Corresponding Author Edwine W. Barasa, Child and Newborn Health Group, KEMRI-Wellcome Trust Research Programme, PO Box 43640-00100, Nairobi, Kenya. Tel.: +254 20 2715160; Fax: +254 20 2711673; E-mail: email@example.com
Objective To describe out-of-pocket costs of inpatient care for children under 5 years of age in district hospitals in Kenya.
Methods A total of 256 caretakers of admitted children were interviewed in 2-week surveys conducted in eight hospitals in four provinces in Kenya. Caretakers were asked to report care seeking behaviour and expenditure related to accessing inpatient care. Family socio-economic status was assessed through reported expenditure in the previous month.
Results Seventy eight percent of caretakers were required to pay user charges to access inpatient care for children. User charges (mean, US$ 8.1; 95% CI, 6.4–9.7) were 59% of total out-of-pocket costs, while transport costs (mean, US$ 4.9; 95% CI, 3.9–6.0) and medicine costs (mean, US$ 0.7; 95% CI, 0.5–1.0) were 36% and 5%, respectively. The mean total out-of-pocket cost per paediatric admission was US$ 14.1 (95% CI, 11.9–16.2). Out-of-pocket expenditures on health were catastrophic for 25.4% (95% CI, 18.4–33.3) of caretakers interviewed. Out-of-pocket expenditures were regressive, with a greater burden being experienced by households with lower socio-economic status.
Conclusion Despite a policy of user fee exemption for children under 5 years of age in Kenya, our findings show that high unofficial user fees are still charged in district hospitals. Financing mechanisms that will offer financial risk protection to children seeking care need to be developed to remove barriers to child survival.
Objectif: Décrire les frais de soins hospitaliers payés directement de la poche pour les enfants de moins de cinq ans dans les hôpitaux de district au Kenya.
Méthodes: 256 gardiens d’enfants admis à l’hôpital ont été interrogés au cours d’une enquête de deux semaines menée dans 8 hôpitaux dans 4 provinces du Kenya. Les gardiens ont été invités à signaler les comportements de recours aux soins et les dépenses liées à l’accès aux soins hospitaliers. Les statuts socioéconomiques des familles ont étéévalués sur base des dépenses déclarées pour le mois précédent.
Résultats: 78% des gardiens ont été tenus de payer les frais d’utilisation pour accéder à des soins hospitaliers pour les enfants. Les frais d’utilisation (moyenne: 8,1 US$; IC95%: 6,4–9,7) constituaient 59% du total des coûts payés de la poche, tandis que les coûts de transport (moyenne: 4,9 US$; IC95%: 3,9–6,0) et les coûts des médicaments (moyenne: 0,7 US$; IC95%: 0,5–1,0) étaient de 36% et 5% respectivement. Le total moyen des frais payés de la poche par admission pédiatrique était de 14,1 US$ (IC95%: 11,9–16,2). Les dépenses payées de la poche pour la santé sont catastrophiques pour 25,4% (IC95%: 18,4–33,3) des gardiens interrogés. Les dépenses payées de la poche étaient régressives, avec une charge plus lourde pour les ménages ayant un statut socioéconomique plus faible.
Conclusion: En dépit d’une politique d’exonération des frais d’utilisation pour les enfants de moins de cinq ans au Kenya, nos résultats montrent que des frais d’utilisation élevés et non officiels sont encore facturés dans les hôpitaux de district. Des mécanismes de financement qui offrent une protection du risque financier pour les enfants en quête de soins devraient être développés pour éliminer les obstacles à la survie de l’enfant.
Objetivo: Describir los gastos de bolsillo de las familias con niños menores de cinco años ingresados en hospitales distritales de Kenia.
Métodos: Se entrevistó a 256 cuidadores de niños ingresados, mediante encuestas de dos semanas realizadas en 8 hospitales en 4 provincias de Kenia. Se pidió a los cuidadores que reportaran los comportamientos de búsqueda de cuidados sanitarios y los gastos relacionados con el acceder a cuidados hospitalarios durante un ingreso. Se evaluó el nivel socioeconómico de la familia mediante los gastos reportados para el mes anterior.
Resultados: A un 78% de los cuidadores se les pidió que pagaran para acceder a los servicios de hospitalización para los niños. Las tasas de usuario (media US$ 8.1, IC 95% 6.4–9.7) eran un 59% del total de los gastos de bolsillo, mientras que los costes derivados del transporte (media US$ 4.9, IC 95% 3.9–6.0) y los costes de medicamentos (media US$ 0.7 IC 95% 0.5–1.0) eran del 36% y 5% respectivamente. La media total de los gastos de bolsillo por admisión pediátrica eran de US$ 14.1 (IC 95% 11.9–16.2). Los gastos de bolsillo en salud eran catastróficos para un 25.4% (IC 95% 18.4–33.3) de los cuidadores entrevistados. Los gastos de bolsillo eran regresivos, con una mayor carga en los hogares con un menor estatus socioeconómico.
Conclusión: A pesar de una política de exención de tasas de usuario para niños menores de cinco años en Kenia, nuestros resultados muestran que en los hospitales distritales aún existe un alto número de cobro de tasas no oficiales. Deberían desarrollarse mecanismos de financiación que ofrezcan protección contra el riesgo financiero para las familias de niños que buscan ayuda sanitaria, quitando así barreras que impiden la supervivencia infantil.
Access to hospital care plays an important role in improving child survival (Schellenberg et al. 2004), and costs have been identified as a significant barrier to access (Perkins et al. 2009). In Kenya, public hospitals operate on a cost-sharing arrangement, where the government provides healthcare services at a subsidised rate, financed by central government budgetary allocations to health and supplemented by out-of-pocket payments from users (Management science for health 2001). The cost-sharing policy provides for exemptions of user charges for children under the age of 5 (Carrin et al. 2007). Experience in developing countries shows that policies on removal of user fees or exemptions are often poorly implemented (Chuma et al. 2009). We examined out-of-pocket payments incurred by caretakers arising from inpatient episodes and explored their magnitude in relation to household expenditures on essential items so as to determine whether out-of-pocket expenditures were catastrophic.
The study was conducted in 2008 as part of a larger study to improve the quality of paediatric inpatient care in district hospitals in Kenya (Ayieko et al. 2011). Data were collected prospectively, in a survey over 2 weeks, by administering a questionnaire to caretakers of children admitted in eight district hospitals in Kenya. The sampling and selection of hospitals has been described elsewhere (Ayieko et al. 2011). We interviewed 256 caretakers (range, 23–36 per hospital). For this descriptive analysis, the data from the eight hospitals were pooled. Total out-of-pocket costs were defined as the sum of transport costs to and from the hospital, user fees and any charges levied for medicines and laboratory services. Caregivers were asked to recall household expenditures they incurred in the previous month. Households were considered to have incurred catastrophic expenditures if their total out-of-pocket costs exceeded 40% of their monthly non-subsistence expenditure (Xu et al. 2003). Households were grouped into socio-economic quintiles (1 = lowest to 5 = highest) and two broader groups, higher and lower socio-economic status, based on household monthly expenditures. Out-of-pocket costs and household expenditures were converted from Kenya shillings to US dollars and inflated to 2010 prices using GDP deflators for Kenya. The data were skewed, and so, we explored presenting them as medians and interquartile ranges. Household expenditure categories (rent, food and education) and out-of-pocket cost categories (medicine, transport and user charges) were heavily zero-inflated and hence could only meaningfully be presented as means. Total out-of-pocket costs are presented as both means and medians with non-parametric tests used to test for associations.
The characteristics of the admitted children are shown in Table 1. The median time of the journey to the hospital by caregivers and the sick children under their care was 45 min (IQR, 30–105). 79.7% (95% CI, 74.2–84.5) of the caregivers used public means to get to hospital while 18.3% (95% CI, 13.7–23.7) walked to the hospital and 2.0% (95% CI, 0.6–4.6) used private means. The median number of visits to a healthcare provider before the child was admitted was 3 (IQR, 1–6).
Table 1. Characteristics of admitted children
Median age in months
Proportion (95% CI)
Diarrhoea and dehydration
Higher socio-economic group
Lower socio-economic group
Employment status of caregivers
Household monthly expenditure on food was 79% of the total monthly expenditure on essentials (food, rent and education); expenditures on rent constituted 11%, and on education 10% (Table 2). The majority of caretakers interviewed (77.7% (95% CI, 72.1–82.7)) reported paying user fees for admission care of sick children. 66.7% (95% CI, 59.3–73.4) used their savings to meet hospital out-of-pocket costs, and 33.3% (95% CI, 26.6–40.7) borrowed money. User charges contributed to the greatest proportion (59%) of out-of-pocket costs, followed by transport (36%) and medicine (5%). Table 3 shows hospitalization expenditure. Figure 1 shows the relationship between out-of-pocket expenditures and their burden (out-of-pocket expenditure as a percentage of total household expenditure) and socio-economic quintiles. Non-parametric tests showed that the household out-of-pocket costs were significantly higher in lower socio-economic strata households (median US$ 11.0 (IQR (4.7–19.3)) than in higher socio-economic strata households (median, US$ 7.0 (IQR, 2.6–14.2)) (rank sum P = 0.025) but did not vary significantly with diagnosis. Out-of-pocket costs were catastrophic for 25.4% (95% CI, 18.4–33.3) of households. The proportion of catastrophic health expenditures was higher in households in the lower socio-economic group (41.3% (95% CI, 29.0%–54.4%)) than in the higher socio-economic group (11.5% (95% CI, 5.4%–20.7%)).
Table 2. Household monthly expenditures
Number of caretakers
As% of total expenditure
Mean expenditure US$ (95% CI)
*Does not include self-produced food.
Table 3. Out-of-pocket expenditures associated with inpatient care for children
Median US$ (IQR)
Mean US$ (95% CI)
As % of total
Total in-patient out-of-pocket costs
Out-of-pocket costs as a percentage of total household expenditure
Proportion of households with catastrophic expenditures
25.4 (95% CI 18.4–33.3)
Our findings show that 78% of caretakers paid user fees for inpatient paediatric care for sick children. The mean out-of-pocket costs for inpatient paediatric care were US$ 14.1 (95%CI 11.9–16.2), which is almost three times higher than reported out-of-pocket costs for district hospital paediatric inpatient care in Tanzania (US$ 5.5) (Saksena et al. 2010). Given that children under 5 years of age are officially exempted from user fees in public health facilities in Kenya, it is apparent that this policy is not well implemented in practice. The violation of the user fee exemption policy in Kenya has been reported in previous studies (Ayieko et al. 2009; Chuma et al. 2009), with funding gaps given as a reason for this violation (Chuma et al. 2009). The poor implementation of exemption and waiver mechanisms within cost-sharing policies is likely to introduce inequities in access to child health.
The mean percentage of out-of-pocket costs to total household expenditure on essentials was 36.5% (95% CI 27.4–45.5). This level of healthcare expenditure is arguably high and comparable to findings in Tanzania of 35.4% (Saksena et al. 2010). A cost burden >40% of household non-subsistence expenditure is likely to be catastrophic to the household (Xu et al. 2003). Based on this definition, out-of-pocket expenditures on health were catastrophic for 25.4% (95% CI, 18.4–33.3) of caretakers interviewed. The mean percentage of out-of-pocket costs to non-subsistence household expenditure was 46.9% (95% CI, 28.5–65.2). When households were divided into 2 socio-economic groups, catastrophic costs were incurred by 41.3% (95% CI, 29.0–54.4) in the lower socio-economic group and 11.5% (95% CI, 5.4–20.8) in the higher socio-economic group. As expected, out-of-pocket expenditures in these households appear to be regressive with a greater burden being experienced by households in lower socio-economic groups given that their capacity to pay is diminished compared to households in higher socio-economic groups (Figure 1).
This study has a number of limitations. Data on household monthly expenditures depend on recalled expenditure over a period of a month and are subject to recall biases. Out-of-pocket costs reported are lower than actual costs, given that we did not collect information on other direct costs (e.g. expenses for food and upkeep of any accompanying relatives). The study also only focused on out-of-pocket costs associated with hospital admissions and left out costs associated with outpatient visits and with those who did not seek care in hospitals; these costs would be important in giving a comprehensive picture of out-of-pocket healthcare costs associated with child illnesses. Also, whereas catastrophic health expenditures are conventionally calculated based on annual health expenditure and annual consumption expenditure, the data available allowed for a calculation based on monthly household and health expenditures. These limitations notwithstanding the data presented are potentially useful as inputs in costing and/or cost-effectiveness models that require patient cost and suggest there are significant-out-of pocket costs associated with paediatric admission care in district hospitals in Kenya, which offer a barrier to access to care.
Policy makers need to mitigate the adverse effects of such expenditures, for example, by extending insurance coverage to the uninsured or by increasing efforts to implement exemption mechanisms. Our findings reinforce observations of tension between policy makers and health facility managers in resource-limited settings, with managers trying to provide services with limited resources and hence forced to disregard policies that threaten their revenue streams.
The authors are grateful to the staff of all the hospitals included in the study and colleagues from the Ministry of Public Health and Sanitation, the Ministry of Medical Services and the KEMRI/Wellcome Trust Programme for their assistance in the conduct of this study. This work is published with the permission of the Director of KEMRI and was funded by the Wellcome Trust.