The proportion and determinants of appropriate health seeking behavior for febrile illness among caregivers of children under-five years in Butula sub-county, Busia county, Kenya Authors: Jean L. Kananura, Betsy C.Rono, Kamija S.Phiri, Prashant Kulkarni, Oluseye Ademola Okunola, Jean Louis Kananura, Prashant Kulkarni, Jean Louis Kananura, Chijioke Amara Ezenyeaku, Jean Louis Kananura Published: 2024 ABSTRACT BackgroundAlmost 10 million children under the age of five in Kenya, die due to fever-related diseases. In Busia, a county in Kenya, malaria accounts for 13% of all child fatalities under the age of five, a number higher than any other county. This study aimed to determine (a) proportion of appropriate health-seeking behavior and (b) determinants of health-seeking behaviors (HSBs) among their caregivers in Butula Sub-County, Busia County, as reported by the caregivers.MethodsThis cross-sectional mixed-method study included 271 caregivers, 11 community health volunteers, and health facility workers in Butula Sub-County. Systematic random sampling for participants and purposive sampling for key informants were used. A questionnaire collected the data that was analysed using frequency and percentages and logistic regression.ResultsAround seventy percent of caregivers reported seeking care for a child's fever within 24 hours. Individual factors that caregivers reported to influence appropriate health-seeking HSB were unemployment (adjusted odds ratio (aOR) = 0.49, 95% CI: 0.217 – 0.593, p = 0.018), self-medication preference (aOR = 0.14, 95% CI: 0.054 – 0.363, p < 0.001), had atleast two children (aOR = 0.63, 95% CI: 0.425 – 0.937, p = 0.042), and confidence in identifying fever (aOR = 7.0, 95% CI: 2.200 – 22.439, p = 0.001). Health-system factors reported to influence HSBs were facility too far (aOR = 0.86, 95% CI: 0.526 – 0.914, p = 0.027), getting health education (aOR = 1.8, 95% CI: 1.201-4.122, p=0.015), and facility level (aOR = 4.4, 95% CI: 2.015 – 9.750, p < 0.001). Qualitative findings found health system factors related to HSB as stockouts, facility distance, and staff workload.ConclusionsEmployment, multiple children, preferences to self-medicate, confidence in fever identification, facility level, distance to facility, and education are ky factors affecting caregiver’s HSB. Policy and practice efforts should focus on these significant individual and health system determinants for HSBs among caregivers of children under five with febrile illness. Introduction Febrile illness continues to be a significant global health problem, causing approximately 75% of mortality in children under five years old 1 . Although the World Health Organization (WHO) does not classify febrile illness as a disease, it is associated with infectious conditions such as pneumonia, malaria, and typhoid fever, which are linked with high morbidity and mortality in children under five 1, 2 . Each year, it is estimated that about 10 million children under five years in developing countries die due to febrile illnesses 3 . In Kenya, more than 10 million children under five years require treatment for acute febrile illnesses 4 . In Busia County (where the study site is located), the highest rate of febrile illness and related mortality is reported in all the Kenyan counties. In 2022, the Kenya Health Information System (KHIS), for instance, reported that the prevalence of under 5 years cases with fever was 20% in Butula Sub-County Busia County 5 . Febrile illness-related deaths are high because caregivers of children under five fail to seek timely medical care. In 2020, the Kenya Malaria Indicator Survey (KMIS) showed that the proportion of children with fever who received treatment dropped from 72% to 64% between 2015–2020 6 . Research shows that the behavior of seeking timely treatment significantly impacts febrile illness 7 . In the current study, these behaviors are referred to as health-seeking behaviors (HSBs). Health-seeking behaviors are defined as any attempt by an individual (in this case a caregiver) to obtain the proper and timely treatment to address a health condition upon symptom manifestation 8 . UNICEF and WHO, in their global action plans, recommend that caregivers obtain adequate treatment for their children to reduce the impact of febrile illness 9 . Many factors may determine HSBs among caregivers of children under five, including individual and health-system factors. Research on health-seeking behavior for febrile illness has revealed regional variations in health-seeking rates. According to Min et al. 10 , geographic location is linked to the decision not to seek treatment for fever, with only 32.0% obtaining treatment within 24 hours. In sub-Saharan Africa, the percentage of caregivers who sought treatment for children with febrile illnesses was 67.3%, 56.8%, and 27.0% in Malawi, Tanzania, and Zambia 11 . Odimbe and Atuhairwe, in a study conducted in the Ugandan town of Busia Municipality, revealed that factors such as illness severity and caregivers' age influenced caregivers' health-seeking behaviors for children under five years with malaria 12 . Additionally, they identified the quality of health services and health workers' behaviors as factors influencing caregivers' HSBs to treat malaria in children under five 12 . Liyew et al. 11 reported that knowledge, educational status, and age influenced caregivers' decision to seek treatment for febrile illness. They also reported that the availability of health facilities influenced caregivers' HSBs for children with fever 11 . In Kenya, poor HSBs among caregivers are demonstrated by the significant decline in the proportion of children receiving treatment for fever 6 . Individual and health system factors may contribute to these poor HSBs. Still, there are limited comprehensive studies done in Kenya, particularly in the Western region that is prone to febrile illness, to support these claims. Therefore, this study aimed to achieve two objectives, (a) the proportion of appropriate health-seeking behavior and (b) determinants (individual and health-related factors) of health-seeking behaviors among caregivers of children under 5 years with febrile illness in Butula Sub-County, Busia County, Kenya, as reported by caregivers. Methods Study type and period A community-based analytical cross-sectional, explanatory sequential mixed methods design was employed in the study. The research assistants first collected quantitative data and later qualitative data to support the quantitative data. The study period was between May 22, 2023 and July 28, 2023. Study setting The study site was Butula Sub-County, Busia County. It is one seven sub-counties in Busia County, the other being Teso North, Teso South, Nambale, Bunyala (Bundalangi), Samia (Funyula) and Matayos sub-county administrative units. Not much is published about the sub-county. Busia together with Kakamega, Bungoma, and Vihiga counties constitute the western region of Kenya. It borders the Republic of Uganda to the North, Siaya to the south, Kakamega to the east, and Bungoma to the Northeast. Busia County covers an area of approximately 1,683 sq. km and is located between latitudes 00° 01' and 00° 47' North of the Equator and longitudes 33° 57' and 34° 26' East of Greenwich Meridian. Butula sub-county is prone to febrile illness-associated diseases because of being in the Victoria region basin, hence why it was purposively selected as the study area. Target Population. Caregivers of children under five years old in the study setting. Inclusion Criteria. Caregivers over 18 years, with a child under five, residing in Butula, and willing to consent. Exclusion Criteria. Caregivers who had lived in Busia County for less than two weeks before study period as they have not resided in the county to experience febrile diseases and engage in appropriate HSB. Sampling Butula sub-county has six wards. Three (Marachi West, Marachi Central, and Kingandole) were selected using simple random sampling. The sample for the quantitative phase of the study was selected from homes with children under five using systematic random sampling. A sample size of 271 participants was used in the study. This sample was obtained using Fisher's formula with an acceptable confidence level of 95% 13 . The prevalence or proportion used was the prevalence of fever among under-five children 14 . Where: n 0 is the sample size Z 2 is the abscissa of the standard curve that cuts off an area α at the tails (1 – α), which equals the desired confidence level at 95%). e is the desired level of precision p = Expected prevalence or proportion of an attribute/disease in the population from previous or pilot studies. Data on the prevalence of 20% of children below five with a fever in the two weeks before the malaria survey in Butula sub-county 5 . q is 1-p. Z = 1.96 for a 95 % level of confidence = 217 the sample is adjusted upwards by 10% to cushion for attrition cases The sampling frame was determined from the sub-county register containing the number of households for caregivers under under-five. A selection interval (k) was determined by dividing the total number of households and participants. The randomly generated numbers on a computer were used to select the first household in the study each day. The 271 caregivers were allocated across the three wards at a ratio of 2:2:1. Healthcare workers who work in the pediatric department at Mayo Sub-county Hospital and community health volunteers providing care to children and caregivers in Butula sub-County were purposively selected as crucial informants. Eleven key informants were found to saturate the qualitative data. Data collection and tools Two research assistants collected quantitative data from 271 caregivers between May and July 2023 using a semi-structured questionnaire. The semi-structured questionnaire developed by the primary investigator was used to collect data for the quantitative phase of the study. It contained five parts that collected data on social demographic, social and economic data of caregivers, characteristics of children, the prevalence of febrile, HSBs for febrile illness, and factors influencing HSBs. Its validity and reliability were tested using 10% of the sample in the Nyalenda ward in Kisumu County, which has similar geographical characteristics and experiences high cases of febrile illness as Butula Sub-County. Cronbach's alpha was 0.86, indicating high reliability. Before distributing the questionnaire, the research assistants explained the study's purpose and objectives and obtained a signed informed consent form. An unbiased witness was present when a caregiver was incapable of reading or writing to make sure they received the correct information regarding the study. An in-person (face-to-face) approach was used to administer the questionnaire. The choice of research assistants was based on their educational background; prerequisites for inclusion were (a) a minimum degree of medical training, (b) fluency in English and Kiswahili, and (c) speaking the native language. A critical informant interview guide, developed by the study investigator, helped collect data from healthcare workers at the Mayo Sub-County Hospital through semi-structured interviews. Interviews with health workers lasted 30–45 minutes and targeted institution-level factors such as turnaround time (TAT), availability of commodities, availability of human resources, distance, hospital fee, and distance. The interview session ended after reaching data saturation—FGDs with CHVs collected data on health-system-related determinants of HSBs. Tape recordings were used to capture and maintain the informants exact words. A focus group discussion (FGD) guide, also developed by principal author, collected data from two FGDs, each lasting 30–60 minutes., The aim of these FGDs was to explore more into the determinants of HSB among caregivers of children under five with for febrile illness. They were conducted until data saturation. Data management and analysis Quantitative data was cleaned in Excel and imported into Statistical Package for Social Scientists (SPSS) version 25 for descriptive and inferential analysis. Descriptive statistics included frequency, percent, and measures of central tendency and presented as figures and tables. A binominal logistic regression analysis was used to determine the association between individual or health-system factors and HSBs using odds ratio (OR) at 95% confidence interval and p-value less than 0.05. Qualitative data analysis began with transcribing the voice recording and transcripts. Concepts, words, phrases, and sentences were coded to identify themes through categorization. These themes formed the study results. All qualitative data analysis was manually done in Microsoft Excel 365. Results In our study, 271 caregivers were enrolled between June and December 2023, and all of them responded to the questionnaire. There were no refusals, however 14 households had no one available during the survey and were omitted from the study altogether. Participants characteristics The mean age of participants was 33.4 ± 12.4 years, with most (42.1%) being between 21.3 years. 84.1% were female, and 44.6% had secondary-level education. More than half (59%) were self-employed and earned an income of less than KSH 5000 (75.3%). Most (79.3%) lived in a nuclear family, had two or fewer children (48.7%), and had one child under five years (67.5%). ( Table 1). Table 1. Demographic characteristics of study participants.TotalProportion (%) Age (years) ≤ 20 21 – 30 31 – 40 41 – 50 51 – 60 ≥ 61 26 114 80 20 18 13 9.6 42.1 29.5 7.4 6.6 4.8 Mean age M = 33.4 year (SD = 12.4) Sex Male Female 43 228 15.9 84.1 Education Level None Primary Secondary Tertiary 13 81 121 56 4.8 29.9 44.6 20.7 Occupation status Unemployed Self Formal 90 160 21 33.2 59.0 7.7 Religion Christians Others 268 03 98.9 1.1 Residency Ward Marachi Central Kingandole Marachi West 110 106 55 40.6 39.1 20.3 Income (KSH) < 5,000 5,000 – 10,000 10,001 – 20,000 20,001 – 50,000 > 50,000 204 36 13 14 4 75.3 13.3 4.8 5.2 1.5 Type of Household Grandparent Family Single-parent Nuclear Family Extended Family 12 23 215 21 4.4 8.5 79.3 7.7 Age of the Child ≤ 6 months 7 – 12 Months 13 – 24 Months 25 – 36 Months 37 – 48 Months 49 – 60 Months 52 63 60 49 35 12 19.2 23.2 22.1 18.1 12.9 4.4 The mean age of the child M = 21.5 months (SD =15.6) Number of Other Children Two or less 3 – 5 Six or More 132 89 50 48.7 32.8 18.5 The mean number of other children Mean = 2.9 (SD = 2.1) Number of Children < 5 years One Two Three Four or More 183 64 6 2 71.8 25.1 2.4 0.8 The mean number of children under 5 years M = 1.4 (SD = 2.0) Demographic characteristics of study participants. Prevalence of febrile illness among under-five children Febrile illnesses were assessed on how many children experienced fever-related illnesses in the last two weeks before the study. Figure 1 shows that 64.6% of children experienced fever episodes, and 35.4% did not. The proportion of children with fevers as reported by caregivers. Proportion of appropriate Health Seeking Behavior The health-seeking behavior was assessed on how long the caregiver took to seek health after detecting the child had a fever. Figure 2 shows that 70.1% of caregivers reported that they did seek healthcare within 24 hours of detecting fever in their child, while 29.9% did not. The period within which caregivers seek health care services after detecting fever in the child. Determinants of Health-Seeking Behavior among caregivers Individual-level factors. The bivariate analysis of individual factors and HSB within 24 hours showed a significant association for some factors. Table 2 shows that unemployment, self-medication, personal beliefs about care, confidence in fever identification, fever severity, using a hand to identify fever, and having multiple children were significantly associated with health-seeking behavior. Being unemployed lowered the likelihood of a caregiver seeking care within 24 hours by 69% (OR = 0.41, 95% CI: 0.212 – 0.807, p = 0.011). Those who addressed fever by self-medication rather than seeking professional health had odds of caring for fever within 24 hours reduced by roughly 84% (OR = 0.16, 95% CI: 0.093 – 0.290, p < 0.001). The belief that it is good to take care of fever within 24 hours was associated with a threefold higher likelihood of seeking health within 24 hours (OR = 3.4, 95% CI: 1.441 – 8.217, p = 0.007). Those confident in identifying fever in under-five-year-old children had close to 9 times higher likelihood of having the child's fever taken care of within 24 hours (OR = 8.8, 95% CI: 3.331 – 23.049, p < 0.001). The severity of the fever increased the chances of care for fever being sought within 24 hours by two times (OR = 2.4, 95% CI: 1.401 – 4.083, p = 0.002). Using the hand to identify fever was associated with 85% chances of low HSB within the first 24 hours (OR = 0.15, 95% CI: 0.019 – 0.693, p = 0.045). Other children seemed to calm the caregiver and lower the tendency to seek care within 24 hours by 46% (OR = 0.54, 95 CI: 0.294 – 0.981). Table 2. A bivariate association between individual factors and caregivers' health-seeking behavior.VariableGood HSB N (%)Poor HSB N (%)Odds Ratio (OR) (95% CI)P -Value Caregiver Age ≤ 25 Yrs. > 25 Yrs. 53 (27.9) 137 (72.1) 20 (24.7) 61 (75.3) 1.2 (0.650 – 2.142) 0.655 Level of Education Primary or None Secondary or Higher 63 (33.2) 127 (66.8) 31 (38.3) 50 (61.7) 0.8 (0.466 – 1.373) 0.486 Occupation Unemployed Employed (Self/Formal) 130 (68.4) 60 (31.6) 68 (84.0) 13 (16.0) 0.41 (0.212 – 0.807) 0.011 Marital Status Single/Divorced/Widowed Married 150 (78.9) 40 (21.1) 61 (75.3) 20 (24.7) 1.2 (0.761 – 1.749) 0.525 Relationship with Child Parent Others 156 (82.1) 34 (17.9) 67 (82.7) 14 (17.3) 0.9 (0.483 – 1.902) 1.000 Monthly Income ≤ Ksh. 5000 > Ksh. 5000 142 (74.7) 48 (25.3) 62 (76.5) 19 (23.5) 0.9 (0.493 – 1.667) 0.878 How they Often Address Fever Self-Medication CHV or Clinician 49 (25.8) 141 (74.2) 55 (67.9) 26 (32.1) 0.16 (0.093 – 0.290) < 0.001 Good to Take Care with 24 Hrs. Yes No 180 (94.7) 10 (5.3) 68 (84.0) 13 (16.0) 3.44 (1.441 – 8.217) 0.007 Can Identify Fever in Under 5 Yes No 184 (96.8) 6 (3.2) 63 (77.8) 18 (22.2) 8.8 (3.331 – 23.049) < 0.001 Severity Influenced HSB Yes No 133 (70.0) 57 (30.0) 40 (49.4) 41 (50.6) 2.4 (1.401 – 4.083) 0.002 Culture Influence HSB Yes No 7 (3.7) 183 (96.3) 4 (4.9) 77 (95.1) 0.74 (0.209 – 2.588) 0.738 How do you Identify fever? Feeling by Hand Thermometer 175 (92.1) 15 (7.9) 80 (98.8) 1 (1.2) 0.15 (0.019 – 0.693) 0.045 Has Two or More Children Yes No 124 (65.3) 66 (34.7) 63 (77.8) 18 (22.2) 0.54 (0.294 – 0.981) 0.045 Has other Under-five Children Yes No 179 (94.2) 11 (5.8) 76 (93.8) 5 (6.2) 1.071 (0.360 – 3.186) 1.000 Legend: CHV = community health volunteer, HSB-health seeking behavior, OR = odds ratio, p = p-value A bivariate association between individual factors and caregivers' health-seeking behavior. Legend: CHV = community health volunteer, HSB-health seeking behavior, OR = odds ratio, p = p-value Controlling Confounding Variables. Table 3 shows the adjusted odds ratio for individual factors influencing HSB within 24 hours after controlling confounders, such as unemployment, self-medication preference, having multiple children, and confidence in disease identification. Unemployment lowered health-seeking behavior rate by 51% (aOR = 0.49, 95% CI: 0.217 – 0.593, p = 0.018). Caregivers who preferred self-medication over professional care for their child's fever had 86% lower chances of seeking care within 24 hours (aOR = 0.14, 95% CI: 0.054 – 0.363, p < 0.001). Those with two or more children witness a 37% lower chance of seeking care within 24 hours than those with one or less (aOR = 0.63, 95% CI: 0.425 – 0.937, p = 0.042). Caregivers who could be confident identifying when their child had a fever were seven times more likely to seek health care services within 24 hours than those who were not confident of detecting fever (aOR = 7.0, 95% CI: 2.200 – 22.439, p = 0.001). Table 3. Multivariable logistic regression on individual-level factors influencing Health-Seeking Behavior.VariablesCrude Odds Ratio (95% CI)P - ValueAdjusted Odds Ratio (aOR) (95% CI)P - value Caregiver occupation Unemployed Employed (Self/Formal) 0.41 (0.212 – 0.807) 0.011 0.49 (0.217 – 0.593) 0.018 How they often address fever Self-Medication CHV or Clinician 0.16 (0.0.93 – 0.290) < 0.001 0.14 (0.054 – 0.363) < 0.001 Good to take care with 24 hours Yes No 3.44 (1.441 – 8.217) 0.037 1.62 (0.525 – 5.017) 0.401 Can identify fever in under 5-year-olds Yes No 8.8 (3.331 – 23.049) < 0.001 7.03 (2.200 – 22.439) 0.001 Severity influenced HSB Yes No 2.4 (1.401 – 4.083) 0.002 1.23 (0.622 – 2.418) 0.556 How do you identify fever? Feeling by Hand Thermometer 0.15 (0.019 – 0.693) 0.045 0.21 (0.025 – 1.801) 0.156 Has two or more children Yes No 0.54 (0.294 – 0.981) 0.045 0.63 (0.425 – 0.937) 0.042 Legend: CHV = community health volunteer. Multivariable logistic regression on individual-level factors influencing Health-Seeking Behavior. Legend: CHV = community health volunteer. Health systems factors. Health-system factors associated with HSB based are presented in Figure 3. Based on quantitative data, most participants (84.1%, N = 228) confirmed that they had received health education from the providers on fever care while 10.0% did not. Furthermore, 71.2% (N = 193) agreed or strongly agreed to considering the facility level when deciding to seek medical attention. 44.6% (N = 121) verified that long waiting had a negative impact on the delay in seeking medical attention while 32.1% (N = 87) thought that their HSB was unaffected by extended waiting time. 37.3% (N = 86) reported that their HSB was limited by inadequate medication. 37% (N = 86) indicated that the expense of febrile treatment was a limiting factor in their HSB while 49% (N = 134) disputed this. Long distance to the facility was mentioned as a problem restricting HSB by 25.8% (N = 70) of the participants while 63.8% (N = 173) considered long distance was not an issue impacting HSB. Participants’ perception on facility system factors influence on HSB. Multivariate analysis showed that a facility being too far was associated with lowering the HSB within 24 hours by 14% (adjusted odds ratio (aOR) = 0.86, 95% CI: 0.526-0.914, p = 0.027). Those offered health education had increased odds of having sought care within 24 hours by 82% (aOR = 1.82, 95% CI: 1.201-4.122, p = 0.015). The type/level of the facility was also a factor for consideration, increasing the access to fever care within the first 24 hours by more than four times (aOR = 4.43, 95% CI: 2.015-9.750, p < 0.001). However, cost of care (p = 0.154), long waiting time (p = 0.689), and lack of drugs in the facility (p = 0.891) had no statistical significance association ( Table 4). Table 4. Bivariate and multivariate association between health seeking behavior and health system factors.Good HSB N (%)Poor HSB N (%)OR (95% CI)pAdjusted OR (95% CI)p Facility too Far No (reference) Yes 150 (78.9) 40 (21.1) 51 (63.0) 30 (37.0) 0.45 (0.256 – 0.802)0.010 0.86 (0.526 – 0.914)0.027 Cost of Care No (reference) Yes 135 (71.1) 55 (28.9) 50 (61.7) 31 (38.3) 0.66 (0.380 – 1.135) 0.154-- Long Waiting Time No (reference) Yes 112 (58.9) 78 (41.1) 36 (44.4) 45 (55.6) 0.56 (0.330 – 0.942) 0.033 0.85 (0.378 – 1.902)0.689 Lack of Drugs in Facility No (reference) Yes 120 (63.2) 70 (36.8) 50 (61.7) 31 (38.3) 0.94 (0.550 – 1.609) 0.891-- Staff Offered Health Education No (reference) Yes 19 (10.0) 171 (90.0) 24 (29.6) 57 (70.4) 3.8 (1.934 – 7.423) < 0.001 1.82 (1.201 – 4.122) 0.015 Facility Type/Level No (reference) Yes 44 (23.2) 146 (76.8) 34 (42.0) 47 (58.0) 2.4 (1.378 – 4.182) 0.003 4.43 (2.015 – 9.750)< 0.001 Legend: HSB = health seeking behavior, OR = odds ratio, p = p-value Bivariate and multivariate association between health seeking behavior and health system factors. Legend: HSB = health seeking behavior, OR = odds ratio, p = p-value Findings from critical informant interviews Based on the interview with health workers, the following themes emerged as impacting HSB among caregivers of children under five years; medication stockouts, staffing issues, facility level, behavior of health professionals, distance from the facility, and lack of funding for outreach activities. These findings supported the quantitative findings. One health care informant said "… caregivers preferred facilities with specialists, those that do not charge, where they can get medication, where staffs are more hospitable, those close to caregivers," but "… lack of funding for sensitization campaign does limit our reach to the caregivers creating a problem on identifying them". Another one argued that "… some prefer facilities with minimum to zero charges for services, those nearer to them, public facilities" and further added that " the facility has enough staff, engage the community outreaches programs, but most caregivers tend to ignore the severity of the medical issues." The issue of drug stock out was further stressed with one interviewee saying "… caregiver prefer to not go to hospitals since they would not be given any medication even after having been seen". FGDs with community health volunteers generated the following themes: caregivers' ignorance, facility accessibility, drug unavailability, poor staff motivation, and care costs. One of the FGD participants highlighted that "… in some cases and just disturbed by the level of ignorance among caregivers not taking fever seriously in a region prone with malaria cases…." with a fellow participant adding, "… sometimes you feel for them, especially after knowing how bad the situation can be praying that it does not turn out to be malaria." In terms of facility accessibility, another FGD participants noted "… from my interaction, I think the long distance they have to cover to the facility has demotivated them to seek the health care services." Participants agreed that the county government should do more to motivate the staff, emphasizing the im-portance of fever care and "… authorities to recognize the significant role CHVs can have in primary care." The social economic activity for most of the people in the region is farming, " cost of care lowers health seeking behavior due to poverty" and that "… creates a problem to dispensaries and level three facilities not charging by having long waiting time and drug stock out." Discussion The aim of this study was to determine the determine the proportion of appropriate health-seeking behaviour for febrile illness among caregivers of children under five years and the individual factors and health-system factors associated with health-seeking behaviors among their caregivers in Butula sub-county, Busia County. The findings have revealed that the proportion of HSB is high in this community. This finding is consistent with the survey findings by KMIS who reported a prevalence of 67% 15 and Liyew et al. 11 who found a prevalence of 67.3% in Malawi. These is significantly higher than reported in O’Meara et al. 4 at 22.4% but slightly lower than reported by Nyaoke et al. 16 at 84%. The study findings have revealed that individual factors influencing appropriate health-seeking behavior are unemployment, preferring self-medication over professional care, having two or more children, and confidence in identifying fever. Caregivers who reported not being unemployed were 0.49 less likely to seek appropriate healthcare. These findings support Chauhan et al. 17 , who reported that occupation influences health-seeking behavior. Caregivers who could identify fever signs were seven times more likely to Have appropriate HSB. This is consistent with Zenebe et al. 9 , who reported that the proportion of mothers with good knowledge of childhood illnesses was around two times more likely to seek healthcare than those with poor knowledge. Similarly, Guntur et al. 7 reported that understanding fever signs positively influences health-seeking behavior. It is unclear why these factors significantly influenced HSB and not the others. Unemployment might affect caregiver’s financial ability to support HSB. Caregivers who prefer not to give their children medication for febrile illness on their own may have poor HSB. Having more than two children and confidence in identifying fever may give caregivers the perceptions that they have the experience to care for their sick children without professional care. However, further enquiry is needed to support these assertions. Additionally, the findings revealed that factors that significantly influence HSBs included staff offered health education, distance to facility, care costs, and facility type or level. There was a high probability that caregivers would engage in HSB when they received health education about febrile diseases from the facility staff. This could be explained by perceived higher confidence levels among caregivers in the nursing staff's knowledge or increased levels of trust for them to provide accurate information during the education. The healthcare facility, being level three, increased HSB twofold. Level three hospitals provide more services and have better medical equipment and other resources. According to Getahum 18 , a lack of adequate resources is a factor determining HSB. Medication stockouts, personnel shortages, and insufficient funding for community outreach programs were associated with poor HSB 18 . Our results are consistent with previous research. According to Muriithi et al. 19 , even when a patient desired medical attention, getting there would be difficult due to the distance to the institution. Kamat 20 made similar observations that caregivers tend to choose nearby institutions when seeking medical attention, mostly due to the reduced cost of treatment. Ease of access to healthcare facilities is a motivator to HSB 20 . Demand for healthcare reduces as the distance to the facility increases due to limited means of transportation and lack of transport fees 19 . Health institutions that promoted illness education to moms were favored above those that did not 18 . Oladigbolu et al. 21 also revealed that knowledge, medicine availability, and the high cost of care were some factors limiting HSB. Limitations A major limitation of the study is that it was cross-sectional and therefore limited to a specific period. It does not allow for further analysis of how HSB changes over time. Another limitation was that the investigators did not confirm whether caregivers’ health seeking behavior was within the indicated time and therefore relying on sincerity of the caregivers. There was potential for data collection bias. Conclusion In conclusion, the proportion of HSB in the investigated community was high. Individual factors significantly associated with appropriate health-seeking behavior were occupation, how they address fever, ability to identify fever in under-five, and number of children per household. Furthermore, we concluded that health system factors associated with HSB include facility being too far, staff workload, drug stock out, inadequate sensitization due to shortage of community outreach program, facility type, and health education. Interventions or initiatives to improve HSBs should consider targeting these factors if efforts to address febrile illnesses is realized in this community and similar neighboring communities. Further research should focus on longitudinal prospective cohort studies to help establish a causal-effect relationship between HSBs and individual or health-system factors, targeting Busia County in its entirety to understand the problem at the county level, possibly opening room for further investigations at the national level. Ethics and consent The Ethics Review Board at Jomo Kenyatta University of Agriculture and Technology approved the study (ref: JKU/ISERC/02316/0834) on March 9, 2023. Busia County government approved Butula Sub-County as the study site. Study participants signed the informed consent form before enrolling in the study. Privacy and confidentiality were maintained by omitting personal identifiers and storing the data in password-protected devices. 1. Please clarify if it's semi-structured questionnaire or semi-structured interview schedule? "The respondent fills out the questionnaire on their own. A trained interviewer asks questions and records the responses. Schedules are often used to record activities or events, such as a day's schedule." 2. Please provide details of quotes (anonymous) to bring more weight to the response. 3. "Health-seeking behavior (HSB) is the actions people take when they perceive themselves to be ill or have a health problem".The conclusion says - the proportion of HSB was high. What does it mean? Does it mean correct HSB?? Is the study design appropriate and does the work have academic merit? Partly Is the work clearly and accurately presented and does it cite the current literature? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? No Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Community based research; qualitative; quantitatie research; health system research I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. This study explores the patterns and determinants of health-seeking behavior for febrile illnesses among caregivers of children under five in Butula Sub-County, addressing a critical gap in child health outcomes in rural Kenya. Its relevance is grounded in the high prevalence of febrile illnesses, such as malaria, which significantly contribute to morbidity and mortality in this vulnerable population. However, please the author should kindly check the prevalence used in calculating the sampling size in the study, p used was 0.20 instead of 0.17 according to the malaria survey (KDHS,2022). Also, the study design adopted should be cross-sectional, explanatory sequential mixed-methods. The study provides valuable insights into a specific rural setting, offering context-specific data that can inform targeted interventions. By examining both individual and systemic factors, the study captures the multifaceted nature of health-seeking behavior. Findings emphasize the need for community-level education and health system strengthening to improve timely care-seeking. The determinants highlighted from the study were only individual factors and health system factors other contextual factors were not succinctly elucidated. In conclusion , the article is well written and fills the research gap in its study context, however noting the corrections will enhance the credibility of the work. Is the study design appropriate and does the work have academic merit? Partly Is the work clearly and accurately presented and does it cite the current literature? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes Are sufficient details of methods and analysis provided to allow replication by others? Yes Reviewer Expertise: Health Systems and Services; Health behavior research; and Health policy I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. All comments have been addressed in the manuscript version 2 Reviewer’s comments The proportion and determinants of appropriate health seeking behavior for febrile illness among caregivers of children under-five years in Butula sub-county, Busia county, Kenya Authors mentioned use of questionnaire method but considered interviews with community and stakeholders. Please check if it’s questionnaire or interview schedule used for the enquiry. Were qualitative findings congruent with survey findings? “… quantitative data from 271 caregivers between May and July 2023 using a semi-structured questionnaire”. How only quantitative data generated using semi-structured interviews? Participant characteristics did not bi-furcate providers and caregivers. Was the proportion of febrile illnesses considering all the seasons? In Figure-3 kindly differentiate between perception and knowledge/awareness-based data. The findings from healthcare workers’ interviews need to be corroborated with the survey findings. The discussion should be made stronger and more precise. Currently it is mixed of discussion and conclusion. Conclusion part includes results. Make it more concise and specific. It also includes recommendations which are not mentioned anywhere in the result section. Is the study design appropriate and does the work have academic merit? Partly Is the work clearly and accurately presented and does it cite the current literature? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? No Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Community based research; qualitative; quantitatie research; health system research I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. Answer1: I mentioned questionnaire for caregivers while the key informants interview were used for the key informants from the pediatric departments, and yes the qualitative findings were congruent with survey findings(and it has been addressed in manuscript) Answer2: Semi-structured questionnaires were used on quantitative data collection and semi-structured interview were used on key informant (key informant interview),However it wasn’t specified in manuscript, but it now has been addressed. Answer3:The participants characteristics did not bi-furcate providers and caregivers because health care providers were not main part of the study , they were care giver. Health care provider served for gathering qualitative data in order to support the study Answer4: No. The study was conducted in a Kenyan winter season (June, July) This The study is standard cross section study , proportion of febrile illness data was collected by considering if they had a fever within previous two weeks Answer5: Depending on how questions were framed, all variables were perception based-data N.B:I have updated my conclusion and recommendation Reviewer’s report Title: Appropriate and relevant Authors name: Not uniformly written. All the three names of the first author were written fully while the other two authors’ second names were abbreviated. The authors should choose and stick to only one format for writing all their names. Abstract: Background and methods were well written. The 70.1% (The proportion of caregivers that reported seeking care for a child's fever within 24 hours) under results should be written in words since sentences are not started with figures. The conclusion contained only the recommendation for improving the situation without a summary of the findings. The authors should summarize the findings of the study in one sentence before giving the recommendation. Introduction: Well written and captured the background information, the statement of the problem, the justification for the study and the objectives of the study. Methods - Study setting: Line 1- ‘Its one seven sub-counties in Busia County…’ should be corrected to It is one of the seven … Line 4 – ‘ It border the Republic of Uganda…’ should be corrected to It borders the …. Exclusion criteria: Authors should note that exclusion criteria are not the opposite of inclusion criteria rather they are the characteristics found in those included in the study that will lead to their being disqualified from participating in the study. The authors should correct the exclusion criteria appropriately and also provide their explanation for excluding care-givers who had lived in Busia County for less than two weeks before the study period from the study. Sampling: The authors should clarify whether the three out of the six wards in Butula sub-county were selected at random or using simple random sampling technique. Sample size determination: p should be the proportion or prevalence obtained for the outcome variable (ie proportion of appropriate health seeking behavior for febrile illness in this instance) in a previous study rather than the prevalence of an attribute/disease (ie fever) in the population from previous studies used by authors. The authors should note this for subsequent studies. The authors mentioned that they used the prevalence of 17% of children below five with a fever in the two weeks before the malaria survey (KDHS, 2022) for determining the sample size. But in the formula, p was equal to 0.20 instead of 0.17. The authors should provide an explanation for this. Data collection and tools Paragraph 2 line 5 – ‘Tape recordings were used capture …’ should be corrected to Tape recordings were used to capture …. Data management and analysis: Line 1 – ‘Quantitative data was cleaed in Excel ..’ should be corrected to Quantitative data was cleaned in Excel… Line 4 - A binominal logistic regression analysis was used …… and p-value less than 0.1. The authors should correct this or provide an explanation for why they used p-value less than 0.1 here to denote statistical significance instead of the desired level of precision at 0.05 that they mentioned in their sample size determination. Results: The authors mentioned that 14 households had no one available during the survey. They should provide an explanation on what they did with those 14 households. Table 1: Demographic characteristics of the respondents The authors should explain why some of the figures were highlighted or written in bold in this table. For example – 44.6% for secondary education, 59% for self- employed, etc 16 (5.9%) of the respondents did not have any child less than 5 years, yet they were included in this study when the inclusion criterium was care-givers over 18 years with a child under five. The authors should provide an explanation for this. Discussion: Line 1 – ‘The aim of this study was to determine the prevalence of febrile illness among children under five years…’ should be corrected to … to determine the proportion of appropriate health-seeking behaviour for febrile illness among children under five years …. Conclusion: Well written and drawn from the findings from the study. Reviewer’s impression – The article is well written and indexable. However, the authors need to address the issues raised above to make it more scientific. Is the study design appropriate and does the work have academic merit? Yes Is the work clearly and accurately presented and does it cite the current literature? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes Are sufficient details of methods and analysis provided to allow replication by others? Yes Reviewer Expertise: Public health, Health Systems Management I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. All comments are addressed in version 2(new version)