Timely treatment-seeking for diarrheal diseases among caregivers with under-five children in public hospitals of Sidama region, Ethiopia: Unmatched case-control study

Abstract

Introduction:

A significant proportion of under-five diarrhea-related morbidity and mortality can be reduced by timely initiation of treatment. This study assessed determinants of timely treatment-seeking behavior for diarrheal disease among caregivers with under-five children in public hospitals in Sidama region, Ethiopia.

Methods:

A facility-based unmatched case-control study was conducted among public hospitals. Data were collected using KOBO Collect with a smartphone and analyzed using SPSS 26.

Results:

Three hundred two cases and 302 controls were included. Being urban resident (AOR = 0.251 (95% CI, 0.157, 0.401, p = 0.000)), mothers with no formal education (AOR = 0.397 (95% CI, 0.222, 0.713, p = 0.002)), child age < 24 months (AOR = 0.210 (95% CI, 0.121, 0.364, p = 0.000)), above 120 min walking distance from nearby health facility [AOR = 0.426 (95% CI, 0.211, 0.861, p = 0.017)] were negatively associated. Whereas, reported cost of treatment easy to pay (AOR = 7.988 (95% CI, 3.734, 17.091, p = 0.000)), community-based health insurance (CBHI) membership (AOR = 4.940 (95% CI, 3.124, 7.812, p = 0.000)), and a history of previous diarrhea (AOR = 1.702 (95% CI, 1.021, 2.837, p = 0.041)) were positively associated with timely treatment seeking behavior.

Conclusion:

Being urban residents, caregivers’ educational status, child age <24 months, and long walking distance from nearby health facility, low-cost of treatment, being a CBHI member, and a child with a history of previous diarrhea were independently associated with timely treatment-seeking. Therefore, it is important to design strategies to improve these factors by involving relevant stakeholders. Future researchers should consider strong designs like prospective cohort involving multiple health facilities to identify causal factors for timely treatment-seeking behavior.

Introduction

Diarrhea is a passage of three or more watery or loose stools per day with or without accompanying signs and symptoms, such as vomiting, nausea, fever, or abdominal pain. ¹ There are three types of diarrhea: acute watery diarrhea, which lasts several hours to days; and persistent diarrhea that lasts 14 days or longer. ² Diarrheal disease occurs in nearly 1.7 billion children and kills approximately 525,000 under-five children annually. ¹ It is the second most common cause of mortality for children under-five worldwide, accounting for 1 in 9 deaths.^3,4 The majority of these deaths were among under-five children living in low- and middle-income countries, with the highest prevalence in sub-Saharan Africa, with 76 fatalities for every 1000 live births. ⁵ Ethiopia, with a prevalence that ranges from 23% to 31.7%, ⁶ accounts for 8% of under-five mortality.^7,8

Deaths from diarrhea are largely preventable if adequate treatment is provided timely in the course of the illness. Timely treatment-seeking is defined as seeking care or treatment within 24 h from recognizing signs and symptoms of diarrhea in under-five children. ⁹ Timely initiation for treatment of diarrhea is a quality indicator that reduces diarrhea-associated morbidity and mortality by preventing dehydration. ¹⁰ The Ethiopian government initiated the Health Extension Program (HEP) ³ to deliver preventive education, health promotion (health-seeking behaviors), and curative service at the health post through the Integrated Community Case Management (ICCM) program. Ethiopia has made significant progress in reducing under-five child mortality, with 67 under-five children dying per 1000 live births, and 480 children die every day from easily preventable diseases in 2018. ¹¹ Despite this, diarrhea continues to threaten the well-being of children below 5 years of age and contributes to substantial morbidity and mortality in developing countries like Ethiopia. ¹² In addition, the low rate of immediate treatment within 24 h, an inability to identify situations that endanger life, and poor care-seeking behaviors caused delayed treatment-seeking of the caregivers. This delay may affect children's health and result in complications that make medical treatment less safe and ineffective. ⁴

Factors for diarrheal disease occurrence are multifactorial, and it is frequently documented that diarrheal disease is linked to several socio-demographic, environmental, and behavioral factors (8) (Figure 1). The documented risk factors for diarrhea include younger age, malnutrition, early weaning, seasonal patterns, low maternal education, lack of piped water supply, poor water-storage practices, lack of vigilant hand washing, poor sanitation and not treating water in the home type of occupation, household income level, age of the child, history of vaccination and waste disposal systems.^4,8,13,14 In addition to this, the incapacity to recognize life-threatening circumstances and poor care-seeking habits resulted in unnecessary delays in treatment-seeking, which could contribute to worsening of disease conditions and mortality. Having up-to-date evidence for the factors affecting the timely treatment-seeking behavior of mothers of under-five children with diarrheal diseases is essential for planning intervention strategies, improving treatment compliance, and evaluating healthcare services in the study area. Therefore, this study aimed to identify determinants of timely treatment-seeking for diarrheal diseases among caregivers with under-five children attending public general hospitals in the Sidama Region.

Figure 1.

Conceptual framework of determinants of timely treatment-seeking for diarrheal disease among under-five children paired with their caregivers. Adapted by authors.

Methods and materials

Study design, area, and study period

A facility-based unmatched case-control study was used from April 1 to May 31, 2024. Sidama Regional State is one of 13 regions of Ethiopia. Its state headquarters is Hawassa, which is located 272 km to the southeast of Addis Ababa, the capital city of Ethiopia. The region has a total area of 6981.8 square kilometers and is administratively divided into four zones and six town administrations. According to the 2018 estimate, the region has a total population of 4,294,730 (49% females and 51% males). Currently, the region has 26 hospitals providing service for the population (1 comprehensive, 5 general and 20 primary hospitals). Among the general hospitals, three general hospitals (Bona, Yirgalem, and Daye) were included in this study. Bona Hospital is located 486 km Southwest of Addis Ababa and 112 km Southwest of Hawassa in Bona Zuria woreda. The Hospital was upgraded from a health center to a district Hospital in 1999 E.C. The Hospital provides service for about 1.6 million people, including the southern part of the Oromia regional state. Which gives services for under-five children at 2 regular outpatient departments (OPD) and one ear-nose and throat) (ENT), One pediatric ward with one pediatrician, four General practitioners, four pediatric nurses, and six clinical nurses. The Yirgalem General Hospital is one of the oldest hospitals in the region, which is located 45 km from Hawassa city. It provides services for 2.5 million. The other hospital is Daye General Hospital, which is located 131 km from Hawassa, and it provides service for under-five children with one regular under-five OPD, one ENT & one pediatric ward.

Source and study population

The source population was all under-five children with diarrhea attending selected general hospitals. All children under the age of 5 years, paired with caregivers who sought health facilities due to diarrhea within the study area during the data collection period, were the study population. Cases were under-five children paired with caregivers, and with signs/symptoms of diarrhea who sought treatment within 24 h of the recognition of diarrhea. Controls were under-five children paired with caregivers, and with signs/symptoms of diarrhea who sought treatment after 24 h of the recognition of diarrhea.

Inclusion and exclusion criteria

All children under five-children paired with their caregivers who sought selected health facilities due to diarrhea during the data collection period, were included. However, children whose caregivers did not mention the exact dates of onset of diarrhea, and children admitted for other diagnoses and who developed a new onset of diarrhea at the ward were excluded.

Sample size determination and sampling procedures

Sample size determination

The sample size was calculated using Epi-info software for an unmatched case-control study. It was calculated by considering significant determinant factors from previous studies (maternal educational status, child’s age in months, perceived diarrhea cure-rate without treatment & sex of child). Finally, the highest sample size will be taken. During sample size calculation, assumptions of 80% power, 95% confidence level, 1:1 case-to-control ratio and 5% margin of error, and 10% nonresponse rate were used. By considering the percent of control to exposure 48.9% for children below 24 months, AOR = 1.64, and percent of the cases to exposure 61.7%, the estimated sample size was 610 (305 cases and 305 controls). ¹⁵ Similarly, considering the percent of control to exposure 21% for the perceived diarrhea cure rate without treatment, AOR = 2.1, and percent of the cases to exposure 37%, the estimated sample size was 276 (138 cases and 138 controls). ⁴ By considering the percent of control to exposure, 19.5% for mothers with no formal education, AOR = 4.61, the percent of the case to exposure was 25.8%, and the estimated sample size was 78 (39 cases and 39 controls). ¹⁵ By considering the percent of control to exposure 45% for being female children, AOR = 1.93, and percent of the cases to exposure 64.5%, the estimated sample size was 350 (175 cases and 175 controls). ⁹ The final sample size using the age of children < 24 months as a determinant factor with an AOR of 1.64 after considering a 10% non-response rate was 610 (305 Cases and 305 controls).

Sampling technique

Among five general hospitals in the Sidama region, three public hospitals (Bona, Daye and Yirgalem) were selected by the simple random sampling technique using a lottery method. In the year preceding the study, 354, 242, and 392 under-five diarrhea cases were reported in a two-month period in Bona, Daye, and Yirgalem general hospitals respectively. During the same duration, about 988 under-five children visited each hospital for health services (diarrhea being a reason for 33.4% of under-five health facility visits). Depending on this number, a proportional allocation of respondents was done to each facility. Then, 218, 150, and 242 cases and controls from Bona General Hospital, Daye General Hospital, and Yirgalem General Hospital were included. Concerning the sampling interval, every third (i.e., the average of 2.7, 4.08, 2.52) cases were included. The first participant of the case group (caregivers with children visiting the hospital within 24 hours of the onset of diarrheal illness) was randomly selected by the lottery method. Similarly, third controls (caregivers with children visiting the hospital after 24 h of the onset of diarrheal illness) were included. The first participant of the control group was randomly selected by the lottery method. When caregivers complained of diarrhea in their child and completed their consultation with a health care professional, they were moved to a private room for an interview until the total required sample size was obtained.

Study variables

Dependent variable

Timely treatment-seeking

Independent variables

Sociodemographic factors: (age of caregivers, marital status, religion, residence, educational status, occupation, age of child, and sex of child. Contributing factors: monthly income, treatment costs, health insurance membership, distance to the nearest health facility, and preferred nearby health facility. Disease-related factors: diarrhea symptoms, dehydration status, history of diarrhea in the last 6 months, and type of diarrhea, treatments used at home (traditional medicine, self-medication). Health-system related factors include distance from the health facility, cost of treatment, health insurance membership, and counseling service.

Data collection methods and tools

Data was collected using a pre-tested, structured questionnaire through face-to-face interviews that was adapted from previous literature. The data were collected using KOBO Collect with a smartphone. Appropriate confidentiality (restricting server access to authorized personnel, protecting all KOBO collect passwords and using limited code information for creating tool box), and data availability and security details (daily backup of the database to a separate, remote location) were considered. The questionnaire contains four main parts: socio-demographic factors, enabling factors, disease-related factors, and health-system-related factors (Available in supplementary file questionnaire used in this study). Data was collected by face-to-face interviews with caregivers using structured questionnaires prepared for this purpose. We used a patient chart to check dehydration status. The interview was conducted in a private room to create an atmosphere of empathy and confidence within a secure environment. The patient satisfaction with the service provided in the health facility was measured by using 10 questions designed for this purpose. Each question has five responses on a Likert scale: 1 (very poor), 2 (poor), 3 (neutral), 4 (good), and 5 (very good). The mean value was calculated, and those who scored above the mean were considered to have good satisfaction, and those who scored below the mean were considered to have poor satisfaction (10). The data was collected by three BSc nurses with experience in data under the supervision of general practitioners at the pediatrics wards of respective hospitals.

Data processing and analysis

Data was exported from Kobo Collect to SPSS Software version 26 for further analysis. Descriptive statistics were used to report sociodemographic variables, contributing factors, disease factors, and health system-related factors. In addition, the Chi-square test was computed to see group variation for all variables. Bivariate and multivariate logistic regressions were computed to identify the presence and strength of associations. Variables with a p-value < 0.25 in binary logistic regression were taken to multiple logistic regression for further analysis. Before taking variables to multiple logistic regressions, multicollinearity was checked for all candidate variables using the collinearity diagnostic test in the linear regression model. The result of the collinearity diagnostic test showed that no multicollinearity, as the maximum Variance Inflation Factor (VIF) result was 3.73 (1.52 ± 0.82, ranging from 1.04 to 3.73). Odds ratio with 95% CI was computed, and variables having p-values less than 0.05 in the multiple logistic regression models were considered as significantly associated with the dependent variable (timely treatment-seeking behavior). Finally, the model was fitted using enter methods, and Hosmer-Lemeshow goodness of fit test was used to test the model fitness, which provided a p-value of 0.214 that indicates the model fitted well.

Data quality assurance

The questionnaire was adapted from previously conducted related studies, and some changes were made after reviewing relevant literature according to local context. The English version of the questionnaire is translated into Sidamu Afoo, a commonly used local language, for better understanding by the respondents. A pre-test of the questionnaire on 5% (15 cases and 15 controls) of the respondents was done in Leku General Hospital to know whether a respondent understood the questionnaire in the same way and to find out how long it takes to complete the survey in real-time. The result of the pre-test showed that, out of the total respondents, 14 (93%) among cases and 15 (100%) among controls responded to all the questions fully without missing any question, and only 1(6.6%) among cases responded incompletely. Again, the pre-test showed that 25–30 min were needed to complete one questionnaire and no question made discomfort to the respondent. After that, a two-day training for data collectors and supervisors was given to ensure the quality of data. Data collectors were supervised continuously by the supervisor, and they reviewed every questionnaire for completeness and logical consistency. The same interviewer was used to interview both cases and controls to control for information bias. A mark (√) was put on the medical chart of the interviewed patients to avoid repeated inclusion of study participants upon revisiting the health facility due to failure of recovery or experiencing another episode of diarrhea within the data collection period. After entry, data clearance and exploration were done to see outliers.

Operational definition

Timely treatment seeking: Different studies that evaluated treatment seeking time for under-five children with diarrhea revealed variable results ranging from 1- 30 days (median time of 2 days).^16,17 In this study, timely treatment seeking is defined as treatment that was sought from health facilities within 24 h from the recognition of the presence of diarrhea in under-five children, and otherwise delayed.

Caregivers: Mothers, fathers, or any person above 18 years of age who at the time of the study was directly responsible for the care of the child. ¹⁷

Results

Socio-demographic characteristics of study participants

In this study, 604 (302 cases and 302 controls) children, along with their caregivers, were included, providing a response rate of 99%. More than half, 189(62.6%) of caregivers with cases and 186 (61.6%) of controls were falls in the age group of 15–25 years. The mean age for caregivers with cases was 24.65, and the Standard Deviation (SD) of ±4.21, and for that of controls was 24.94 (±4.023 SD). Out of 604 children included in the study, 266 (88.1%) among cases and 171 (56.6%) in that of controls were in the age group of <24 months. The mean age for cases was 16.69 (±10.54 SD) and 25.10 (±13.07 SD) months for the controls. About 195(64.6%), respondents were female children among cases, and 127 (42.1%) were male among cases. The majority, 235 (77.8%) of cases and 125 (41.4%) of controls resided in a rural area (Table 1). Regarding contributing factors, 113 (37.4%) cases and 88 (29.1%) controls reported greater than 120 min walking distance to a preferred health facility. The current study also found that 148 (49%) of caregivers with cases and 204 (67.5%) of controls earn monthly income ≥ 2500 Ethiopian Birr (ETB). Similarly, the majority of 190 (62.9%) cases and 171 (56.6%) controls preferred to treat their child with diarrhea at the government. Three-fourths of cases, 221 (73.2%), and 114 (37.7%) controls were not members of community health insurance (Table 1).

Table 1.

Socio-demographic and contributing factors of caregivers with under five children attending public hospitals in Sidama Region, Ethiopia, 2024 (n = 604).

Variables		Patient category		Pearson Chi-Square (two-tailed)	p-value
		Cases (n = 302)	Controls (n =302)
Age of mothers/caregivers	15–25 years	189 (62.6%)	186 (61.6%)	0.974	0.614
	26–34 years	90 (29.8%)	98 (32.5%)
	≥35 years	23 (7.6%)	18 (6.0%)
Marital status of mothers/caregivers	Married	246 (81.4%)	250 (82.8%)
	Divorced	21 (7%)	15 (5%)	2.511	0.473
	Widowed	35 (11.6%)	37 (12.3%)
Residence area of mothers/caregivers	Urban	67 (22.2%)	177 (58.6%)
	Rural	235 (77.8%)	125 (41.4%)	83.2	0.000
Religion of caregivers	Orthodox	87 (28.8%)	81 (26.8%)
	Protestant	137 (45.4%)	156 (51.7%)
	Muslim	50 (16.6%)	44 (14.6%)	2.829	0.419
	Others	28 (9.3%)	21 (7.0%)
Occupation of caregivers	Housewife	158 (52.3%)	137 (45.4%)
	Merchant	78 (25.8%)	84 (27.8%)	3.249	0.355
	Gov’t employee	43 (14.2%)	53 (17.5%)
	Others	23 (7.6%)	28 (9.3%)
Average monthly income of caregivers	<=1250	90 (29.8%)	45 (14.9%)
	1250–2500	64 (21.2%)	53 (17.5%)	25.849	0.000
	>=2501	148 (49%)	204 (67.5%)
Educational status of caregivers	No formal education	39 (12.9%)	19 (6.3%)
	Primary	148 (49%)	110 (36.4%)	24.580	0.000
	Secondary	55 (18.2%)	76 (25.2%)
	College and above	60 (19.9%)	97(32.1%)
Age of child	<24 months	266 (88.1%)	171 (56.6%)
	≥ 24 months	36 (11.9%)	131 (43.4%)	74.694	0.000
Sex of child	Male	107 (35.4%)	175 (57.9%)
	Female	195 (64.6%)	127 (42.1%)	1.715	0.190
Birth order of the child	First	105 (34.8%)	118 (39.1%)
	Second or more	197 (65.2%)	184 (60.9%)	1.201	0.273
Contributing factors
Distance from nearby health facility	< 15min	37 (12.3%)	35 (11.6%)
	15–30min	38 (12.6%)	30 (9.9%)	9.039	0.060
	30–60min	71 (23.5%)	88 (29.1%)
	60–120min	43 (14.2%)	61 (20.2%)
	>120min	113 (37.4%)	88 (29.1%)
Cost of treatment	Easy to pay	156 (51.7%)	182 (60.3%)	52.391	0.000
	Difficult to pay	65 (21.5%)	103 (34.1%)
	Very Difficult to pay	81 (26.8%)	17 (5.6%)
Preferred place to treat a child with diarrhea	Government facility	190 (62.9%)	171(56.6%)
	Private clinic	46 (15.2%)	94 (31.1%)	25.622	0.000
	Traditional Healer	66 (21.9%)	37 (12.3%)
Community-based Health insurance member	Yes	81 (26.8%)	188 (62.3%)	76.737	0.000
	No	221 (73.2%)	114 (37.7%)

Disease and health-system related factors

According to this study, 119 (39.4%) of cases and 72 (23.8%) controls presented with watery diarrhea. Also, record review showed that 157 (52%) of cases and 203(67.2%) of controls were presented with severe dehydration. In addition to this, 50 (24.2%) cases and 41(19.8%) controls had fever in addition to diarrhea. Most of the caregivers 207, 68.5%) among cases and 216 (71.5%) among controls responded to diarrhea in the child by taking them to the health facility at the first episode. More than three-fourths of cases, 226 (74.8%), and controls 221(73.2%) had a previous history of diarrhea. The majority of respondents were satisfied with the service provided at the health facility, 259 (85.8%) in the case group and 275 (91.9%) controls. Additionally, this study also showed 22 (7.3%) cases and 19(6.3%) controls did not receive counseling on the importance of timely visiting of health facility for children with diarrhea (Table 2).

Table 2.

Disease-related factors among caregivers with under five children attending public Hospitals in Sidama Region, Ethiopia, 2024 (n = 604).

Variables		Patient category		Pearson Chi-Square (two-tailed)	p-value
		Cases (n = 302)	Controls (n = 302)
Status of dehydration	No dehydration	54 (17.9%)	65 (21.5%)	32.887	0.000
	Some dehydration	91 (30.1%)	34 (11.3%)
	Severe dehydration	157 (52%)	203 (67.2%)
Type of diarrhea	Watery	119 (39.4%)	72 (23.8%)	21.768	0.000
	Bloody	68 (22.5%)	62 (20.5%)
	Mucoid	115 (38.1%)	168 (55.6%)
First Response to diarrhea	Take to a health facility	207 (68.5%)	216 (71.5%)
	Take to a traditional healer	41 (13.6%)	39 (12.9%)	0.727	0.695
	Treat with drug from a pharmacy	54 (17.9%)	47 (15.6%)
Who decides first	Mother	71 (23.5%)	107 (35.4%)
	Father	59 (19.5%)	79 (26.2%)	21.946	0.000
	Both	117 (38.7%)	85 (28.1%)
	Grand Parents	55 (18.2%)	31(10.3%)
Previous history of diarrhea	Yes	226 (74.8%)	221(73.2%)	0.215	0.643
	No	76 (25.2%)	81(26.8%)
Satisfied with the service given	Yes	259 (85.8%)	275 (91.9%)	0.894	0.345
	No	43 (14.2%)	27(8.9%)
Counseled on the importance of timely visits	Yes	280 (92.7%)	283 (93.7%)	0.235	0.627
	No	22 (7.3%)	19 (6.3%)

Concerning severity signs and symptoms caregivers can identify, 54 (17.9%) of cases and 83 (27.5%) of controls reported change in child behavior, 96 (31.8%) of cases and 74 (24.5%) of controls reported increased frequency and duration of diarrhea, 68 (22.5%) of cases and 67 (22.2%) of controls reported presence of fever, 60 (19.9%) of cases and 49 (16.2%) of controls reported presence of bloody diarrhea, and 24 (7.9%) of cases and 29 (9.6%) of controls reported presence of sunken fontanel as signs and symptoms of severe diarrhea (Figure 2).

Figure 2.

Severity signs and symptoms of diarrhea caregivers can identify among caregivers with under-five children in public hospitals of Sidama region (n = 604).

Determinants of timely treatment Seeking behavior

About determinants of timely treatment-seeking behavior of mothers, binary logistic regression revealed urban residence (COR = 4.06 (95% CI, 2.47, 6.695, p = 0.000)), being merchant caregiver (COR=2.102 (95% CI, 0.904, 4.887, p = 0.084)), and being government employed caregiver (COR = 1.320 (95% CI, 0.506, 3.446, p = 0.57)) were associated with timely treatment seeking for under-five child with diarrhea. Similarly, having average monthly income 1250 ETB or less (COR = 2.525 (95% CI, 1.367, 4.463, p = 0.003)), average monthly income 1250–2500, (COR = 1.654 (95% CI, 0.939, 2.912, p = 0.081)), no formal education of caregivers (COR = 2.993 (95% CI, 1.224, 7.317, p = 0.016)), having a child less than 24 months old (COR = 6.03 (95% CI, 3.348, 10.863, p = 0.000)), being female child (COR = 0.605 (95% CI, 0.371, 0.986, p = 0.044)), caregivers from a walking distance of 15–30 min from a health facility (COR = 3.505 (95% CI, 1.534, 8.005, p = 0.003)), 30–60 min walking distance from a health facility (COR = 0.497 (95% CI, 0.230, 1.075, p = 0.076)), 60–120 min walking distance from a health facility (COR = 2.474 (95% CI, 1.245, 4.915, p = 0.010)), and above 120 min walking distance from a health facility [COR = 2.705 (95% CI, 1.263, 5.792, p = 0.010)] were associated with timely treatment seeking for an under-five child with diarrhea. In addition, cost of treatment easy to pay (COR = 0.1 (95% CI, 0.044, 0.225, p = 0.000)), cost of treatment difficult to pay (COR = 0.158 (95% CI, 0.063, 0.396, p = 0.000)), being member of CBHI (COR = 0.189 (95% CI, 0.117, 0.305, p = 0.000)), and having history of previous diarrhea (COR = 0.609 (95% CI, 0.358, 1.033, p = 0.066)) were associated with timely treatment seeking for under-five child with diarrhea (Table 3).

Table 3.

Factors associated with timely treatment-seeking for under-five children with diarrhea among caregivers with under-five children attending public Hospitals in Sidama, Ethiopia (n = 604).

Variables		Patient category		COR (95%CI)	AOR (95%CI)	P-Value
		Cases (n = 302)	Controls (n = 302)
Residence area of caregivers	Urban	67 (22.2)	177 (58.6)	4.066 (2.47, 6.695)	0.251(0.157, 0.401)	0.000*
	Rural	235 (77.8)	125 (41.4)		1
Occupation of caregivers	Housewife	158 (52.3)	137 (45.4)		1
	Merchant	78 (25.8)	84 (27.8)	2.102 (0.904, 4.887)	0.496 (0.222, 1.108)	0.087
	Gov’t employee	43 (14.2)	53 (17.5)	1.320 (0.506, 3.446)	0.779 (0.311, 1.954)	0.595
	Others	23 (7.6)	28 (9.3)	1.575 (.638, 3.887)	0.625 (0.269, 1.450)	0.274
Average Monthly income of caregivers	≤1250	90 (29.8)	45 (14.9)	2.525 (1.367, 4.663)	0.397 (0.222, 0.713)	0.002
	1250–2500	64 (21.2)	53 (17.5)	1.654 (0.93, 2.912)	0.622 (0.363, 1.068)	0.085
	≥2501	148 (49)	204 (67.5)		1
Educational status of caregiver	No formal education	39 (12.9)	19 (6.3)	2.993 (1.224, 7.317)	0.291 (0.127, 0.67)	0.004*
	Primary	148 (49)	110 (36.4)	1.870 (1.060, 3.300)	0.532 (0.306, 0.93)	0.025*
	Secondary	55 (18.2)	76 (25.2)	0.832 (.427, 1.624)	1.257 (0.66, 2.38)	0.484
	College and above	60 (19.9)	97 (32.1)		1
Age of child	<24 months	266 (88.1)	171 (56.6)	6.030 (3.348, 10.86)	0.210 (0.12, 0.364)	0.000*
	≥ 24 months	36 (11.9)	131 (43.4)		1
Sex of child	Male	159 (52.6)	175 (57.9)		1
	Female	143 (65.2)	127 (42.1)	0.605 (0.371, 0.986)	1.562 (0.98, 2.498)	0.062
Walking distance from the nearby health facility	< 15 min	37 (12.3)	35 (11.6)		1
	15–30 min	38 (12.6)	30 (9.9)	3.505 (1.534, 8.01)	1
	30–60 min	71 (23.5)	88 (29.1)	0.497 (0.230, 1.075)	1.978 (0.953, 4.10)	0.067
	60–120 min	43 (14.2)	61 (20.2)	2.474 (1.245, 4.92)	0.474 (0.245, 0.92)	0.027*
	> 120 min	113 (37.4)	88 (29.1)	2.705 (1.263, 5.79)	0.426 (0.211, 0.86)	0.017*
Cost of treatment	Easy to pay	156 (51.7)	182 (60.3)	0.10 (.044, 0.225)	7.98 (3.734, 17.09)	0.000*
	Difficult to pay	65 (21.5)	103 (34.1)	0.158 (0.063, 0.396)	5.575 (2.359, 13.178)	0.000*
	Very difficult to pay	81(26.8)	17(5.6)		1
CBHI member	Yes	81(26.8)	188(62.3)	0.189 (.117, 0.305)	4.940 (3.124, 7.812)	0.000*
	No	221(73.2)	114(37.7)		1
Previous history of diarrhea	Yes	226(74.8)	221(73.2)	0.609 (0.358, 1.033)	1.702 (1.021, 2.837)	0.041
	No	76(25.2)	81(26.8)

The reference category is cases (delayed treatment), and controls (treatment instituted within 24 hours, no delay); * = statistically significant at p <0.05.

These variables were further subjected to multiple logistic regression to avoid multiple variables and control for confounding effects. Urban residents were 25% (AOR = 0.251 (95% CI, 0.157, 0.401, p = 0.000)) less likely to seek timely treatment for under-five children with diarrhea when compared to their counterparts. Caregivers with no formal education were 39.7% (AOR = 0.397 (95% CI, 0.222, 0.713, p = 0.002)) less likely to seek timely treatment for under-five children with diarrhea when compared to caregivers who attended college and above. Similarly, caregivers who attended primary school were 53.2% (AOR = 0.532 (95% CI, 0.306, 0.925, p = 0.025)) less likely to seek timely treatment for under-five children with diarrhea. Caregivers having children less than 24 months were 21% (AOR = 0.210 (95% CI, 0.121, 0.364, p = 0.000)) less likely to seek timely treatment for under-five children with diarrhea when compared to caregivers having children above 24 months. Caregivers from 60 to 120 min walking distance from nearby health facility distance from health facility were 47% (AOR = 0.474 (95% CI, 0.245, 0.918, p = 0.027)), and caregivers from above 120 min walking distance from nearby health facility were 42.6% (AOR = 0.426 (95% CI, 0.211, 0.861, p = 0.017)) less likely to seek timely treatment for under-five child with diarrhea when compared with caregivers in 15 min and less walking distance. Whereas, caregivers who reported the cost of treatment as easy to pay were 7. 8 times (AOR = 7.988 (95% CI, 3.734, 17.091, p = 0.000)), and caregivers who reported the cost of treatment as difficult were 5.5 times (AOR = 5.575 (95% CI, 2.359, 13.178, p = 0.000)) more likely to seek timely treatment for under-five child with diarrhea when compared with caregivers who report treatment cost as very difficult to pay. Similarly, caregivers with CBHI membership were 4.9 times (AOR = 4.940 (95% CI, 3.124, 7.812, p = 0.000)) more likely to seek timely treatment for under-five children with diarrhea when compared with their counterparts. Caregivers of children with a history of previous diarrhea were 1.7 times (AOR = 1.702 (95% CI, 1.021, 2.837, p = 0.041)) more likely to seek timely treatment for under-five children with diarrhea when compared with their counterparts (Table 3).

Discussion

This study revealed, determinants of early treatment seeking for diarrheal diseases among caregivers with under-five children attending selected public hospitals of Sidama Region by involving 604 (302 cases and 302 controls) participants. Being rural residents, maternal educational status, having a child less than 24 months, monthly income, distance from the nearest health facility, cost of treatment, being a community-based health insurance member, and previous history of diarrheal disease were independently associated with timely treatment-seeking for an under-five child with diarrhea. Urban residents were 25% (AOR = 0.251 (95% CI, 0.157, 0.401, p = 0.000)) less likely to seek timely treatment for under-five children with diarrhea when compared to their counterparts. However, the findings from Kenya, ¹⁸ Meta-analysis from Ethiopia, ¹⁹ and a case-control study from southwest Ethiopia, ⁴ which states, children from rural areas had higher odds of delay being taken to health facilities for medical care for diarrheal diseases than children from urban areas. In addition, a study from Gambia, ²⁰ central Ethiopia, ¹⁷ and northwest Ethiopia ¹⁵ did not show an association between residence and timely treatment-seeking among caregivers with under-five children for diarrheal disease. The possible explanation could be that rural residents might have better information about diarrheal illness through health extension workers in the study area. Since management of diarrhea illness is one of the 18 health extension service packages which are organized under 4 major areas (hygiene and environmental sanitation, disease prevention and control, family health services, and health education and communication). ²¹

The other determinant of timely treatment-seeking behavior was the age of the child. Caregivers having children less than 24 months were 21% (AOR = 0.210 (95% CI, 0.121, 0.364, p = 0.000)) less likely to seek timely treatment for under-five children with diarrhea when compared to caregivers having children above 24 months. This is in parallel with studies from India, ²² Central Ethiopia, ¹⁷ Arba Minch Southern Ethiopia, ⁹ and Northwest Ethiopia, ¹⁵ which suggest the caregivers of younger children (<24 months) were more likely to delay than those with older (>24 months) children. Similarly study from China showed that the age of children was negatively associated with treatment seeking in such a way that older children were protected against late treatment seeking for diarrheal disease. ²³ Again, the global reports on the trend in care-seeking and access to health service utilization justify our findings. It is stated that older children are more likely to seek early health care than young children. ²² Also, this finding implies that the implementation of government programs aimed at reaching children under 2 years was not met. According to the strategic approach, the first two years were the honeymoon period for children's development. Thus, strategic approaches like treating children with diarrhea with oral rehydration salt (ORS) and zinc early, treating pneumonia with antibiotics, and under two years Growth Monitoring and Promotion (GMP) were among some strategies to tackle the problem related to that disease if treatment was sought early.

In this study, caregivers’ education is associated with the timely treatment-seeking behavior of mothers. Caregivers with no formal education were 39.7% (AOR = 0.397 (95% CI, 0.222, 0.713, p = 0.002)), and caregivers who attended primary school were 53.2% (AOR = 0.532 (95% CI, 0.306, 0.925, p = 0.025)) less likely to seek timely treatment for under-five child with diarrhea. This is consistent with studies from central Ethiopia,^17,24,25 Yemen, ²⁶ and Chad. ²⁷ Multi-level analysis from developing countries revealed that a higher level of maternal education is positively associated with treatment-seeking behavior of mothers, whereas poverty is negatively associated with treatment-seeking behavior of caregivers with under five children. ²⁸ A similar study from Zimbabwe showed that higher wealth quantile is positively associated with seeking for recent diarrhea episode among children less than 6 years. ²⁹ This variation might be due to, caregivers who attended school are thought to have a better opportunity to learn health information than those who did not attend school. Moreover, illiterate mothers may not have basic knowledge of the impacts of the potential risk of delay in seeking timely treatment. In addition, mothers or caregivers with better education could have better job opportunities and can afford their seek child care easily.

Distance from the health facility is also associated with timely treatment-seeking. Caregivers from 60 to 120 min walking distance from nearby health facility distance from health facility were 47% (AOR = 0.474 (95% CI, 0.245, 0.918, p = 0.027)), and caregivers from above 120 min walking distance from nearby health facility were 42.6% (AOR = 0.426 (95% CI, 0.211, 0.861, p = 0.017)) less likely to seek timely treatment for under-five child with diarrhea when compared with caregivers in 15 min and less walking distance. This is supported by evidence from a study conducted in Nigeria that showed a statistically significant association between health-care-seeking behavior and respondents' level of education, occupation, monthly income, as well as distance taken to reach a health facility within 30 min. ³⁰ A similar study conducted among mothers with under five children in Chad showed that mothers who reported that distance to the health facility was a barrier were less likely to seek healthcare for childhood illnesses, compared to those who faced no geographical barrier to healthcare access. ²⁷ A modeling study conducted in Sub-Saharan Africa also showed a negative association between distance from a health facility and healthcare-seeking behaviors of mothers with children. ³¹ This is because geographic access is one of four dimensions of access to healthcare (availability, affordability, accessibility, and acceptability). ³² Therefore, improving geographical access to health facilities is important to enhance the early treatment-seeking behavior of caregivers with under-five children.

In this study, caregivers who reported the cost of treatment as easy to pay were 7. 8 times (AOR = 7.988 (95% CI, 3.734, 17.091, p = 0.000)), and caregivers who reported the cost of treatment as difficult were 5.5 times (AOR = 5.575 (95% CI, 2.359, 13.178, p = 0.000)) more likely to seek timely treatment for under-five child with diarrhea when compared with caregivers who report treatment cost as very difficult to pay. Similarly, caregivers with CBHI membership were 4.9 times (AOR = 4.940 (95% CI, 3.124, 7.812, p = 0.000)) more likely to seek timely treatment for under-five children with diarrhea when compared with their counterparts. This is consistent with the study done in northwest Ethiopia. ¹⁵ The recent systematic review and meta-analysis conducted in Ethiopia revealed a positive association between the income of mothers and healthcare-seeking behaviors. ³³ This is because improving the financial protection of households through health insurance improves access and utilization of health services by reducing point-of-care costs to members. This is also supported by a study conducted in the East Gojjam Zone that revealed a positive association between the health-seeking behavior of mothers for childhood illness and community-based health insurance membership. ³⁴ Therefore, it is important to design strategies that encourage health insurance coverage to improve the healthcare-seeking behavior of mothers with under-five children.

Caregivers of children with no history of previous diarrhea were 1.7 times (AOR = 1.702 (95% CI, 1.021, 2.837, p = 0.041)) more likely to seek timely treatment for under-five children with diarrhea when compared with their counterparts. A similar study conducted in Arba Minch, Ethiopia, showed a positive association between mothers who did not visit a health facility for first-episode diarrhea and treatment-seeking behavior. ⁹ This could be due to increased awareness created because of previous exposure to health facilities for the treatment of diarrheal illness. During facility visits for treatment of childhood diarrheal illness, mothers or caregivers receive education about disease prevention and treatment, which could enhance their knowledge and related treatment-seeking for their children.

Strengths and limitations of this study

The strengths of this study rely on its methodology (i.e., using an adequate sample size) to identify determinants of timely treatment-seeking behavior of mothers or caregivers. However, the limitations of the study were recall bias linked to the difficulty of remembering the onset of diarrhea and selection bias (due to the facility-based nature of the study). Being a facility-based study, the use of the information generated by this study beyond the source population should be used with caution. In addition, including children with diarrhea attending general hospitals, without including under-five children attending health centers, primary hospitals, and private health facilities, may also limit the generalizability of the findings.

Conclusion

This study revealed that being an urban resident, caregivers’ educational status, caregivers having children less than 24 months, and long walking distance from nearby health facilities were negatively and independently associated with the timely treatment-seeking behavior of mothers. However, caregivers reported the cost of treatment as easy to pay, being a member of community-based health insurance (CBHI), and caregivers of children with a history of previous diarrhea were positively and independently associated with the timely treatment-seeking behavior of mothers. Therefore, the regional health bureau and relevant stakeholders should design strategies like community awareness creation and increasing health insurance coverage to improve the timely health-seeking behaviors of mothers. Health professionals and health extension workers should design strategies to address urban caregivers. Regional health bureau and relevant stakeholders should design strategies including enhancing the education and employment status of caregivers, increasing CBHI membership coverage, building health facilities nearby, and raising awareness in the community about the importance of timely treatment-seeking for under-five children with diarrheal illness. For future researchers, it is imperative to consider more strong study designs like prospective cohort involving multiple health facilities to identify causal factors for the timely treatment-seeking behavior of mothers with under-five children having the diarrheal disease.