Inequalities in Catastrophic Health Expenditures in Conflict-Affected Areas and the Colombian Peace Agreement: An Oaxaca Blinder Change Decomposition Analysis

doi:10.21203/rs.3.rs-528612/v1

Download PDF

Research

Inequalities in Catastrophic Health Expenditures in Conflict-Affected Areas and the Colombian Peace Agreement: An Oaxaca Blinder Change Decomposition Analysis

https://doi.org/10.21203/rs.3.rs-528612/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

The present study analyzes inequalities in catastrophic health expenditures in conflict-affected regions of Meta, Colombia and socioeconomic factors contributing to the existence and changes in catastrophic expenditures before and after the sign of Colombian Peace Agreement with FARC-EP guerilla group in 2016.

Methods

The study uses the results of the survey Conflicto, Paz y Salud (CONPAS) conducted in 1309 households of Meta, Colombia, a territory historically impacted by armed conflict, for the years 2014 and 2018. We define catastrophic expenditures as health expenditures above 20% of the capacity to pay of a household. We disaggregate the changes in inequalities in catastrophic expenditures through the Oaxaca-Blinder change decomposition method.

Results

The incidence of catastrophic expenditures slightly increased between 2014 to 2018, from 29.3–30.7%. Inequalities in catastrophic expenditures, measured through concentration indexes (CI), also increased from 2014 (CI: -0.152) to 2018 (CI: -0.232). Results show that differences in catastrophic expenditures between socioeconomic groups are mostly attributed to an increased influence of specific sociodemographic variables such as living in rural zones, being a middle-aged person, living in conflict-affected territories, or presenting any type of mental and physical disability.

Conclusions

Conflict-deescalation and the peace agreement may have facilitated lower-income groups to have access to health services, especially in territories highly impacted by conflict. This, consequently, may have led to higher levels of out-of-pocket expenditures and, therefore, to higher chances of experiencing catastrophic expenditures for lower-income groups in comparison to higher-income groups. Therefore, results indicate the importance of designing policies that guarantee access to health services for people in conflict -affected regions but also, that minimize health care inequalities in out-of-pocket payments that may arouse between people at different socioeconomic groups.

Health Policy

Colombia

peace treaty

catastrophic expenditures

health inequalities

For almost 60 years, Colombia has experienced one of the most long-lasting armed conflicts in the world. Colombia's conflict has resulted in approximately 262,000 deaths, 80,000 forced disappearances, 15,000 victims of sexual assault, and more than 7 million internally displaced people (1; 2). Colombia's conflict has affected in different ways and at various levels of conflict intensity, several regions of Colombia, specific communities, and some of the most vulnerable population groups (3). Nevertheless, these armed struggle's consequences go beyond the direct impacts on security and losses in human lives and influence other dimensions of social, political, and economic outcomes (4).

One of the dimensions where conflict has had important consequences is in health. Conflict, directly and indirectly, impacts health outcomes and opportunities (5). Asides from general physical and mental health consequences, conflicts worsen health provision, and complicate health services operations and processes (6). These effects are a consequence of direct damages to health facilities and health care infrastructures, conflict-related threats to health professionals, or structural problems such as inadequate institutions for promoting people's rights and opportunities (7). Health services may be forced to stop or to operate under challenging circumstances, which, in the long run, may lead health services to be unstable or rugged to maintain operations overtime (8).

In some regions, inadequate health services or difficulties to access these health care services translate into high costs for the individuals and difficulties in the coverage and the financial sustainability of health care systems (9). Health systems worldwide usually appeal to public resources, taxes, or pre-payment services to finance health care (10). Nevertheless, some of these health expenses are charged directly to healthcare users both as a mechanism for facilitating health financing and limiting moral hazard problems in health (11). These direct payments covered by individuals and not the health system, such as certain medicines or specific health services, are usually called out-of-pocket expenses (OOP).

OOP may be an adequate mechanism for transferring some of the costs of health provision to the people that benefit from these services. Nevertheless, when these expenses surpass people's capacity to pay, these costs become catastrophic expenses or "payments that exceed a household's ability to pay, once food and basic consumption costs are deducted" (12). In the long run, high OOPs and continuous catastrophic expenditures may lead to financial ruin or difficulties in maintaining an adequate quality of life (13).

The presence of catastrophic expenditures in a region or a country usually reflects several types of economic system difficulties that restrain or limit economic development (14). No matter the level of out-of-pocket expenditures, households are at risk of incurring in catastrophic payments if there are high levels of poverty, some social groups are excluded from financial risk protection mechanisms, and health care utilization is high, situations that are common in middle and low-income countries (15).

People experiencing financial difficulties or suffering from poverty are significantly affected by these circumstances. They may be more vulnerable to adverse health shocks and, ultimately, greater health care costs (16). Even though there is a growing use of health care systems, developing countries usually have weak social institutions, inadequate risk pooling mechanisms, and inadequate tax-financed health care systems, limitations that manifest themselves in high levels of household health care expenditures (15). These problems may be more severe in conflict-affected territories where households have financial limitations to generate income, and simultaneously, health provision is problematic within largely unregulated health markets (17). The risk of experiencing bad health in conflict-affected regions increases as well as the financial burdens and limitations to which people may be exposed to, such as loss of jobs, destruction of public infrastructure in their communities (including health facilities) and, ultimately, greater risks of being sick (18). If, simultaneously, these people are obligated to cover expensive health care treatments or belong to highly vulnerable population group,s who may have specific limitations to generate stable income streams, like elderly people or with certain health conditions, catastrophic expenditures may lead to long-run financial ruin (19).

The health and financial risks outlined above, a key concern for remedial policies, may be distributed heterogeneously across the population. Exposure to different levels of conflict incidence across different socioeconomic groups may lead to differences in the incidence of health payments and, ultimately, in the risk of experiencing catastrophic health expenditures. Moreover, in conflict-affected regions there is often variations in the exposure to conflict violence, leading to inequalities in health expenditures. Ultimately, health expenditures may impose greater financial burdens over certain socioeconomic groups, leading to inequities in health, and problems in overall quality of life and wellbeing (20). Analyzing inequalities in catastrophic expenditures is essential to identify contributing factors that sustain these differences over time and, therefore, important to improve the design of public policies that reduce health financing disparities.

In 2016, Colombia signed a peace accord with Fuerzas Armadas Revolucionarias de Colombia (FARC-EP), one of the guerilla groups that, for years, dominated several of the territories of the Meta region (21). The treaty led to the establishment of the Espacios Territoriales de Capacitación y Reincorporación (ETCR), created to facilitate the gradual reincorporation of demobilized guerilla groups to civil society. These processes, initially, may have contributed to a reduction in direct conflict violence, which may have led to a reduction of the physical and psychological consequences that direct armed struggle has on health outcomes, and simultaneously, to reductions in health expenditures. Nevertheless, conflict's health consequences may have long-run impacts and may be more severe in certain population groups that more impacted by armed conflict.

To contribute to our understanding of the issues above, our study investigates the evolution of catastrophic expenditures over time and between socioeconomic groups in the Colombian region of Meta, an area that was intensely affected by conflict violence, mostly related to actions of FARC-EP. To evaluate the prevalence and change of inequalities in health expenditures, we measure the change in the incidence of catastrophic expenditures for the years 2014 and 2018 and analyze, through a decomposition method, the extent to which inequalities in specific socioeconomic factors contribute to inequalities in health expenditures in each period. First, we present our methodological approach, followed by our main results and concluding with a discussion of our findings.

Our research uses primary data from the survey Conflicto, Paz y Salud (CONPAS), designed by our research team. The survey was conducted in year 2018 in 1309 households of the department (province) of Meta, Colombia and includes responses for year 2014 collected through recall questions answered by survey respondents. The survey sample was selected through a probabilistic method using a two-stage sampling approach, being representative at the level of rural and urban areas and for groups of Meta municipalities affected by different levels of conflict violence (see below).

To calculate the incidence of catastrophic expenditures, we use the methodology proposed by Xu (22). A household incurs catastrophic expenditures if the ratio of out-of-pocket expenditures over a household's capacity to pay exceeds a certain threshold. In the literature, this threshold ranges between 20–40% and is usually established depending on a country's income level. We use a threshold of 20% to make our results comparable with previous national studies conducted in Colombia about catastrophic expenditures (23; 24). Out-of-pocket expenditures were directly measured by asking respondents what amount of the total household expenditure in the previous month was used to pay for health-related services or products, including health insurance copayments but excluding insurance premia. Capacity to pay is defined as the household's available expenditure capabilities once minimum subsistence expenditures are deducted – the difference between total expenditure and minimum required food expenditures (22). Total household expenditure was measured directly in the survey. Minimum required food expenditures are calculated following the procedure proposed by Xu (22).

We first conduct some descriptive analysis of our data. Then, we use a multivariate logistic regression to analyze sociodemographic factors that affect the risk to experience catastrophic health expenditures for both 2014 and 2018.

We estimate inequalities in catastrophic expenditures by calculating concentration curves and the health concentration index (HCI) for both years, 2014 and 2018. A concentration curve plots the cumulative percentage of a variable (in our study catastrophic expenditures) in the y-axis, against the cumulative percentage of the population ranked from poorest to richest, measured through a specific socioeconomic ranking variable, in the x-axis (25). The concentration curve allows us to measure if catastrophic expenditures are disproportionally occurring in a specific socioeconomic group and, consequently, if inequalities in health outcomes are present in this population. The HCI, defined as twice the area between the concentration curve and the perfect equality line, measures how severe this inequality is. A HCI of zero represents perfect equality, below zero that inequality is against the poor, and above zero that inequality is against the rich. To calculate these HCIs, we estimate, with data of household characteristics, the Household Wealth Index (HWI) (26) which is a measure of household non-monetary income based on the possession of different assets.

We estimate the HCI for catastrophic health expenditures for both years 2014 and 2018. We then analyze to what extent changes in inequalities in catastrophic expenditures between 2014 and 2018 may be explained by changes in inequalities in sociodemographic variables related to health expenditures. Two mechanisms may explain changes in inequalities in catastrophic expenditures: a change in the distribution of an explanatory variable across socioeconomic groups between these two time periods (changes in distribution) or changes in the influence of a socioeconomic variable over time (changes in marginal effects). To separate these two components, we use the Oaxaca-Blinder decomposition method (27), which separates changes in the health concentration index between two time periods as a linear function of the changes in explanatory socioeconomic variables.

Socioeconomic factors for the decomposition analysis were initially selected through a literature review to identify variables correlated with catastrophic expenditures in previous studies (28; 29) and, more specifically in previous research in post-conflict settings (17). In the final decomposition model, the following variables were included: gender, age, residence (urban or rural), currently experiencing internal displacement, type of work, ethnicity, marital status, education level, conflict incidence of the municipality of residence, number of household members, number of children under six years old, health insurance scheme, being recently sick or hospitalized, and measurements of possible mental health disorders and mental or psychical disabilities. Internal displacement status was self-reported by survey respondents. Possible mental health disorder was measured using the Self Report Questionnaire (SRQ-20) (30), and physical and mental health disabilities were measured through the World Health Organization Disability Assessment Schedule (WHODAS). The municipality's conflict incidence was estimated using the classification developed by the Conflict Analysis Resource Center (31), with municipalities categorized as being with "high”, "low" or "no conflict", in addition to Meta's largest capital city (Villavicencio), mostly highly affected, but analyzed separately because of its size in comparison to other municipalities.

All calculations were conducted using robust standard errors and 95% confidence intervals. We estimate the models using the software Stata 15 MP, using previously assigned survey sample weights.

Table 1 shows a summary of the characteristics of study participants and the incidence of catastrophic expenditures using a threshold of 20% . In 2018 all people surveyed had the legal age (18 years or over) and lived in Meta, but, in 2014, year for which retrospective questions were conducted, some of these people were minors or lived outside Meta. In the first row of the age variable, this small group (N:63) is also considered.

The greatest group of people surveyed is 18-44 years old both in 2014 (51.8%) and in 2018 (45.7%). Most people are women (54.2%) and live in urban areas (59.7%). Approximately 42.0% of the population report being a victim of internal displacement. The majority of people work in informal jobs, both in 2014 (71.2%) and 2018 (74.6%), live in lowly conflict-affected territories (69.2% in 2014 and 74.6% in 2018), and on average, 78.5% belong to an ethnic group considered majority in the territory (mestizo and white). Most of the population surveyed attained up to primary school education (43.1% and 40.9%). Finally, 15.1% and 32.4% of people have a possible mental health disorder, measured through the Self Report Questionnaire, indicated, in this paper as SRQ+ (i.e a SRQ positive case). Approximately 29.3% and 30.7% of households faced catastrophic health expenditures in 2014 and 2018, respectively.

Logistic regression analysis

Table 2 shows the odds ratios (OR), calculated from a multivariate logistic regression, to analyze sociodemographic factors that increase the risk of specific populations to have catastrophic expenditures for 2014 and 2018. OR quantify the association between “exposure” to a specific factor and an outcome, representing the odds that this outcome occurs given a particular exposure, compared to the odds of the outcome occurring in the absence of that exposure (32).

For 2014, the odds of experiencing catastrophic expenditures are, for elderly people (65 years or more), 1.99 higher in comparison to younger population groups. The odds of experiencing catastrophic expenditures for people living in rural zones are 1.68 higher than in urban areas. People with any type of physical or mental disability were 1.06 higher chances of experiencing catastrophic expenditures than others without these health conditions. People living in municipalities with low conflict intensity had a higher chance (OR: 1.50) of incurring catastrophic expenditures than people living in the capital city. Internally displaced people had a higher chance (OR: 1.33) of experiencing catastrophic expenditures than those not displaced. Finally, people who did not had any type of illness in the previous 12 months had a fewer chance (OR: 0.61) of facing catastrophic expenditures.

In 2018 similar socioeconomic factors contributed to a higher chance of experiencing catastrophic health expenditures, albeit with slight differences. Older people (OR:1.68) and middle-aged people in the 45-64 age range (OR: 1.38) had higher chance of experiencing catastrophic health expenditures than younger age groups. People in both low and highly affected areas had a higher chance (OR: 2.20 and 2.80, respectively) of experiencing catastrophic health expenditures than people living in other municipalities. Finally, people with any type of physical or mental disability (measured with WHODAS) had a higher chance of experiencing catastrophic expenditures (OR: 1.04). On the contrary, those who did not experience any type of health problem in the previous 12 months had a lower chance of experiencing catastrophic expenditures (OR: 0.63).

Oaxaca Decomposition Analysis

Even though catastrophic health expenditures are similar between 2014 and 2018 (29.3 and 30.7%), there are differences in the incidence of these expenditures between socioeconomic groups. Figures 1 and 2 show concentration curves for catastrophic expenditures for both 2014 and 2018.

Inequalities in catastrophic health expenditures were higher in 2018 than in 2014. For 2014 and 2018, the health concentration indexes (HCI) are, respectively, -0.152 and -0.232 (Figures 1 and 2), which means that catastrophic health expenditures were unevenly distributed among different socioeconomic levels and, more specifically, that inequalities against the poorest groups were observed in both years.

The change in the slope of the concentration curve slope shows the areas where inequalities in catastrophic expenditures are growing faster. For the year 2018, more people in the 0-20% income group are experiencing catastrophic expenditures compared to 2014, as shown by the higher slope of the concentration curve in the 0-20% of the x-axis. Catastrophic expenditures in 2018 reached its peak at the 30% percentile, point at which the cumulative percentage of catastrophic expenditures is of 50% and, for 2014 is, approximately 42%. In the 30%-50% income group, the rate of people with catastrophic expenditures grows at similar rates for both years. Nevertheless, while in the 50%-70% income group, there is a steady reduction in the number of people with catastrophic expenditures for the year 2018, in 2014, catastrophic expenditures continue growing at similar rates to those found at the lower-income groups. In the 80% -100% income group, catastrophic expenditures grow at much lower rates for 2014 and 2018 than the most impoverished populations.

In summary, higher inequalities in 2018 are mostly explained by higher levels of catastrophic expenditures at the middle-income groups (Quintile 3), in comparison to 2014. Inequalities in 2018 increase even more due to a reduction in catastrophic expenditures in the 50-70% (Quintile 4), in comparison to 2014. These results are consistent with those found in the multivariate logistic regression (Table 2). Here, the chances of facing catastrophic expenditures for Quintile 3 are higher in 2018 (OR: 0.91 ) than in 2014 (OR: 0.75) , when compared to the lowest income groups. Similarly, the chances of facing catastrophic expenditures for Quintile 4 are lower in 2018 (OR: 0.60 ) than in 2014 (OR: 1.13).

Table 3 shows the results of applying the Oaxaca decomposition method to identify the socioeconomic factors contributing to changes in inequalities in catastrophic expenditures. The Oaxaca method decomposes the change in the health concentration index into changes in the distribution of determinant factors (distributional effect in Table 3) and changes in the influence of a socioeconomic variables over time (coefficient effect in Table 3). In our table, the term ηΔC represents changes in the levels or values of a socioeconomic variable between socioeconomic groups from 2014 to 2018 (distributional effect), while the term CΔη captures changes in the influence of a specific variable for determining health inequalities between these two years (coefficient effect).

The decomposition results for changes in catastrophic expenditure inequalities show that most of the differences in inequalities over time are explained as a consequence of changes in the elasticities, i.e. in the influence of sociodemographic characteristics for explaining health expenditures inequalities (-0.104), instead of changes in the distribution of levels of these determinants (-0.007). Working in the informal sector (-0.067) and living in municipalities highly affected by conflict violence (-0.050) are the two factors with the highest individual contributions to the increase in catastrophic health expenditure inequalities, primarily due to the more significant influence of these determinants over time. Changes in the distribution of populations between rural and urban residence areas (-0.018) are the major contributor to changes in catastrophic expenditure inequalities because of a distribution effect.

Robustness check

In Table 4, we perform a robustness check of the decomposition results using different threshold levels for defining catastrophic expenditures. Results show that while the distributional effects are mostly robust to the threshold used, the coefficient effect estimates vary more widely with the specific threshold adopted. Overall, however, the conclusions from our main analyses remain largely unchanged in qualitative terms.

Main Conclusions

Inequalities in catastrophic expenditures increased from a concentration index of -0.152 in 2014 to -0.232 in 2018, meaning that, in both years but more sharply in 2018, it is the population groups at lower socioeconomic levels that are experiencing catastrophic health expenditures more frequently. Our results show that changes in catastrophic expenditure inequalities over the period 2014-2018 are mostly explained by changes in the importance of specific sociodemographic determinants over the same period and, to a much lesser extent, by changes in inequalities in socioeconomic factors. For both years, people living in rural areas, at middle or older age ranges, with health disabilities, being previously sick, or living in conflict-affected municipalities were more prone to experience catastrophic health expenditures than other groups.

Comparison with Colombian and international studies

Few international studies have explored changes in catastrophic expenditure inequalities in (post-)conflict settings. Edoka et al. (17) analyze changes in catastrophic health expenditures after the end of the eleven-year civil armed conflict in Sierra Leone for the years 2003 to 2011. In contrast to our study, where catastrophic expenditure incidence increased from 29.3% to 30.7% during the pre/post-conflict period examined, Edoka et al. (17) found a decrease from 50% to 32% in the incidence of catastrophic expenditure, which the authors attribute to higher use of regional health facilities instead of NGO health services, a reduction of ill-health, and relocation of households to areas with better living conditions. Nevertheless, post-conflict levels of catastrophic expenditures are similar in both countries (30.7% and 32%). Differences may be attributed to the more severe impairment (to virtual collapse) of health services and infrastructure during the armed conflict in Sierra Leone in comparison to Colombia. This post-conflict recovery of health the most basic health services in Sierra Leone from their nearly inexistence, may have contributed to the significant drop in catastrophic expenditure levels in this country . The longer time frame of the Sierra Leone analysis (almost eight years) may further explain the different conclusions of our study, as a consequence of the more extended recovery period for health services. As indicated by our results, major contributors to health inequalities and catastrophic expenditures in Colombia are much more related to individual characteristics (age, mental and physical disabilities, for example) than health service provision or insurance affiliation. These contextual differences may require health policies to be targeted at specific populations to achieve reductions in the overall levels of catastrophic expenditures in Colombia.

Even though catastrophic expenditures remained almost constant (29.3% and 30.7%), and inequalities increased, our results do not necessarily indicate negative outcomes after the Colombian peace treaty. As in Edoka et al. (17), results show how area-related factors (urban/rural household location and conflict incidence) were key for determining changes in health expenditure patterns. Results show higher chances of catastrophic expenditures in highly affected territories only for 2018. Conflict-deescalation and the peace agreement may have facilitated lower-income groups' access to health services, especially in territories highly impacted by conflict. This, consequently, may have led to higher levels of out-of-pocket expenditures and, therefore, to higher chances of experiencing catastrophic expenditures for this population. Also, conflict de-escalation in post-conflict scenarios may lead, at least in the short run, to a reduction in violent actions and, consequently, facilitate the recovery (or even establishment) of healthcare services in the region.

Consequently, this may lead to quality improvements and favor the provision of healthcare supplies and operation of health professionals in areas where only humanitarian brigades used to provide some healthcare. Simultaneously, conflict de-escalation may facilitate economic recovery in conflict-affected regions, which may improve household income for some populations , reduce financial vulnerability, and, therefore, explain the drop in catastrophic expenditures in 2018 for some of the highest income groups (Quntile 4). Both mechanisms leave levels of catastrophic expenditures inaltered and increase inequality but, ultimately, reflect improvements in health care utilization. Therefore, in conflict-affected territories, two health policy priorities become especially important: first, guarantee the recovery, provision, and maintenance of adequate and permanent health services over time and, second, design better insurance and financial protection policies that guarantee continuous health care utilization at low costs to the most vulnerable groups.

It is noteworthy that the incidence of catastrophic expenditures in the conflict-affected region that we examine (30.7%) is much higher than the average incidence figures for Colombia reported in other studies. Recent estimates from national household surveys point to an incidence of catastrophic spending between 2.2% and 8.2% for Colombia in 2016, depending on the threshold adopted (33). Amaya-Lara (23) estimates an average incidence of catastrophic health expenditures of 9.6% for a 20% threshold. The importance of the level of conflict violence for explaining health inequalities, found through our decomposition approach, shows how different areas may be exposed to different circumstances that increase households' risk of incurring catastrophic health payments. This highlights the perils of relying on average figures to set priorities for national health policy, under the risk of masking important inequities against vulnerable groups that deserve urgent attention. Conflict-affected regions and rural areas (the latter which also tended to be disproportionally affected by conflict violence in Colombia) may have more difficulties in providing sustained health services, experience higher care provision costs and limited service capacity, all of which may increase the costs of medical care to families and call for remedial policies targeted explicitly to these population groups.

Main contribution, strengths, and limitations of the study

Even though our results apply primarily to the region of Meta, Colombia, this research is based on a large-scale survey that is representative of rural and urban areas exposed to different levels of conflict incidence. This characteristic allows us to draw more general insights about the relationship between inequalities in catastrophic health expenditures and the distribution of socioeconomic characteristics in conflict-affected settings. Crucially, our data's panel structure allows us to estimate changes right before and after one of the most critical episodes in the history of Colombian armed conflict: the signing of the Peace Accord with the FARC guerilla group in 2016. Our research contributes to understanding changes in catastrophic expenditures in post-conflict scenarios, shedding light also into the influence of conflict violence patterns to explain changes in health expenditure inequalities. Rather than establishing causal links between catastrophic expenditure inequalities and the Colombian Peace Agreement, our research focuses instead on offering more general insights that can assist with designing relevant health policies in conflict-affected regions.

Access to health expenditure and other detailed representative data for more than 1000 households has enabled us to perform more fine-grained analyses of the relationship between conflict and catastrophic expenditures, compared to the previous studies focused on specific population groups or cities. Unfortunately, we do not have information on health expenditures for our study population before 2014, which would have been informative to determine how household expenditures evolved during earlier stages of the Colombian conflict. Our analysis is also not geared towards allowing firm conclusions to be drawn about the causal effect of the peace agreement on catastrophic expenditures in Meta, or about long-run trends in health expenditures, due to our methodological strategy and the relatively short period covered by our data. The Oaxaca decomposition method does not allow us to identify when health inequalities started to increase or the role of other unobserved contributing factors that may explain changes in these inequalities, such as personal preferences or risk attitudes.

We found that our quantitative estimates of the distributional effects are robust to changes in the threshold of catastrophic expenditures but less so for the coefficient effects. This may be due in part to possible measurement errors in health expenditures due to recall bias. Even though the survey expenditure questions referring to 2014 and 2018 were the same, longer recall periods may lead to imprecisions in total expenditure figures reported and, usually, will lead to an underestimation of health expenses and/or overestimation of other everyday expenses, like food expenditures, which may, in turn, lead to some imprecision in the estimation of catastrophic expenditures (Beegle et al. 2010). Our catastrophic spending results do seem consistent with those from previous studies (e.g. Edoka et al.(17)), although it must be noted that such studies have generally faced the same limitations in applying a similar estimation methodology to ours, thus highlighting the fact that recall and measurement bias are likely to be present in any analyses of self-reported health expenditures.

Health care provision and coverage is a difficult task in any part of the world. However, conflict and its negative consequences in health give origin to major complexities for guaranteeing health care services. Adequate and sustained health coverage is difficult in conflict-affected regions as well as ensuring health equity in outcomes and opportunities. In this order, health policies not only must improve health provision but contribute to minimizing social differences that sustain inequalities and, ultimately, reduce overall wellbeing for some of the most vulnerable populations groups of the world. Post-conflict Colombia is a significant opportunity and scenario to reduce inequalities and promote capacity building in one of the countries where socioeconomic inequalities have historically sustained health care differences. However, a new social reality is possible, as well as improvements in health care outcomes and capabilities.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Ethics approval and consent to participate

Research authorized by the ethics committees of Alberto Lleras Camargo, School of Government, Universidad de los Andes, Colombia and of University of York, UK

Consent for publication

Not applicable

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare that they have no competing interests

Funding

Project funded by the UK Medical Research Council, Economic and Social Research Council, DFID and Wellcome Trust (Joint Health Systems Research Initiative). Grant code: MR/R013667/1.

Authors' contributions

GC, psychiatrist, selected the mental health instruments and contributed to discussion

SLG led the writing process, the research design and selected the methodology; JSCS performed data cleaning and econometric analysis , GC and CGU revised the manuscript and contributed to data analysis and interpretation of results; RMS and OB contributed to discussion, funded survey design and facilitated contact with relevant stakeholders for data collection in the field. NK and SL revised methodological and econometric coherency and made final revisions and reviews.

Acknowledgements

We would like to thank the public officers from the department of Meta Governor’s Office and, specially, the Secretary of Health, partners that facilitated and supported data collection in the territory.

Instituto Nacional de Salud 2017. Consecuencias del Conflicto Armado en Salud en Colombia; Noveno Informe Técnico Bogotá, D.C., 2017
Grupo de Memoria Histórica 2013. ¡Basta Ya! Colombia: Memorias de guerra y dignidad. CN Histórica, informe General del Grupo de Memoria Histórica.
Calderón Rojas J. Etapas del conflicto armado en Colombia: hacia el posconflicto. Latinoamérica. Revista de Estudios Latinoamericanos. 2016 Jun(62):227-57.
Devkota B, van Teijlingen ER. Understanding effects of armed conflict on health outcomes: the case of Nepal. Conflict and health. 2010 Dec;4(1):1-8.
Ergo A, Htoo TS, Badiani-Magnusson R, Royono R. A new hope: from neglect of the health sector to aspirations for Universal Health Coverage in Myanmar. Health policy and planning. 2019 Oct 1;34(Supplement_1):i38-46.
Franco S, Suarez CM, Naranjo CB, Báez LC, Rozo P. The effects of the armed conflict on the life and health in Colombia. Ciência & Saúde Coletiva. 2006;11:349-61.
Druce P, Bogatyreva E, Siem FF, Gates S, Kaade H, Sundby J, Rostrup M, Andersen C, Rustad SC, Tchie A, Mood R. Approaches to protect and maintain health care services in armed conflict–meeting SDGs 3 and 16
Ataullahjan A, Gaffey MF, Tounkara M, Diarra S, Doumbia S, Bhutta ZA, Bassani DG. C’est vraiment compliqué: a case study on the delivery of maternal and child health and nutrition interventions in the conflict-affected regions of Mali. Conflict and Health. 2020 Dec;14:1-7.
Qirbi N, Ismail SA. Health system functionality in a low-income country in the midst of conflict: the case of Yemen. Health policy and planning. 2017 Jul 1;32(6):911-22.
Grosskopf S, Self S, Zaim O. Estimating the efficiency of the system of healthcare financing in achieving better health. Applied Economics. 2006 Jul 20;38(13):1477-88.
Victoor A, Noordman J, Potappel A, Meijers M, Kloek CJ, de Jong JD. Discussing patients’ insurance and out-of-pocket expenses during GPs’ consultations. BMC health services research. 2019 Dec;19(1):1-7.
World Health Organization. Distribution of health payments and catastrophic expenditures methodology. World Health Organization; 2005.
Mukherjee J. An introduction to global health delivery: practice, equity, human rights. Oxford University Press; 2018.
Nghiem SH, Connelly LB. Convergence and determinants of health expenditures in OECD countries. Health economics review. 2017 Dec;7(1):1-1.
Xu K, Evans DB, Kawabata K, Zeramdini R, Klavus J, Murray CJ. Household catastrophic health expenditure: a multicountry analysis. The lancet. 2003 Jul 12;362(9378):111-7.
Kimani DN, Mugo MG, Kioko UM. Catastrophic health expenditures and impoverishment in Kenya. European Scientific Journal. 2016 May 1;12(15).
Edoka I, McPake B, Ensor T, Amara R, Edem-Hotah J. Changes in catastrophic health expenditure in post-conflict Sierra Leone: an Oaxaca-blinder decomposition analysis. International journal for equity in health. 2017 Dec;16(1):1-5.
Flórezi CE, Giedionii U, Pardoiii R. Risk factors for catastrophic health expenditure in Colombia. 1olume. 2011;53:155.
Li Y, Wu Q, Xu L, Legge D, Hao Y, Gao L, Ning N, Wan G. Factors affecting catastrophic health expenditure and impoverishment from medical expenses in China: policy implications of universal health insurance. Bulletin of the World Health Organization. 2012;90:664-71.
Kavosi Z, Rashidian A, Pourreza A, Majdzadeh R, Pourmalek F, Hosseinpour AR, Mohammad K, Arab M. Inequality in household catastrophic health care expenditure in a low-income society of Iran. Health policy and planning. 2012 Oct 1;27(7):613-23
González KX. Construcción de paz en escenarios de violencia intracomunitaria. Estudio de caso Sierra de la Macarena (Meta-Colombia). Estudios Políticos. 2017(51):196-217
Xu K. Distribución del gasto en salud y gastos catastróficos: metodología. Organización Mundial de la Salud; 2005.
Amaya-Lara JL. Catastrophic expenditure due to out-of-pocket health payments and its determinants in Colombian households. International journal for equity in health. 2016 Dec;15(1):1-1.
Lara JL, Gómez FR. Determining factors of catastrophic health spending in Bogota, Colombia. International journal of health care finance and economics. 2011 Jun 1;11(2):83-100.
O'Donnell O, Van Doorslaer E, Wagstaff A, Lindelow M. Analyzing health equity using household survey data: a guide to techniques and their implementation. The World Bank; 2007 Oct 27.
Filmer D, Pritchett LH. Estimating wealth effects without expenditure data—or tears: an application to educational enrollments in states of India. Demography. 2001 Feb;38(1):115-32.
Oaxaca R. Male-female wage differentials in urban labor markets. International economic review. 1973 Oct 1:693-709.
Buigut S, Ettarh R, Amendah DD. Catastrophic health expenditure and its determinants in Kenya slum communities. International journal for equity in health. 2015 Dec;14(1):1-2.
Abolhallaje M, Hasani SA, Bastani P, Ramezanian M, Kazemian M. Determinants of catastrophic health expenditure in Iran. Iranian journal of public health. 2013;42(Supple1):155.
Beusenberg M, Orley JH, World Health Organization. A User's guide to the self reporting questionnaire (SRQ. World Health Organization; 1994.
Conflict Analysis Resource Center (CERAC) 2020. Análisis de Conflictos y Violencia Política. Tipología por municipios del conflicto armado. Disponible en: http://www.cerac.org.co/es/l%C3%ADneas-de-investigaci%C3%B3n/analisis-conflicto/tipologia-por-municipios-del-conflicto-armado.html
Szumilas M. “Explaining odds ratios.” Journal of the Canadian Academy of Child and Adolescent Psychiatry 19 (3): 227.
World Bank 2016. Data Bank. World Development Indicators. In: https://databank.worldbank.org/source/world-development-indicators
Beegle K, De Weerdt J, Friedman J, Gibson J. Methods of household consumption measurement through surveys: Experimental results from Tanzania. The World Bank; 2010 Dec 1.

Table 1: Descriptive Statistics, CONPAS 2014 – 2018

Variable	CONPAS 2014 (N:1309)	CONPAS 2018 (N:1309)
	N (%)	N (%)
Concentration index	-0.151	-0.232
Catastrophic expenditure
No	925 (70.7)	907 (69.3)
Yes	384 (29.3)	402 (30.7)
Age
Under 18 years	63 (4.8)	0 (0.0)
18 – 44 years	564 (51.8)	598 (45.7)
45 – 64 years	412 (37.8)	506 (38.7)
65 years or more	113 (10.4)	205 (15.6)
Gender
Male	600 (45.8)	600 (45.8)
Female	709 (54.2)	709 (54.2)
Residence
Urban	782 (59.7)	782 (59.7)
Rural	527 (40.3)	527 (40.3)
Internal displacement
No	759 (58.0)	759 (58.0)
Yes	550 (42.0)	550 (42.0)
Work Type
Formal	280 (21.4)	204 (15.6)
Informal	906 (69.2)	977 (74.6)
Out of Labor Force	123 (9.4)	128 (9.8)
Marital Status
Married	284 (21.7)	281 (21.5)
Consensual Union	564 (43.1)	536 (40.9)
Divorced	220 (16.8)	299 (22.8)
Widow/er	72 (5.5)	94 (7.2)
Single	169 (12.9)	99 (7.5)
Conflict Level
Lived outside Meta	163 (12.4)	0 (0.0)
Capital city	282 (22.9)	300 (22.9)
Highly affected	143 (11.4)	306 (23.4)
No conflict	183 (14.3)	294 (22.5)
Lowly affected	481 (38.8)	409 (31.2)
Ethnicity
Majority	1027 (78.5)	1027 (78.5)
Minority	282 (21.5)	282 (21.5)
Education
None	79 (6.0)	79 (6.0)
Primary School	564 (43.1)	535 (40.9)
Secondary School	454 (34.7)	439 (33.5)
Undergraduate	212 (16.2)	256 (19.6)
SRQ+
No	1111 (84.9)	885 (67.6)
Yes	198 (15.1)	424 (32.4)
Sick in the previous 12 months
No	728 (55.6)	525 (40.1)
Yes	581 (44.4)	784 (59.9)
Health insurance scheme
Not affiliated	68 (5.2)	70 (5.3)
Contributive	410 (31.3)	350 (26.7)
Subsidized	771 (58.9)	829 (63.3)
Other (exception, special schemes)	60 (4.6)	60 (4.6)
Children under 5 years old (mean)	0.267	0.267
Log household size (mean)	1.078	1.078
WHODAS (mean)	3.127	4.734

Source: Own elaboration based on CONPAS 2014 – 2018

Abbreviations: SRQ+: person has a possible mental health disorder (i.e., a SRQ positive case).

WHODAS: Disability score in World Health Organization Disability Assessment Schedule

Table 2: Logistic regression model for catastrophic expenditures, 2014-2018

Variable	OR (2014)	OR (2018)
Internal displacement
No	1.00	1.00
Yes	1.33* [1.01-1.75]	1.20 [0.91-1.59]
Age group
Less than 18 years	1.77 [0.87-3.61]	N. A
18 – 44 years	1.00	1.00
45 – 64 years	1.19 [0.86-1.65]	1.38* [1.01-1.92]
65 years or more	1.99** [1.18-3.34]	1.68* [1.05-2.70]
Gender
Male	1.00	1.00
Female	1.12 [0.82-1.53]	1.21 [0.91-1.61]
Residence
Urban	1.00	1.00
Rural	1.68** [1.22-2.30]	1.47* [1.03-2.09]
Work
Formal	1.00	1.00
Informal	0.99 [0.67-1.46]	1.38 [0.87-2.18]
Out of Labor Force	0.61 [0.33-1.11]	1.34 [0.75-2.41]
Conflict level
Lived outside Meta	1.38 [0.86-2.22]
Capital city	1.00	1.00
Highly affected	1.50 [0.90-2.51]	2.80** [1.74-4.49]
No conflict	1.43 [0.91-2.26]	1.55 [0.99-2.42]
Low conflict intensity	1.50* [1.02-2.20]	2.20** [1.45-3.36]
Ethnicity
Majority	1.00	1.00
Minority	0.90 [0.66-1.23]	1.03 [0.76-1.40]
Education
None	1.00	1.00
Primary School	1.32 [0.75-2.32]	0.83 [0.49-1.42]
Secondary School	1.21 [0.65-2.26]	0.83 [0.46-1.51]
Undergraduate	1.56 [0.78-3.13]	1.19 [0.62-2.32]
Marital Status
Married	1.00	1.00
Consensual Union	0.98 [0.69-1.38]	0.83 [0.59-1.18]
Divorced	0.90 [0.58-1.38]	0.76 [0.51 - 1.16]
Widow/er	1.10 [0.60-2.03]	0.68 [0.38 -1.22]
Single	0.73 [0.42-1.26]	0.51* [0.26-0.97]
Health insurance scheme
Not affiliated	1.09 [0.59-2.02]	1.24 [0.66-2.33]
Contributive	1.00	1.00
Subsidized	0.74 [0.51-1.05]	0.77 [0.52-1.12]
Other	1.52 [0.83-2.79]	1.83 [0.98-3.39]
Sick in the previous 12 months
Yes	1.00	1.00
No	0.61** [0.47-0.80]	0.63** [0.47-0.83]
Children under 5 years old	1.09 [0.84-1.42]	1.02 [0.77-1.33]
Household size	1.11 [0.84-1.46]	1.03 [0.77-1.38]
WHODAS	1.06** [1.03-1.09]	1.04** [1.01-1.06]
SRQ+
No	1.00	1.00
Yes	1.14 [0.78-1.67]	1.20 [0.87-1.66]
Poverty Quintiles
Quintile 1	1.00	1.00
Quintile 2	0.98 [0.66-1.47]	0.94 [0.62-1.41]
Quintile 3	0.75 [0.48-1.16]	0.91 [0.57-1.46]
Quintile 4	1.13 [0.70-1.83]	0.60 [0.34-1.04]
Quintile 5	0.78 [0.46-1.35]	0.82 [0.45-1.49]

Source: Own elaboration based on CONPAS 2014 – 2018 (** pvalue<=0.01; * pvalue <=0.05)

Abbreviations: SRQ+: person has a possible mental health disorder (i.e., a SRQ positive case).

WHODAS: Disability score in World Health Organization Dissability Assessment Schedule

Table 3: Oaxaca-blinder decomposition of changes in inequalities of catastrophic expenditures, 2014-2018

Variable	Distributional effect	Coefficient effect	Total
Variable	ηΔC	CΔη	Total
Displaced	-	-	-
No	-	-	-
Yes	0.002	0.010	0.012
Age group	-	-	-
Less than 18 years	NA	-0.001	-0.001
18 – 44 years	-	-	-
45 – 64 years	0.001	-0.006	-0.005
65 years or more	-0.002	-0.000	-0.002
Gender	-	-	-
Male	-	-	-
Female	0.003	0.006	0.009
Residence	-	-	-
Urban	-	-	-
Rural	-0.018	0.010	-0.03
Work	-	-	-
Formal	-	-	-
Informal	0.013	-0.067	-0.053
Out of Labor Force	-0.002	0.016	0.014
Conflict level	-	-	-
Outside Meta	0.000	-0.002	-0.002
Capital city	-	-	-
Highly affected	-0.007	-0.050	-0.057
No conflict	0.002	0.003	0.005
Lowly affected	0.014	-0.012	0.002
Ethnicity	-	-	-
Majority	-	-	-
Minority	0.000	-0.005	-0.005
Education	-	-	-
None	-	-	-
Primary School	-0.002	0.056	0.054
Secondary School	0.004	-0.020	-0.016
Undergraduate	-0.000	-0.013	-0.013
Marital Status	-	-	-
Married	-	-	-
Consensual Union	-0.005	0.003	0.002
Divorced	0.003	0.004	0.007
Widow/er	-0.016	0.002	0.001
Single	-0.000	-0.001	-0.001
Health insurance scheme	-	-	-
Not affiliated	-0.000	-0.000
Contributive	-	-	-
Subsidized	-0.006	-0.009	-0.016
Other	0.001	0.002	0.003
Sick in the previous 12 months	-	-	-
Yes	-	-	-
No	-0.004	0.008	0.004
Children under 5 years old	0.000	-0.001	-0.001
Log household size	0.001	-0.001	0.000
WHODAS	0.003	0.000	0.003
SRQ+	-	-	-
No	-	-	-
Yes	0.003	-0.005	-0.002
Quintiles	-	-	-
Quintil 1	-	-	-
Quintil 2	0.000	0.003	0.003
Quintil 3	-0.000	0.000	-0.000
Quintil 4	-0.000	-0.038	-0.038
Quintil 5	0.000	0.004	0.004
Total	-0.007	-0.104	-0.111

Source: Own elaboration based on CONPAS 2014 – 2018

Abbreviations: SRQ+: person has a possible mental health disorder (i.e., a SRQ positive case).

WHODAS: Disability score in World Health Organization Dissability Assessment Schedule

Table 4: Robustness check – Oaxaca decomposition at different thresholds

Variable	Distributional effect – ηΔC
Variable	5%	10%	15%	25%	30%
Displaced	-	-	-	-	-
No	-	-	-	-	-
Yes	0.002	0.002	0.002	0.000	-0.000
Age group	-	-	-	-	-
Less than 18 years	0.000	0.000	0.000	0.000	0.000
18 – 44 years	-	-	-	-	-
45 – 64 years	0.000	0.000	0.001	0.001	0.001
65 years or more	-0.001	-0.001	-0.002	-0.002	-0.003
Gender	-	-	-	-	-
Male	-	-	-	-	-
Female	0.001	0.002	0.002	0.000	0.001
Residence	-	-	-	-	-
Urban	-	-	-	-	-
Rural	-0.010	-0.013	-0.014	-0.013	-0,016
Work	-	-	-	-	-
Formal	-	-	-	-	-
Informal	0.012	0.006	0.008	0.004	-0.001
Out of Labor Force	-0.002	-0.002	-0.002	-0.002	-0.000
Conflict level	-	-	-	-	-
Outside Meta	0.000	0.000	0.000	0.000	0.000
Capital city	-	-	-	-	-
Highly affected	-0.006	-0.006	-0.007	-0.008	-0,008
No conflict	0.003	0.002	0.002	0.001	0.001
Lowly affected	0.013	0.013	0.012	0.015	0.014
Ethnicity	-	-	-	-	-
Majority	-	-	-	-	-
Minority	-0.000	-0.000	-0.000	-0.000	-0.000
Education	-	-	-	-	-
None	-	-	-	-	-
Primary School	0.001	0.001	-0.002	-0.001	-0.001
Secondary School	0.000	0.000	0.005	0.003	0.002
Undergraduate	-0.001	-0.001	-0.000	-0.000	-0.000
Marital Status	-	-	-	-	-
Married	-	-	-	-	-
Consensual Union	-0.002	0.000	-0.000	-0.009	-0.011
Divorced	0.002	0.001	0.002	0.004	0.005
Widow/er	-0.000	0.000	-0.001	-0.001	-0.001
Single	-0.000	-0.000	-0.000	-0.000	-0.000
Health insurance scheme	-	-	-	-	-
Not affiliated	-0.000	-0.000	-0.000	-0.000	-0.000
Contributive	-	-	-	-	-
Subsidized	-0.008	-0.006	-0.006	-0.004	-0.000
Other	0.000	0.001	0.000	0.002	0.002
Sick in the previous 12 months	-	-	-	-	-
Yes	-	-	-	-	-
No	-0.004	-0.004	-0.004	-0.004	-0.005
Children under 5 years old	0.001	0.001	0.000	0.000	0.000
Log household size	0.000	0.000	0.001	0.005	0.005
WHODAS	0.002	0.003	0.003	0.003	0.003
SRQ+	-	-	-	-
No	-	-	-	-	--
Yes	0.001	0.001	0.002	0.003	0.005
Quntiles	-	-	-	-	-
Quintile 1	-	-	-	-	-
Quintile 2	-0.000	-0.000	-0.000	0.000	0.000
Quintile 3	0.000	0.000	0.000	-0.000	-0.000
Quintile 4	0.000	0.000	-0.000	-0.000	-0.000
Quintile 5	-0.000	-0.000	-0.000	0.000	0.000
Total
Variable	Coefficient effect – CΔη
Displaced	-	-	-	-	-
No	-	-	-	--
Yes	0.004	0.008	0.004	0.028	0.044
Age group	-	-	-	-	-
Less than 18 years	-0.000	-0.001	-0.001	-0.001	-0.002
18 – 44 years	-	-	-	-	-
45 – 64 years	-0.005	-0.004	-0.008	-0.007	-0.010
65 years or more	-0.002	-0.000	-0.001	-0.002	-0.003
Gender	-	-	-	-	-
Male	-	-	-	-	-
Female	-0.002	-0.002	0.000	0.003	0.006
Residence	-	-	-	-	-
Urban	-	-	-	-	-
Rural	-0.010	-0.016	0.009	0.000	-0.028
Work	-	-	-	-	-
Formal	-	-	-	-	-
Informal	-0.050	-0.027	-0.029	0.013	0.025
Out of Labor Force	0.005	0.007	0.014	0.015	0.015
Conflict level	-	-	-	-	-
Outside of Meta	-0.004	-0.004	-0.003	-0.002	-0.002
Capital city	-	-	-	-	-
Highly affected	-0.031	-0.034	-0.040	-0.055	-0.071
No conflict	0.004	-0.000	0.001	0.005	0.006
Lowly affected	0.005	-0.002	-0.000	-0.011	-0.013
Ethnicity	-	-	-	-	-
Majority	-	-	-	-	-
Minority	-0.004	-0.003	-0.003	-0.003	-0.006
Education	-	-	-	-	-
None	-	-	-	-	-
Primary School	0.017	0.018	0.056	0.027	0.060
Secondary School	-0.009	-0.010	-0.018	-0.019	-0.023
Undergraduate	0.002	-0.002	-0.016	-0.000	-0.007
Marital Status	-	-	-	-	-
Married	-	-	-	-	-
Consensual Union	-0.001	0.000	-0.001	0.006	0.006
Divorced	-0.002	-0.002	-0.001	0.005	0.009
Widow/er	0.002	0.001	0.002	0.003	0.005
Single	-0.000	0.002	0.002	-0.003	-0.001
Health insurance scheme	-	-	-	-	-
Not affiliated	-0.000	0.000	-0.000	0.000	-0.001
Contributive	-	-	-	-	-
Subsidized	-0.008	-0.026	-0.031	-0.033	-0.068
Other	0.002	0.002	-0.001	0.003	0.002
Sick in the previous 12 months	-	-	-	-	-
Yes	-	-	-	-	-
No	0.007	0.006	0.009	0.003	0.004
Children under 5 years old	0.000	0.001	-0.001	-0.001	-0.001
Log household size	0.001	0.001	0.000	0.000	-0.003
WHODAS	-0.002	-0.003	-0.002	0.000	0.001
SRQ+	-	-	-	-	-
No	-	-	-	-	-
Yes	-0.003	-0.003	-0.003	-0.001	-0.008
Quintiles	-	-	-	-	-
Quintile 1	-	-	-	-	-
Quintile 2	-0.018	-0.018	-0.009	0.004	0.007
Quintile 3	0.000	0.000	0.000	0.000	0.000
Quintile 4	0.021	0.023	-0.013	-0.042	-0.029
Quintile 5	0.037	0.035	0.008	-0.028	-0.007
Total

Source: Own elaboration based on CONPAS 2014 – 2018

Abbreviations: SRQ+: person has a possible mental health disorder (i.e., a SRQ positive case).

Download PDF

Editorial decision: Major Revision
15 Jul, 2021
Review #2 received at journal
11 Jul, 2021
Review #1 received at journal
24 Jun, 2021
Reviewer #2 agreed at journal
21 Jun, 2021
Reviews received at journal
04 Jun, 2021
Reviewers invited by journal
04 Jun, 2021
Reviewer #1 agreed at journal
03 Jun, 2021
Editor assigned by journal
16 May, 2021
Submission checks completed at journal
15 May, 2021
Editor invited by journal
15 May, 2021
First submitted to journal
14 May, 2021

You are reading this latest preprint version

Inequalities in Catastrophic Health Expenditures in Conflict-Affected Areas and the Colombian Peace Agreement: An Oaxaca Blinder Change Decomposition Analysis

Status:

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Conclusion

Declarations

References

Tables

Status:

Version 1