Content uploaded by Sola Asa
Author content
All content in this area was uploaded by Sola Asa on Jun 08, 2018
Content may be subject to copyright.
Health Transition Review 7, 1997, 221–234
Mothers' management of childhood diseases in
Yorubaland: the influence of cultural beliefs*
Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Department of Demography and Social Statistics, Obafemi Awolowo
University, Ile-Ife, Nigeria
Several studies have noted that, besides inadequate availability of health care services in many
areas, especially the less developed countries, certain disease-specific and non-disease-
specific cultural beliefs may influence people's health seeking behaviour1. It has even been
noted that health services may be underutilized and several health and child care instructions
may be ineffective or ignored in traditional and transitional societies where people's ideas and
behavioural patterns conflict with the knowledge being passed to them (Feyisetan and
Adeokun 1992; Feyisetan 1992). Feyisetan and Adeokun (1992) argued that non-adoption of
modern preventive and curative measures cannot be attributed to poverty alone since the costs
of some preventive and curative measures are not exorbitant in several of these societies2.
Rather, they suggested that the gap between awareness of modern health measures and health
seeking behaviour must be sought in the social and cultural determinants of behaviour in such
matters as child care and disease management.
Earlier studies have noted that children in Nigeria die mainly from malaria, diarrhoea,
measles, neonatal tetanus, whooping cough, tuberculosis, and bronchopneumonia (Morley and
MacWilliam 1961; Ogunlesi 1961; Morley, Woodland and Martin 1963, 1966; Baxter-Grillo
and Leshi 1964; Animashaun 1977; Tomkins 1981)3. Because these diseases are preventable
at low cost to the individual, there is a need to investigate why large percentages of children
are still subjected to many episodes of these diseases.
* The authors are grateful to the Rockefeller Foundation, New York, for providing the financial resources to
undertake this study. We are particularly grateful to Mrs Evelyn Majidi of the Population Sciences Division
for her kind assistance at every stage of the project. All correspondence regarding this paper should be
addressed to Dr. Bamikale J. Feyisetan at the above address.
1 Three important dimensions of disease management have usually been noted: perceptions of aetiology;
perceived, or adopted, preventive measures; and perceived, or adopted, curative measures. While it is
expected that people's perceptions of disease aetiology and their choice of preventive or curative measures
will be correlated, exposure to several health care instructions and ideas as well as socio-economic
constraints may create a situation in which they are not (Feyisetan 1992).
2 Until the introduction of the Structural Adjustment Programme in Nigeria, health services were mainly
provided by the public sector at highly subsidized rates; thus the individual paid little or nothing for these
services. However, with the introduction of the Structural Adjustment Programme, the costs of receiving
treatment from the health facilities seem to have gone beyond the reach of many people with the result that
health seeking behaviour of many people may have been altered.
3 The 1990 Nigeria Demographic and Health Survey also shows that many under-five children suffered from
fever (32.6%), diarrhoea (17.9%), and acute respiratory infection (6.7%) in the two weeks preceding the
survey.
222 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
In this paper, we examine (1) the mothers' perceptions of the aetiology of the three most
cited childhood diseases in our study areas, measles, diarrhoea and fever, and the effect of
these perceptions on the mothers' suggested curative measures; and (2) the persistence of the
belief in abiku4 and how this cultural belief can influence mothers’ management of childhood
diseases. Since, for most mothers, perceptions of the aetiology of the childhood diseases are
rooted in cultural beliefs, a brief review of disease-specific cultural beliefs is undertaken. In
order to determine the effect of socio-economic factors, the mothers' perceptions of the
aetiology of the childhood diseases, their recommended curative measures and the belief in
abiku are examined according to selected socio-economic variables5.
Cultural beliefs and management of childhood diseases
Two aspects of disease management are examined: perceptions of causation and the types of
curative measures that are perceived as adequate by mothers when children have episodes of
the three childhood diseases.
Several authors have emphasized the need to consider the cultural beliefs and practices of
people when designing measures and programs aimed at improving their health (Suchman
1983; Ubomba-Jaswa 1988; Feyisetan 1988, 1992; Feyisetan and Adeokun 1992). Suchman
(1983) noted that the need for health planners to understand the culture of their population
arises from the fact that the meaning of illness, and behavioural responses to illness, are basic
factors influencing the reactions of the public to public health programs.
Feyisetan and Adeokun (1992) also noted that the simplicity of available modern curative
measures or the relative availability of such measures may not be sufficient conditions for
their adoption on a large scale. The extent to which modern methods are adopted may still
depend on the people's conception of the causes of ill-health and on their level of conviction
about the efficacy of such methods. Where conflicting views are held about the causes of
ailment, people may be confused as to which of the traditional or modern methods is the
appropriate treatment or preventive regime. Modern health services may be under-used or
used in conjunction with traditional methods in societies with fatalistic views about certain
diseases.
Two diseases, measles and diarrhoea, are particularly important in understanding the role
of cultural beliefs in disease management among the Yoruba. In traditional Yoruba society,
episodes of measles are usually attributed to a variety of causes which have no link with the
concept of a virus (Odebiyi and Ekong 1982). Each episode of measles is traditionally
considered as a punishment for breaching family taboos or as an evil deed from witches or
enemies. The belief that disease episodes are caused by enemies is usually stronger in
polygynous households where co-wives are natural suspects. With this kind of fatalistic view
about the cause of measles, parents, especially in the rural areas with limited access to health
facilities, may not seek modern medical care when children have measles.
4 Abiku is the Yoruba word for children believed to have come from the spirit world, who can die at will
unless certain rituals are performed.
5 Suggested curative methods do not necessarily measure the actual curative methods adopted by the
mothers. The suggested curative methods measure, at best, what the mothers perceived to be the ideal way to
manage the disease episodes; they reflect what the mothers would have done in an ‘ideal’ situation.
However, the actual measures adopted by the mothers when their children are sick may depend on many
factors which include availability of resources and accessibility to modern health facilities.
Mothers’ management of childhood diseases in Yorubaland 223
Health Transition Review
Traditional belief systems may become less important for diseases for which there are
cheap and compulsory immunization programs. If cheap or free immunization is made
available to all children and there is public enlightenment on its benefits, parents are likely to
obtain vaccinations for their children even when their perceptions of the aetiology of the
diseases have not changed6.
While measles has usually been perceived to be a deadly disease among the Yoruba,
diarrhoea is, in most cases, perceived merely as a means of getting rid of body impurities or as
a sign of ‘teething’, ‘crawling’ or ‘stretching’. Some mothers also believe that diarrhoea is
caused by consumption of sweet things, and are less likely to administer oral rehydration
solution (ORS) to their children since ORS contains sugar and salt. Since exposure to the risk
of infection is not usually recognized by parents, diarrhoea is not perceived as a major
problem until the baby is almost completely dehydrated. It is not uncommon to find mothers
who recommend less feeding during diarrhoea since everything the baby takes will be ‘passed
out anyway’.
The effects of disease-specific beliefs on health seeking behaviour are usually
compounded by the non-disease-specific beliefs. Of the non-disease-specific beliefs, the
Yoruba belief in abiku (see note 4) is relevant to our understanding of disease management of
the mothers. Yoruba people believe that some children come from the ‘spirit world’ and will
return there after a short time on earth unless certain rituals are performed. It is important in
any study of disease management among the Yoruba to investigate the persistence of this
belief and also to examine the effect of such beliefs on curative measures likely to be adopted
when children, believed to have come from the spirit world, are sick.
Source of data
Data for this study were obtained from a survey on the ‘Impact of Cultural Beliefs and
Practices on Child Health among the Yoruba’. The survey was undertaken in the rural and
urban areas of selected local government areas (LGAs) of Ondo and Ekiti States. The
headquarters of the selected LGAs served as the urban areas while two villages, one near and
one far from the headquarters, were selected to represent the rural areas. Two rounds of the
survey were undertaken, the first in September 1991 and the second in April-May 19927.
In preparation for the 1991 population census, the National Population commission
divided Nigeria into Supervisory Areas, each of which consisted of four to six Enumeration
Areas. In the urban areas three supervisory areas were randomly selected and from each
supervisory area, two to three enumeration areas were selected. Three supervisory areas were
also selected in villages where there were more than three supervisory areas8. We
systematically selected housing units from the ones already identified by the National
6 It is, however, unfortunate that, in many African countries, the dwindling economic fortunes have limited
the availability of vaccines with the result that vaccination coverage is falling.
7 When the survey was conducted in 1991-92, Ekiti State had not been created; it was still part of the former
Ondo State. The choice of the Local Government Areas where the study was undertaken was dictated by the
desire to have a sample that reflects all the major sub-ethnic groups in the former Ondo State. Thus, two local
government areas, Ijero and Isokan, were selected from the present Ekiti State and three, Akoko North,
Akure and Ikale, were selected from the present Ondo State. The numbers of women interviewed in each
local government area are: Ijero 196, Isokan 375, Akoko North 281, Akure 315 and Ikale 285.
8 Sometimes we had to go to two or three hamlets that constitute a village.
224 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
Population Commission as part of the preparation for the 1991 census exercise. Based on the
decision to select one household per housing unit, a random selection of one household was
undertaken where there were two or more households in a housing unit. Women aged 15-49
years with at least one live birth constituted the study population.
Information was collected through three sources: the formal interview, in-depth
interviews coupled with ethnographic observations, and focus-group discussions. This paper
is, however, primarily based on information collected through formal interviews but the
results are sometimes supported by findings from the qualitative data. Two interview
schedules were used: the household and the individual schedules. The household schedule
contained questions on every member of the household and was used to determine eligible
women for the individual schedule. The individual schedule contained questions on maternal
and household characteristics; biological attributes of under-five children at birth and at the
time of survey; maternal and child nutrition; measures of malnutrition among under-five
children (weights and heights); sanitation and environmental conditions; incidence and
prevalence of childhood diseases; perceptions of disease aetiology; diagnosis of diseases;
preventive and curative measures; antenatal and postnatal child care practices, including
immunization; child feeding practices; fertility, contraception and mortality.
To obtain information on different aspects of childhood disease management, women
were first asked to mention the common childhood diseases in their community. For each
disease mentioned, they were asked to state what they perceived as the cause and the
preventive and curative measures that should be adopted. These responses were later
compared with medically established causes, prevention and cure of these diseases. Stated
causes that differed from medically established causes or had mythological underpinnings
were labelled ‘inadequate’ while those which implied causative agents already identified
through medical research were labelled ‘adequate’. Suggested curative measures were
classified as ‘non-medical’ when they differed from medically established curative measures
and ‘medical’ when any of the medically established curative measures was mentioned or the
use of modern health facilities such as hospitals or health centres was suggested. Specific
questions were also asked of the mothers on their belief in abiku.
Four interviewers and a supervisor were assigned to each local government area. All the
interviewers were females with some exposure to population or health education, and were
fluent in the local dialect. Supervision of field work was continuous. In addition to
supervising field activities, the supervisors checked completed schedules on the spot.
Interviewers were asked to revisit households for which they had inadequately completed
schedules. The research team also paid regular visits to survey areas to ensure adherence to
survey procedures, determine survey problems and monitor the progress of work. Internal
consistency checks suggest that the data from this survey are generally of good quality.
Method of Analysis
Besides simple cross-tabulations, the major tool of analysis is the logistic regression model.
The logistic regression procedure was adopted to estimate the net impact of socio-economic
variables on our dichotomously measured dependent variables: whether or not a mother would
know the correct causes of each of the childhood diseases; and whether or not a mother
believed in abiku.
For the regression models, the independent variables are dichotomous and as such assume
the value one or zero. To avoid having a singular matrix that would yield no unique solution,
especially with the inclusion of a constant term, one category of each binary independent
variable is thus omitted in the estimation of the parameters of the equation. The excluded
categories are depicted as RC (reference category) in Tables 2 and 5.
Mothers’ management of childhood diseases in Yorubaland 225
Health Transition Review
Results
The results are presented in two parts. In the first part, we present the mothers' perceptions of
the aetiology of the three childhood diseases, measles, diarrhoea and fever,9 and the effect of
such perceptions on the type of curative measures suggested by the mothers. Questions on
aetiology, prevention and cure were only asked of respondents who identified these diseases
as major childhood diseases in their community. Of the 1559 ever-married respondents
selected for this study10, 1125, 1248 and 1404 identified measles, diarrhoea and fever,
respectively, as major childhood diseases. In the second part, we present our findings on the
persistence of the belief in abiku and the likely effect of this belief on mothers' management
of childhood diseases.
Perceived aetiology of the childhood diseases and suggested curative measures
Perceived aetiology
The percentages of mothers who had adequate perceptions of the causes of measles, diarrhoea
and fever are presented in Table 1. Panel I of Table 1 shows that 4.4 per cent of the mothers
have adequate knowledge of the cause of measles, 55.8 per cent of diarrhoea, and 66.4 per
cent of fever. With respect to causes of measles, there has been a considerable shift from
supernatural to natural causes even though the stated natural causes are inadequate. A
majority of the mothers gave ‘too much heat’ or ‘dry season’ as the cause of measles. Almost
all mothers with inadequate perceptions of the causes of diarrhoea attributed it to ‘teething’.
Causes of diarrhoea that are classified as adequate include ‘drinking contaminated water’ and
‘consumption of contaminated food’. Responses that attribute malaria fever to mosquito bites
are classified as adequate while inadequate responses include ‘working/playing too much in
the sun’ and ‘dry season’.
9 While measles and diarrhoea are well recognized, fever is a symptom rather than a disease (Feyisetan and
Adeokun 1992). Although fevers are often caused by malaria in the study areas, they may also be caused by
other diseases. However, our respondents were clear about the kind of fever they referred to: fever caused by
malaria.
10 Selection was based on the woman's ability to identify at least one childhood disease.
226 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
Table 1
Percentages of women with adequate knowledge of the causes of childhood diseases
FactorsaWomen with adequate knowledge of the cause of:
Measles
%NDiarrhoea
%NFever
%N
I. All women 4.4 1086 55.8 1202 66.4 1357
II. Education
None 1.6 317 45.6 371 57.7 411
Primary 4.6 389 56.4 424 68.1 480
Secondary &
higher 6.6 380 64.6 407 72.3 466
III. Age (years)
15-24 3.3 180 62.6 206 62.2 249
25-29 5.2 231 61.0 259 71.6 278
30-34 7.5 226 53.3 246 71.1 276
35-39 4.8 186 60.4 192 73.1 219
40-49 1.8 223 48.2 257 58.8 289
Age unknown 0.0 40 31.0 42 41.3 46
IV. Rural-urban
residence
Rural 1.4 507 44.7 562 56.3 631
Urban 7.1 579 65.6 640 75.3 726
V. Religion
Catholic 1.2 163 48.6 181 65.5 203
Protestant 5.5 544 49.8 576 62.9 677
Other Christian 4.0 273 70.4 321 72.9 343
Islamb4.7 106 56.5 124 68.7 134
Notes:
a All the variables are significantly correlated with perceived causes of childhood diseases at the 5%
level
b Includes 16 women who were adherents of traditional religion
To determine the factors that influence the likelihood that a mother would have accurate
knowledge of the aetiology of the three childhood diseases, responses were examined
according to selected characteristics of the mothers: education, age, urban-rural residence and
religion. The results, presented in panels II to V of Table 1, indicate that mother's education
and the likelihood that she would know the correct causes of the diseases are positively
correlated; mothers in the urban areas are more likely than rural mothers to have adequate
knowledge of the causes of the diseases; and while the ‘Other Christian’ mothers (Christian
mothers who were neither Catholic nor Protestants) are most likely to have adequate
perceptions of the aetiology of diarrhoea and fever, Protestant mothers are most likely to have
adequate knowledge of the causes of measles. With respect to age, the table shows that
mothers below the age of 40 years have more adequate knowledge of the causes of these
diseases than those above 40 years of age; within the age range 15 to 39 years, there is little
variation in the proportions of mothers who have adequate knowledge of the causes of these
diseases.
Mothers’ management of childhood diseases in Yorubaland 227
Health Transition Review
Since each of the variables is significantly related, at the five per cent level, to the
probability that a mother would know a correct cause of an identified childhood disease, a
logistic regression model was then estimated to determine the net effect of each variable on
the likelihood that a mother would have an accurate knowledge of the cause of each disease11.
The estimated coefficients are presented in Table 2, which shows that knowledge of the
correct causes of these diseases increases with mother's education; urban residence is
associated with more adequate knowledge of the causes; the effect of age diminishes
significantly on controlling other variables; and ‘Other Christian’ women are more likely to
have an adequate knowledge of the causes of diarrhoea and, to some extent, fever. For
measles, knowledge of correct causes is independent of religion once other factors are
controlled. The reason for the positive effect of education and urban residence on knowledge
of disease aetiology may not be surprising. More educated mothers and urban residents are
usually more exposed to current ideas than less educated mothers and rural residents.
Unfortunately, it has been difficult to explain the differences in knowledge by religion. The
qualitative data do not provide any reason for the differences, especially since the focus
groups were not constituted on the basis of religion.
Table 2
Effects of socio-economic factors on the likelihood that a woman will have accurate knowledge of
the causes of childhood diseases: the logistic regression modela
Measles Diarrhoea Fever
Coefficients of:
Constant -5.571 0.142 0.227
Education
None RCbRC RC
Primary 0.744 (0.528) 0.289 (0.159) 0.328
*
(0.155)
Secondary & higher 1.158* (0.553) 0.482**(0.188) 0.583
**
(0.189)
Age (years)
15-24 0.001 (0.705) 0.198 (0.224) -0.322 (0.214)
25-29 0.478 (0.643) 0.137 (0.212) 0.144 (0.211)
30-34 1.056 (0.586) -0.059 (0.192) 0.339 (0.189)
35-39 0.668 (0.626) 0.351 (0.204) 0.502
*
(0.201)
40-49 RC RC RC
Age unknown -4.505 (14.957) -0.812* (0.373) -0.721
*
(0.334)
Rural-urban residence
Rural RC RC RC
Urban 1.409**(0.423) 0.748**(0.126) 0.733
**
(0.124)
Religion
Catholic -0.824 (0.784) -0.767**(0.201) -0.218 (0.200)
Protestant 0.446 (0.370) -0.903**(0.155) -0.475
**
(0.152)
Other Christian RC RC RC
Islam 0.232 (0.566) -0.742**(0.228) -0.275 (0.231)
N of cases 1086 1202 1357
Model Chi-Square 43.434 116.665 99.655
Degree of freedom 11 11 11
11 Bivariate relationships may be spurious and hence may disappear on controlling other variables.
228 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
Notes: a Standard errors are in parentheses; b Reference category
** Significant at 1% level; * Significant at 5% level
Perceived curative measures
Suggested curative measures are classified into two categories: ‘modern medical’ and ‘non-
medical’12. Panel I of Table 3 shows that 72.3 per cent, 77.4 per cent and 90.2 per cent of all
women suggested modern medical curative measures for measles, diarrhoea and fever,
respectively13. For all the diseases, the percentages of women who suggested modern medical
curative measures are much higher than the percentages who knew the correct causes of these
diseases.
Table 3
Percentages of women indicating modern medical curative methods for childhood diseasesa
Factors Childhood Disease
Measles Diarrhoea Fever
I. All women 73.0 (3.2) 77.4 (19.6) 90.6 (27.9)
II. Education
None 74.1 (2.5) 79.0 (11.3) 91.0 (24.3)
Primary 75.1 (4.6) 79.0 (20.0) 91.0 (29.0)
Secondary & higher 70.0 (2.4) 74.2 (24.6) 89.7 (30.0)
III. Age (years)
15-24 65.0 (2.2) 73.8 (27.7) 86.3 (33.3)
25-29 71.4 (0.4) 77.6 (19.3) 93.2 (29.9)
30-34 75.7 (5.8) 80.1 (21.1) 92.4 (25.4)
35-39 76.3 (4.3) 80.7 (19.3) 90.3 (24.2)
40-49 73.5 (4.0) 72.8 ( 9.3) 88.9 (27.7)
Age unknown 85.0 (0.0) 90.5 (16.7) 93.5 (21.7)
IV. Rural-urban residence
Rural 77.5b (2.4) 77.2 (13.3) 91.6 (23.0)
Urban 69.1c (4.0) 77.5 (23.8) 89.7 (32.2)
12 Most of the responses in the category ‘modern medical’ are very imprecise. Initially, responses on
curative measures were grouped into four categories: (1) ‘medically recommended’, if a specific medically
recommended measure was mentioned; (2) ‘traditional/less precise/inadequate’, if response implies the
adoption of a traditional, or totally inaccurate measure; (3) ‘health facility’, if respondent merely suggests the
use of the hospital/health centre; and (4) ‘drugs/medicine’, if respondent merely recommends the use of drug
or medicine without specifying the type of drug. However, given the generally low level of medical
knowledge in our society, we thought it not feasible to expect most women to know precisely the
recommended curative measures or what the doctors or nurses offer to cure these diseases in the hospital. In
addition, precise knowledge of curative measures may depend on whether a woman has a child who has
experienced an attack of measles. We therefore decided to lump (1), (3) and (4) together as ‘modern
medical’. Also, rather than say the actual word ‘immunization’, many women just said ‘take the child to the
hospital for necessary attention’ when asked how measles could be prevented.
13 Our classification into ‘modern medical’ and ‘non-modern medical’ has nothing to do with the efficacy of
the methods.
Mothers’ management of childhood diseases in Yorubaland 229
Health Transition Review
V. Religion
Catholic 73.6b (4.3) 84.0b (19.3) 90.6 (31.5)
Protestant 73.5 (2.9) 74.3c (16.3) 90.3 (26.4)
Other Christian 76.2 (3.7) 79.1 (16.5) 92.7 (25.7)
Islam 61.3c (1.9) 78.4 (36.3) 86.6 (35.8)
VI. Perceptions of disease aetiology
Adequate 97.9b77.5 93.0b
Inadequate 71.9c77.2 85.7c
Notes: a The percentages of women who mentioned a specific curative measure that is medically
recommended are in parentheses
b and c indicate values of a variable that are significantly different from each other and thus make the
variable significantly related to the likelihood that a woman will know (and adopt) a modern medical
curative measure at 5% level
The big difference between the percentage of women who knew the correct cause of
measles (4.4) and the percentage who suggested modern medical curative measures (73.0)
reinforces our earlier contention that levels of adequacy of knowledge about disease causation
may be irrelevant in the choice of curative measures when appropriate preventive and curative
measures are available at low cost. The availability of primary health care centres, and the
enlightenment campaigns associated with launching the expanded program of immunization
in these communities, must have given the women an opportunity to know more about
modern preventive and curative measures for measles irrespective of their beliefs about
causation. This observation was reinforced by the qualitative data which show that large
numbers of mothers are already aware that measles can be prevented by vaccination.
Curative measures that imply isolation of the patient, use of medical lotion to reduce
itching, and use of modern health facilities, especially the hospital, are classified as ‘modern
medical’ for measles. Of the curative measures classified as ‘non-medical’ for measles,
‘consultation with traditional healers’ is predominant. For fever, curative measures classified
as ‘modern medical’ include the use of anti-malaria drugs, injections, and modern health
facilities, and curative measures classified as ‘non-medical’ include the use of concoctions
prepared from agbo, locally available herbs, and consultation with traditional healers14. For
diarrhoea, any response that does not indicate oral rehydration solution, whether or not self-
prepared, nor imply the use of a modern health facility, is regarded as ‘non-medical’.
To determine whether suggested curative measures differ by socio-economic factors,
responses were also examined by respondents' age, education, urban-rural residence, religion
and perceptions of disease aetiology. The results, presented in panels II to VI of Table 3,
indicate that the percentage of women who suggested modern medical curative measures for
each of the three diseases does not vary by age, education, and, except for measles, by urban-
rural residence. Religion is moderately associated with the likelihood that a woman would
report modern medical curative measures for measles (panel V), and women who have
14 Some locally available herbs, agbo, have been reported as efficacious especially in the treatment of
malaria fever. However, it would be presumptuous to assume that all types of agbo are efficacious and hence
could be placed on the same level of efficacy as the modern medical methods. Until new findings indicate the
contrary, we feel safer to assume that the well tested modern medical methods are generally more efficacious
than agbo and the other traditional methods.
230 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
adequate knowledge of the causes of measles and fever are more likely to report modern
medical curative measures (panel VI).
Care should be taken in interpreting the lack of significant correlation between the socio-
economic factors and the likelihood that a woman would report modern medical curative
measures. Because the category ‘modern medical’ is very broad, the relationship between
suggested curative measures and socio-economic factors may be confounded. This realization
prompted us to examine the percentages of women who suggested specific medically
recommended curative measures by socio-economic factors. We found that for diarrhoea and
fever, the percentages of women who recommended specific modern medical curative
measures vary significantly by the socio-economic variables (see figures in parentheses in
Table 3)15.
Belief in abiku and the effect of this belief on disease management
In addition to disease-specific questions, the respondents were also asked to indicate whether
they believed that there are abiku children. Panel 1 of Table 4 suggests that the majority of the
women believe in abiku children. While 56.2 per cent of the women believed in abiku
children, 30.6 per cent did not share this belief and 13.2 per cent were unsure of their feelings.
Asked how an abiku can be identified, mothers who believed in abiku gave such responses as
repeated deaths; evidence of deformity from past deaths; frequent indisposition; non-
responsiveness of their illnesses to modern medical care; and verification from traditional
healers or soothsayers. Both the focus-group discussions and in-depth interviews point to high
mortality rates among abiku children. In a focus-group discussion among women aged 40
years and above in Ijero, a participant exclaimed:
Why do you think we call them abiku? It is because they die at will...they really don't have
to be sick for long before they die.
Table 4
Percentages of women believing in abikua
Factors Percentage N of cases
I. All women 56.2 (13.2) 1502
II. Education
None 61.5 (14.1) 455
Primary 57.2 (12.2) 540
Secondary & higher 50.3 (13.4) 507
III. Age (years)
15-24 48.6 (15.9) 276
25-29 53.1 (15.6) 307
30-34 56.3 (12.0) 300
35-39 55.6 (10.1) 248
40-49 62.6 (13.1) 321
Age unknown 78.0 (6.0) 50
IV. Rural-urban residence
15 A logistic regression model was not estimated for the probability that a woman would recommend
modern curative methods since most of the variables are not significantly correlated to the dependent
variable at the bivariate level.
Mothers’ management of childhood diseases in Yorubaland 231
Health Transition Review
Rural 55.0 (12.7) 724
Urban 57.3 (13.6) 778
V. Religion
Catholic 54.1 (12.4) 218
Protestant 51.0 (16.4) 781
Other Christian 69.3 ( 8.9) 358
Islam 55.2 ( 7.6) 145
a Percentages of ‘don't know’ responses in parentheses
Table 4 also shows that education is significantly negatively correlated with the
percentage of women who believe in abiku (panel II); age is positively correlated with the
likelihood that a woman will believe in abiku (panel III); urban-rural residence is not related
to the belief in abiku (panel IV); and ‘Other Christian’ mothers are more likely than any other
religious group to believe in abiku (panel V).
To determine the net effect of these characteristics, we estimated a logistic regression
model that relates the likelihood of believing in abiku to the socio-economic factors. The
coefficients of the model which are presented in Table 5 indicate that education has no
significant impact on the belief in abiku once other variables are controlled; mothers between
15 and 24 years of age are less likely than older mothers to believe in abiku, and mothers over
24 are equally likely to share this belief; the belief in abiku is equally strong in both the rural
and urban areas; and religion has an effect on this belief with the ‘Other Christian’ women
still more likely than any other religious group to retain this belief.
232 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
Table 5
Effects of socio-economic factors on the likelihood that a woman will believe in abiku: the
logistic regression modela
Coefficients of:
Constant 1.132
Education
None RCb
Primary -0.046 (0.142)
Secondary & higher -0.307 (0.169)
Age (years)
15-24 -0.484* (0.196)
25-29 -0.345 (0.190)
30-34 -0.275 (0.172)
35-39 -0.309 (0.178)
40-49 RC
Age unknown 0.643 (0.367)
Rural-urban residence
Rural RC
Urban 0.154 (0.119)
Religion
Catholic -0.706**(0.181)
Protestant -0.804**(0.138)
Other Christian RC
Islam -0.641**(0.2060)
N of cases 1502
Model Chi-Square 65.851
Degree of freedom 11
Notes: a Standard errors are in parentheses; b Reference category
**Significant at 1% level; * Significant at 5% level
To determine the effect of this belief on disease management, we asked mothers who
believe in abiku to indicate whether a suspected abiku should be subjected to the same
treatment as a non-abiku, and where treatment should be sought for a suspected abiku when
he or she is sick. The responses are presented in panels I and II of Table 6. Panel I shows that
62 per cent reported that a suspected abiku should not be treated like an ordinary child when it
is sick. To the question on where a suspected abiku should be treated, 70.3 per cent answered
traditional healers' home, church or mosque (panel II of Table 4); 91.5 per cent of women who
stated that an abiku child should not be treated like the others suggested the traditional healer's
home, the church or mosque as the place for treatment. These findings reflect the mothers' belief
that illnesses of abiku are not caused by natural but by supernatural forces. Thus, such
illnesses are believed to be incurable by ‘mere administration of drugs or injections in the
hospitals’, as was also reported in the focus-group discussions.
Table 6
Health seeking for abiku (%)
Mothers’ management of childhood diseases in Yorubaland 233
Health Transition Review
I. Treat abiku like non-abiku child when it is sick?
Yes 35.2
No 61.8
Don't know 3.0
No. of cases 844
II. Where should abiku be treated?
Hospital/health centre 28.5
Traditional healer 60.7
Church/mosque 9.6
Other 1.2
No. of cases 844
Summary and conclusion
In the preceding analysis, an attempt has been made to examine how perceptions of disease
aetiology and the persistence of the belief in abiku could influence disease management. Data
collected from Ondo and Ekiti States, Nigeria were used. It has been demonstrated that high
percentages of Yoruba mothers do not have an accurate knowledge of the causes of the
selected childhood diseases, especially measles; many of the mothers recommended modern
curative methods in spite of the high level of ignorance about disease causation; the belief in
abiku is still strong among mothers in Yorubaland; and the curative measures likely to be
adopted by a mother may depend on whether the sick child is believed to be an abiku.
An important issue that emerges is the apparent contradiction between what the mothers
perceived to be the ideal curative measures and the curative measures that they might end up
adopting as a result of their belief in abiku. The same group of women who recommended the
use of modern curative methods, even when they were ignorant of disease cause, also opined
that modern curative methods would be inadequate for a sick abiku. Thus, they recommended
the use of traditional healers and religious institutions irrespective of the nature of illness. The
probability, therefore, that a child will be given adequate medical care depends on whether or
not it is believed to be an abiku. This is a serious matter in a society where over half16 of the
women who are mainly responsible for seeking adequate health care for the children still
retain the belief in abiku.
This study has demonstrated, once more, the need to take into consideration people's
beliefs and practices when implementing health policies. Such consideration demands that
implementation of health policy is localized. The success of any health policy depends very
much on how much effort is put into streamlining beliefs that are at variance with medical
principles. For instance, the use or non-use of modern health facilities in a Yoruba community
may be affected by how much the people believe in abiku. People who believe that illnesses
of abiku are unresponsive to modern medical care are not likely to use health facilities for
children thought to be abiku. The fact that considerable proportions of Yoruba mothers with
some education still retain this belief shows how strong the belief is and thus it cannot be
dismissed as unimportant. There is, therefore, a need for more public enlightenment on this
issue.
16 About two-thirds of the women probably share this belief since 13.1 per cent of the women were confused
on the issue at the time of the interview.
234 Bamikale J. Feyisetan, Sola Asa and Joshua A. Ebigbola
Health Transition Review
References
Animashaun, A. 1977. Measles and blindness in Nigerian children. Nigerian Journal of Pediatrics 4,
1:10-13.
Ayeni, O. and S.O. Oduntan. 1980. Infant mortality rates and trends in Nigerian rural populations.
Journal of Tropical Pediatrics and Environmental Child Health 26,1:7-10.
Baxter-Grillo, D.L. and F.E.A. Leshi. 1964. Factors influencing the occurrence of neonatal tetanus in
Ibadan. West African Medical Journal 13:23-8.
Feyisetan, Bamikale. 1988. ‘Understanding the relationship between child care practices and child
mortality: the need for ethnographic approaches in child mortality studies. Pp. 7.2.37-7.2.54 in
Proceedings of the African Population Conference. Liege, Belgium: International Union for the
Scientific Study of Population.
Feyisetan, Bamikale J. 1992. The need to investigate disease-specific and other health related cultural
beliefs in child mortality studies. Pp. 125-140 in Proceedings of the International Seminar on
Morbidity, Mortality and Social Policy: A Focus on the Young and the Elderly.
Feyisetan, Bamikale J. and L. Adeokun. 1992. Impact of child care and disease treatment on infant
mortality. Pp. 145-159 in Mortality and Society in Sub-Saharan Africa, ed. E. Van de Walle, G.
Pison and Mpembele Sala-Diakanda. Oxford: Clarendon Press.
Morley, D.C. and K.M. MacWilliam. 1961. Measles in a Nigerian community. West African Medical
Journal 16:246-253.
Morley, D., M. Woodland and W.J. Martin. 1963. Measles in Nigerian children. Journal of Hygiene 61,
1:115-134.
Morley, D., Margaret Woodland and W.J. Martin. 1966. Whooping cough in Nigerian children. Tropical
and Geographical Medicine 18:169-182.
Odebiyi, A.I. and S.C. Ekong. 1982. Mother's concept of measles and attitudes towards the measles
vaccine in Ile-Ife, Nigeria. Journal of Epidemiology and Community Health 36, 3:209-213.
Ogunlesi, T.O. 1961. Respiratory infections in the preschool child: a review of 435 cases admitted to
Adeoyo Hospital, Ibadan. West African Medical Journal 10:231.
Suchman, E. 1983. Sociology and the Field of Public Health. New York: Russell Sage Foundation.
Tomkins, A. 1981. Nutritional status and severity of diarrhoea among preschool children in rural
Nigeria. Lancet 18 April:860-862.
Ubomba-Jaswa, S.R. 1988. Culture and health: lessons from data collection on child health in Ghana.
Pp. 100-101 in Proceedings of a Workshop Held in Accra on Research Issues in Child Health and
Child Care. Ottawa: IDRC.
