Prevalence and factors associated with domestic accidents in children in the district of Orodara, Burkina Faso

Abstract

Background

In West Africa, domestic accidents are generally underestimated, and only few scientific population-based studies have assessed their prevalence. When available, existing data often focus only a single type of accident. This study aimed to determine the prevalence of domestic accidents and their determinants among children aged 0–14 years in the Orodara health district in Burkina Faso.

Methods

We surveyed a fixed number of households across all 37 villages of the Orodara health district, Kénédougou Province, Burkina Faso. The dependent variable was the occurrence of a domestic accident in a child aged 0 to 14 years, defined as an accident in the past six months that required treatment (coded 1 for “yes”, and 0 for “no” otherwise). Independent variables included sociodemographic characteristics of the household and child. Data were analyzed using Stata software version 17.0, applying logistic regression with a 5% significance threshold.

Results

Data from 1578 children from 802 mothers in 805 households were included in the analysis. We recorded 99 domestic accidents, corresponding to a prevalence of 6.27%. The main determinants of domestic accidents were the place of residence (adjusted odds ratio (ORa) = 2.9, 95% Confidence Interval (CI) = 1.44–5.84, p = 0.003), and the child’s age, with children over 12 months being more likely to experience domestic accidents than those under 12 months.

Conclusion

Compared to the literature, domestic accidents are relatively uncommon in the Orodara health district of Burkina Faso, a result likely influenced by our study design. The main sociodemographic factors associated with higher injury risk were younger age and rural residence. The development of public awareness-raising strategies and campaigns may reduce the occurrence of domestic accidents in children.

Introduction

In Africa, the under-five mortality rate is the highest worldwide, with 72 deaths per 1’000 live births, nearly ten times higher than the risk faced by children in Europe [1]. According to a study by Adeloye et al., 2015, the worldwide injury mortality rates were 29.6 and 32.7 per 100’000 children aged 0 to 11 months and 1 to 4 years, respectively. Road accidents and drowning were the leading causes of death [2]. Among children aged over 5 years, the situation remains problematic: According to a study published in 2022, worldwide, accidents are responsible for 32.9% of deaths among children aged 5 to 19. Although a large proportion of these accidents are due to road accidents (7.8% of all deaths), other types of accidents are no less significant (drowning 5.2% and self-inflicted injuries 4.2%) [3].

In light of these alarming findings, the United Nations Sustainable Development Goal 3 (SDG3) focuses on healthy lives and well-being at all ages and sets targets for reducing premature mortality from maternal, communicable, and noncommunicable causes and injuries for the period 2016-30 [4]. Similarly, the Global Strategy for Women’s, Children’s, and Adolescents’ Health (2016-30) sets survival as its main objective and adolescent mortality as one of its 16 key-indicators [5]. In light of these figures, achieving SDG3 will require the implementation of effective mechanisms to prevent all types of injuries, including domestic accidents among children and adolescents.

In the literature, most studies on domestic accidents rely on hospital data [6–10], which are not suitable for estimating prevalence and incidence, as many children with injuries never reach health centers. Indeed, studies carried out in Rwanda [11] and Burkina Faso [12] have shown that the first point of first referral for care is not always a health center. This suggests that hospital-based studies not only lack an appropriate denominator for calculating prevalence, but are also likely to underestimate the true incidence or prevalence of domestic accidents. Population studies in several African countries, such as Egypt, Ghana, Sudan, and Uganda, have shown that domestic accidents are underreported [13–18], with prevalences ranging from 10% to 25%, although not all types of accidents were consistently taken into account. Further, authors did not use the same recall period: while some studies used an observation period of six months [14, 16, 18], others used a recall period of one year [15, 17].

In Burkina Faso, domestic accidents are still underestimated as well, and no scientific studies have determined their general prevalence. Existing data only cover one particular type of accident, as for example Koueta et al. [19] and Cessouma et al. [20] who focused on accidental acute poisoning in children in Burkina Faso. Apart from these studies, Burkina Faso has annual data from the Statistical Yearbook of the Ministry of Health, which reported 2020 2’732 cases of burns in primary health facilities, including 1’717 cases (73.54%) in children aged 0–14 years [21]. Of note, no specific interventions currently exist in this country to prevent domestic accidents in children.

Given the scarcity of available data, our study aimed to generate evidence on domestic accidents among children aged 0–14 years in one region of the country, the Orodara health district, by assessing the prevalence and identifying associated factors.

Methods

Study setting

The study was conducted in the Orodara health district located in Kénédougou province. It is one of the seven health districts in the Hauts-Bassins region, situated in western Burkina Faso, and covers an area of 8’307 km². In 2022, the Orodara health district had a population of 267’645, of whom 116’022 (43.3%) were children under 14 years of age. The district has one medical center with a surgical unit, and 39 primary health centers [22].

Operational definition

To define “domestic accident”, we applied the following operational definition, inspired by the one used in the study by Méndez et al. in 2020 in Tanzania [23]: “an event occurring in or near the home during the six months preceding the study and resulting in intoxication, an animal or insect bite, a fall, a burn, a cut or any other injury, whether intentional or not, requiring some form of treatment.”

Study population

We included mothers and their children aged 0 to 14 years at the time of the domestic accident, and living in the Orodara health district.

Sample size

The key indicator used to calculate the sample size was the proportion of children who had experienced a domestic accident. The hypothetical prevalence was set at 25%, based on data from Egypt (20.6% [15]) and Ghana (16%, respectively [24]). We set the margin of error at 3% and a cluster effect at 2, since several children were included per household. This resulted in a target of 1’600 children, corresponding to approximately 800 households. We stratified the cohort by place of residence, surveying 200 households (selected through a reasoned choice method) in urban areas (specifically, the sectors of the town of Orodara), and 600 households in rural areas (specifically, the villages).

Inclusion criteria

The inclusion criteria were: (i) lived in the village/town for the previous six months, (ii) had at least one child aged 0 to 14 years, and (iii) provided informed consent. If informed consent was not given, individuals were excluded from the study.

Sampling

We surveyed households in 37 villages with a primary health center in the Orodara health district. Within each village, we surveyed a fixed number of households. We chose a starting point in each village using the random “walk” method, as we had no map of the village [25, 26]. When the household is screened using the random walk method, the interviewer makes two checks: (1) “Is there a woman in the household?”, and (2) “Does the woman have a child under the age of 14 years old?”. If the response to these two questions was “yes”, the interviewer proceeded with the consent process by providing the household with all the necessary information about the study. If the woman agreed, she signed the consent form before administering the questionnaires. The interviewer recorded any children under the age of 14 of the selected mother who had consented to the study.

Study type

This study is a cross-sectional study.

Period and duration of the study

Data were collected from August to October 2022.

Variables

The dependent variable was the occurrence of a domestic accident. To assess this, mothers were asked a series of questions based on our operational definition: First, the interviewer verified whether the child had permanently lived in this household since January 1, 2022. If yes, the interviewer listed the different types of accidents to the mother that could occur in the household and asked specifically about each, to avoid omissions. The mother was then asked whether the child had experienced an injury or accident requiring treatment since that date. If at least one domestic accident was recorded, the interviewer further inquired about the number of accidents requiring treatment and the approximate accident date. A child was classified as having had a domestic accident (dependent variable = “yes”) if all three conditions were met: permanent residence, occurrence of a domestic accident requiring treatment, and an accident within the last six months. Otherwise, the variable was coded “no”.

The independent variables mainly were the sociodemographic characteristics of the households, the mothers, the fathers, and the children (Table 1).

Table 1.

Independent variables

Group of variables	Variables	Categories
Mothers’ characteristics	Age (in years)	< 25*, 25–34, ≥ 35
	Education level	None*, Primary, Secondary and above
	Occupation	Domestic work*, Student or employee
	Marital status	Single/Separated*, In union
Fathers’ characteristics	Education level	None*, Primary, Secondary and above
	Occupation	Farmer/Planter*, Tradesman, Student or employee
Children’s characteristics	Age	< 12 months*, 1–4 years, 5–9 years, ≥ 10 years
	Sex	Male*, Female
	Education level	None*, Primary or secondary school
	Job or usual activity	Not applicable (< 3 years old)*, Pupil, None, Learning work
Household variables	Place of residence	Urban*, Rural
	Household size	< 5 people*, 5–9 people, ≥ 10 people

*: Reference categorie

Data collection sources, tools, and techniques

We used data extraction, interviews, and observation to collect data on the occurrence of domestic accidents in children. A questionnaire was developed based on those used by previous authors [27, 28]. To design it, we reviewed the scientific literature to identify existing definitions and classifications of domestic accidents and based our instrument on the selected framework. For each reported accident, we collected information on the circumstances of occurrence, the body part affected, and the care-seeking behavior.

Interviewer training and pretesting

We trained the data collectors for two days on the questionnaire and survey techniques. During this training, the questionnaire was translated into the local language (mainly Dioula). We tested the questionnaire in one village not included in the study to check for consistency and, moreover, to standardize the understanding and translation of the questionnaire and the study methodology. Two interviewers carried out the data collection, each pretested the questionnaire with about ten mothers, before ifs final deployment in the field. Thus, although the questionnaire was not formally validated by local experts, is was pre-tested to ensure clarity and usability prior to data collection.

Data management, processing, and analysis

A data quality assurance plan was developed during the training phase. The study was implemented under the supervision of the research team. Data were collected electronically using tablets. The questionnaire was designed on CSPro software [29] and exported to CSEntry for Android for electronic data collection. Collected data were then exported to Stata software version 18.0 [30] for correction, data management, and analysis. The unit of analysis was the child. Firstly, descriptive statistics using frequencies, percentages, and confidence intervals were performed to determine the prevalence of domestic accidents and types of accidents. Then, binary logistic regression was used to determine the factors associated with home accidents, taking into account clustering. We considered that observations are independent across groups (clusters), but not necessarily within groups, since a mother could have one or more children. For this analysis, we first performed a bivariate analysis to examine the association between each independent variable, and the outcome variable using the variable “Mother’s ID” as a clustering variable. The multivariable analysis was performed by including the independent variables (with p < 0.2 in the bivariate analysis), in the model using a bottom-up stepwise procedure. The final multivariable model had the lowest Akaike Information Criterion value. All statistical tests were two-sided; a p-value of < 0.05 was considered statistically significant. Results of logistic regression are presented using crude and adjusted odds ratios and their respective 95% confidence intervals, indicating the significance level.

Results

We included 802 mothers and 1578 children from 805 households. No household refused participation.

Characteristics of children’s mothers

The sample description showed that the mothers were mainly housewives (73.3%), 58.6% had never attended school, and 28.8% were under 25 years. Table 2 shows detailed characteristics of the mothers.

Table 2.

Characteristics of children’ mothers and households (N = 802)

Characteristics	N = 802	%
Place of residence
Urban	208	25,9
Rural	594	74,1
Mother’s age (in years)
Under 25	231	28,8
25–34	331	41,3
35 and over	240	29,9
Mother’s education level
None	470	58,6
Primary	164	20,4
Secondary and above	168	20,9
Father’s education level
None	461	57,5
Primary	209	26,1
Secondary and above	132	16,5
Mother’s occupation
Domestic work	588	73,3
Student or employee	214	26,7
Father’s occupation
Farmer/Planter	601	74,9
Tradesman	133	16,6
Student or employee	68	8,5
Mother’s marital status
Single/Separated	61	7,6
In union	741	92,4
Household size
< 5 people	382	47,6
5–9 people	366	45,6
≥ 10 people	54	6,7

Characteristics of children

Of the 1578 children included in this study, 63.9% had no education level; 595 (37.7%) were aged between 1 and 4 years; 35.2% were pupils. Children’s characteristics are shown in Table 3.

Table 3.

Characteristics of the children (N = 1578)

Characteristics	N = 1578	%
Child’s age
< 12 months	185	11,7
1–4 years	595	37,7
5–9 years	544	34,5
≥ 10 years	254	16,1
Child’s sex
Male	817	51,8
Female	761	48,2
Child’s education level
None	1009	63,9
Primary or secondary school	569	36,1
Child’s job or usual activity
Not applicable, < 3 years old	902	57,2
Pupil	555	35,2
None	93	5,9
Learning work	28	1,8

Prevalence of domestic accidents among children

Among the 1578 children who met our selection criteria, 147 accidents were reported. After applying our operational definition of domestic accidents, 99 of these were classified as domestic accidents, corresponding to a prevalence of 6.27%. In urban areas, nine domestic accidents were recorded in 360 children, corresponding to a prevalence of 2.5%, versus 90 in 1218 children in rural areas, with a prevalence of 7.4% (Fig. 1). By age group, the prevalence of domestic accidents was 2.7% in children under 12 months, 7.4% in those aged 1–4 years, 7.0% in those aged 5–9 years, and 4.7% in children aged 10 years and older. The flow diagram of the study is shown in Fig. 1.

Fig. 1.

Study flow diagram showing domestic accidents by place of residence

Type and mechanisms of domestic accidents

As for the time of the day of occurrence of the accidents, most of them (49%) occurred in the afternoon. Burns were the most common domestic accidents (34%), followed by transport-related accidents (19%), penetrating wounds (14%), and falls (14%). With respect to injury location, the upper limbs were the most frequently affected (44%), followed by the lower limbs (33%). Results are detailed in Table 4.

Table 4.

Type. time of occurrence. and body area injured

Type. time of occurrence. and body areas injured	n = 99	%
Time of injury occurrence
Before 12:00 am	42	42.4
Between 12:00 am and 07:00 pm	49	49.5
After 7:00 pm	8	8.1
Injury mechanisms
Burns	34	34.3
Transport	19	19.2
Penetration	14	14.1
Fall	14	14.1
Foreign body	4	4.0
Bite	3	3.0
Crushing	3	3.0
Blow/projectile	3	3.0
Other mechanism	3	3.0
Poisoning	2	2.0
Location of injuries
Lower limb	44	44.4
Upper limb	33	33.3
Chest	5	5.1
Face	5	5.1
Back	5	5.1
Head	4	4.0
Abdomen	2	2.0
Neck	1	1.0

Factors associated with domestic accidents

Sociodemographic factors significantly associated with domestic accidents were the place of residence and the children’s age. After adjusting for other variables, children in rural areas were 3.39 times more likely to have domestic accidents compared to children living in urban areas (adjusted odds ratio (aOR) = 2.9; 95% CI [1.44–5.84]; p = 0.003). Further, compared with the under one year old, children in the age-group 1–4 years and 5–9 years were significantly more prone to experience domestic accidents, with, respectively, a 2.87 times and 3.82 times more likelihood (aOR = 2.87; 95% CI [1.1–7.46]; p = 0.031 and aOR = 3.82; 95% CI [1.37–10.65]; p = 0.011, respectively). The detailed results are presented in Table 5.

Table 5.

Factors associated with the occurrence of domestic accidents in children in the Orodara district, Burkina Faso (N = 1578)

Characteristics	Crude odds-ratio (95% CI)	p-value	Adjusted odds-ratio (95% CI)	p-value
Child’s age
Under 12 months	Reference		Reference
1–4 years	2.88(1.12–7.38)	0.028	2.87(1.1–7.46)	0.031
5–9 years	2.7(1.04–7.03)	0.041	3.82(1.37–10.65)	0.011
≥ 10 years	1.79(0.6–5.31)	0.298	3.55(1.01–12.47)	0.048
Mother’s age (in years)
< 25	Reference		Reference
25–34	0.64(0.39–1.04)	0.073	0.61(0.37–1.03)	0.066
≥ 35	0.8(0.46–1.37)	0.409	0.76(0.44–1.34)	0.343
Place of residence
Urban	Reference		Reference
Rural	3.11(1.57–6.18)	0.001	2.9(1.44–5.84)	0.003
Child’s education level
None	Reference		Reference
Primary or secondary school	0.68(0.43–1.08)	0.100	0.56(0.3–1.05)	0.073
Mother’s occupation
Domestic work	Reference
Student or employee	0.64(0.37–1.09)	0.102
Mother’s marital status
Single/Separated	Reference
In union	1.66(0.51–5.44)	0.402
Household size
< 5 people	Reference
5–9 people	0.86(0.56–1.31)	0.476
≥ 10 people	0.81(0.3–2.18)	0.677
Child’s sex
Male	Reference
Female	1.26(0.83–1.89)	0.276
Mother’s education level
None
Primary	0.69(0.37–1.3)	0.250
Secondary and above	0.89(0.52–1.54)	0.677
Father’s education level
None	Reference
Primary	1.08(0.66–1.77)	0.746
Secondary and above	0.64(0.33–1.23)	0.178
Father’s occupation
Farmer/Planter	Reference
Tradesman	0.68(0.36–1.28)	0.233
Student or employee	0.87(0.38–1.99)	0.744
Child’s job or usual activity
Not applicable, < 3 years old	Reference
Pupil	0.7(0.44–1.1)	0.121
None	0.74(0.29–1.91)	0.539
Learning work	1.01(0.23–4.39)	0.992

Discussion

Prevalence of domestic accidents

We report, for the first time in the literature, data on the prevalence of domestic accidents in Burkina Faso. We found that the prevalence of domestic accidents among children aged 0–14 years reached 6.27% in the Orodara health district. Surprisingly, this proportion is lower than most of the values reported in the literature, where prevalence ranges from 10% to 25% [13–17, 31].

Our study’s low prevalence might be explained by different reasons:

• First, we notice that there are differences in operational definitions of domestic accidents. In contrast to others, we measured domestic accidents requiring treatment, which is not the case in other studies. For example, Gyedu et al., in Ghana, defined a domestic accident as an injury in a child under the age of 18 years that occurred within 200 m of the home, and either prevented school or work attendance, or for which treatment was sought [18]. Also, others have studied domestic accidents without limiting to those treated [9, 32]. We deliberately included only domestic accidents requiring treatment, as our goal was not only to use a clear definition that mothers could easily apply, but also to restrict the prevalence estimate to significant accidents, i.e. those with an impact on the health care system. Yes indeed, this makes direct comparison with previously reported prevalence rates challenging.

• Second, the prevalence of domestic accidents might be influenced by seasonality and/or duration of observation: some studies assessed it over different time frames than ours, either longer, i.e. 12 months for the Ghana study [33], or shorter, i.e. 3 months in Mali [32]. And indeed, seasonal factors likely influence domestic accidents: During the rainy season, slippery terrain and storms increase risks of falls and electrical injuries, while the dry season might be more associated with animal-related accidents such as snake or scorpion bites. Harvest and planting periods may add further risks, as children accompany parents and face hazards from tools or climbing trees. Further, associated fatigue in parents may reduce supervision at home. School holidays can also raise domestic accident rates, as children spend more time at home or outdoors without close oversight. Our observational time frame from August to October in Burkina Faso includes the rainy season, gradually transitioning to the dry season, and children are still on school break in August. Thus, many environmental factors influencing the rate of domestic accidents are included in our investigation period.

• Third, as discussed below, age groups greatly influence the prevalence of domestic accidents, which most probably also explains differences between studies: some focused only on children under five years of age, such as those conducted in Egypt [13–15], Sudan [17] or Uganda [34], while others included children up to 18 years of age, such as Gyedu et al. in Ghana [16, 18], Mendez et al. in Tanzania [35] and Padonou et al. in Benin [9]. In our study, we included children up to 14 years of age, both to ensure broad inclusivity and because pediatric care in Burkina Faso extends to this age group.

• Lastly, the relatively low prevalence of domestic accidents observed in this study may also be linked to the level of development of residential infrastructures and housing conditions. The more modern the home, the greater the likelihood of domestic accidents from glass doors, falls from beds if households use them, injuries from electric shocks, falls from multi-story buildings, or falls down the stairs, among others. However, to our knowledge, all these potential sources of domestic accidents are absent in Orodara, which clearly distinguishes it from other African or European settings.

Factors associated with domestic accidents

Age

Analysis of associated factors showed that younger age was significantly less associated with the occurrence of domestic accidents. Compared with children aged less than one year, those aged 1–4, 5–9 years, and 10–14 years were more likely to experience domestic accidents. This is in line with the study from Ghana, where Gyedu et al. reported that children aged 1–4 and 5–9 years were more likely to be injured than infants under 12 months [18]. Similarly, other studies have identified young children (1–4 years) as being particularly prone to domestic accidents [13, 23, 36]. The explanation seems to be linked to the developmental status of these children, who are learning to walk, combined with a natural curiosity with the urge to explore their environment, and who have, at this age, a low-risk perception and inability to perceive danger [37]. Indeed, in the study by Halawa et al. in Egypt [14], children under two years living at home had a greater probability of burns caused by hot water, liquids, and flames. This situation is probably common in most developing countries, where the population relies heavily on charcoal and wood fires for cooking [38]. In fact, a study carried out in Ghana reported that out of 6’520 households, 3’856 used charcoal for cooking (59%), and in 4’544 households (70%), the stove/cooking surface was within reach of children under the age of five (i.e., less than one meter) [39]. These conditions are still prevalent in developing African countries such as Burkina Faso and are, therefore, sources of exposure to burns among children. And indeed, in our study as well, burns were the most common domestic accident, accounting to 34% of cases. Further, this finding may also be explained by young children’s limited awareness at their developmental stage and their inability to recognize household risks and hazards, as well as by possible neglect by parents or guardians. We therefore strongly recommend implementing targeted interventions for this younger age group and their mothers to reduce the frequency of domestic accidents.

Rural environment

Our results showed that rural children experience significantly more domestic accidents than urban children. Explanations may lie in a combination of environmental and social factors. As mentioned above, household activities may vary depending on the area of residence and, for example, rural households rely more heavily on open fires, charcoal or wood for cooking, which increases the risk of burns. Indeed, in our study, burns emerged as the most common type of domestic accident. This difference in heat sources may therefore help explain the disparities observed between rural and urban areas.

Further, rural housing structures often lack protective features such as railings or barriers around wells, while proximity to rivers or ponds raises the risk of drowning. In addition, children are frequently exposed to agricultural tools and animals, and may be left unsupervised when parents are engaged in farming or manual labor. Also, cultural practices that involve children in household or agricultural tasks from an early age may also contribute to higher accident rates in rural areas.

Study limitations

Limitations of this study may be related to the technique of data collection (interviews): this requires respondents to have an excellent ability to answer the questions accurately, which may vary from individual to individual. This said, we assume that a mother naturally remembers the details of her child’s accident needing treatment and thus conclude that we have accurately gathered most data. Further, it was not possible to verify the reported information with the primary health centers, and we therefore had to assume that instances where the child required treatment were accurately reported. Finally, although the parent’s occupations can serve as proxies for the household’s socioeconomic status, the absence of a direct measure of this variable could be considered a limitation of the study, as it could help explain the occurrence domestic accident.

Conclusion

Compared to the literature, this study found domestic accidents to be relatively uncommon in the Orodara health district of Burkina Faso, a result likely influenced by our study design. The main sociodemographic factors associated with higher injury risk were younger age and rural residence. The development of public awareness-raising strategies and campaigns may reduce the occurrence of domestic accidents in children.

Authors’ contributions

AC conceptualized the study and analyzed and interpreted the results. AS was the data manager. AB, SI, and BW contributed significantly to the study conception and manuscript writing. All the authors read and approved the final manuscript.

Funding

The University Hospitals of Geneva funded this study.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The parents of the children selected to participate in the research were reassured that their decision would not affect how they would be treated in the primary health centers. In addition, they were given information about the study in the language of their choice, in line with the ethical requirements for research involving human subjects. Those who consented to participate in the study were asked to sign the informed consent form. Participants who could not read (or sign) used their fingerprints on the consent form (using an inkwell). The telephone contacts of members of the research team were made available to parents if further information or assistance was needed. Ethical approval for the study was sought and obtained from the Burkina Health Research Ethics Committee (number 2022-08-182), and all methods were performed in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.