HEALTH CARE-SEEKING BEHAVIOUR FOR CHILD ILLNESSES AMONG RURAL MOTHERS IN SOUTH AFRICA: A PILOT STUDY

The aim of this study was to examine the health care-seeking behaviour of mothers when their children under five years suffer from common childhood illnesses such as diarrhoea, fever, cough and worms. The study was conducted in a rural community in the Limpopo Province of South Africa. The sample consisted of 100 rural mothers attending a clinic. The results indicated that the complaint most commonly reported was fever (95%), followed by diarrhoea (91%); worms were reported by only 25% of the mothers. The most common health care-seeking behaviour was a visit to the clinic: 79% of the mothers took their children to the clinic for coughing, 68% for fever, 50% for diarrhoea and 11% for worms. The second most common form of health care-seeking behaviour was self-care: for diarrhoea (20%) and for fever (13%); a private doctor was consulted for coughing (11%) and drug vendors were used for the treatment of worms (8%). Most mothers (76%) used home remedies for the treatment of diarrhoea and modern drugs for the treatment of fever (91%), for coughing (98%), and for worms (22%). Among mothers in the age group 31 to 49 years, 52.9% had experienced the death of a child, followed by 13.3% in the age group 15 to 19 years, and 9.8% in the age group 20 to 30 years.


INTRODUCTION
Health care-seeking behaviour is that action taken by an individual in response to a stimulus (such as the perception of a symptom) that he or she decides is indicative of a condition needing evaluation by a health professional. This behaviour is influenced by personal, physical, and psychological characteristics and by sociocultural and environmental factors (Gilliland, Phillips, Raczynski, Smith, Cornell & Bittner, 1999:95).
Health care-seeking involves a series of stages or phases, beginning with the patient becoming aware of a need and ending with medical assessment and treatment, if warranted. Delay can occur at any stage, including delay in patient care-seeking and delay in treatment once contact is made with the health care system. While structural-environmental issues may play a role in treatment delay (that is access to care, poor emergency medical service and doctor-instituted delays in diagnosis or treatment), the largest component of delay for acute problems occurs before the patient contacts the health care system (Gilliland et al. 1999:96). Several studies have noted that, besides inadequate availability of health care services in many areas, especially the less developing countries, certain disease-specific and non-disease-specific cultural beliefs may influence people's health care-seeking behaviour (Feyisetan, Asa & Ebigbola, 1997:221).
In South Africa, government policies have placed considerable emphasis on improving access to health care (especially primary health care) and reducing racial disparities in access to health care since 1994 (Department of Health, 1999:5). In the Limpopo Province Primary Health Care (PHC) seems underutilised, with an average PHC utilisation of an estimated 2.0 consultations/year (national goal: 3.5) (Limpopo Province Department of Health, 2001:15). Smith, Solanki and Kimmie (1999:12-19) found that factors influencing access to health care of South Africans were cost of care (travel cost, consultation cost), time cost (traveling time, waiting time), availability of health care services (new clinics in areas, days of clinic operation, hours of clinic operations), and health care access when ill.
According to a 1993 WHO report, the major causes of mortality in children under five years of age in developing countries, either as a single cause or associated with other causes, are acute respiratory infections (34%), malnutrition (29%), diarrhoea (25%), measles (9%), and malaria (8%). The mean underfive mortality rate of sub-Saharan Africa was 256/1000 live births in 1960. It dropped to 204 in 1980 andto 177 in 1994 (Martin, 1998:103). In 1998 in the Limpopo Province, South Africa, the under-five mortality rate was 51.3/1000 (Department of Health, 1998:7) indicating a steady decline in childhood mortality. However, there may not be a direct relationship between the decline in childhood mortality and PHC services since the utilisation of PHC services is low in the Limpopo Province. This raises the question about the influence of health care-seeking behaviour and drug use in child survival. In recent years, epidemiologists and social scientists have devoted increasing attention to studying health care-seeking behaviour associated with the two leading causes of child mortality, namely diarrhoeal illness and acute respiratory infection (ARI). Yet, the knowledge about how and when families in developing countries seek treatment for these prevalent illnesses remains seriously incomplete (Goldman & Heuveline, 2000:145). No study was found investigating health care-seeking for common child illnesses among mothers in South Africa.
To investigate how mothers take care of their ailing children (and thus may contribute to the under-five mortality rate), a pilot-study was undertaken to examine mothers' health care-seeking behaviour when their children under five years of age present with common childhood illnesses. The illnesses covered in the study included diarrhoea, fever, cough and worms in a rural community in the Limpopo Province, South Africa. The goal of this pilot-study was to identify possible problems in health care-seeking behaviours for common childhood illnesses and whether or not the health interview survey would be a good method for analysing health care-seeking behaviour in South Africa.

Design and setting
A cross-sectional survey was carried out among mothers of children aged five years and younger in the village of Eisleben (Botlokwa area) in the Limpopo Province, South Africa. The biomedical health care services in the village, Eisleben, were one clinic and one private doctor. The services at the clinic were provided free for every user. The nearest hospital was about 30 km from the village under study.

Sample and procedure
This study was conducted among a sample of 100 rural mothers. A non-probability convenience sampling method was used to select mothers consecutively attending the clinic (taking their child for check up or because their child was sick) until a sample of 100 was reached. Inclusion criteria were: 1) mothers had to have at least one under-five year old child (of any age below five) who had suffered from either diarrhoea, fever, coughing or worms during the previous year, 2) they have to have lived in Eisleben for at least five years, and 3) they should not currently be suffering from a psychiatric illness.
Interviews were conducted in private after permission and informed informal consent had been obtained.
Anonymity and confidentiality were assured. A trained research assistant conducted interviews in Northern Sotho and answers were entered into the questionnaire.

Measure
For the interview, the study used semi-structured and open-ended questions that were based on a study done among Guatemalan mothers (Van der Stuyft, Sorensen, Delgado & Bocaletti, 1996:163).
The interview schedule consisted of six major parts.
The first part sought personal data and household characteristics: age, number of children alive, ethnicity, number of years of formal education, cohabitation and living with in-laws, mother's occupation, father's occupation, and mother's income.
The second part of the interview schedule dealt with the mother's health care-seeking behaviours for their children (under five years) for the last illness episode (of diarrhoea, fever, coughing, or worms). These behaviours were 1) self-care, 2) drug vendor, 3) health services (divided into private doctor, clinic and hospital), 4) traditional healer, and 5) faith healer.
The third part was about the type of treatment used for their child's last illness episode (of diarrhoea, fever, coughing or worms). The treatment mode was categorised into four items: 1) modern drugs, 2) herbal treatment, 3) home remedies, and 4) faith healer.
In the fourth part, a question was asked about who chose the remedy for the child's illness. The response options included the self, husband/partner, parent and neighbours/friends.
In the fifth part, mothers were asked to indicate the name of the remedy and explain the effect of it, rated from 1= not very effective to 5 = very effective.  (Boyatzis, 1998:20-30). Sotho. The educational level of mothers in this study was as follows: primary school 15%, secondary school 66%, and tertiary education 19%. Forty-two mothers were living with husbands or partners and 58 were single mothers. Sixty-eight mothers were unemployed, 11 were students and only 21 were employed [profession-als (9), domestic worker (6), self-employed (6)]. How ever, when asked about the income, 28 % responded that they had income. The employment status of the child's father was: 51% employed, 39% unemployed, and 10 % students. The occupation of the employed fathers was: factory worker (17), business (6), teacher (6), taxi driver (6), security guard (5), police/soldier (4), and farmer (2). The mothers were asked to rate their economic family background. More than half (53%) rated themselves as "quite poor", 42% "not very well off", 4% "quite well off" and 1% "wealthy".
The number of live children that the mothers had ranged from one to seven: one child 47%, two children 24%, three children 15% and more than three children 14%.

Personal data and household
The distance from the households of the mothers in this study to the clinic was between 1 and 5 km: about 1 km 48%, 2 km 47%, and 3 to 5 km 5%. Among the mothers in this study only 12% were members of a medical aid scheme; 88% were not.

Health care-seeking behaviours for previous illness episodes in children
All of the mothers (100%) reported that their children had experienced coughing, 89% had experienced their child having diarrhoea, 95% reported fever, and only 25% had experienced their children having worms. The majority of the mothers took their child to a clinic: for Diarrhoea diarrhoea (50%), fever (68%), coughing (79%), and worms (11%). The second priority of health careseeking behaviour was self-care: for diarrhoea (20%) and for fever (13%). A private doctor was consulted for coughing (11%) and the drug vendor was used for the treatment of worms (8%). These results are summarised in Table 1.
The most common treatment used for fever in children was modern drugs (91%), followed by home remedies (2%), faith healer (2%) and herbs (1%). Most of the mothers (84%) chose the type of treatment for fever by themselves, followed by advice from a parent (8%), neighbours/friends (6%), and husband/partner (2%). Comments indicating lack of faith in the effectiveness of the treatment are: "It makes the fever go down but few hours later the fever comes back". "It only works after some days".
Medicines like Liquid paraffin and Actifed were also accepted by mothers as very effective in rapidly reducing fever. The data also found some misconceptions: Buscopan was believed to be very effective in reducing fever from umbilical cord pain. ZCC tea was also perceived as very effective after an hour. One mother said that taking a shower was effective to stop fever and make the body temperature go down.

History of dead child
Twenty-five (25%) of the mothers had had a dead child (including abortion and stillbirth) under the age of 5 years. The duration of the illness before death (excluding five abortions and five still births) was one week (n=12), and more than one month (n=3). The causes of death were abortion (5), still birth (5), illness (13), and in two (2) cases the mothers did not know because there were no signs or symptoms. Illnesses mentioned seem to be related to four categories (1) the digestive system such as vomiting, diarrhoea, no appetite, and sore in the stomach, (2) congenital diseases such as brain and heart problems, (3) Lekoni (red spots on the back of the neck) and eczema and (4) pulsating fontanelle (Hlogwana).
Causes of the children's death were described as follows:

Still birth
Five mothers reported stillbirth.

Health care-seeking behaviour for the dead child
Excluding the five cases of abortion and five stillbirths, nine of the mothers took their child to hospital. Of those nine, six went to the clinic and/or private doctor first before going to the hospital and three went straight to hospital. Three took their child to the clinic and/or private doctor. One went with her child to the Zion Christian church and two did not seek treatment as they found their child dead.
than mothers living without a husband or partner.
Education, occupation and income were not found to be associated with having a dead child.

DISCUSSION
The study found that most mothers used modern treatment (clinic, private doctor, and hospital) for their children's illnesses (cough, fever, diarrhoea, and  had a dead child, followed by 13.3% in the age group 15 to 19 years, and 9.8% in the age group 20 to 30 years.
Older mothers had significantly more dead children than younger women. Mothers with more than three children had also significantly more dead children than mothers with three and fewer children. Mothers living with husband or partner also had more dead children Table 2 shows the socio-demographic characteristics of mothers who had a dead child and those who did not have a dead child.

The socio-demographic characteristics of the mothers having a dead child
worms) and only a few of them used a drug vendor or faith healer. The clinic was found in this study to be the major health-care provider. Two major reasons for this were that clinic services are free of charge (only 12% were members of a medical aid scheme) and are accessible, being only between one and five kilometres from the homes of the mothers. The result that only a few consulted alternative practitioners is similar to the study of health-seeking behaviour and self-treatment for common childhood symptoms in rural Guatemala (Delgado, Sorensen & Van der Stuyft, 1994:161), which found that traditional healers were hardly consulted (range: 0%-3%). In a study in rural Zimbabwe on the perceptions of childhood diarrhoea and its treatment, it was also found that utilisation rates of the formal health services were unexpectedly high with a low demand for indigenous herbalists (Zoysa, Carson, Feachem, Kirkwood, Lindsay-Smith & Loewenson, 1984:727).
In this study most mothers (more than 80% for each illness studied) selected the type of treatment themselves. This result is contrary to a study by Delgado et al. (1994:161) among Guatemalan mothers, who generally sought help and treatment advice from an older woman in the family; they did so more often for diarrhoea (82%), followed by fever (64%), cough (43%) and worms (28%).
Regarding the cause of under-five child death, 13 mothers mentioned an illness, which could be grouped into four categories (1) the digestive system such as vomiting, diarrhoea, no appetite, and sore in the stomach, (2) congenital diseases such as brain and heart problems, (3) Lekoni (red spots on the back of the neck) and eczema and (4) pulsating fontanelle (Hlogwana). Only one mentioned that she had consulted a faith healer. The two "traditional" illnesses Lekoni and Hlogwana mentioned as cause of death are commonly treated by traditional and faith healers in the Limpopo Province (Peltzer, 1998:193, Peltzer, 1999. Older mothers were found to have more dead children than younger women. Mothers with more than three children had also significantly more dead children than mothers with three and fewer children. Mothers who had children at an older age tend to have more risk pregnancies than younger mothers. Possible reasons for the above are that older mothers and mothers with four and more children are more at risk to have an abnormal or dead baby, which should be investigated further. These problems could be easily addressed by giving advice to pregnant women at antenatal clinics. Africa. The interview schedule should further include detailed information about the nature and timing of illness and treatment behaviour (Goldman & Heuveline, 2000:145).

ACKNOWLEDGEMENT
The University of the North funded the study. We would also like to thank the anonymous reviewers for their