village headmen and verifications

made in the field by the supervisors. ,This procedure lasted

during a11 the phase 1. It was dropped later because it was found

to be too costly and not as accurate as the records obtained by

the annual census. Compared to the annual census, the dual record

evaluation showed that the officia1 vital registration covered

only 32% of births, the village registers about 50% and tte

visits by supervisors about 90% (Cantrelle, 1969).

The systematic continuous recording of events was resumed in

January 1987, for the vaccine trial, and has been maintained

since. This time it was based on weekly visits to each household

of the study area. Despite the very freq-uent visits to

households, each year some missing births and deaths are found

by the annual census. Missing births are nostly births to

migrants women who were not yet entered in the file. Missing

deaths are deaths of any age which were not reported, most likely

because the field worker did not ask the correct questions.

Reuistration and Questioninq Method

The way questions are organized in the questionnaire and asked

to the families seems to be a major determinant of data quality.

In the first project (62-66), the base was the annual census

which was recorded on a,household sheet. During the next round,

21

F

. . - r:

_ .--...-~-.-..._

I

the color of the ink was changed (blue in 1962, red in 1964,

green in 1965) and changes in household composition were noted on

the same sheet. Obviously, this procedure, which was introduced

for the 3 years of the first project, was not appropriate for a

long lasting surveillance.

During the 1968 round, an individual tard was introduced. The

individual tard system was tested first in Khombole and was used

in Ndemene and Ngayokheme until 1983. The individual tard

contained a11 the necessary information on residence, socio-

economic status, marital status, and other relevant information

(pregnancies, still births, measles, vaccination, weaning, cause

of death). The individual tard system was very flexible. First,

it contained a11 the information on an individual in a single

place. Second, it enabled the field workers to recompose

households by just binding individual cards with a rubber band,

therefore adjusting the structure of the population at any time

by just moving the cards within the household (deaths and

outmigrants were put at the end) or among households when

individuals moved from one household to another. Third, the tard

system was suitable to manual counting of events every year.

In 1983, computer print-cuts were introduced as the support

for guestioning and recording events. The idea was to recompose

the exact situation of the household every year. This way, only

22

the previous situation of each individual was displayed which was

sometimes a handicap when compared ta the tard system. However,

the computer could print the exact questions to be asked at the

next census as a function of the age of the person. This had

major advantages: it simplified and standardized the work of the

field workers, it provided a11 the information on a household, at

the same place; it ensured that proper questions were asked every

year (Garenne, 1984); it allowed 'to do permanent checks in the

computer files and to correct errors; it made easy to add new

questions when needed. In addition to the printed household file,

events had separate questionnaires, which allowed the gathering

of more specific information and the binding of series of annual

events separately.

Definition of residence

The definition of residence is key for a follow-up study and

for the definition of births and deaths. The definition of

residence raises necessarily difficult questions in areas where

there is a large amount of in- and out-migration and imposes

choices in the function of the objectives of the demographic

surveillance. The definition of residence was always based on &

lure criteria and on long durations of stay, the idea being not

to lose demographic information that was available and relevant

for the study, and not to introduce events relative to short ter-m

23

?

I

.

visitors. In the first project, strong criteria was used:

migrants were defined by either the intention to move permanently

or an absence of at least 4 years. In phase 11, the intention

criteria were kept and the 4 years rule was reduced to one year,

that is absence at two successive censuses with no return in

between. This corresponds guite closely to tempora,ry workers who

do not corne back for one rainy season, since the censuses are

taken in the dry season. The definition was later refined (in

1981) for special categories of migrants, in particular school

children, teachers and visiting husbands, to best fit the local

situations. The fact that censuses were conducted in the dry

season never allowed the proper recording of presence of the

pavetanes, the temporary workers who came only for the harvest

in the rainy season. On the contrary, the noranes, the temporary

workers who moved out during the dry season, when there is no

agricultural work, were systematically recorded. Note that the

definition of residence for birth (birth to a resident mother)

differs from a regular administrative definition that would apply

in a vital registration system, where place of birth is the basis

for counting the event. In the demograpt,ic surveiliance, a birth

in Dakar to a resident rr,other was counted and a birth in the

study villages to a non-residenr mother was not counteci. The

opposite was true for the vital registration officer.

24

Haovfnq

Large scale maps (l/SO,OOO) were available when the study

started in 1962. In addition, rough drawings of situations. of

compounds in each village were done at that time. The compounds .

were first numbered according to their location in the village.

The maps were not used after the team of field workers became

permanent and were acq-uainted uith the population. Locutions of

compounds in villages were redrawn in 1983 and in 1985 for the

extension of the study area, using in particular air photographs,

and a complete mapping of a11 compunds was recently completed

from a new set of air photographs by Lericollais and B-ecker.

Comnuterization

The series of demographic events were put on computers since

the first project. In the 196O's, for the phase 1 project, the

mechanography was used for sorting and counting punched cards but

this system was slow. Computers of the third generation were used

to enter data of phase II in 1974 and in 1981. Prom 1981 to 1986,

events were entered every year after the census and cleaned

immediately., In January 1987, data were transferred on an IBK-

compatible network of micro-computers and data were entered

continuously in a relational database system, vhich is explained

in another document (Garenne, forthcoming).

25

?

, I’

Discussfon

Working in-a traditional society , where nothing is fixed by

law and almost' everything is negotiable, includlng name and age,

raises numerous questions for scientiffc work. Sereer people have

several first names that they cari use. In certain cases they even

change..names or age class. Horeover, mothers m.ay gfve the same

first name to more than one Child. This is particularly common

for a new born when à older sibling died recently. It has

happened that a man had two wives with the same first and last

names. Although these questions cari be resolved, they take time,

skiil, knowledge and attention. There are many ways to estimate

age with reasonable accuracy in this context, which are better

explained in the instructions for field workers. The strategy

used in Senegal has always been to appropriately use ail of the

availàble information (traditional counting, seasons, historical

events) and to not rely on a single standard method.

The demographic surveillance system produces a population

register, in which names have to be record&. This raises an

Lmportant issue of confidentiality. The study is midway between

a demographic survey, for which confidentialitywould be required

by not recording names, and access would b-e vide, and a health

registér for which information would be complete but access would

b-e restricted.

26

.

VI IKL'ERV-RNTIONS

The study area was the site for several specific health

interventions. The first intervention was vaccination with the

further attenuated live measlee vaccine in 1966. Hortality and

morbidity after vaccination was monitored. A more forma1 vaccine

tria1 of two hfgh t-ter measles vaccines (Edmonston Zagreb and

Szhwarz) was recently conducted. This study was followed up by

another tria1 of an acellular vaccine against pertussio, which is

currently in course. The efficacy of the tetanus vaccination

durinq pregnancy was also monitored. Other interventions

organized by other institutions were also monitored: the Basic

Health Services project and the more recent Primary Health Care

system, a food supplementation proqram (PPKS) and a proqram to

detect and refer hiqh risk pregnancies. A system to screen and

refer malnourished chiidren was set up and a small tenter to

treat malnutrition was opened in one of the local dispensaries

(Toucar) . The use of essential drugs was pronoted and the

research team participated at various points in time in the

orqanization and management of the drug supply, in particular for

the prevention and treatment of malaria. Cases of poliomyelitis

and other physical handicaps were recorded and twice a year

handicapped people received the necessary assistance and

protheses. Cases of epilepsy were also recorded and treated when

possible.

27

” ’

#. .

r

The main issue that was faced for major interventions such as

vaccine trials, was the issue of informed consent, We found it to

be extremely difficult to explain the issues raised by randomized

trials to people who had spent a11 of thelr lives in a

traditional remote society such as the Sereer.

Almost three decades of observation, fntemention and research

is a major achievement in and of itself. This continuity was

possible only thanks to the firm Will of Cantrelle througbout the

entire period. The series of demographic data in Niakhar is

almost unique in Tropical Africa, with the exception of the three

villages of Keneba, Randuar and Jali followed-up by the British

Hedical Research Council (B-XRC) in the Gambia. The prospective

6tudies in Senegal have served as a training tenter for

researchers from Senegal and many other countries. Niakhar was

also a demonstration area for visitors who cams from a11 over the

world. The site has been the focus of numerous multidisciplinary

research projects and has been the site for research that would

never have occured at first to the founders of the study. The

studies have produced a large body of data, which has been used

by various researchers with different backgrounds and which Will

be the source of future researches.

28

The age pattern of mortality and the seasonality of deaths

remain a source of interesting controversies in the literature

(Cantrelle, 1974). The study of the relationsNp between birth

intervals, breastfeeding and Child survival vas a pioneer work

and is one of the most quoted pieces of work in the field

(Cantrelle and Leridon, 1971). The study on the relationship

between malnutrition and Child survival is the largest ever

conducted in Tropical Africa (Garenne et a1 1987). It has lead to

better knowledge of malnutrition as a process, it has pinpointed

the value of ans circumference as a basic tool for screening at

risk children, and has lead to an analysis of the causes of death

that are related vith poor nutritional status. The nuptiality

regime based on videspread polygamy and its relationship vith

high fertility is also a classic study (Garenne and Van de h'alle,

1988). The studies on the transmission of mezsles (Garenne and

Aaby, 1990), choiera (forthcoming) and neonatal tetanus (Leroy

and Garenne, 1991) produced new insights on the relationship

between factors of exposure to infectious diseases and Child

survival. Studies of the impact of measles vaccination are almost

unique in the world and the results of the recent vaccine tria1

(Garenne et al. 1991) are still a matter of controversy and may

have major research and policy implications,

29

?

. I’

. ’

’

,__._ -... ..- _~., .-. .-- --- - .- --

REFERENCES

.

Beau, .JP, Fontaine 0, Garenne H. 1989. Hanagement of

Ralnourished Children with Acute Diarrhoea and Sugar

Jntolerance. Journal of Tropical Pecliatrics. 35 (Decembre

'1989) : 281-264.

Becker C, Collignon R. 1989. Sante et Population en Sendgambie

des origines b 1960. INED. Paris.

Boutillier et al. 1961. La moyenne vallee du Séndgal. PUF.

Paris..

Brass et al.. 1968. The demography of tropical Africa. Princeton

University Press.

Briend A., Garenne H., H-aire, B., Fontaine O., Dieng K. 1989.

Nutritional Status, Age and Sunrival: the Huscle Mass

Hypcthesis. European Journal of Clinical Rutrition 43 :

715-726.

Caldwell and Okonjo, 1968. La population de l'Afrique

tropicale. Population Council. New-York.

Calwell et al. 1973. Croissance démographique et évolution socio-

30

economigue en Afrique de l'ouest. Population Council.

P. CANTRELLE. 1964. L'État-civil et les autres sources de

renseignements SUI l'&olution demographigue au Senegal.

(Rapport de fin de mission ONU), Dakar, mal, 112 p. multigr.

P. CANTRELLE. 1968. Hortalite par rougeole dans la region du

Sine-Saloum (Sénegal) 1963-1965. In : Conditions de vie de

l'enfance en milieu rural en Afrique, pp 156-158, Centre

International de l'Enfance, Paris.

P. CANTRELLE, A. LERICOLLAIS. 1968. Evolution de la scolarisation

dans une zone rurale du Senegal (Arrondissement de Niakhar,

1949- 1956). In : Conditions de vie de l'enfance en milieu

rural en Afrique, pp 226-232, Centre International de

l'Enfance, Paris.

P. CANTRRLLE. 1969. Etude démographique dans la region du Sine-

Saloum (Sénegal). Etat civil et observation démographique.

Travaux et documents de ~'ORSTOM, n* 1, ORSTOK, Paris, 121 p.

P. CANTRELLE. 1971. Etude de cas : population et ressources dans

une zone rurale du Sénégal. Dot. CEA. Po~u. Conf. 1/9. 21 p.

multigr. Conférence africaine sur la population, Accra, Déc.

Texte publie egalement in : Afrique médicale. tic. 1972, n*

105, pp 1029-1036, et Janv. 1973, n' 106, pp 47-52.

31

. I’

__-._.

P. rXNw.ELLÈ, H. LERIWN. 1971. Breast Feeding Mortality in

Childhood and Fertility in a Rural Zone,of Senegal. Population

Studies, Xxv, 3, pp 505-533.

P. CANTRELLE, H. LERIDOH, P. LIVENAIS. 1980. Fkonditd, allaitement

et mortalit6 infantile t diffkences inter-ethniques dans une

meme region (Saloum, Senegal). Population II' 3, 623-648.

P. CANTRELLE. 1980. Hortalitb infanto-juvenile d'hivernage dans le

Sine-Saloum. Environnement africain, vol IV, n* 14-16, 1980,

pp 413-428.

P. CANTRELLE. 1967. Afrique Noire, Madagascar, Comores. Démogra-

phie Compar&e. Fascicule 6. Mortalité : facteurs. Delégation

Générale a la Recherche Scientifique et Technigxe, Paris, 65

P*

P. CANTRELLE. 1969. 3- Connaissance de la rougeole parmi les

populations africaines (pp 13-14) and 4- Mortalité par

rougeole au Senegal. P.17-19. In : Conditions de vie de

l'enfant en milieu rural en Afrique, pp 128-146. Centre

International de l'Enfance, Paris.

F. CANTRELLE. 1971. Mortalite périnatale et infantile au Sénégal.

32

In : Congrès International de la Population, Londres, 1969,

UIESP. Li&ge, pp 1032-1042.

P. CANTRELLB. 1973. Niveaux, types et tendances de la mortalite.

In t CALJWELL J.C. ed. Croissance demographique et évolutton

socio-economique en Afrique de l'Ouest. ThePopulatin Council,

New-York, pp 137-165. Version] anglaise 'Xortalfty, levels,

patterns and trends' publié en 1975 in 'Population Growth and

Socio-economic change in Africa‘, The Population Council, New-

York, pp 98-118.

P. CANTRESLE. 1974. La mgthode d'observation suivie par enguete

81 passages r6#tés. OS/EPR. Laboratories for Population

Statistics. The University of Eiorth Carolina at Chapel Bill,

Scientific Report, Series n' 14, 36 p.

P. CANTRBLLB. 1974. 1s there a standard pattern of tropical

mortality 7 In Population in African development. Drdina

Ed.Li&ge vol.1, pp 33-42. Version en anglais de la

communication : Existe- t-il un type de mortalite tropicale

1' Conference africaine sur la population. Accra. Déc.1971.

Texte publie Rgalement dans Afrique medicale. Nov. 1972, no

104, pp 933-940.

P. WTRBLBE, B. FERRY. 1979. Approche de la fecondite naturelle

33

?

__~ _._....__ - _ - ___ ~._-. ._.. __ ..- .._.._ .___ -.. -._.---.-_- -..--_ -.. .- ..-. f

dans les populations contemporaines. In : Natural,Fertility

/ F&ondit6 Naturelle. Ed. H. LERIDON, J. KENFZN, Ordinal Ed.,

Li&ge/ 1979, pp 317-370 (communication présentée au seminaire

sur la fecondFt6 naturelle organisee par l'Institut National

des Etudes Démographiques (INED) et l'Union Internationale

pour 1'Etude Scientifique de la Population (UIESP), Commission

a!Analyse Comparée de la P&ondit@ h Paris, MZB 1977).

P. CANTRELLE, I.L. DIOP, K. GAREXNE, H. GURYE, A. SADIO. 1986. The

prolile of mortality and its determinants in Senegal,

1960-1980. In Deteminants of Hortality Change and

Differentials in Developing Countries. The Pive-Country Case

Study Project. Population Studies, n'94, United Nations, New

York, pp.86-116.

P. CANTRELLE. 1986. Problèmes poses par l'étude des causes de

décès. Estimation de la Kortalite du jeune Enfant (O-5 ans)

pou= guider les Actions de Santé dans les Pays en

Développement. Skdnaire INSERM, vol 145, pp.241-154.

P. CANTRELLE, T.LOCOH. 1989. Social and cultural factors affecting

health in West Africa. Rockfeller Foundation Exploratory

Health Transition Prog-ram Workshop 1. Canberra, Mai 1989.

Chiao CM, Thompson WS, Chen Dl?. An experiment in the registration

34

of vital statistics in China. Scripps Foundation for Research

in Population Problems, Oxford, Ohio, 1938.

Fontaine, 0.; H. Garenne; J.P. Beau et E. Paye. 1984. La

Horbidite par Diarrhée Aigüe en KLlieu Rural au Senégal.

Colloque IRSERH: ia diarrhée du jeune. Vol 121 I 295-300.

(Acute diarrhea morbidity in rural Senegal)

H. GARENNE, P. CM?l'RELLE. 1983. La baisse de la mortalite A

Ngayokhhe 1963-1982 ou quelle transition demographigue dans

les villages du Sine-Saloum (SOnégal)? Journees demographigues

ORSTOH sur la transition dknographigue. Paris, sept. 1983, 13

p.multigr.

r,. CXRENNE, P. CAKTRELLE, I.L. DIOP. 1985, Le cas du Senegal. In

La lutte contre la mort. Influence des politiques sociaies et

des politiques de sant0 sur l'evolution future de la

mortalite. Ed. par VALLIN J.etLOPEZ A., PKJP Paris,pp 307-329.

La version anglaise est parue dans : Health policy, Social

policy and mortality prospects, VALLIN 3. et LOPE2 A, éd.

Ordina, Li&ge, pp 335-340.

H. GARENNE, P. CANTFCELLE. 1986. HortalitB des enfants ayant par-

ticipe à un programme de protection nutritionnelle (Diohine,

Sénegal). Estimation de la Hortalite du jeune Enfant (O-5 ans)

35

_’

r ,’

c

pou= guider les Actions de Sante dans les Pays en

Développement. Seminaire INSERM, vol 145, pp.541-544.

M. Garenne, P. Cantrelle. 1989. Prospective atudies of communitfes:

their unique potentiel for etudying the health transition

reflections from the orstom experience in Senegal. Rockfeller

Poundation Exploratory Health Transition Program Workshop 2.

Londres, June 1989.

Garenne, H. 1985. Le concept de l'etude longitudinale et ses

implications pour la collecte des donnees: example d'un

questionnaire informatise pour ameliorer l'enregistrement

des decès precoces au Senegal. 17 p. Actes du Seminaire

de l'Institut du Sahel, Bamako, 2C-24 Août 04. (en

anglais) The concept of follow-up suvey and its

implications for Data Collection: exmple of using a

computerized questionnaire for iznproving the recording of

early deaths in rural Senegal. IUSSP seminar, Canberra

7-12 Sept-r, 1984. 10 p.

Garenne; H. and P. Van de Walle. 1985. Rnowledge, Attitudes and

Practices Related to Child Health and Hortality in

Sine-Saloum, Senegal. Proceeding of the TOSSP conference.

Florence June, 1985. Vol 4 : 267-278. (reprinted in:

Selected readings in the cultural, social and behavioural

36

determinants of health. J.C. Caldwell and G. Santow ed.

Health Transition Series N'l. Hfghland Press. Canberra.

1989: 164-173).

Garenne, H. 1985. Do Women Forget thefr Births? A study of

Birth Histories in Rural Senegal. Proceedings of the .

seninar at the Antwerpen school of Tropical Hedicine.

12-14 December, 1985. 12 p.

Garenne, M. and 0. Fontaine. 1986. Assessing Probable Causes of

Deaths Using a Standardized Questionnaire. A study in Rural

Senegal. Proceedings of the IUSSP seminzu on morbidity and

Hortality, Sienna 7-10 July, 1986 : 123-142. (version

française : Detennination des causes probables de deces

d'après un guestionnaire standardise: une etude au Senégal en

milieu rural).

Garenne, M. et P. Cantrelle. 1986. Rougeole et mortalite au

Séndgal. Etude de l'impact de la vaccination effectuee a

Khombole 1965-1968 sur la survie des enfants. In:

Estimation de la mortalité du jeune enfant (O-5 ans) pour

guider les actions de sante dans les pays en

developpement.. Stinaire INSERM. Vol 145 : 515-532.

Garenne, M. et P. Cantrelle. 1986. Mortalite des enfants ayant

37

. . .._ . I c

< l

participe B un programme de protection nutritionnelle

(Diohine; Senégal). In: Estimation de la mortalite du

jeune enfant (O-S ans) pour guider les actions de sante

:dans les pays en d6veloppement. Shninaixe IBSERBL Vol 145

.I 541-544.

Garenne, H.; B. Haire; 0. Fontaine; K. Dieng et A. Briend.

1987. Risques de deces associes b differents Etats

nutritionnels chez l'enfant d'&ge prescolaire. ORSTOH.

Dakar. Septembre 1987. ,246 p. Accepte pour rMdition dans

la série Etudes et Theses, ORSTOH, Paris.

GareMe, H.; B. Maire; 0. Fontaine; X. Dieng and A. Briend.

1987. Un critère de prévalence de la malnutrition: la

survie de l'enfant. Actes des 3' Journ&s Scientifiques

Internationales du GERK, Saly 6-10 octobre, 1987: in D.

Lemmonier et Y. Ingenbleek ed. Les carences

nutritionnelles dans les pays en voie de développement.

Xarthala. Paris. 1989 : 12-19.

Gareruie, X. et J. Lombard. 1988. La migration dirigée des

Sereer vers les Terres neuves. Actes des 'TroisiPmes

Journees démographiques de 1'ORSTOH'. Paris 20-22

septembre 1988. 25 p.

38

Garenne, H. et P. Aaby. 1990. Pattern of exposure and measles

mortality in Senegal. Journal of Infectious Diseases, 161

: 1088-1094.

Garenne, H. and E. Van de Walle. 1989. Polygyny and fertility

among the Sereer of Senegal. Population Studies, 43 (2) t

267-203.

Garenne, H.; 0. Leroy; JP. Beau; H. Whittle; 1. Sene; AR. Sow.

1990. Efficacy, Immunogenicity and Safety of two high

titer measles vaccines. Report to the Task Force for Child

Survival.

Garenne H., Becker C., Cardenas R. 1990. Heterogeneity, Life Cycle

and the Potential Impact of Aids in a Rural Area of Africa.

Paper presented at che IUSSP seminar on the Anthropological

Studies Relevant to the Sexual Transmission of HIV,

Sonderborg, Derunark, 19-22 Novernber 1990. Accepted for

publication.

Garenne, H.; 0. Leroy; JP. Beau; 1. Sene. 1991. Child mortelity

after high titer measles vaccination: a prospective study in

Senegal. (Submitted for publication).

Lericollais, A. et al. 1991. Forthcoming.

39

?- ‘ . . . . -~-- . . - . . - . . - _ . - . - - . . . . . . - - . . - . - - - - A._- _ . . . . __I_.- I - . . - . . -

Leroy, 0. èt H. Garenne. 1987. La mortalité par tétanos

neonatal: la situation h Niakhar au Sénegal. in G. Pison,

E. Van de Halle, X. Sala Diakanda ed. Hortaliti et SociétO

i en Afrique. PUP. Paris. 1989 t 153-167.

Leroy, 0. and H. Garenne. 1990. Risk factora of Neonatal

Tetanus in Senegal. Accepted by the International Journal

of Epidemiology.

Lorimer F. XXXX

,

i

Ominde SH and Ejiog-u CN. 1972. L'accroissement de la population

et l'avenir economigue de l'Afrique. Population Council.

H. REY, 1. DIOP PAR, R. BAYLET, P. CiWTRELLE, J. P. ANCELLE.

Réaction clinique au vaccin rougeoleux vivant attenue

(Edmonston B) en milieu coutumier Sénegalais. Bull. Soc. Red.

Afr. Noire lgue. frse. 1964 - 9 tt. 255-271.

H. RBY, R. BAYLET, P. CANTRELLE, 1. DIOP FM, S. DAUCHY. 1965. La

vaccination contre la rougeole par vaccin vivant. 'Deux

experiences en Afrique tropicale (SénBgal). La Presse

Médicale, n* 73, pp 2729-2734.

40

i

Hap 1 :

Administrative divisions of Senegal in 1980.

41

t

. ,’

3

_- .-- _ ---. _ ._ -._- ..- ..____ - _.___. ~ ___-

Hap 2 :

Situation ofyresearch sreas in rural Senegal, 1963-1991.

42

.

i

t ‘_ t i

:!

‘! t

Map 3 :

h’iakhar study area, 1983-1991 (phase III).

43

t

.

t

international union for the scientific study of population

IUSSP COMMIll-EE ON ANTHROPOLOGICAL DEMOGRAPHY

and

ORSTOM

Seminar on

SOCIO-CULTURAL DETERMINANTS OF MORBIDITY AND MORTALITY IN DEVELOPING COUNTRIES:

THE ROLE OF LONGITUDINAL STUDIES Saly Portudal, Senegal, 7-l 1 October 1991

Three decades of research on population and health: The ORSTOM experience in rural Senegal: 1962-l 991

Michel Garenne and Pierre Cantrelle

. - I