1. health and fitness
2. drugs

TMIH489
Tropical Medicine and International Health
volume 4 no 12 pp 836–845 december 1999
Maternal responses to childhood fevers: a comparison of rural
and urban residents in coastal Kenya
C. S. Molyneux1,2, V. Mung’ala-Odera1,T. Harpham3 and R.W. Snow2,4
1
2
3
4
Kenya Medical Research Institute (KEMRI)/Wellcome Trust Centre for Geographic Medicine Research, Kilifi, Kenya
Nuffield Department of Medicine, University of Oxford, John Radcliffe Hospital, Oxford, UK
The Kenya Medical Research Institute/Wellcome Trust Collaborative Programme, Nairobi, Kenya
Department of Urban Policy and Planning, South Bank University, London, UK
Summary
Urbanization is an important demographic phenomenon in sub-Saharan Africa, and rural-urban migration
remains a major contributor to urban growth. In a context of sustained economic recession, these demographic processes have been associated with a rise in urban poverty and ill health. Developments in health
service provision need to reflect new needs arising from demographic and disease ecology change. In malariaendemic coastal Kenya, we compared lifelong rural (n 5 248) and urban resident (n 5 284) Mijikenda mothers’ responses to childhood fevers. Despite marked differences between the rural and urban study areas in
demographic structure and physical access to biomedical services, rural and urban mothers’ treatment-seeking patterns were similar: most mothers sought only biomedical treatment (88%). Shop-bought medicines
were used first or only in 69% of the rural and urban fevers that were treated, and government or private
clinics were contacted in 49%. A higher proportion of urban informal vendors stocked prescription-only
drugs, and urban mothers more likely to contact a private than a government facility. We conclude that
improving self-treatment has enormous potential to reduce morbidity and mortality in low-income urban
areas, as has frequently been argued for rural areas. However, because of the underlying socio-economic, cultural and structural differences between rural and urban areas, rural approaches to tackle this may have to be
modified in urban environments.
keywords malaria, maternal treatment-seeking, shops, health facility access, Kenya
correspondence Sassy Molyneux, Kenya Medical Research Institute (KEMRI)/ Wellcome Trust Centre
for Geographic Medicine Research, P.O. Box 230, Kilifi, Kenya. E-mail: [email protected]
Introduction
The majority of the sub-Saharan African population live in
rural areas. However, the region currently has the highest
rates of urbanization in the developing world (Brockerhoff &
Brennan 1998), two-fifths of which has been attributed to
rural-urban migration (Chant & Radcliffe 1992). With mounting international concern about the health status of rapidly
growing low-income urban communities (Harpham 1997;
Gould 1998), an improved understanding of the way in which
these demographic processes influence treatment-seeking
patterns is urgently needed (Tanner & Harpham 1995).
Household responses to illness are influenced by socioeconomic and cultural factors (Kleinman 1980; CunninghamBurley 1990) and by ease of access to treatment sources
(Stock 1985; Kloos et al. 1987; Mburu et al. 1987; Glik et al.
1989). In sub-Saharan Africa, rural and urban populations
836
differ demographically, in socio-economic and cultural composition, and in proximity to formal and informal treatment
sources. Urban populations are generally younger, better educated, and more ethnically heterogeneous than rural populations. Government health services, private practitioners,
pharmacies and shops selling over-the-counter medications
are concentrated in urban areas. These rural-urban differences imply that urban residents use biomedical (or ‘modern’)
services and drugs more often and at an earlier stage than
rural residents, and that rural residents more readily turn to
biocultural (or ‘traditional’) therapy.
Urbanization patterns observed all over sub-Saharan Africa
are reflected in Coastal Kenya, where urban growth has
centred on Mombasa, the second largest city in the country.
Coastal Kenya is a malaria-endemic area, and malaria is the
principal cause of childhood mortality. Fever is a key symptom of uncomplicated malaria, but can rapidly develop into
© 1999 Blackwell Science Ltd
Tropical Medicine and International Health
volume 4 no 12 pp 836–845 december 1999
C. S. Molyneux et al. Maternal responses to childhood fevers in Kenya
severe, life-threatening malaria if not treated with appropriate antimalarials. This study aimed to compare the
treatment-seeking patterns of rural and urban residents of
the dominant ethnic group in Coastal Kenya, the Mijikenda.
The focus was on Mijikenda mothers’ responses to uncomplicated childhood fevers.
Materials and methods
Study populations
Using a standard definition of ‘urban’, i.e. ‘a proclaimed
urban centre/town/city with a population of 10 000 or more’
Figure 1 Location of the study area.
(Parry 1995), one rural and one urban area were selected for
the study.
A 446.6-km2 area to the north and south of Kilifi Town was
selected as the rural study area (Figure 1). The population
consists largely of Mijikenda farmers who supplement subsistence incomes with cash crops and remittances from family
members working in the nearby urban centres of Kilifi,
Mombasa and Malindi. The Mijikenda family system is
patriarchal and polygamous. The geography, demography,
anthropology and disease ecology of the area has been
described in detail elsewhere (Parkin 1991; Snow et al. 1994).
Ziwa la N’gombe, a 0.6-km2 area on the outskirts of
Kilifi District
Rural study area
sub-section
Rural study area
(444.6 km2)
Kilifi creek
Mtwapa creek
Mombasa
District
Urban study area
(0.6 km2)
25 km
© 1999 Blackwell Science Ltd
837
Tropical Medicine and International Health
C. S. Molyneux et al.
volume 4 no 12 pp 836–845 december 1999
Maternal responses to childhood fevers in Kenya
Mombasa City, was selected as the urban study area
(Figure 1). Ziwa la N’gombe is approximately 60 km from the
rural study area, and has the low-income, densely populated
housing characteristics common to Mombasa and other
urban areas of East Africa. Household incomes are derived
largely from formal and informal work in and around
Mombasa, much of which is linked to the tourism industry.
The history, geography and demography of Mombasa city
have been described elsewhere (Kindy 1972; Government of
Kenya 1990; Bambrah 1996), and a number of studies offer an
insight into the lives of the city’s residents (Parkin 1979;
Mazrui 1996; Malombe & Kanyonyo 1997).
Mapping and enumeration
The position of every household, shop, pharmacy and health
facility in both study areas was recorded using hand-drawn
maps. Maps were used to conduct a census of every resident
in each household. To ensure consistency in mapping and
enumeration, a household was defined as ‘a person or group
of people living in the same house, who are answerable to the
same head and share a common source of food and/or
income’. People were regarded as resident if they had a sleeping place within a household and had already or intended to
remain there for a total of at least six months. The six-month
‘cut-off’ in defining residency was based on extensive pilot
work in and around both study areas.
To identify the relative position of households to each
other, and to every possible treatment source, digitized maps
were developed for a subsection of the rural study area
(Figure 1) and the entire urban area using a combination of
scanned hand-drawn maps and a hand-held satellite navigational system (SatNav® Trimble Navigation, Europe Ltd,
UK). Digitized maps were computerized using the software
package MapInfo® (Troy Ltd, USA). All shops and pharmacies on the digitized maps were revisited, and an audit
taken of all antimalarials stocked. Semi-structured interviews
were also held with a range of health care providers.
Mothers’ survey
Consenting mothers from the samples were interviewed. In
addition to socio-economic and demographic data, illness
histories over the last two weeks were taken for all resident
children aged 6 months up to the 10th birthday. For each illness episode, details were documented on symptoms and
their perceived severity, duration, types of treatment sought,
and from whom and at what stage of the illness treatment
was sought. Data were entered onto closed and coded questionnaires developed through pilot work. Several questions
included specified prompts to facilitate data accuracy and
comparability.
Data quality and analysis
Ten Mijikenda fieldworkers were carefully trained using participatory approaches in communication skills, form-filling
and demographic techniques. All data were double-entered
and verified in Foxpro (Version 2.6) and subjected to range
and consistency checks. Inconsistencies were verified in the
field through re-interview. Rural and urban mothers’ characteristics and treatment-seeking patterns were compared using
x2-tests (for proportions) and 2 sample t-tests (for continuous
variables) in SPSSV6.
Results
Basic demographics
Estimated coverage of the census work in both study areas
was over 95% of the total population. The demographic
structure of the populations is presented in detail elsewhere
(Molyneux et al. 1999a). Briefly, 82 594 people were enumerated in 7788 households in the rural area surrounding Kilifi
town (446.6 km2). 14 885 people were enumerated in 6458
households on the outskirts of Mombasa city (0.6 km2).
Urban households were generally smaller (mean 2.34; median
2; interquartile distance 2) than rural ones (mean 10.6;
median 9; interquartile distance 8).
Sampling of mothers
Following the census, random samples of rural and urban
resident women aged 15 years or above with the prime
responsibility for at least one resident child aged 6 months up
to the 10th birthday were interviewed in order to obtain more
detailed information. Rural resident mothers were initially
sampled from a survey of all women’s birth histories. In
follow-up interviews Mijikenda mothers who had never lived
in an urban area were identified. Urban resident Mijikenda
mothers were sampled from the census, which was structured
to enable the identification of lifelong urban resident
mothers, and mothers who had immigrated from a rural area.
838
Physical access to health services
GIS software was used to identify the ratio of residents to the
shops selling antimalarials, pharmacies, trained community
health workers (CHWs), private clinics and government dispensaries in and around the subsection of the rural and the
entire urban study area (Table 1). Shops were defined as
structures (ranging from temporary stands to permanent
buildings) which sell over-the-counter drugs as well as a number of other everyday goods. The majority of shops dispensed
antipyretics; 73% of rural and 96% of urban shops sold antimalarials, and a review of stocks was allowed in 92% of these
© 1999 Blackwell Science Ltd
Tropical Medicine and International Health
volume 4 no 12 pp 836–845 december 1999
C. S. Molyneux et al. Maternal responses to childhood fevers in Kenya
Table 1 Availability and access to treatment
sources in the study areas
Shops
Shops selling antimalarials
Pharmacies
Community Health Workers
Private clinics
Government dispensaries
Total interviewed
Rural study area*
Number of Ratio **
facilities
facility:people
Urban study area
Number of Ratio **
facilities
facility:people
159
116
000
022
009
003
203
143
137
003
007
009
000***
162
1 : 218
1 : 299
–
1 : 1714
1 : 3857
1 : 11570
–
1 : 109
1 : 114
1 : 5210
1 : 2232
1 : 1736
–
–
* based on mapping and enumeration in the rural subsection (Figure 1). ** based on total population estimates of rural subsection and urban area (n 5 34 712 and 15 629, respectively)
*** the nearest government dispensary is situated approximately 1.5 km from the urban study
area.
shops (Table 2). Of note is that 7% rural and 40% urban
shops had in stock either sulphadoxine-pyrimethamine (proprietary name Fansidar) and/or amodiaquine (proprietary
name Malaratab), over the study period prescription-only
drugs in Kenya. In addition to greater access to these drugs
through the informal sector, urban residents also have access
to them from the three pharmacies.
Community-based health care was evident in both study
areas. CHWs have been trained to supply a limited range of
drugs at subsidized rates from their homes, and to refer
patients to other facilities when necessary. Interviews revealed
that CHWs in the rural area varied in availability and activity,
and that in town they were not operating at all. In both areas
problems were largely due to failures in drug supply.
Alternative therapy sources include the peripheral government services (or dispensaries) and private clinics. Dispensaries operate on a cost retrieval basis, charging
approximately 20–40KSH (20–40 pence) to treat a febrile
child. Complications are referred to the nearest District
Hospital. Private clinics are relatively expensive: rural and
urban practitioners charge approximately 150 and 250KSH,
respectively, to treat a febrile child.
Traditional healers were not mapped but discussions revealed good physical access to a variety of healers (waganga)
in both study areas. Waganga are distinguished primarily
according to whether they are wa ramli (who diagnose) or
wa kutibu (who treat), and secondarily by their field of
specialization, of which there is a wide range. Waganga in
both areas differ greatly from each other in their reputation
and charges.
The urban population has a lower ratio of people to shops,
pharmacies and private clinics than the rural subsection population, and the latter has a lower ratio of people to CHWs
and government dispensaries (Table 1). However, all urban
residents live within 1 km of all facilities, and the nearest
© 1999 Blackwell Science Ltd
government dispensary is a short bus ride or half-hour ‘quick
walk’ away. 87% of the rural study area homesteads live
within 1 km of a shop, 55% within 2 km of a CHW, and only
32% within 2 km of a private clinic and/or government dispensary. The urban population therefore has better physical
access to informal and formal biomedical services in and
around their area of residence than the rural population.
In addition to these services, the rural study area subsection is transected by the main Malindi-Mombasa road
which, together with a number of tarred roads, provides the
population access to the administrative township of Kilifi,
where the district hospital and 5 private clinics are situated.
The central point of the rural subsection is approximately
10 km from the district hospital, and over 66% of households
Table 2 Availability of proprietary antimalarials in the study area
shops
Proprietary drug
name
Rural shops
(n 5 108)*
Urban shops
(n 5 125)
Chloroquine generic†
Dawaquin – adult†
Dawaquin – junior†
Malaraquin†
Homaquin†
Oraquin†
Rohoquin†
Laraquin†
Maladrin†
Fansidar‡
Malaratabs§
Camoquin§
01
32
12
91
26
13
00
00
03
00
08
00
005
061
031
118
061
025
001
002
015
002
050
003
* based on mapping and enumeration in the rural subsection
(Figure 1). † chloroquine-based; ‡ sulphadoxine–pyrimethaminebased; § amodiaquine-based.
839
Tropical Medicine and International Health
C. S. Molyneux et al.
volume 4 no 12 pp 836–845 december 1999
Maternal responses to childhood fevers in Kenya
Mothers’ characteristics
Age
, 35 years
$ 35 years
Marital status
Unmarried
Married
Religion
None/traditional
Christian/Muslim
Formal education
None
$ standard 1
Earns any cash
None
Yes
Children , 10
means (SD)
Household type
Nuclear
Extended
Husband’s residence
Household
Elsewhere
Table 3 Characteristics of lifelong rural and
urban resident sample of mothers
Rural percentage
(n 5 248)
Urban percentage
(n 5 284)
64
36
81
19
, 0.0001
13
87
20
80
, 0.05
70
30
32
68
, 0.0001
74
26
44
56
, 0.0001
44
56
02.5 (1.2)
40
60
01.9 (1.0)
5 0.455
20
80
68
32
, 0.0001
68
32
93
07
, 0.0001
P–value*
, 0.0001
* Derived using x2-tests and (for the continuous variable) a two sample t-test.
are within 2 km of a bus stop. The urban study area is situated alongside the Malindi-Mombasa road, and a constant
flow of buses and taxis ensures good access to health services
available elsewhere in the city. Mombasa District Hospital is
situated about 3 km from the urban study area.
Mothers’ characteristics
248 lifelong rural and 284 urban resident mothers participated in the detailed survey interviews. Their mobility patterns and the ties maintained between rural and urban areas
are described in detail elsewhere (Molyneux et al. 1999a). It
should be noted, however, that most (87%) of the urban
sample were born in a rural area, and had initially migrated
to town to stay with relatives (83%), primarily husbands. The
majority were familiar with the urban environment before
moving, were supported socially, economically and with some
form of housing on arrival in town, and intended to return to
live in a rural area in the future. One third had spent at least
10% of nights elsewhere in the year preceding the interview
or since migration into their current household. Much of this
‘circulation’ (76%) was to rural areas.
The rural and urban samples differed significantly in many
of their characteristics (Table 3). Rural mothers were
generally older, cared for a greater number of children
840
, 10 years, were more likely to be married, more likely to
follow traditional beliefs, and less likely to have received any
formal education than urban resident mothers. They were
also more likely to live in extended households, and to have
husbands resident elsewhere, primarily in town. Rural and
urban mothers did not differ significantly in their likelihood
to be earning some form of income. The majority of those
earning money were involved in small-scale businesses in and
around their communities.
Treatment-seeking for uncomplicated childhood fevers
During the course of the interviews, childhood fevers in the
two weeks preceding the interview were identified. A total of
317 childhood fevers were reported, 95% of which received
some form of therapy. The types of action taken and the
stage at which they were taken are shown in Table 4.
The most common actions were the purchasing of shopbought drugs (47%), visits to private clinics or hospitals
(20%), and other forms of therapy in the home (12%). Visits
to government hospitals constituted 10% of the actions
taken. Home treatment and the purchasing of shop-bought
drugs generally took place on the same day or the day after
symptoms were first observed (67%). Clinics, dispensaries
and hospitals were more likely to be contacted on the second
© 1999 Blackwell Science Ltd
Tropical Medicine and International Health
volume 4 no 12 pp 836–845 december 1999
C. S. Molyneux et al. Maternal responses to childhood fevers in Kenya
Table 4 When and where mothers responded to fevers
Home treatment
Shop drugs
Private clinic/hospital
Government clinic/dispensary
Government hospital
Healer
Other
Lifelong rural resident mothers
——————————————––––––––
Actions (%)
% actions taken on
n 5 147
day 0 and 1*
Urban resident mothers
—————————————––––––––—
Actions (%)
% actions taken on
n 5 236
day 0 and 1*
09
48
15
06
16
03
03
13
46
24
03
07
06
01
69
79
43
67
27
00
–
59
64
40
28
25
07
–
* Day 0 is the day the symptom was first observed.
day or later (62%). The proportion of each action taken on
day 0 or day 1 was not significantly associated with rural or
urban residence. However, for some actions the numbers are
very small.
Mothers’ treatment-seeking patterns for each fever were
grouped in several ways (Table 5). Distinguishing between
biomedical (‘modern’) and biocultural (‘traditional’) actions,
most responses to fever were biomedical: only 12% of fevers
were treated with herbs and/or by healers instead of or as
well as with biomedical therapy. Of the group who administered biocultural therapy, a healer was consulted in 4
instances by rural mothers and in 11 by urban mothers. A
total of 13 treated fevers were not treated outside the household using biomedical therapy. Of those that were (91% of
all reported fevers), the most common responses were to
administer shop-bought drugs only (49%), and to consult a
private practitioner (29%) or government service (19%). The
sequence of biomedical actions taken shows that most fevers
treated outside the household were treated using shop-bought
medications only or first (69%).
Lifelong rural and urban resident mothers’ treatmentseeking patterns were compared in three ways (patterns
defined as ‘other’ were excluded):
biomedical only (n 5 176), or biomedical and/or biocultural actions (n 5 36);
biomedical actions outside the household were shops
only (n 5 141), and/or government/private services
(n 5 139);
•
•
Table 5 Treatment-seeking for fevers in the
last two weeks
Type of therapy used
Biomedical only
Biomedical and/or biocultural
Total
Type of biomedical service used*
Shop-bought drugs (SDs) only
SDs and/or government service only
SDs and/or private practitioner only
Other
Total
Sequence of biomedical actions*
Shop-bought drugs only
Shop drugs r clinic/hospital
r Clinic/hospital
Other
Total
Lifelong rural resident
mothers n (%)
Urban resident
mothers n (%)
119 (90)
013 (10)
132 (100)
147 (86)
023 (14)
170 (100)
065
034
024
003
126
(52)
(27)
(19)
(2)
(100)
076
020
061
006
163
(47)
(12)
(37)
(4)
(100)
065
022
038
001
126
(52)
(17)
(30)
(1)
(100)
076
036
046
005
163
(47)
(22)
(28)
(3)
(100)
* Does not include fevers which were not treated outside the household, or were only treated by
a healer (6 rural and 7 urban fevers).
© 1999 Blackwell Science Ltd
841
Tropical Medicine and International Health
C. S. Molyneux et al.
•
Maternal responses to childhood fevers in Kenya
sequence of biomedical actions taken outside the household was shops only/first (n 5 199), or direct contact of
a government/private clinic or hospital (n 5 84).
With respect to therapeutic approach and sequence of biomedical actions taken, patterns were similar with no significant differences between rural and urban mothers. However,
there was a significant difference between in the type of biomedical services contacted (x2 16.71; P , 0.001): urban
mothers were almost twice as likely as lifelong rural resident
mothers to consult a private clinic, and less than half as likely
to consult a government service.
To explore variation in treatment-seeking patterns, rural
and urban mothers’ responses were combined and compared
by a range of demographic and socio-economic variables.
These included child’s age, mothers’ age, religious affiliation
and education, whether or not the mother was earning some
form of income, and (for urban mothers) length of urban
residence. The only differential demonstrating a significant
association was age of sick child. Younger children
(, 5 years) were more likely to be taken directly to a clinic or
hospital than children (x2 5 years (> 12.41; P 5 0.002). In
informal discussions, mothers explained that younger children are less able to take tablets, less able to explain what
they are suffering from, and more vulnerable to illness.
Discussion
Urbanization is an important demographic phenomenon in
Africa, and the health of a growing low-income urban population a cause of mounting international concern. Of interest
to health planners is the extent to which health service provision is meeting the changing health needs of urban populations, and whether there are rural and urban differences in
responses to potentially life-threatening illnesses such as
malaria.
On the malaria-endemic Kenyan Coast, urbanization is
centred around the city of Mombasa. Through our study
framework we identified a group of rural resident mothers
who had never lived in an urban area, and a group of urban
resident mothers who were mostly rural-urban migrants.
Both groups were Mijikenda, the dominant ethnic group in
Coastal Kenya. We compared their responses to uncomplicated childhood fevers reported in the two weeks preceding
the surveys. Despite dramatic demographic differences
between the rural and urban study populations (Molyneux
et al. 1999a), in access to formal and informal biomedical
services (Tables 1 and 2), and in the samples’ characteristics,
responses to uncomplicated fevers were similar (Tables 4
and 5).
One key finding is the importance of shop-bought medications as an initial response to the onset of fever in both
842
volume 4 no 12 pp 836–845 december 1999
areas. The frequent and early use of shop-bought antimalarials in household responses to fever/malaria has been
shown in numerous rural populations in sub-Saharan Africa
(Deming et al. 1989; Barnish et al. 1993; Foster et al. 1993;
Yeneneh et al. 1993; Vundule & Mharakurwa 1996).
Relatively few studies have compared rural and urban residents’ responses, and where they have, authors have noted
both striking differences and great similarity in use of shopbought drugs (Glik et al. 1989; Makubalo 1991; Mwenesi
1993; Kilian 1995). Nevertheless, all studies of urban populations have shown that the purchasing of shop-bought drugs
constitutes over 40% actions taken, in most cases as a firstline response (Glik et al. 1989; Makubalo 1991; Carme et al.
1992; Mwenesi 1993; Kilian 1995; Chiguzo 1999). As in rural
areas, the attraction of shop-bought drugs in urban areas is
‘rational’: there are numerous shops offering a cheap and
convenient service relative to even nearby government and
private facilities; and drugs are readily available and in many
cases cure or alleviate the patient’s symptoms (Molyneux
1999).
We did not document drugs purchased from shops, or
dosages administered to children. Researchers in the rural
study area have shown, however, that drugs are generally
administered
‘with little information on correct use, leading to random choices of drugs [being] used on a trial and error
basis, with rapid alterations in treatment if no immediate improvement is seen, and no concept of a need to
continue treatment in the absence of symptoms’ (Marsh
1998).
In low-income settlements of Mombasa, only 16–39% of
mothers know the correct chloroquine dose for a three-yearold (Government of Kenya 1992), and the inappropriate
administration of antimalarials has been noted in other
urban areas (Stein et al. 1988; Makubalo 1991; Lipowsky
et al. 1992; Massele et al. 1993; Chiguzo 1999). Greater access
to second-line antimalarials from informal vendors (as we
demonstrated in this paper), and the ease in some areas of
obtaining drugs from pharmacies without prescription
(Indalo 1997), may be a source of particular concern in urban
areas, with important implications for the regulation of
prescription-only antimalarial drugs. However, given that
chloroquine resistance is widespread in Kenya, the urban
informal sector may be responding to a high clinical demand
for second-line drugs. The very recent shift from chloroquine
to sulphadoxine-pyrimethamine as Kenya’s first-line antimalarial suggests that national policy-makers have recognized
and are responding to this demand. The way in which this
policy change will translate into practice, and the subsequent
implications of this shift, remain unclear.
We noted that the only significant difference in rural and
© 1999 Blackwell Science Ltd
Tropical Medicine and International Health
volume 4 no 12 pp 836–845 december 1999
C. S. Molyneux et al. Maternal responses to childhood fevers in Kenya
urban mothers’ reported responses was in the use of private/
government clinics: rural mothers were more likely to seek
government services, and urban mothers to consult private
practitioners. Informal discussions suggested greater use of
private practitioners by urban resident mothers was due to a
number of factors including the relative proximity and quality of care offered at these facilities; the cost in time, money
and (in some cases) work lost incurred in attending a government facility; and more immediate access than in rural areas
to cash from work, a husband, or a neighbour (Molyneux
1999; Molyneux et al. 1999b).
A number of study-specific factors may have biased our
findings, including recall period and the position of the rural
study area. Regarding recall period, mothers in both study
areas reported that healers are generally contacted if biomedical practitioners ‘fail’ to cure a fever (Molyneux et al.
1999b). Switch in form of therapy has been noted elsewhere
in sub-Saharan Africa as a response to severe, chronic and
recurring illnesses (Good 1987). In malaria, this may include
what biomedical practitioners would term re-infections,
relapses and recurrence of symptoms (Hausmann Muela
et al. 1998). Larger sample sizes or a longer recall period may
therefore have revealed a greater proportion of rural mothers
visiting healers. Conversely, or in addition, use of biocultural
therapy may have been under-reported in the rural area.
Nevertheless, in both areas mothers openly reported far
greater use of herbs and healers in response to childhood
convulsions (Molyneux et al. 1999b), suggesting such biases
were minimized.
Regarding the position of the rural study area, proximity
to Kilifi town ensures the population relatively good access to
the town’s health facilities. This includes high-quality care
provided in the Wellcome-KEMRI research wards at Kilifi
District Hospital. Study area mothers may therefore use biomedicine more often and more promptly than would be the
case in the more interior areas of Kilifi District. Although this
could be considered particularly problematic in a rural-urban
comparison, the differences between the rural and urban
study populations in demography and in access to biomedical
services suggest the comparison is valid. This is supported by
the documentation of similar treatment-seeking patterns in
rural and in urban areas elsewhere in sub-Saharan Africa.
One explanation for the similarity in rural and urban
mothers’ treatment-seeking patterns may be the exchange of
‘rural’ and ‘urban’ information and ideas about different illnesses and appropriate therapy. In coastal Kenya such
exchanges are the inevitable outcome of chain rural-urban
migration, the maintenance of rural-urban ties, high mobility
between rural and urban populations, and return migration.
These demographic features have been observed all over subSaharan Africa (Todd 1996). However, perhaps a more
important explanation is related to the high use of shop-
© 1999 Blackwell Science Ltd
bought antimalarials in both areas. In a context of growing
intra-urban inequity in access to services, an increasing proportion of urban residents living in absolute poverty, the
introduction of user-fees, and deteriorating quality of care in
government facilities, the attraction of shops in providing a
relatively low-cost, convenient and reliable (in terms of drug
availability) service cannot be overstated.
The enormous potential there is to reduce malaria morbidity and mortality through the improvement of home
treatment of childhood fevers has led to a number of
community-based initiatives, including the training of
mothers, community health workers, or shop-keepers in diagnosis, appropriate antimalarial use, and referral (Spencer et
al. 1987; Greenwood et al. 1988; Delacollette et al. 1996;
Pagnoni et al. 1997; Marsh et al. 1999). These initiatives have
largely taken place in rural areas, and have demonstrated
changes in treatment-seeking practice and, less frequently,
reductions in morbidity.
Although the data presented in this paper shows that such
community-based interventions are at least as relevant for
low-income urban populations, a number of efficiency and
equity issues arise from rural-urban differences in socioeconomic, cultural and demographic factors, and health
facility organization and accessibility. Should time and
resources be channelled into improving socio-economic
access to and drug availability in existing services as opposed
to developing new interventions? Would this have any impact
on the use of shop-bought medicines? Which interventions
would have the greatest impact on reducing mortality? To
what extent do high mobility, the breakdown of traditional
family structures, changing gender relations and the ethnic
and socio-cultural diversity prevalent in most cities (Silimperi
1994) affect community-based initiatives, and indeed the
development of an urban ‘community’ identity?
The popularity of self-medication all over the world suggests the use of over-the-counter drugs will remain a norm in
household treatment-seeking. However, there is still much to
learn about reasons for and the appropriateness of drug use
in sub-Saharan African towns. Clearly, research and interventions aimed at reducing urban morbidity and mortality
will have to consider multiple and competing health
providers, as well as the perspectives of a diverse range of
community members. This may become increasingly feasible
with the recent shift in approaches to urban health from ‘topdown’ disease-targeted projects to ‘bottom-up’, processorientated programmes aimed at improving the overall
well-being of urban citizens.
Acknowledgements
This research was funded by The Wellcome Trust, UK (Grant
number 033340) and the Kenyan Medical Research Institute.
843
Tropical Medicine and International Health
C. S. Molyneux et al.
Maternal responses to childhood fevers in Kenya
We wish to thank Dr Norbert Peshu and Professor Kevin
Marsh for their support and Dr Vicki Marsh her useful comments on the manuscript. We are grateful to the fieldworkers,
data entry clerks and administrative staff at the KEMRI Unit
for their hard work and dedication, and special thanks to the
rural and urban study area residents whose assistance and cooperation made the study possible. RWS is supported by The
Wellcome Trust as part of their senior fellowships programme in Basic Biomedical Sciences. This paper is published
with the permission of the director of KEMRI.
References
Bambrah CK (1996) Draft Proceedings of the Municipal
Consultation on Urban Poverty Reduction. 10th-12th April,
Polana Hotel, Mombasa District.
Barnish G, Maude GH, Bockarie MJ, Eggelte TA, Greenwood BM &
Ceesay S (1993) Malaria in a rural area of Sierra Leone. 1. Initial
results. Annals of Tropical Medicine and Parasitology 87,
125–136.
Brockerhoff M & Brennan P (1998) The poverty of cities in developing regions. Population and Development Review 24, 75–114.
Carme B, Koulengana P, Nzambe A & Du Bodan G (1992) Current
practices for the prevention and treatment of malaria in children
and in pregnant women in the Brazzaville region (Congo). Annals
of Tropical Medicine and Parasitology 86, 319–322.
Chant S & Radcliffe AS (1992) Gender and Migration in Developing
Countries. Belhaven Press, London.
Chiguzo AN (1999) Factors associated with the occurrence of
malaria epidemics. The case of Hurumasa peri-urban, Eldoret
Municipality, Kenya. Poster presented at the MIM African
Malaria Conference. 14–19 March, Durban, South Africa.
Cunningham-Burley S (1990) Mothers’ beliefs about and perceptions
of their children’s illness. In Readings in Medical Sociology (eds
S Cunningham-Burley & N McKeganey). Routledge, London.
Delacollette C, Van der Stuyft P & Molima K (1996) Using community health workers for malaria control: experience in Zaire.
Bulletin of the World Health Organization 74, 423–430.
Deming M, Gayibor A, Murphy K, Jones T & Karsa T (1989) Home
treatment of febrile children with antimalarial drugs in Togo.
Bulletin of the World Health Organization 67, 695–700.
Foster SO, Spiegel RA, Mokdad A et al. (1993) Immunization, oral
rehydration therapy and malaria chemotherapy among children
under five in Bomi and Grand Cape Mount Countries, Liberia
1984 and 1988. International Journal of Epidemiology 22 (Suppl.),
S50–S55.
Glik D, Ward W, Gordon A & Haba F (1989) Malaria treatment
practices among mothers in Guinea. Journal of Heath and Social
Behaviour 30, 421–435.
Government of Kenya (1990) Mombasa District Development Plan
1989–93. Ministry of Planning and National Development,
Government Printers, Nairobi.
Government of Kenya (1992) Baseline survey information for
Kongowea, Freretown/Kisauni, Likoni and Chaani sites.
Unpublished report by the Mombasa Municipal Council Bamako
Initiative Programme, Mombasa.
844
volume 4 no 12 pp 836–845 december 1999
Gould STW (1998) African mortality and the new ‘urban penalty’.
Health and Place 4, 171–181.
Greenwood BM, Bradley AK, Greenwood AM et al. (1988)
Comparison of two strategies for control of malaria within a primary health care programme in the Gambia. Lancet i, 1121–1127.
Good CM (1987) Ethnomedical systems in Africa. Patterns of
Traditional Medicine in Rural and Urban Kenya. Guilford Press,
New York.
Harpham T (1997) Urbanisation and health in transition. Lancet
349, 11–13.
Hausmann Muela S, Muela Ribera J & Tanner M (1998) Fake
malaria and hidden parasites – the ambiguity of malaria. Anthropology and Medicine 5, 43–61.
Indalo AA (1997) Antibiotic sale behaviour in Nairobi: a contributing factor to the antimicrobial drug resistance. East African
Medical Journal 74, 171–173.
Kilian AHD (1995) Malaria control in Kabarole and Bundibugyo
Districts, Western Uganda. Report on a comprehensive malaria
situation analysis and design of a district control programme.
14–15 March 1995, Fort Portal, Uganda.
Kindy H (1972) Life and Politics in Mombasa. East African Publishing House, Nairobi.
Kleinman A (1980) Patients and Healers in the Context of Culture.
University of California Press, Berkeley.
Kloos H, Etea A & Degefa A (1987) Illness and health behaviour in
Addis Ababa and rural Central Ethiopia. Social Science and
Medicine 25, 1003–1019.
Lipowsky R, Kroeger A & Vazquez ML (1992) Sociomedical aspects
of malaria control in Columbia. Social Science and Medicine 34,
625–637.
Makubalo EL (1991) Malaria and chloroquine use in Northern
Zambia. PhD Thesis, LSHTM, University of London.
Malombe JM & Kanyonyo M (1997) Urban community appraisal
and planning in Nairobi: a review of experience. Unpublished
report prepared for the British Development Division in Eastern
Africa, Nairobi.
Marsh V (1998) Implementation and evaluation of a shop keeper
training programme to optimise the home use of shop-bought
antimalarial drugs for uncomplicated childhood fevers in coastal
Kenya. DFID Grant Proposal R7179 (RD493).
Marsh VM, Mutemi WM, Muturi J et al. (1999) Changing home
treatment of childhood fevers by training shop keepers in rural
Kenya. Tropical Medicine and International Health 4, 383–389.
Massele YA, Sayi J, Nsimba DES, Ofori-Adjeyi D & Laing OR (1993)
Knowledge and management of malaria in Dar es Salaam,
Tanzania. East Africa Medical Journal 70, 639–641.
Mazrui AA (1996) Mombasa: three stages towards globalisation. In
Re-Presenting the City: Ethnicity, Capital and Culture in the
Twenty-First Century Metropolis (ed. AD King). New York
University Press, New York.
Mburu FM, Spencer HC & Kasesje DCO (1987) Changes in sources
of treatment occurring after inception of a community-based
malaria control programme in Saradidi, Kenya. Annals of Tropical
Medicine and Parasitology 81 (Suppl. 1), S105–S110.
Molyneux CS (1999) Migration, mobility and health-seeking behaviour of mothers living in rural and peri-urban areas on the
Kenyan Coast. PhD Thesis, South Bank University, London.
© 1999 Blackwell Science Ltd
Tropical Medicine and International Health
volume 4 no 12 pp 836–845 december 1999
C. S. Molyneux et al. Maternal responses to childhood fevers in Kenya
Molyneux CS, Mung’ala-Odera V, Harpham T & Snow RW (1999a)
Migration and circulation among low-income mothers on the
Kenyan Coast: implications for health policy. Submitted to Health
Policy and Planning.
Molyneux CS, Murira G, Masha J & Snow RW (1999b) Intra-household relations and treatment decision making for child illness: a
Kenyan case study. Submitted to Social Science and Medicine.
Mwenesi H (1993) Mothers’ Definition and treatment of childhood
malaria on the Kenyan Coast. PhD Thesis, LSHTM, University of
London.
Pagnoni F, Convelbo N, Tiendrebeogo J, Simon C & Esposit F (1997)
A community based programme to provide prompt and adequate
treatment of presumptive malaria. Transactions of the Royal
Society of Tropical Medicine and Hygiene 91, 512–517.
Parkin JD (1979) Along the line of road expanding rural centres in
Kenya’s coast province. Africa 49, 273–282.
Parkin D (1991) Sacred void. Spatial Images of Work and Ritual
Among the Giriama of Kenya. Cambridge University Press,
Cambridge.
Parry C (1995) Quantitative measurement of mental health problems
in urban areas: opportunities and constraints. In Urbanisation and
Mental Health in Developing Countries (eds T Harpham & I
Blue), Aldershot, Avebury.
Silimperi D (1994) Delivering child survival services to high risk
populations (draft), BASICS Strategy Paper. pp 1–57.
Snow RW, Mung’ala VO, Foster D & Marsh K (1994) The role of the
© 1999 Blackwell Science Ltd
district hospital in child survival at the Kenyan Coast. African
Journal of Health Sciences 1, 71–75.
Spencer HC, Kaseje DCO, Mosley WH, Sempebwa EKN, Huong AY
& Roberts JM (1987) Impact on mortality and fertility of a community-based malaria control programmes in Saradidi. Kenya. Annals
of Tropical Medicine and Parasitology 81 (Suppl.), S36–S45.
Stein CM, Gora NP & Macheka BM (1988) Self medication with
chloroquine for malaria prophylaxis in urban and rural
Zimbabwians. Tropical and Geographical Medicine 40, 264–268.
Stock R (1985) Health care for some: a Nigerian study of who gets
what, where and why? International Journal of Health Services 15,
469–483.
Tanner M & Harpham T (1995) Action and research in urban health
development – progress and prospects. In Urban Health in
Developing Countries: Progress and Prospects (eds T Harpham &
M Tanner). Earthscan Publications Ltd, London, pp. 216–220.
Todd A (1996) Health inequalities in urban areas: a guide to the literature. Environment and Urbanisation 8, 141–152.
Vundule C & Mharakurwa S (1996) Knowledge, practices, and perceptions about malaria in rural communities of Zimbabwe:
Relevance to malaria control. Bulletin of the World Health
Organization 74, 55–60.
Yeneneh H, Gyorkos TW, Joseph L, Pickering J & Tedla S (1993)
Antimalarial drug utilization in women in Ethiopia, results from a
KAP study. Bulletin of the World Health Organization 71,
763–772.
845