ABSTRACT

Background
The number of orphans in Rakai district, Uganda is estimated to be 34,902 (OCBO, 2000) which translates into 28% of children under the age of 18 years. Young people who have been orphaned and as a result became heads of households must look after themselves and their siblings. These children are likely to be faced with several health problems and have to take crucial life decisions without parental/adult guidance.
Objectives
This study was conducted in order to understand how child-headed households, Rakai district in Uganda recognize malaria, their health-seeking behavior when malaria is suspected and reasons for the type of behavior compared to the adult-headed households.
Methods
A comparative cross-sectional study was conducted in 300 households in Rakai district, Uganda, in which 8/23 (35%) of subcounties and 150 child-headed households were included. The closest neighborhood adult-headed household to each childheaded household was selected for comparison. Individual interview was carried out with the respondents using semi-structure questionnaire. Key informant interview and focus group discussions were also conducted.
Results
The main findings were that the respondents in child-headed households had less knowledge on signs and symptoms of simple and severe malaria compared to adult heads of households. Respondents in child-headed households were less likely to seek health care from health facilities (OR=0.59, CI=0.36-0.97, p-value=0.028). There was no significant difference in the time lag before taking first action in the two types of households (OR=0.72, CI=0.42-1.22, p-value=0.194). The respondents in childheaded households were six times (OR=5.70,CI=2.75-11.91, p-value<0.001) more likely to use local herb for treatment of malaria than the adult heads households. Major reasons stated by the respondents for choosing where health care is sought included distance to source of health care, cheap or free treatment, availability of drugs, and quick services to patients.
Conclusion
The respondents in child-headed households had less knowledge on signs and symptoms of simple and severe malaria and receive too little or late health care from health professionals compared to the adult heads of households probably due to lack of knowledge and money. Information Communication and Education programs should be designed and target the child-headed households and supply home packs.

Keywords: Health seeking behavior, child-headed households, adult-headed households, AIDS, Orphans, Rakai, Uganda.

Malaria has been historically a very serious health problem in Uganda. Currently it accounts for 24-40% of all out patient visits at health facilities and 20% of hospital admissions⁵. In Rakai district malaria is currently the leading cause of morbidity contributing to 36.7% of Out Patients Department (OPD) attendance⁶.

Using malaria as a tracer condition, we conducted a comparative cross-sectional study to describe the health-seeking behavior among childand adult-headed households. We also learned reasons for choosing where health care is sought from heads of households in the two categories.

The objective of the study was to compare the knowledge of the child- and adult-headed households on signs and symptoms of simple and severe malaria in Rakai-Uganda, their health-seeking behavior when malaria is suspected and the reasons for choosing where health care is sought for malaria.

METHODS

A comparative, cross-sectional study design was used to collect information on health-seeking behavior from sample of 300 households (150child- and 150 adult-headed households). Pocock’s formula for comparing two proportions was used to estimate the sample size⁷. Two of the four counties in the district (Kooki and Kakuuto) were purposively selected for inclusion on the basis of their high numbers of child-headed households. In the selected counties, 50% of the sub-counties were also selected purposively. Lists of child headed-households in the selected sub-counties were obtained from the Orphaned Community Based Organization (OCBO) and Lutheran World Federation (LWF) offices in the district where one combine list was created. Every third household on the list was systematically selected until 150 child-headed households were obtained. The closest neighborhood adult-headed household to each childheaded household was selected for comparison.

The respondents in child-headed households were the household heads or any child who was found present and could answer the questions. In adult-headed households, the head or spouse was interviewed. Twenty-six key informant interviews and six focus group discussions were purposively selected and conducted.

Informed consent was obtained from all respondents prior to inclusion in the study. A semistructured questionnaire was then administered to obtain each respondent’s socio demographic and behavioral data. The socio-demographic data included age, sex, tribe, religion, education level and occupation. The behavioral data included knowledge of signs and symptoms of simple and severe malaria, first action taken when malaria is suspected, time lag before first action is taken, first treatment given and reasons for choosing where health care is sought for malaria.

Data was cleaned, coded and entered into EPIinfo version 6. It was checked for consistency, analyzed and associations were measured using Odds Ratio test at < 0.05 significance level.

RESULTS

Socio-demographic characteristics of respondents in child- and adult-headed households

These results were generated from 150 respondents in each type of household. Of the respondents in childheaded households 63/150 (42%) were within the age range of 11-15 years. Out of 150 respondents, 99 (66%) were male. Majority of the respondents 128/150 (85%) were Baganda which is the dominant tribe in Rakai district. By religion, 145/150 (97%) were both Catholics and protestants. Of 150, 114 (76%) who formed the majority had primary education and only 7/150 (5%) had no formal education. By occupation, 77/150 (51%) were pupils and students, and 73/150 (49%) were peasants. Of the respondents in adult-headed households, 73/150 (49%) were within the age of 19-35 years. Out of 150 respondents, 62 (41%) were male, The dominant tribe 112/150 (75%) was Baganda. Majority 139/150 (93%) of the respondents were both Catholics and protestants. Of 150, 82 (55%) had primary education and 46/150 (31%) had no formal education. By occupation 66/150 (44%) were peasants and 65/150 (43%) were housewives.

Respondents knowledge of the signs and symptoms of malaria

The common sign and symptom of simple malaria, which was mentioned by the respondents in both child- and adultheaded households, was fever (Table 1). There was a significant difference (OR=0.43, CI=0.25-0.75, p-Value=0.001) between the respondents in child- and adultheaded households on the knowledge of fever as the major sign and symptom of simple malaria. There was also a significant difference in mentioning headache, joint pains, muscle ache, sweating, vomiting and loss of appetite as signs and symptoms of simple malaria. There was no significant difference in knowledge concerning nausea and shivering between the two types of households.

Table 1. Comparison of respondents’ knowledge of the signs and symptoms of simple malaria

Signs and		Percentage of	Percentage of		Odds Ratio	P-Value
symptoms		respondents in	respondents in (95% CI)
of	simple	child- headed	adult-headed
malaria		Households	households
		n=147	n=148
Fever		61.9	79.1	0.43 (0.25 – 0.75)		0.001*
Shivering		57.8	60.8	0.88 (0.54 – 1.45)		0.601
Headache		51.0	65.5	0.55 (0.33 – 0.90)		0.012*
Vomiting		49.0	63.5	0.55 (0.34 – 0.91)		0.012*
Joint pains		37.4	54.1	0.51 (0.31 – 0.83)		0.004*
Other**		28.6	11.5	3.08 (1.59 – 6.04)		0.000*
Loss of appetite		19.7	37.8	0.40 (0.23 – 0.71)		0.001*
Muscle ache		6.1	14.2	0.39 (0.16 – 0.95)		0.022*
Nausea		5.4	11.5	0.44 (0.17 – 1.14)		0.062
Sweating		4.1	14.9	0.24 (0.08 – 0.67)		0.001*

* Statistically significant association (P=<0.05) between the type of household and knowledge of the signs and symptoms of simple malaria.
**Other signs and symptoms were cough, skin rash, weakness and diarrhea

On signs and symptoms of severe malaria, the respondents in child-headed households were less likely to mention convulsions, fast breathing, jaundice and pallor as signs and symptoms of severe malaria than the adult heads of households (Table 2). The respondents in child-headed households were 2 times more likely not to know any sign and symptom of severe malaria than the adult heads of households.

Table 2. Comparison of respondents’ knowledge of the signs and symptoms of severe malaria

Signs and	Percentage of	Percentage of	O dds Ratio	P-Value
symptom s of	respondents in	respondents in	(95% C I)
severe m alaria	child- headed	adult-headed
	households	households
	n=147	n=148

Vomiting

52.4

58.1

0.79 (0.49 – 1.30)

0.323

Other**

47.6

13.5

5.83 (3.15- 10.81)

0.000*

Don’t know any sign

15.6

1.3

2.00 (1.68 – 2.38)

0.000*

Jaundice

12.2

35.1

0.26 (0.13 – 0.49)

0.000*

Pallor

6.8

34.5

0.14 (0.06 – 0.30)

0.000*

Convulsions

2.7

43.9

0.04 (0.01 – 0.11)

0.000*

Fast breathing

1.4

7.4

0.17 (0.03 – 0.85)

0.011*

* Statistically significant association (P=<0.05) between the two types of households and knowledge of the signs and symptoms of severe malaria.
** Other signs and symptoms mentioned were diarrhea, thirst and very weak

First action taken when malaria is suspected

The first actions taken included (a) treating the sick person at home, (b) taking the sick person to health unit, (c) taking the sick person to a herbalist, and (d) taking no action (Figure 1). The respondents in child-headed households were less likely than the adult heads of households (OR=0.59, CI=0.36-0.97, P value=0.028) to seek care at health units. During focus group discussion and key informant interview, it was commonly reported that the majority of people take home treatment and health facilities are used incase of failure.

(Figure 1). The respondents in child-headed households treatment and health facilities are used incase of failure.

Time lag before first action is taken

The respondents in both types of households had no significant difference in time lag before first action taken after malaria is suspected, but the type of action taken varied. The respondents in child-headed households were mainly using home treatment, which may be easily available and affordable. Both respondents in child- and adultheaded households mentioned using Western medicines and local herbs in treatment of malaria. However, the respondents in child-headed households were six times more likely to mention the use of local herbs than the adult heads of households. This is likely to reflect the much lower level of financial capacity in child-headed households to purchase drugs. Locally available herbs can probably be gathered at no cost. The use of local herbs may not be effective for treatment of malaria and may result in serious illness, and ultimately death, it also causes delay in obtaining effective treatment. The respondents in child-headed households may also not know how these local herbs should be used, their doses, and the actual efficacy of most of these herbs has not been determined. A study conducted in Mulago revealed that mothers were treating children for malaria with herbs¹¹. This demonstrates further that the use of herbs is still common in many areas in Uganda. It is important to sensitize the community about use of local medicines and it’s dangers.

The major reasons mentioned for choosing where health care is sought was distance to source of health care, free and cheap treatment. The respondents in child-headed households were less likely to mention distance to source of health care than adult heads of households. This prob-ably reflects the more common practice in childheaded households of relaying on home treatment involving self-medication and could not think of distance as a problem. The community tends to use nearby places to eliminate transport costs. Implementation of the Ministry of Health, Uganda, policy on household (home) management of malaria and supply of home pack program should target the child-headed households.

Study limitations

Selection of the counties and sub-counties basing on their high numbers of child-headed households in the district could have created selection bias. The counties and sub-counties with low numbers of the child-headed households may have different characteristics.

This study catered for only children who are registered by Non-governmental organizations (NGOs) in Rakai district. Other child heads of household may exhibit a different health-seeking behavior pattern and may have different reasons to account for that behavior pattern.

ACKNOWLEDGEMENTS

I acknowledge the Rockefeller Foundation, for financial support throughout the course of training and initial funding of this study. Sincere thanks go to the Uganda Malaria Surveillance Project (CDC Award No: S1932-021/21 entitled; “Malaria Surveillance and Control in Uganda”), which is a collaboration between the Makerere University Institute of Public health, the School of Public Health at University of California-Berkeley and the Ministry of Health Uganda for the financial support towards completion of this study. I must thank the Rakai District Health Team, various Non-Governmental Organizations in Rakai districts, field guides and respondents for their maximum co-operation throughout data collection.