Does free of charge policy increase care-seeking for children under ve in Sierra Leone? Evidence from population-based studies from 2008-2016

In 2010, the Government of Sierra Leone has implemented the Free Health Care Initiative (FHCI) in the country with the objective to reduce under ve mortality. The biggest share of this mortality in the country is attributable to infectious diseases. Care-seeking for children under ve is a key point to reduce mortality due to infectious diseases as it permits early diagnosis and prompt and correct treatment. The objective of this study was therefore to assess the trend in the prevalence of care-seeking as well as to identify determinants of utilization of healthcare services by under-ve (U5) caregivers under the policy initiative. The analysis of care-seeking behavior was done using data from three population-based studies in Sierra Leone (2008-before FHCI and 2013, and 2016-After FHCI). Care-seeking behavior was assessed through care-seeking of caregivers of children U5 with a history of fever in the 2 weeks prior to the survey. We compared the percentages of care-seeking and performed modied Poisson modeling to evaluate the determinants of care-seeking during the period of FHCI (2013–2016).


Abstract Background
In 2010, the Government of Sierra Leone has implemented the Free Health Care Initiative (FHCI) in the country with the objective to reduce under ve mortality.
The biggest share of this mortality in the country is attributable to infectious diseases. Care-seeking for children under ve is a key point to reduce mortality due to infectious diseases as it permits early diagnosis and prompt and correct treatment. The objective of this study was therefore to assess the trend in the prevalence of care-seeking as well as to identify determinants of utilization of healthcare services by under-ve (U5) caregivers under the policy initiative.

Methods
The analysis of care-seeking behavior was done using data from three population-based studies in Sierra Leone (2008-before FHCI and 2013, and 2016-After FHCI). Care-seeking behavior was assessed through care-seeking of caregivers of children U5 with a history of fever in the 2 weeks prior to the survey. We compared the percentages of care-seeking and performed modi ed Poisson modeling to evaluate the determinants of care-seeking during the period of FHCI . Care-seeking was low when the household head was a man, when the head of the household was 15-24 years old, when the child was living in a poorest household, in north or west region and when the child was more than one year old.

Conclusions
The increase of care-seeking in children under ve with fever coincided with the initiation of the FHCI in Sierra Leone. Effective interventions to improve the health facilities' visitation of the children under ve should target the identi ed factors. Background A substantial progress was made in the control of the global under-ve mortality rate with a 49% decline achieved in 18 years, from 2000 (76‰) to 2018 (39‰). The biggest share of this mortality (41.3%) is attributable to infectious diseases among which the leading killers -pneumonia (12%), diarrhea (8.3%), sepsis (7%) and malaria (5%) -are preventable and treatable febrile diseases [1]. These diseases in children under the age of ve remain a public health problem and constitute a handicap of the good growth of children in low incomes countries, particularly in sub-Saharan Africa countries [2]. It was reported 779 cases of respiratory infections per 100,000 population in 2017, and 1028 million cases of diarrhea every year in childhood [3] in sub-Saharan Africa (SSA).
Malaria also affected 213 million people in Africa in 2018 [4]. Sub-Saharan Africa, which housed 50% of the global under-ve deaths in 2018, still ranks last with a mortality rate (at 78‰) slightly above the global estimates two decades ago. Sierra Leone at 105‰, that being 1.4 times above the regional estimate of mortality rate in children under ve [1].
To reduce infectious diseases burden, World Health Organization (WHO) recommends through the integrated management of childhood illnesses, the capacity building of health personnel, the improvement of family and community practices and the strengthening of health systems [5], [6]. In addition, for respiratory and diarrheal diseases, a batch of vaccines has been recommended for children under ve years. Other measures among which, the improvement of access to safe drinking water, the utilization of improved sanitation, personal and food hygiene were part of the African countries' programs to address these pathologies. For malaria, there is vector control using insecticide treated mosquito nets (ITNs), indoor home spraying, the prevention of malaria infection using the intermittent preventive treatment in pregnancy (IPTps) and seasonal malaria chemoprevention (SMC) in children under ve years [7]. These interventions contribute to reducing the burden of infectious diseases. To be more e cient in the reduction of morbidity and mortality in children under ve years, earlier care-seeking at the appropriate place with an adapted treatment is necessary for children under ve years. Despite these control and prevention measures implemented in most Sub-Sahara Africa countries including Sierra Leone, the sustainable development goal 3.2 aiming to reduce the mortality of children under ve to 25 for 1000 live births by 2030 is far from being achieved [8].
Another implemented strategy to reduce burden of these infections is to increase care seeking behavior for febrile children through universal health coverage as part of the sustainable goals aiming to make available essential health care services for all by reducing out-of-pocket expenses of the household. However the universal health coverage, covers only 50% of the population in the world [9]. User fees may account for up to 30 % of household income [10]; as a mean to nance the health system, it has been shown to signi cantly worsen low-and middle-income countries (LMIC) households' nancial hardship and reduce access to healthcare services. Other barriers have been reported to reduce care-seeking. It is health facility deterrents, distances of the location to facilities, socio-cultural and gender, knowledge, and information about healthcare. Thus, in Nigeria, it was reported that access to health facility was by foot on di cult practicable roads in some areas. In Kenya, it was reported mothers perceived the need to be empowered to be able to seek care for their children due to genderrelated barriers. Knowledge about causation and prevention of childhood illnesses, prolonged waiting time, poor communication between staff and patients were also reported to be barriers to health care seeking in Africa and Asia [11,12].
The commitment of African countries to implement universal health coverage leads to the initiative of free health care for the vulnerable population [9]. Sierra Leone implemented free health care for pregnant women, lactating women and children under the age of ve in April 2010, covering essential care. This policy by reducing the barrier of health care cost [13], aimed to increase earlier health care seeking and management of illness in children and to reduce mortality rate in children under ve years.
Although the FHCI was launched ten years ago, there is limited data on the pattern of care-seeking in Sierra Leone and there are no published reports describing factors associated with care-seeking behavior in Sierra Leone even when known that care and treatment of the main infectious diseases are free.
Such knowledge is crucial in re ning public health interventions and therefore needed urgently. Using fever as an indicator of morbidity in children under ve this study aims to evaluate the trend in the prevalence of care-seeking for fever as well as to identify determinants of utilization of healthcare services by U5 caregivers in the context of FHCI. Independent variables included those for the description of the socio-demographic and those for the assessment of the determinants of the care-seeking for fever: the respondent's age ( [15][16][17][18][19][20][21][22][23][24], [25][26][27][28][29][30][31][32][33][34], and ≥35 years); the level of education of the mother of the child (none formal level of education, primary, and secondary or higher level of education); the number of children ever born in the household ( [1][2], [3][4] and 5≥children); the gender of the household head of the child; the age of the head of the household of the child ( [15][16][17][18][19][20][21][22][23][24], [25][26][27][28][29][30][31][32][33][34], and ≥35 years); the wealth index (richest, richer, middle, poor and poorest); the regions (east, north, south, and west); the place of residence (urban or rural); the religion (Christian, Muslim, traditional and others); the gender of the child; the age of the child (<12,  and [36-59] months); and the place to seek health care (public, private, traditional and others). Statistical methods: Statistical analyses were performed using Stata version 15.0. We rst described the characteristics of the parents of the children and those of the children with fever during the two weeks preceding the survey. Chi-square for trend was used to assess differences between participants' characteristics over the surveys. Analyses were done separately to evaluate the trend of the care-seeking for fever in the three surveys; 2008, 2013, and 2016

Methods
for children under ve. The descriptive analyses were weighted for probability sampling and considering strati cation and clustering, as is standard in all surveys. We compared the percentages of care-seeking between the three surveys and performed a modi ed Poisson regression model using a Generalize Estimating Equations (to take into account the clustering effect) to evaluate the determinants of care-seeking under the FHCI (2013-2016).

Results
Characteristics of the study population: Table 1 shows the characteristics of febrile children and those of their parents. In the three surveys, most (more than 45%) of the women were aged from 25 to 34 old years. The majority (more than 59%) of women had no level of education across the surveys. The percentage of women who were Muslim was about 75 % and those who lived in the rural area were more than 65%. At least 40 % were in the poor and poorest groups through the three surveys. For most (more than 75%) households, the head was a man and was aged over 35 years old. The percentage of children aged between 12 to 35 months was at least 45% over the surveys. About 38% of children were in the north region. Trends of the prevalence of the care-seeking for fever: Figure 1     Care-seeking prevalence was 1.5 times higher in 2013 than in 2008 (prevalence ratio,1.48; 95% CI,1.33 to 1.65; p<0.001), and 1.4 times higher in 2016 than in 2008 (prevalence ratio, 1.43; 5% CI, 1.28 to 1.59; p<0.001) after adjustment for mothers age, mother 's education level, the gender of the household head, religion and the age of the children (Table 2).  Table 3 Determinants of the care-seeking for fever under the free-of-charge policy: The results presented in table 3 indicated that the gender of the household head, the age of the household head, the socio-economic status, the geographical regions and the age of the child were significantly associated with care-seeking. Care-seeking was low when the household head was a man, when the head of the household was 15-24 years old, when the child living in a poorest household, living in north or west region and when the child was more than one year old.

Discussion
Using nationally representative health data on children under ve living in Sierra Leone, we found that the implementation of the FHCI in 2010 resulted in a 20 percentage points increase in the prevalence of care-seeking for febrile children. In addition, we identi ed several determinants of care-seeking behavior relative to children and caregivers of which, the household head gender, the region of residence, and the age of children were the most signi cant.
Our study revealed that 7 children under ve in 10 sought care during fever after the FHCI when it was only 5 children under ve in 10 before. The rise we observed in the prevalence of care-seeking for febrile children following the implementation of the FHCI in Sierra Leone is in keeping with the ndings of Garchitorena et al. in Madagascar, where they found similar increase in the care-seeking in children U5 after two years of free health care initiation [20].
In the present study, considering the period following the implementation of the FHCI, the prevalence of care-seeking for febrile children was lower for households in the poorest quintile. The limited ability of the poorest household to bear the indirect costs of health care is a likely contributor to this difference.
In addition, recurrent stock-out of drugs and supplies and the resulting out-of-pocket payment incurred by caregivers for health services despite the FHCI [21,22], contribute to explain this difference between wealth quintiles in terms of care-seeking. A similar nding was reported in a previous study where the increase in the prevalence of care-seeking observed after the abolition of user fees left the most nancially vulnerable people behind [13]. Another nding was the higher prevalence of care-seeking when the head of the household was aged over 25 years compared with that of household heads aged less than 25 years. A possible explanation could be a limited experience of young caregivers in the manifestations of childhood diseases as reported in studies conducted in Sierra Leone and Nigeria where the failure to recognize the symptoms of childhood febrile illnesses led to a delay in seeking care or not seeking care at all [23,24].
We also found that children over one year of age had a low prevalence of care-seeking for fever. The same association was found in other studies in Africa [25,26]. This could be explained by the fact that the caregivers of older children, having dealt with several episodes of childhood febrile illnesses as the child was aging, may have got experienced in treating fevers without resorting to a public health center.
In the present study, the prevalence of care-seeking for febrile children was lower in male compared to the female-headed households. These results highlight the gender in uence at the household decision level in care-seeking decision/behavior for sick children. Care seeking behavior in caregivers of children was reported by Arthur E. in Ghana, Kenya and Zambia to be improved when both parents of the child were involved in the care-seeking decision [27].
In addition, the northern and western regions had low prevalence of care-seeking compared to the eastern region. The population-based surveys in 2013 and 2016 coincided with periods of outbreaks in Sierra Leone. It was the cholera outbreak from 2012 and the Ebola outbreak from 2015. During these outbreaks, the northern and western regions were the most affected [28,29], and population in these regions could have avoided health centers for fear of contamination, this could explain the low prevalence of care-seeking in children under ve in the context of free health care observed in our study. In addition, poor road conditions, availability of affordable and reliable transport means, and physical inaccessibility in the rains may contribute to explain the persisting low prevalence in care seeking in these regions despite the free-of-charge policy [30].
We acknowledge some limitations in our study. The study data collection might have been subject to a social desirability bias. Also, some aspects like social networks of the caregivers, the perception of the necessity of the healthcare services, and the causes of the illnesses of the child contributing to explain the care-seeking behavior could be better addressed by the qualitative research which was not included in the population-based surveys we used data for our analyses.
The strengths of this study include the large size of the sample and its representativeness of the population of children under ve living in Sierra Leone.

Conclusion
Our study revealed that the care-seeking for children under ve increased signi cantly after the FHCI implementation in Sierra Leone. Numerous factors including wealth index were found to be associated negatively with care-seeking. Maintaining the FHCI with adequate strategies to address other barriers beyond nancial ones are urgently needed.