Danger Signs of Childhood Pneumonia: Caregiver Awareness and Care Seeking Behavior in a Developing Country

Background. Efforts to reduce child mortality especially in Africa must as a necessity aim to decrease mortality due to pneumonia. To achieve this, preventive strategies such as expanding vaccination coverage are key. However once a child develops pneumonia prompt treatment which is essential to survival is dependent on mothers and caregiver recognition of the symptoms and danger signs of pneumonia. Methods. This community based cross-sectional study enrolled four hundred and sixty-six caregivers in Enugu state. It aimed to determine knowledge of caregivers about danger signs of pneumonia and the sociodemographic factors that influence knowledge and care seeking behaviour of caregivers. Results. There is poor knowledge of the aetiology and danger signs of pneumonia among caregivers. Higher maternal educational attainment and residence in semiurban area were significantly associated with knowledge of aetiology, danger signs, and vaccination of their children against pneumonia. Fast breathing and difficulty in breathing were the commonest known and experienced WHO recognized danger signs while fever was the commonest perceived danger sign among caregivers. Conclusion. Knowledge of danger signs and health seeking behaviour among caregivers is inadequate. There is need for intensified public and hospital based interventions targeted at mothers to improve their knowledge about pneumonia.


Introduction
Pneumonia causes almost 1 in 5 under-five deaths worldwide and is responsible for more than 2 million deaths of children each year [1]. Nigeria, with a predicated incidence of 6.1 million pneumonia cases and 0.38 episodes of pneumonia per child-year in 2008, ranks 5th among the 15 countries with highest pneumonia burden in the world [2]. It is a leading cause of childhood morbidity and mortality in developing countries including Nigeria [3]. In 2010, pneumonia accounted for 868,000 deaths in under-5 children, which was 14% of all causes of deaths in children [4]. However when detected early, pneumonia can be treated with easily available medication at a low cost [5]. Unfortunately, only one in five caregivers in developing countries can recognize the danger signs of pneumonia like rapid breathing and chest indrawing, and only half of children under 5 with pneumonia are taken to an appropriate healthcare provider [6]. In a population based survey carried out in six countries in Sub-Saharan Africa, only about 18% and 30% of children with suspected pneumonia in Nigeria and Ethiopia, respectively, were taken to hospital for healthcare [7]. In a systematic review of ninetyone studies that reported recognition and/or care seeking for diarrhoea, pneumonia, and malaria in low and middle income countries (LMICs), it was shown that the median sensitivity of recognition of pneumonia among caregivers was low (45.8%) and care seeking from community health worker for pneumonia was 4.2% [8]. The World Health Organization 2 International Journal of Pediatrics (WHO) and UNICEF have recommended strengthening family's capacity to recognize danger signs and prompt care seeking as one of the interventions for controlling pneumonia in children under five [9]. This study seeks to determine the maternal perception and care seeking behaviour concerning danger signs of childhood pneumonia. It is hoped that the findings of this study will help reduce mortality from pneumonia in children and will contribute to achieving the sustainable development goal on child mortality.

Study Area.
This study was conducted in Enugu state, the host town of Enugu State University of Science and Technology. It is located in South East Nigeria on latitude 6 ∘ 27N and longitude 7 ∘ 30E [10]. The economy of Enugu state is dependent mainly on national oil revenue and commerce. Enugu state is made up of 17 local government areas with its capital carved from Enugu North, Enugu South, and Enugu East LGAs. The majority of the inhabitants are of Igbo ethnicity, and Christianity is the dominant religion. The minimum monthly income is similar to the national average of 118,000 (90 US dollars; National Minimum Wage Act 2011). Literacy rate is 66%, which is higher than the national literacy rate of 45%, and there are 955 males per 1,000 females [11].

Study Subject and Sampling
Technique. This was a community based cross-sectional descriptive and analytical study. The study was conducted over a 10-month period in 4 of the 17 local government areas (LGAs) of Enugu state. Multistage sampling method was used to select study participants. The LGAs were divided into rural and semiurban categories based on level of development and population in the LGAs. Simple random sampling using balloting method was used to select 2 LGAs from each category.
In the second stage, one community was selected from each LGA using simple random sampling (i.e., one community from each of the two rural and two semiurban LGAs). In each of the communities visited, through adequate mobilization using the community leaders, the women were asked to gather at a convenient square. Women who were caring and/or had cared for a child in the past 2 years and consented to participate were consecutively enrolled. This process was repeated for each of the four communities visited. Three hundred and nineteen (319) caregivers who met the inclusion criteria and gave consent to participate were interviewed in the rural communities while one hundred and forty-seven (147) were interviewed in the semiurban communities.

Measures
3.1. Sociodemographic Characteristics. The following sociodemographic variables of respondents were assessed: (i) age grouped into 20-30, 31-40, and >40 years, (ii) highest educational attainment grouped into tertiary, secondary, primary, and no education, (iii) number of living children grouped into none, 1-4, and >4 children, and (iv) socioeconomic class: defined as the wealth index of the household derived using maternal and paternal highest educational attainment and occupation based on Oyedeji classification [12].

Knowledge of Pneumonia and Health Seeking Behaviour of Respondents.
Pneumonia is inflammation of one or both lungs caused by variety of organisms and chemical substances characterized by cough, shortness of breath, fever and difficult breathing, cyanosis, and death in severe cases. Pneumonia is also called "oyi" in the local Igbo language and this name was used for respondents that do not understand English. The World Health Organization (WHO) recognizes four features as danger signs in pneumonia. They include stridor, fast breathing, chest wall indrawing, and difficulty in breathing (labored breathing). Knowledge of danger signs of pneumonia was assessed using open ended questions based on WHO definition [9]. Based on their responses, the respondents were grouped into no knowledge of danger signs and knowledge of at least one danger sign. Also knowledge and uptake of vaccine against pneumonia were assessed. The respondents were grouped into "Yes" for those who were aware of the danger signs and had vaccinated their children and "No" for those who had not heard about pneumonia and had not vaccinated their children. Health seeking behavior of respondents was assessed to determine where care was sought for a child with suspected pneumonia and the treatment given.

Data Cleaning and Analysis.
Quality control check to detect errors in questionnaire administration and data recording was done by researchers on daily basis after enrollment. Where there are errors detected, the interviewers were asked to clarify them accordingly with the interviewed mother or caregiver. Microsoft excel 2007 was used to input the raw data. Data cleaning to remove grossly incomplete and/or inconsistent data was also done by researcher assistants and the study researchers. SPSS version 20 was used for data analysis. Chi-square and Fisher exact test where appropriate were used in data analysis. Results were presented as percentages and 95% confidence intervals where appropriate. Statistical significance was set at value < 0.05.

Ethical Consideration.
Ethical clearance was obtained from the Enugu State University Teaching Hospital Ethics Committee. Informed consent (written) was obtained from every mother in her own right and on behalf of her child before recruitment. Participation in the study was entirely voluntary and no financial inducement whatsoever was involved. All information was handled with strict confidentiality.
The WHO/UNICEF recognised danger signs most commonly known by respondents were fast breathing (60.5%) and difficulty in breathing (22.1%) while chest indrawing (12.4%) and stridor, that is, noisy breathing (5.0%), were less known to respondents ( Table 3). Knowledge of at least one WHO/UNICEF recognised danger sign was seen in 304 (65.2%) of respondents while knowledge of 2, 3, and 4 recognised danger signs was seen in 219 (47.0%), 37 (7.9%), and 22 (4.7%) of respondents, respectively. One hundred and sixtytwo (34.8%) had knowledge of none of the WHO/UNICEF danger signs. Just like cause and vaccine knowledge, maternal level of educational attainment ( = 0.04), and place of residence ( = 0.00) were significantly associated with knowledge of the WHO/UNICEF recognised pneumonia danger signs in children among respondents (Table 2).

Discussion
This study investigated the knowledge of caregivers and caregivers in Enugu about the aetiology and danger signs of pneumonia. It also sought to determine factors that influenced the knowledge and health seeking behaviour of caregivers for their under-5 children with probable pneumonia. The study showed high awareness (95%) and knowledge of potential fatality of pneumonia disease (97.7%) among respondents but poor knowledge of the aetiology (4.1%) and danger signs of probable pneumonia. A similar study in Thailand also showed inadequate knowledge of danger signs (7%) and causes (21%) of pneumonia among respondents [13]. The relatively poorer knowledge of causes of pneumonia in this study may be related to the higher number of caregivers without any formal education (29.2%) compared to 4.3% among respondents in the Thailand study. Another study in Uganda corroborated the poor knowledge among caregivers. It was noted in that study that none of the caregivers surveyed was able to mention the four standard danger signs of childhood pneumonia and only 9.4% of the respondents 6 International Journal of Pediatrics knew that lower chest wall indrawing was a danger sign in childhood pneumonia [14].
It was also shown in this study that older caregivers, caregivers with higher educational attainment, and those that are resident in semiurban areas had better knowledge of the cause and danger signs of pneumonia. This is hardly surprising as caregivers with these sociodemographic variables are more likely to have more experiences of childhood illnesses and/or are better informed about pneumonia. A hospital based study in Lagos University Teaching Hospital affirmed this finding [15]. It showed that caregivers with lower than tertiary education were more likely to have less knowledge on aspects of pneumonia disease. The Lagos study in addition showed that caregivers whose information source about pneumonia was from health personnel were more likely to have correct information on the cause of pneumonia compared to those from other sources. This fact was supported by this present study which revealed that respondents who obtained their information from medical personnel were more likely to have correct knowledge of aetiology and prevention of pneumonia compared to respondents that got their information from other sources.
Inquiries into the preventive strategies showed that majority of the respondents believed that adequate clothing and avoidance of cold food, drink, or environment prevents pneumonia disease in a child. This belief which is a common misconception among many lay people in Nigeria was also corroborated by the study in Lagos which found that 75.6% of mothers surveyed believed that cold was the main cause of pneumonia [15]. Furthermore, almost half of respondents in the present survey had knowledge of pneumonia vaccine with only 39.9% of these respondents having vaccinated their child against pneumonia. Similarly, only about a tenth of caregivers in our survey listed vaccination as a preventive strategy for pneumonia. This low uptake rate is expected as pneumococcal vaccines, which are relatively new additions to the national immunization schedules, are not widely available in public health facilities. Furthermore, even when available, they are not free like other routine vaccines in Nigeria.
Higher maternal educational attainment and residence in semiurban areas were significantly associated with knowledge about pneumococcal vaccine and uptake rate among children of surveyed caregivers. A study conducted in the Africa Center Demographic Site in South Africa showed that caregivers with secondary and tertiary education were 1.10 and 1.09 times, respectively, more likely to uptake Pneumococcal Conjugate Vaccine (PCV) for their children compared to caregivers with no education. It further showed that those caregivers resident in rural and periurban areas were 0.90 and 0.96 times less likely to uptake pneumococcal vaccine for their child. The author argued that distance from vaccination centers may have accounted for these findings [16].
Though majority of the respondents in our study took their children to hospital on suspicion of pneumonia, nearly 40% either self-medicated and/or consulted traditionalist or spiritualist as the first line of action. The finding that caregivers with lower education were more likely to engage in this behaviour is easily rationalized. Because these caregivers are more likely to be less informed and easily persuaded, they will more likely be influenced by advice from friends and relatives to seek nonmedical solutions for their child's illness (especially in the face of poverty). This is a fairly common practice in our society. It was further noted that caregivers in rural area were more likely to visit hospital, less likely to selfmedicate, and also more likely to visit traditionalist and/or spiritualist compared to caregivers in urban areas. It could be further rationalized that because of the easy accessibility of various category of drug stores and ease of getting drugs over the counter in urban areas caregivers tended to patronize these drug dealers and self-medicate their children during illnesses. Similarly, because of the traditional belief which is inherent in rural communities and poorer accessibility to healthcare facilities, it is not surprising that caregivers in rural areas may more likely seek traditional and/or spiritualist help for their child's illness. Contrary to this finding, a community based comparative study in Dera district of Oromia Regional state, Ethiopia, showed that more caregivers in urban area (75.0%) sought healthcare for acute respiratory tract infection in their children from healthcare institution compared to those in rural areas (34.4%). Although no explanation was proffered by the authors for this finding, one could speculate that the higher number of caregivers with formal education in the urban category (28.7%) compared to caregivers in the rural category (3.6%) may have contributed to the higher healthcare seeking behaviour among caregivers in the urban compared to those in the rural setting [17] unlike our study where there is no significant difference in educational attainment between caregivers' resident in urban and rural area.
Finally, we observed in our study that fast breathing and difficulty in breathing (labored breathing) were the most known and experienced WHO recognized danger signs among respondents. This may be related to the ease of noticing these symptoms in children compared to chest wall indrawing and stridor. The proportion of caregivers who had knowledge of these signs was more than what was observed in a similar study in Uganda [14]. On the other hand, fever, fast breathing, and cough were the most perceived and experienced danger sign among respondents in this study. Khalida et al. [18] in their own study in Mirpur Khas district of Pakistan similarly noted that fever and cough (65.2%), followed by fast breathing and chest wall indrawing (59.4%), were the most perceived symptoms of pneumonia by mothers in the district. As noted earlier, this may be related to the ease of picking up these symptoms in children.

Limitation
Selection bias would have resulted due to the point enrollment because prospective study participants were required to gather at a preagreed location for consenting and enrollment. Household sampling and enrollment which could have significantly reduced selection bias could not be done during selection of respondents in the community due to unorganized household locations and difficult terrains in some communities. Also, recall bias could have led to errors in data measurement, classification, and analysis since some caregivers were interviewed on events that took place in the past.
International Journal of Pediatrics 7

Conclusion
This study demonstrated poor knowledge of the aetiology, danger signs, and prevention of childhood pneumonia amongst caregivers. Health education targeted at caregivers will enhance dissemination of information to fill this knowledge deficit and hopefully reduce morbidity and mortality from childhood pneumonia. Therefore the authors intend to embark on workshops to sensitize medical personnel on the need to educate the caregivers on the danger signs and prevention of childhood pneumonia.