The results of our survey from four rural districts in Sierra Leone found higher rates of healthcare seeking behaviours for all three conditions (>85%) than have been reported in surveys prior to the FHCI. In the rural sample of the 2008 Demographic Health Survey (DHS),  47% of children with diarrhoea, 45% of children with presumed pneumonia and 47% of children with fever were brought for health care. Similarly in the 2009 rural Sierra Leone District Health Services Baseline Survey (SLDHBS),  these rates were 40%, 50% and 52% respectively. In spite of this high rate of healthcare seeking behaviour many children did not receive the recommended treatment and there were still some children who were not brought to health care services, especially children that were given traditional treatments at home.
Our survey sample has similar demographic characteristics to the rural portion of past surveys, using the same sampling scheme, the same survey organization, and the same questions. However, we found a higher than expected prevalence for all three conditions. In the DHS  conducted in 2008 the two week prevalence rate of diarrhoea among the rural sample (similar to our sample) was 13.7%, for presumed pneumonia 7.4% and for fever 24.4% much lower than in our survey. Similarly, in the rural sample of the SLDHBS conducted in 2009  the rates for diarrhoea were 10.5%, presumed pneumonia 5.2% and fever 21.2% again much lower than what we found. Our finding of higher levels of child morbidity than in previous surveys may be because we conducted the survey in the wet season while the other surveys were conducted in the dry season. During the wet season, illnesses, especially malaria, are usually higher. Additionally with the FHCI, respondents may be over-reporting illnesses in anticipation of receiving free services. The location of the questions within the survey and the skip patterns may have differed slightly from DHS and SLDHS which were much larger questionnaires, and we used a PDA, we do not know if these factors resulted in higher rates of reporting symptoms.
With regards to health care seeking, both quantitative survey and qualitative findings found higher levels than what was reported in surveys prior to FHCI. Our finding of higher healthcare seeking behaviours after FHCI is supported by data from the 2010 MICS survey conducted during the dry season 5 months after FHCI which found 74.3% of rural children with presumed pneumonia being brought to a health care provider . These findings suggest that the FHCI resulted in an increased uptake of government health services. Increase in healthcare seeking after removal of user fees has been documented extensively in Africa [17–25]. These increases primarily benefit the poor [19, 26].
Even though we found high healthcare seeking, a large proportion of children did not receive the recommended treatment they should have based on symptom reporting. Although receiving treatment from a peddler or someone other than government health personnel was more common among those not receiving recommended treatments, nurses, health aides, community health officers and doctors still provided the majority of not recommended treatments. In the qualitative study we found that the use of these alternative providers was in part due to unavailability of medicines at facilities and that families retained a critical perspective of providers, valuing those that provided effective treatment. Poor quality treatment and/or stock outs can undermine the potential impact of providing free health care. In Sierra Leone despite FHIC in some instances, drugs and other essential medical supplies was simply not available (Kabano A., UNICEF Sierra Leone, direct communication 2012) and there are claims that sometimes women were still charged for services . In other countries abolition of user fees without proper planning did result in a decrease in overall quality of services, revenues and increases in difficulties meeting recurrent expenses such as purchasing medications [22, 27–30].
Finally among the small proportion of ill children for whom families did not seek health care outside the home, those who used traditional treatments at home, were the most likely not to seek care. Qualitative data detailed households seeking traditional treatment due to preference and due to lack of alternatives. Delays in seeking treatment have been found to be associated with use of home treatments or self-medication in other countries in sub-Saharan Africa [25, 31, 32]. In addition, we found that children with diarrhoea who had multiple symptoms were more likely to be brought to a health care facility. Other studies in sub-Saharan Africa have also found that those with multiple symptoms were more likely to be brought to care [33, 34].
With regards to social determinants, several studies have found an association between health care seeking and socioeconomic status as well as an association with mother’s level of education [33–36] but we did not find such an association except among children with presumed pneumonia where we found an opposite association than expected, children of lowest wealth quintile were actually more likely to seek healthcare. Perhaps the FHIC removed financial barriers to care, but we also must consider the fact that the population in our districts were uniformly poor so differences by wealth quintile may not be very large. Finally, in the household survey we did not find an association with distance from health facilities as has been found in other studies [31, 33, 35]. However, this lack of association may be because we only calculated a straight line distance, and did not take into account geographic obstacles (e.g. lack of roads, rivers, mountains) that could impact healthcare seeking. In the qualitative findings, respondents did report geographic obstacles as reasons for not seeking health care, particularly for those respondents living in more remote villages or that lived across roads and rivers that became impassable during the rainy seasons, living on islands, etc.
There are several further limitations to our research. Our survey was designed to inform and evaluate the CCM program that would be put in place and not the FHCI. It therefore did not include a pre and post FHCI survey design; however, other surveys conducted before and after the FHCI support our findings. We relied on self-reported symptoms and types of treatments; actual diagnosis of conditions may differ from reported symptoms and despite the use of pictures caretakers may not always accurately report what they received. We assumed inappropriate treatment based on symptoms but the actual physical exam may have revealed signs resulting in another diagnosis. As the qualitative research explored all three child morbidities of interest, respondent fatigue may have affected the quality of some responses.