Menu
Menu
Menu
Menu
Out-of-pocket (OOP) payments have severe consequences for health care access and utilisation and are especially catastrophic for the poor. Although women comprise the majority of the poor in Nigeria and globally, the implications of OOP payments for health care access from a gender perspective have received little attention. This study seeks to fill this gap by using a combination of quantitative and qualitative analysis to investigate the gendered impact of OOPs on healthcare utilisation in south-eastern Nigeria. 411 households were surveyed and six single-sex Focus Group Discussions conducted. This study confirmed the socioeconomic and demographic vulnerability of female-headed households (FHHs), which contributed to gender-based inter-household differences in healthcare access, cost burden, choices of healthcare providers, methods of funding healthcare and coping strategies. FHHs had higher cost burdens from seeking care and untreated morbidity than male-headed households (MHHs) with affordability as a reason for not seeking care. There is also a high utilisation of patent medicine vendors (PMVs) by both households (PMVs are drug vendors that are unregulated, likely to offer very low-quality treatment and do not have trained personnel). OOP payment was predominantly the means of healthcare payment for both households, and households spoke of the difficulties associated with repaying health-related debt with implications for the medical poverty trap. It is recommended that the removal of user fees, introduction of prepayment schemes, and regulating PMVs be considered to improve access and provide protection against debt for FHHs and MHHs. The vulnerability of widows is of special concern and efforts to improve their healthcare access and broader efforts to empower should be encouraged for them and other poor households. © 2014 Onah, Govender.
This study was conducted in Nsukka Local Government Area (NLGA), located in the northern part of Enugu State in south-eastern Nigeria. NLGA comprises 1 urban and 14 rural communities, with a population of almost 310,000, comprising approximately 63, 705 households [27]. Agriculture is the main economic activity and the area is predominantly Ibo (i.e., ethnic group) who are mainly Christians with a few traditional believers in the rural areas. Like other parts of Nigeria, women, including FHHs in NLGA, are less educated, engage more in low level subsistence farming and largely employed in informal employments with low income generation abilities [28]. Heads of FHHs are largely older than their MHHs counterpart and headship is mainly as a result of widowhood [29]. The urban community is a university town and has a wider variety of health care providers including public and private hospitals, primary health care providers, PMVs, and pharmacies. In the rural communities, primary health facilities referred to as health centres and PMVs are the main health services providers. If hospital care is required, people travel to the urban hospital which is between 18– 30 kilometres away. All government facilities charge user fees, although charges differ depending on the type of care sought and patients also pay for drugs. There are exemptions for HIV treatment, leprosy and maternal health. The study used both cross-sectional household surveys and focus-group discussions (FGDs) methods to investigate the research questions. A total of 411 households were interviewed (111 in urban and 300 in rural communities). A household is designated as comprising individuals who live in the same house and who have common arrangements for basic domestic and/or reproductive activities such as cooking and eating” ([30] p:22). Household surveys were chosen over facility-based survey since an understanding of access requires considering the views and experiences of both users and non-users of health care. The following approach was adopted in order to determine the sample size. Given that in 2010, NLGA comprised 69,705 households, the sample size for this study was calculated using Taro Yamane (1967) specification (see Ataguba et al. 2008 [29]) given as: where; n = sample size to be estimated, N = population size, and e = error margin at 95% confidence interval. The population and number of households were extrapolated based on the 2006 population census and an annual 3% population growth rate [28] The minimum sample size required to obtain a confidence interval of 5% around this figure was 400 households. The sample size was increased to 411 households to allow for data incomplete questionnaires. A multi-stage sampling method was used to select households for the survey and the urban and 14 rural communities were classified into enumeration areas (EAs) [29]. First, to ensure adequate representation of both urban and rural EAs, NLGA was stratified into urban and rural communities, representing 30% and 70% of the population respectively. A total of 24 EAs were selected (3 urban, 21 rural) based on probability-proportional to size (PPS) and 39 and 21 households were sampled in each of the urban and rural EAs respectively. In the second stage, a simple systematic random sampling method was used to identify survey households from each of the EAs. The sample of households was appropriately weighted in analysis using the inverse probability weighting method which denotes the inverse of the probability that the observation is included in the analysis due to the chosen sample design [29]. Under the method, each household selected from each enumeration area (EA) is weighted to make it representative of the entire EA such that the sum of the weights for each EA should equal the approximate number of households in that EA. The questionnaire was administered to preferably the household head or the spouse and in their absence, a senior household member. The interviews were conducted by9 trained field workers. Six single-sex FGDs (2 urban, 4 rural) were conducted in 3 communities (1 urban, 2 rural). Each FGD consisted of 8 to 11 participants. Single-sex interviews were considered appropriate given the focus of the research on gender, health care access, coping strategies and intra-household decision-making and sensitive issues which are likely to be spoken of more freely and without fear of reproach in a single-sex group. FGDs were organised to ensure that participants were of similar economic background and economic activity (traders, teachers, farmers, women religious and trading groups), besides considerations of gender. Invitations were sent to men and women in advance of their meeting days. All participants were 18 years and older. The discussions were conducted in the village square and community centres. FGDs were audio taped, transcribed and translated into English and the transcripts were thematically coded and analysed. The household surveys investigated households’ socio-economic and demographic status, general household and health care expenditure patterns, household assets, utilisation patterns, healthcare financing, intra-household decision making and coping strategies. The questionnaire was adapted from an earlier survey conducted in the same region [17] and was translated into the local language. While the household survey provided important data for quantifying the differences and similarities in utilisation patterns between male-and female-headed households, it was inadequate in helping us understand why these differences existed. In this study, the gap was filled through the use of FGDs, which aimed to provide more qualitative data around issues of the burden of OOPs and its implications for health care access, coping strategies, household decision-making in general and more specifically around health expenditure from the perspective of men and women. The FGDs were taped-recorded and notes were taken which were then transcribed. The transcripts were read and broad themes relevant to the study objectives were extracted. In addition, new themes which were identified during the review of the transcripts were also captured and presented in the results. The quantitative data were inputted and managed using EpiData software and then exported to STATA software for analysis. Associations between quantitative variables were assessed using the Chi Square test. A bivariate analysis was conducted and variables which were significant at a probability value (p-value) equal to or less than 0.05 were selected and included. The bivariate analysis was specified to examine the associations between the sex of the household head and other variables including utilisation, decision- making relating to general and health care expenditure, insurance ownership, health care payment options, health status, reasons for not seeking care and coping strategies. Options were subdivided into dichotomous responses of “0” for no and “1” for yes. The monthly cost of health care was calculated by the summation of direct costs (i.e., registration/card fees, consultation fees, laboratory tests and drug costs) that a household incurred in the month previous to the interview. This cost was converted to United States Dollar (2010 exchange rate of US$1.00 = 150 naira). This study used asset indexes as a measure of socio-economic status of households. An asset index was chosen over other measures for constructing the socio-economic status of households. This is because it is easier to collect asset data in contexts like the study site, and income and expenditure data would also not fully represent the household socio-economic status [31], [32]. Information on ownership of electronic equipment (e.g. radio, television and fridge), transport (bicycle, motorcycle and motorcar), sources of energy (kerosene lamp, electricity generators and rechargeable lamps) were pooled together to construct the index. In conducting the principal component analysis, the first component factor was used to represent the asset index. On this basis, the study population was classified into four quartiles (i.e., least poor, poor, very poor and poorest).The first component factor is defined statistically as a weighted sum of the various assets used to assess household wealth, in order for that component to explain as much as possible of the variance observed in asset ownership between households. To estimate the proportion of households incurring potentially catastrophic burdens, costs incurred by each household for health care were divided by household monthly expenditure and reported as a percentage. The household total expenditure was derived by annualising household weekly expenditure on food and beverages and household monthly living expenditure on items such as rent, if any, energy and clothing. The total annual expenditure was then divided by 12 to arrive at the household’s monthly expenditure. health care expenditures are deemed catastrophic if the expenditure is 10% or more of household income [33], where catastrophic implies that such expenditure levels are “ likely to force households to cut their consumption of other minimum needs, trigger productive asset sales or high levels of debt and lead to impoverishment”([34] p:149). The study received ethical approval from the University of Cape Town Ethics Committee and permission was also sought from Nsukka LGA authorities. Informed consent (oral and written) was obtained from all respondents in the household surveys and participants in the FGDs. Oral consent for the FGDs were conducted in the first language of the participants and were captured using an audio recorder, while written consents were used for the household survey and were captured as part of the questionnaires. Oral consent was used in the FGDs due to the difficulties experienced during the household survey on respondents’ literacy level; however participants signed an attendance register. The consent forms were in English and the local language and were read out to obtain oral consent for the FGDs. The consent forms, interview guides, questionnaires, and consent procedures were part of the ethical submissions that were approved for the study. Household interviews and FGDs were conducted in the first language of the respondents and participants.
N/A
