You are here

  • Home
  • Archive
  • Volume 12, Issue 4
  • Direct patient costs of maternal care and birth-related complications at faith-based hospitals in Madagascar: a secondary analysis of programme data using patient invoices

Article Text

Article menu
Global health
Original research
Direct patient costs of maternal care and birth-related complications at faith-based hospitals in Madagascar: a secondary analysis of programme data using patient invoices
  1. Mara Anna Franke1,
  2. Rinja Mitolotra Ranaivoson2,
  3. Mahery Rebaliha2,
  4. Sahondra Rasoarimanana3,
  5. Till Bärnighausen4,5,6,7,
  6. Samuel Knauss1,5,8,9,
  7. Julius Valentin Emmrich1,5,8,9
  1. 1 Charité Global Health, Charité Universitätsmedizin Berlin, Berlin, Germany
  2. 2 Doctors for Madagascar, Ejeda, Madagascar
  3. 3 SALFA Hospitals, Antananarivo, Madagascar
  4. 4 Medical Faculty, University of Heidelberg, Institute of Global Health, Heidelberg, Germany
  5. 5 University of Heidelberg, University Hospital, Heidelberg, Germany
  6. 6 Department of Global Health and Population, Chan School of Public Health, Boston, Massachusetts, USA
  7. 7 Africa Health Research Institute, Somkhele and Durban, South Africa
  8. 8 Department of Neurology with Experimental Neurology, Charité — Universitätsmedizin Berlin, Berlin, Germany
  9. 9 Berlin Institute of Health, Berlin, Germany
  1. Correspondence to Dr Julius Valentin Emmrich; julius.emmrich{at}charite.de

Abstract

Objectives We aimed to determine the rate of catastrophic health expenditure incurred by women using maternal healthcare services at faith-based hospitals in Madagascar.

Design This was a secondary analysis of programmatic data obtained from a non-governmental organisation.

Setting Two faith-based, secondary-level hospitals located in rural communities in southern Madagascar.

Participants All women using maternal healthcare services at the study hospitals between 1 March 2019 and 7 September 2020 were included (n=957 women).

Measures We collected patient invoices and medical records of all participants. We then calculated the rate of catastrophic health expenditure relative to 10% and 25% of average annual household consumption in the study region.

Results Overall, we found a high rate of catastrophic health expenditure (10% threshold: 486/890, 54.6%; 25% threshold: 366/890, 41.1%). Almost all women who required surgical care, most commonly a caesarean section, incurred catastrophic health expenditure (10% threshold: 279/280, 99.6%; 25% threshold: 279/280, 99.6%). The rate of catastrophic health expenditure among women delivering spontaneously was 5.7% (14/247; 10% threshold).

Conclusions Our findings suggest that direct patient costs of managing pregnancy and birth-related complications at faith-based hospitals are likely to cause catastrophic health expenditure. Financial risk protection strategies for reducing out-of-pocket payments for maternal healthcare should include faith-based hospitals to improve health-seeking behaviour and ultimately achieve universal health coverage in Madagascar.

  • catastrophic health expenditure
  • maternal healthcare
  • out of pocket payment
  • Sub-Saharan Africa
  • Madagascar

Data availability statement

Data are available on reasonable request. The raw data that support the findings of this study are available from the corresponding author, JVE, on request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjopen-2021-053823

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Strengths and limitations of this study

    • This study is the first to describe the rate of catastrophic health expenditure incurred by women using maternal healthcare services at faith-based hospitals in Madagascar.

    • To eliminate recall bias and increase data quality, we based study outcomes on patient invoices and medical records and not on self-reported costs or conditions.

    • We did not look at healthcare expenditure at the household level.

    • Our study was limited to two faith-based referral hospitals in rural regions of Madagascar.

    Introduction

    Reducing maternal and neonatal mortality is a key element of the United Nations Sustainable Development Goals. In 2017, approximately 295 000 women died from obstetric complications worldwide.1 Women living in sub-Saharan Africa (SSA) are over 300 times more likely to die from obstetric complications than women in the European Union.1 While there has been substantial progress in reducing all-cause maternal mortality in SSA by 39% between 2000 and 2017, this is insufficient to reach the goal of less than 70 maternal deaths per 100 000 live births by 2030.1 The main causes of maternal deaths in SSA are postpartum haemorrhage, infections, hypertensive disorders during pregnancy and abortion.2 Together, these conditions account for more than 50% of all maternal deaths in the region.2

    Among the multitude of individual and health system level obstacles in accessing quality maternal healthcare services, out-of-pocket (OOP) payments are a major reason for not seeking skilled care during pregnancy or delaying the decision of seeking care.3 4 Catastrophic health expenditure (CHE), commonly defined as OOP payments for healthcare exceeding 10% or 25% of a household’s annual income or consumption,5 can have long-term socioeconomic consequences for affected households. Poorer households are more vulnerable to financial hardship after experiencing CHE and face a high risk of being driven deeper into poverty.6 7 Poverty in turn can limit a household’s capacity to pay for future episodes of illness and the negative long-term consequences of poverty on health are widely recognised.8 Despite the widespread introduction of cost-exemption policies and the abolishment of direct patient costs for maternal healthcare in many countries in SSA, expectant mothers still face OOP payments for maternal care and birth-related complications.9–11 Diagnostic procedures, medicines and delivery supplies are the main drivers of OOP payments.9 10 In Ghana, Kenya, Tanzania and Burkina Faso, 30%–90% of women pay at least a fraction of treatment costs OOP, even after the implementation of cost-exemption schemes or the abolishment of direct patient costs for maternal care in these countries.9 12 13

    In Madagascar, 78.8% of the population of 26 million live on less than US$1.90 a day (2011 PPP).14 Poverty rates are especially high among rural communities.15 Maternal mortality remains high (335 deaths per 100 000 live births).1 Eighty-three per cent of Malagasy women report obstacles to access healthcare and women living in rural areas are particularly affected.16 The most common obstacle is a lack of financial resources and women from the poorest wealth quintile are more likely to face obstacles to seeking care (78%) than those from the richest wealth quintile (58%).16 But even in urban regions of Madagascar, 85% of households resort to coping mechanisms such as borrowing money or selling assets to cover treatment costs for maternal care.17

    The healthcare system in Madagascar is organised by different levels of specialisation. Community health workers and health centres provide primary care at the community level and refer more complicated cases to specialised district and regional centres. In 2015, the Malagasy Ministry of Health created a national pooled funding mechanism to reduce user-fees at the point of care and to increase overall healthcare utilisation.18 Maternal healthcare services including antenatal care and caesarean sections (C-sections) should be free of charge at public health facilities.19 Apart from public health facilities, faith-based hospitals play a major role in Madagascar’s healthcare system,20 including health facilities run by the Catholic and Lutheran churches. The health department of the Lutheran Church of Madagascar (SALFA) is the largest non-public healthcare provider in the country, running over 50 healthcare facilities and treating around 250 000 patients a year, mainly in rural regions.

    However, despite severe financial obstacles to accessing maternal healthcare services, data on direct treatment costs from Madagascar are extremely rare. To date, only one study reported costs for maternal care and birth-related complications obtained from a public urban tertiary hospital.17 No data are available from faith-based hospitals, even though they play an important role for providing maternal healthcare services. Thus, the extent to which direct patient costs result in CHE at faith-based hospitals is unknown. Therefore, we aimed to estimate the rate of CHE caused by direct patient costs for maternal care and birth-related complications at faith-based hospitals in two communities in the rural south of Madagascar. We expect the results of this study to guide policy development related to financial risk protection and to promote efforts towards reducing OOP payments and promoting universal health coverage in Madagascar.

    Methods

    Study area and context

    The study area was located in the rural regions of Atsimo-Andrefana and Anosy in southern Madagascar (figure 1). Atsimo-Andrefana (1.8 million inhabitants) and Anosy (833 000 inhabitants) are among the most remote regions of the island.21 The majority of the population in Atsimo-Andrefana (82%) and Anosy (84%) live below the national poverty line of US$129.56 total annual per capita consumption21 and 78.8% of the national population live below the International Poverty Line of US$1.90 per day (2011 PPP).14 The region of Atsimo-Andrefana is divided into nine health districts, served by nine public hospitals, 116 public health centres and 68 public health posts.22 The Lutheran church of Madagascar runs two district-level hospitals and three health centres in the region. Anosy region is divided into nine health districts, served by three public hospitals, 65 health centres and 14 health posts.22 The Lutheran church runs one district-level hospital and three health centres in the region. The study hospitals located in Ejeda (54 beds) and Manambaro (50 beds), run by the Lutheran church of Madagascar, offer non-surgical and surgical maternal healthcare services. Women could either directly seek care at the study hospitals or be referred from a health post or centre for further treatment. There was no public emergency referral system and women usually had to arrange their own transportation. Doctors for Madagascar, a German-Malagasy non-governmental organisation (NGO) offered direct cash support for maternal healthcare services and a referral service from selected health centres and health posts to the study hospitals free of charge.

    Figure 1
    Figure 1

    Study region. Map of the administrative regions Atsimo-Andrefana (left) and Anosy (right, marked on inset). Dots show study hospitals at Ejeda and Manambaro and health centres and health posts which referred women to study hospitals during the duration of the study.

    Data collection

    We collected data on direct patient costs which women incurred for maternal care and birth-related complications between 1 March 201910 and 7 September 2020 at the study hospitals. Our primary data sources were patient invoices and patient records containing information on patient age, diagnosis at admission, type of referral, name of the referring healthcare centre and treatment details. We obtained primary data from an NGO, which implemented a maternal health programme at the study hospitals. Healthcare staff, who were not otherwise involved in this study, replaced patient identifying information with numerical pseudonyms before forwarding digitised patient invoices and records to the research team for analysis. We collected the original data in French and translated it to English. All data were stored in a protected database.

    Exclusion and inclusion criteria

    All women receiving maternal care or care for birth-related complications at the two study hospitals during the study period were included. Women presenting with more than one admission diagnosis or seeking care for themselves and their newborns were excluded from analysis.

    Data analysis

    We coded admission diagnoses and treatment details obtained from patient invoices and medical records according to the International Classification of Diseases (ICD) V.10.23 Procedures were coded according to the Systematised Nomenclature of Medicine (SNOMED, V.2020-07-31).24 In case of uncertainty (ie, ambiguity of language), we sought clarification from two Malagasy physicians who were unfamiliar with the aim of this study. ICD and SNOMED codes of diagnoses and procedures extracted from patient invoices and medical records are summarised in online supplemental tables 1 and 2. Diagnoses not related to ‘pregnancy, childbirth and the puerperium’ (ICD Codes O00–O99) were classified as ‘other’. Direct treatment costs were collected and analysed in Malagasy Ariary and converted to USD for reporting (average annual exchange rate 2019, US$1=Ar3618.32).

    Supplemental material

    We calculated average treatment costs for each diagnosis and procedure defined by ICD and SNOMED codes, respectively. Only primary diagnosis and one procedure per woman were considered for analysis. We then calculated the proportion of women who incurred CHE per diagnosis and per procedure. In addition, we analysed the prevalence of individual diagnoses and procedures among all women incurring CHE. All analyses were performed in STATA (V.14.2, StataCorp). The results were considered statistically significant for p≤0.05.

    We used violin plots to show summary statistics as well as probability densities of the data as some treatment cost distributions had multiple peaks (ie, multimodal distribution).25 Probability densities were smoothed by a Gaussian kernel density estimator, plots were prepared using GraphPad Prism V.9.

    Definition of CHE

    To assess the prevalence of CHE households incurred for maternal care and birth-related complications in the study region, we defined CHE using total direct patient costs and annual household consumption. We used total direct patient costs as the numerator. We adopted a common convention and used total household consumption as the denominator, which better captures the effect of health expenditures on disposable income. We calculated annual household consumption for each study region by multiplying individual annual per capita consumption with the average number of household members according to most recent household survey data.26 For our purposes, a household was defined as having incurred CHEs if the direct patient costs exceeded 10% or 25% of the annual household consumption in the study region.5

    Results

    Data were available for 957 women who received maternal healthcare services at the study hospitals during the study period. There were no missing data.

    Study population

    The mean age of women was 24.3 years (SD: 7.2). The duration of hospitalisation was 5.1 days (SD 5.2) and 75.1% of women received surgical care. 35.6% were referred by ambulance. The mean distance travelled was 23.2 km (SD 18.7 km), whereas the distance was higher for women referred by ambulance (24.3 km, SD: 18.8) than for those who arranged their own means of transport (15.5 km, SD: 17 km; p=0.001). Neither age (p=0.23) nor distance travelled to the hospital were associated with direct patient costs. Sixty-two women (62/957, 6.5%) presented with more than one diagnosis; on five occasions (5/957, 0.5%) both mothers and new-borns required medical care. These cases were excluded from subsequent analysis.

    Diagnoses

    The most common ICD-10 blocks at admission were complications of labour and delivery (346/890, 38.9), followed by encounter for delivery (186/890, 21%), pregnancies with abortive outcomes (113/890, 12.7%) and maternal care related to the fetus and amniotic cavity and possible delivery problems (84/890, 9.4%, figure 2). Only a minority presented with diagnoses related to oedema, proteinuria and hypertensive disorders in pregnancy, childbirth and the puerperium (24/890, 2.7%), complications predominantly related to the puerperium (18/890, 2.0%) and other disorders predominantly related to pregnancy (8/890, 0.9%). A total of 110/890 (12.4%) of women presented with diagnoses not included in the ICD chapter pregnancy, childbirth and the puerperium. Frequency and percentages of ICD chapter blocks and pertaining diagnoses are summarised in online supplemental table 3.

    Figure 2
    Figure 2

    Direct patient costs per ICD chapter block. Violin plot depicts direct treatment costs incurred by women at the study hospitals during the study period per ICD-10 blocks and subcategories for the most common ICD chapter (complications of labour and delivery) in USD, exchange rate US$1=Ar3618.32; n=890. Median: solid line; IQR; dotted lines; kernel density estimation of data: graph surface area. ICD, International Classification of Diseases.

    Direct treatment costs

    Direct treatment costs per diagnostic category

    On average, women spent US$100.39 (SD: 83.23) per case on maternal healthcare services. Direct treatment costs for encounter for delivery ranged from US$9.92 to US$237.46. Direct treatment costs for abortion and eclampsia ranged between US$8.48 and US$389.63 and US$15.70 to US$237.46, respectively. Overall, we observed a high variation of direct treatment costs among individual categories. The highest variation occurred in the category ‘molar pregnancy’ (SD 120.14) and the lowest in the category’“spontaneous vaginal delivery’ (SD 8.63, figure 2).

    Direct treatment costs per procedure

    The most common procedures were C-section (n=280) and spontaneous vaginal delivery (n=247). The mean costs of a C-section and spontaneous vaginal delivery were US$191.29 (SD: 38.47) and US$30.40 (SD: 18.29), respectively. The most expensive procedures were hysterectomy (n=3; US$249.98, SD: 26.89) and laparotomy (n=12) performed due to extrauterine gravidity, uterus rupture or haemorrhage (US$209.84, SD: 32.11, figure 3).

    Figure 3
    Figure 3

    Direct patient costs per procedure. Violin plot depicts direct treatment costs incurred by women at the study hospitals during the study period per procedure in United States Dollars (USD), exchange rate US$1=Ar3618.32; n=645. procedures were assigned according to the Systematised Nomenclature of medicine clinical terms (SNOMED CT, V.2020-07-31). median: solid line; IQR: dotted lines; kernel density estimation of data: graph surface area. procedures are shown if n>5.

    Direct treatment costs for non-surgical care

    Two-hundred and twenty-six women received non-surgical care. The most common diagnostic block in this group were pregnancies with abortive outcomes (n=52), 30 of which were uncomplicated abortions. Complications of labour and delivery (n=28) including placenta retention and postpartum haemorrhage were the second most common diagnostic block. All women receiving non-surgical care for complications related to the puerperium (n=18) had a postpartum infection. The most common non-surgical treatments were antibiotics (n=199, 88.1%) and analgesia (n=131, 58%). Among the same diagnostic block direct patient costs were significantly higher for women requiring surgery than for those receiving non-surgical treatment (p<0.005). Figure 4 illustrates these findings.

    Figure 4
    Figure 4

    Direct treatment costs for surgical and non-surgical care per ICD chapter block. Violin plot depicts direct treatment costs incurred by women at the study hospitals during the study period grouped by ICD-10 blocks and by type of treatment in USD, exchange rate US$1=Ar3618.32 ; n=677. Median: solid line; IQR: dotted lines; kernel density estimation of data: graph surface area. ICD blocks in which all patients received the same type of care (surgical or non-surgical) were excluded from the graph. ICD, International Classification of Diseases.

    Catastrophic health expenditure

    The average annual consumption per household was Ar1 575 840 (US$435.52) and Ar1 613 280 (US$445.86) in Anosy and Atsimo-Andrefana, respectively.26 Patient expenditures were assumed to be catastrophic when exceeding Ar157 584 (US$43.55) and Ar161 328 (US$44.59) using the 10% threshold and Ar393 960 (US$108.88) and Ar403 320 (US$111.47) using the 25% threshold in Anosy and Atsimo-Andrefana, respectively.

    Overall, 486 of 890 women (54.6%) faced CHE at 10% and 366 of 890 (41.1%) at 25% threshold. Four procedures did not cause CHE at 10% or 25% threshold: blood transfusion, vaginal breech delivery, vaginal twin delivery and drainage of pleural cavity. Women who delivered spontaneously faced CHE at 10% in 5.7% (14/247) and at the 25% threshold in 1.2% (3/247) of cases. The most common cause of CHE was C-section at 10% threshold (279/280, 99.6%) and 25% threshold (279/280, 99.6%). The second most likely cause was excision of the adnexa of the uterus at 10% threshold (42/49, 85.7%) and 25% threshold (41/49, 83.7%). The percentages of CHE at 10% and 25% thresholds by ICD-10 blocks and surgical procedures are shown in figure 5.

    Figure 5
    Figure 5

    Catastrophic health expenditure by ICD chapter block and by procedure. (A) Percentage of women incurring catastrophic health expenditure for each ICD-10 chapter block and (B) by procedure (classified according to the systematised nomenclature of medicine clinical terms SNOMED CT (V.2020-07-31)), relative to 10% and 25% of annual household consumption during the study period; n=890. Error bars depict 95% CI of the mean. ICD, International Classification of Diseases.

    Discussion

    This study analysed direct patient costs of maternal healthcare services and estimated the rate of CHE incurred at two district-level faith-based hospitals in rural Madagascar using patient invoices and medical records as the data source. Our study revealed two main findings. First, the rate of CHE was high. Second, we found considerable variation in costs per diagnosis and treatment.

    The overall rate of CHE (10% threshold: n=486/890, 54.6%; 25% threshold: n=366/890, 41.1%) was higher than in most other studies from SSA, which range from 33% in Ghana (relative to 5% of annual household expenditure) to 47% in Cameroon (relative to 20% of total household income).13 27 However, differences in study methodology may limit comparability. Women who delivered spontaneously rarely faced CHE (14/247, 5.7% at 10% threshold; 3/247, 1.2% at 25% threshold). This is in line with findings from a recent review, which reported CHE being caused by vaginal deliveries in only one out of 12 countries in SSA.28 In contrast, 71.4% (247/346) and 63.3% (219/346) of women who were treated for birth-related complications including obstructed labour (208/346), placental retention (32/346), and foetal distress (20/346) faced CHE at 10% and 25% threshold, respectively. In addition, women who required surgical care (n=664) were at high risk of CHE (10% threshold: n=383/664, 57.7%; 25% threshold: n=347/664, 52.3%). Emergency obstetric care is likely to cause CHE in SSA, with C-sections being a particularly expensive component of care.28–30 However, the rate of CHE as a result of emergency obstetric care including surgical care was higher in our study compared with other countries. In Uganda and Mali, only 25% and 21% of households face CHE (relative to 10% of annal household expenditure and 15% of annual household income, respectively) due to a C-section or emergency obstetric care.6 31 In Ghana, 33% of households face CHE because of obstetric complications (relative to 5% of annual household expenditure).13 The high rate of CHE in our study was most likely due to the composition of the patient sample. We used data from two faith-based hospitals serving as reference hospitals, where the majority of women were treated for birth-related complications. These cases often required surgical care, which is usually more expensive. Costs of a C-section in our study were also higher than those reported from an urban public healthcare provider in Madagascar (US$191 vs US$161 (corrected for inflation).17 This is in line with the finding that treatment provided by faith-based hospitals compared with public hospitals may generally be more expensive in SSA.32 The high proportion of people living below the international poverty line of US$1.90 per day in Madagascar (78.8%) compared with Ghana (13%), Uganda (41.5%) and Mali (50.3%) may as well have contributed to this finding.14

    The overall direct patient costs for maternal healthcare services including spontaneous vaginal delivery, C-section, abortion, eclampsia and maternal haemorrhage were similar to other reports from countries in SSA, such as Ghana, Tanzania and Uganda.28 The direct patient costs were lowest for spontaneous vaginal delivery, as it requires few drugs and consumables and brief hospitalisation. Direct patient costs were highest for birth-related complications, which were likely to require a C-section. Direct patient costs for a C-section were more than six times higher than those for a spontaneous vaginal delivery. This ratio was similar to data from a public urban hospital in the DRC33 but lower compared with data from rural and urban primary and secondary health facilities in Burkina Faso where costs for a C-section are up to 27 times higher than those for a spontaneous vaginal delivery.10 More generally, OOP payments for emergency obstetric care—including surgical and non-surgical treatment—are two to six times higher than for spontaneous vaginal delivery in Burkina Faso, Ghana and Benin.10 34

    OOP payments for birth-related complications may pose a significant hindrance for women seeking life-saving care, especially for those from poor socioeconomic backgrounds.35 Two-thirds of Malagasy women report direct patient costs as an obstacle to seeking care.16 Similar findings have been reported from various other settings across SSA.4 Antenatal care visits increased by 25% in a rural district in Madagascar when user fees were abolished, further supporting the notion that pregnant women forgo seeking qualified care due to direct patient costs.18 Interestingly, the treatment costs of some birth-related complications for which women received a surgical intervention were capped at US$237.46 and US$201.75 at Manambaro and Ejeda hospital, respectively. This likely indicates that direct patient costs were capped by faith-based hospitals to lower financial obstacles to accessing care.

    Direct patient costs varied greatly among individual diagnoses and procedures. The costs for spontaneous vaginal delivery in our study ranged from US$9.92 to US$193.46 which was higher than variations reported from Uganda (US$2.7–US$33.90), Malawi (US$10.20 –US$24.00) or Ghana (US$7.70–US$14.60).36 The costs for a C-section varied between US$111.05 and US$415.86, which was similar to those at an urban public healthcare provider in Madagascar where costs for a C-section range from US$5 to US$351.17 The variation of costs could be caused by multiple factors. First, differences in treatment protocols or surgical practice between the study hospitals may have influenced the services provided and the length of hospital stay. Second, the availability of resources including drugs and consumables may have affected direct patient costs. This may be aggravated by drug shortages and rising drug prices during the COVID-19 pandemic, which have been reported in other countries in SSA37 and were empirically observed in Madagascar. Third, identical procedures were performed for a wide range of pathologies, indicating that differences in resource requirements may have contributed to cost variation. Last, the average inflation rate of 5.3% over the study period could additionally influence the variation of costs across cases.38 The low predictability of costs is further compounded by informal payments at health facilities39 or stock-outs at hospital pharmacies.40

    Our study has limitations. First, we used two different data sets to obtain information on household consumption and health expenditure, which may lead to an inaccurate estimate of CHE. This is because we conducted a secondary analysis of programme data, which did not contain information on an individual household’s income or consumption. Instead, we used consumption data from the most recent DHS survey. As the study hospitals were the only secondary-level hospitals providing maternal and newborn healthcare in the study region, DHS data likely reflect the sociodemographic characteristics of patients’ households. Second, we calculated the prevalence of CHE using one-time health expenditure incurred by a single household member. This likely results in an underestimation of the true prevalence of CHE. Third, our analysis did not consider indirect costs of maternal healthcare, such as for transport, food or clothing. However, transport costs play a major role in causing CHE and deter patients from seeking care.13 16 35 Fourth, we obtained our data from patient invoices which were issued by the hospitals, not directly from patients. While healthcare personnel are likely to have a great interest in accurately invoicing expensive treatments, they might not have done so for low-cost treatments as they might have been deterred by the efforts of reporting. Thus, the rate of CHE might be higher than estimated. Fifth, coding of diagnoses was done retrospectively from free text patient invoices and medical records by the data analyst. Therefore, even though we sought clarification with healthcare personnel when necessary, this may have introduced coding errors. Last, data were only available from faith-based hospitals in rural regions and limited to the south of Madagascar. However, faith-based hospitals are an important pillar of the Malagasy healthcare system and no other study has previously reported data on direct patient costs for maternal healthcare from faith-based hospitals in Madagascar.

    Conclusions

    Overall, our findings suggest that direct patient costs of managing maternal care and birth-related complications at faith-based hospitals in rural Madagascar are very commonly catastrophic relative to annual household consumption. OOP payments for maternal healthcare are likely to contribute to high poverty levels in Madagascar and may deter women from seeking care. Effective policies to reduce OOP payments may, in turn, alleviate poverty, promote health-seeking behaviour and ultimately contribute to reducing maternal mortality in Madagascar.

    Data availability statement

    Data are available on reasonable request. The raw data that support the findings of this study are available from the corresponding author, JVE, on request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Ethical approval for this study was obtained from the Ethics Committee of the Medical Faculty of Heidelberg University (S-713/2020). In addition, we obtained formal approval to conduct this secondary analysis of de-identified NGO data from the district health office (a regional subdivision of the Malagasy Ministry of Health) in Atsimo-Andrefana.

    Acknowledgments

    We thank midwives Maminiaiko Jeannot Esmeraldine Razanatiana, Nirina Roseline Rasoanandrasana, Jeannette Michelline Razanamalala, Larissa Faranirina Razakamandimby from Ejeda and Manambaro hospitals for collecting the data that support the findings of this study.

    References

    1. ↵
      1. World Health Organization
      . Maternal mortality: Levels and trends 200 to 2017 [internet]., 2019. Available: Availablefromhttps://www.who.int/publications/i/item/9789241516488 [Accessed 3 March 2021].
    2. ↵
      2. Say L ,
      3. Chou D ,
      4. Gemmill A , et al
      . Global causes of maternal death: a who systematic analysis. Lancet Glob Health 2014;2:e323–33.doi:10.1016/S2214-109X(14)70227-X pmid:http://www.ncbi.nlm.nih.gov/pubmed/25103301
      OpenUrlPubMed
    3. ↵
      2. Kyei-Nimakoh M ,
      3. Carolan-Olah M ,
      4. McCann TV
      . Access barriers to obstetric care at health facilities in sub-Saharan Africa—a systematic review.. Syst Rev6 2017;110.
    4. ↵
      2. Geleto A ,
      3. Chojentaa C ,
      4. Musa A , et al
      . Barriers to access and utilization of emergency obstetric care at health facilities in sub-Saharan Africa: a systematic review of literature. Syst Rev7, 2018;183(:183.doi:10.1186/s13643-018-0842-2 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30424808
      OpenUrlPubMed
    5. ↵
      1. World Health Organization
      . Global Monitoring Report on Financial Protection in Health 2019. In: World Health organization and international bank for reconstruction and development / the world bank Licence: CC BY-NC-SA 3.0 IGO Internet, 2020. https://www.who.int/healthinfo/universal_health_coverage/report/fp_gmr_2019.pdf?ua=1
    6. ↵
      2. Arsenault C ,
      3. Fournier P ,
      4. Philibert A , et al
      . Emergency obstetric care in Mali: catastrophic spending and its impoverishing effects on households. Bull World Health Organ 2013;91:207–16.doi:10.2471/BLT.12.108969 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23476093
      OpenUrlPubMed
    7. ↵
      2. Wagstaff A ,
      3. Eozenou P ,
      4. Smitz M
      . Out-Of-Pocket expenditures on health: a global Stocktake. World Bank Res Obs 2020;35:123–57.doi:10.1093/wbro/lkz009
      OpenUrl
    8. ↵
      1. World Health Organization
      . Closing the gap in a generation: health equity through action on the social determinants of health - Final report of the commission on social determinants of health [internet], 2008. Available: https://www.who.int/publications-detail-redirect/WHO-IER-CSDH-08.1 [Accessed 14 December 2020].
    9. ↵
      2. Dalinjong PA ,
      3. Wang AY ,
      4. Homer CSE
      . The operations of the free maternal care policy and out of pocket payments during childbirth in rural Northern Ghana. Health Econ Rev7, 2017;41:41.doi:10.1186/s13561-017-0180-4 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29168019
      OpenUrlPubMed
    10. ↵
      2. Meda IB ,
      3. Baguiya A ,
      4. Ridde V , et al
      . Out-Of-Pocket payments in the context of a free maternal health care policy in Burkina Faso: a national cross-sectional survey. Health Econ Rev9, 2019;9:2019.doi:10.1186/s13561-019-0228-8 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30919219
      OpenUrlPubMed
    11. ↵
      2. Kaiser JL ,
      3. McGlasson KL ,
      4. Rockers PC , et al
      . Out-Of-Pocket expenditure for home and facility-based delivery among rural women in Zambia: a mixed-methods, cross-sectional study. Int J Womens Health 2019;11:411–30.doi:10.2147/IJWH.S214081 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31447591
      OpenUrlPubMed
    12. ↵
      2. Perkins M ,
      3. Brazier E ,
      4. Themmen E , et al
      . Out-Of-Pocket costs for facility-based maternity care in three African countries. Health Policy Plan 2009;24:289–300.doi:10.1093/heapol/czp013 pmid:http://www.ncbi.nlm.nih.gov/pubmed/19346273
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      2. Dalaba MA ,
      3. Akweongo P ,
      4. Aborigo RA , et al
      . Cost to households in treating maternal complications in northern Ghana: a cross sectional study. BMC Health Services Research15,34 2015;15.doi:10.1186/s12913-014-0659-1
    14. ↵
      1. The World Bank
      . Data from: poverty headcount ratio at (2011 ppp) (% of population), 2021. Available: https://data.worldbank.org/indicator/SI.POV.DDAY?end=2017&locations=1W&name_desc=true&start=1981&view=chart [Accessed 11 March 2021].
    15. ↵
      1. The World Bank
      . The face of poverty in Madagascar [internet]. The World Bank Group 2014 http://documents1.worldbank.org/curated/en/538821468271809604/pdf/781310PRIORITY0English0Apr900May012.pdf
    16. ↵
      1. Institut National de la Statistique Madagascar
      . Enquête nationale sur le suivi des objectifs du millénaire pour le développement Madagascar [internet], 2013. Available: https://www.instat.mg/wp-content/uploads/2016/11/INSTAT_Ensomd_Resume-2012-2013.pdf [Accessed 14 December 2020].
    17. ↵
      2. Honda A ,
      3. Randaoharison PG ,
      4. Matsui M
      . Affordability of emergency obstetric and neonatal care at public hospitals in Madagascar. Reprod Health Matters 2011;19:10–20.doi:10.1016/S0968-8080(11)37559-3 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21555082
      OpenUrlCrossRefPubMed
    18. ↵
      2. Garchitorena A ,
      3. Miller AC ,
      4. Cordier LF
      . In Madagascar, Use of Health Care Services Increased When Fees Were Removed:. In: Health AFF. 36(8). Lessons For Universal Health Coverage, 2017. (published Online First: 1 August 2017).
    19. ↵
      2. Ramamonjisoa DE ,
      3. Lang E
      . Health Financing Innovations in Madagascar on the Path to Universal Health Coverage [internet], 2018. Available: http://www.healthpolicyplus.com/ns/pubs/8211-8373_HPPlusMadagascarUHCBrief.pdf [Accessed 10 January 2021].
    20. ↵
      2. Harimanana A ,
      3. Barennes H ,
      4. Reinharz D
      . Organizational analysis of maternal mortality reduction programs in Madagascar. Int J Health Plann Manage 2011;26:e186–96.doi:10.1002/hpm.1077 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21796678
      OpenUrlPubMed
    21. ↵
      1. Institut National de la Statistique Madagascar
      . Troisième Recensement général de la population et de l’habitation [internet], 2019. Available: https://www.instat.mg/wp-content/uploads/Rapport-Prelim-2019_ver_final.pdf [Accessed 30 January 2021].
    22. ↵
      2. Maina J ,
      3. Ouma PO ,
      4. Macharia PM , et al
      . A spatial database of health facilities managed by the public health sector in sub Saharan Africa. Sci Data 2019;6:134.doi:10.1038/s41597-019-0142-2 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31346183
      OpenUrlPubMed
    23. ↵
      1. World Health Organization
      . Data from: ICD-10: international statistical classification of diseases and related health problems: tenth revision. 2, 2019. https://apps.who.int/iris/handle/10665/42980
    24. ↵
      1. International Health Terminology Standards Development Organisation
      . SNOMED CT. In: 2019.
    25. ↵
      2. Hintze JL ,
      3. Nelson RD
      . Violin plots: a box Plot-Density trace synergism. The American Statistician 1998;52:181–4.
      OpenUrlCrossRefWeb of Science
    26. ↵
      1. Institut National de la Statistique Madagascar
      . Enquête périodique auprès des ménages[internet], 2010. Available: https://www.instat.mg/wpcontent/uploads/2016/11/INSTAT_Epm2010-08-2011.pdf [Accessed 28 November 2020].
    27. ↵
      2. Laisin I ,
      3. Ndamsa DT ,
      4. Njong MA
      . The implications of healthcare utilization for catastrophic health expenditure in Cameroon. Journal of Economics and Finance 2020;11:22–36.
      OpenUrl
    28. ↵
      2. Mori AT ,
      3. Binyaruka P ,
      4. Hangoma P , et al
      . Patient and health system costs of managing pregnancy and birth-related complications in sub-Saharan Africa: a systematic review. Health Econ Rev 2020;10:26.doi:10.1186/s13561-020-00283-y pmid:http://www.ncbi.nlm.nih.gov/pubmed/32803373
      OpenUrlPubMed
    29. ↵
      2. Storeng KT ,
      3. Baggaley RF ,
      4. Ganaba R , et al
      . Paying the price: the cost and consequences of emergency obstetric care in Burkina Faso. Soc Sci Med 2008;66:545–57.doi:10.1016/j.socscimed.2007.10.001 pmid:http://www.ncbi.nlm.nih.gov/pubmed/18061325
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      2. Mengistu T ,
      3. Berruti A ,
      4. Krivelyova A
      . Cost of providing emergency obstetric care in Tanzania’s Kigoma region. Int J of Health Plann Manage 2019;34(4:e1510–9.
      OpenUrl
    31. ↵
      2. Anderson GA ,
      3. Ilcisin L ,
      4. Kayima P , et al
      . Out-Of-Pocket payment for surgery in Uganda: the rate of impoverishing and catastrophic expenditure at a government Hospital. PLoS One;12:e0187293.doi:10.1371/journal.pone.0187293
    32. ↵
      2. Olivier J ,
      3. Tsimpo C ,
      4. Gemignani R , et al
      . Understanding the roles of faith-based health-care providers in Africa: review of the evidence with a focus on magnitude, reach, cost, and satisfaction. Lancet 2015;386:1765–75.doi:10.1016/S0140-6736(15)60251-3 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26159398
      OpenUrlCrossRefPubMed
    33. ↵
      2. Ntambue AM ,
      3. Malonga FK ,
      4. Dramaix-Wilmet M , et al
      . Commercialization of obstetric and neonatal care in the Democratic Republic of the Congo: a study of the variability in user fees in Lubumbashi, 2014. PLoS One 2018;13:e0205082.doi:10.1371/journal.pone.0205082 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30304060
      OpenUrlPubMed
    34. ↵
      2. Borghi J et al
      . Costs of near-miss obstetric complications for women and their families in Benin and Ghana. Health Policy Plan 2003;18:383–90.doi:10.1093/heapol/czg046
      OpenUrlCrossRefPubMedWeb of Science
    35. ↵
      2. Ravit M ,
      3. Philibert A ,
      4. Tourigny C , et al
      . The hidden costs of a free caesarean section policy in West Africa (Kayes region, Mali). Matern Child Health J 2015;19:1734–43.doi:10.1007/s10995-015-1687-0 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25874875
      OpenUrlPubMed
    36. ↵
      2. Levin A ,
      3. Dmytraczenko T ,
      4. McEuen M , et al
      . Costs of maternal health care services in three anglophone African countries. Int J Health Plann Manage 2003;18:3–22.doi:10.1002/hpm.690 pmid:http://www.ncbi.nlm.nih.gov/pubmed/12683270
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      2. Ahmed SAKS ,
      3. Ajisola M ,
      4. Azeem K , et al
      . Impact of the societal response to COVID-19 on access to healthcare for non-COVID-19 health issues in slum communities of Bangladesh, Kenya, Nigeria and Pakistan: results of pre-COVID and COVID-19 lockdown stakeholder engagements. BMJ Glob Health 2020;5.doi:10.1136/bmjgh-2020-003042 pmid:http://www.ncbi.nlm.nih.gov/pubmed/32819917
    38. ↵
      1. Institut national de la statistique
      . Compte nationaux rebases. In: 2020. https://www.instat.mg/telechargements/publications-mensuelles/comptes-nationaux-annuels/. (16 January 2021).
    39. ↵
      2. Spangler SA ,
      3. Bloom SS
      . Use of biomedical obstetric care in rural Tanzania: the role of social and material inequalities. Soc Sci Med 2010;71:760–8.doi:10.1016/j.socscimed.2010.05.025 pmid:http://www.ncbi.nlm.nih.gov/pubmed/20579797
      OpenUrlPubMed
    40. ↵
      2. Bayo P ,
      3. Belaid L ,
      4. Tahir EO , et al
      . "Midwives do not appreciate pregnant women who come to the maternity with torn and dirty clothing": institutional delivery and postnatal care in Torit County, South Sudan: a mixed method study. BMC Pregnancy Childbirth 2020;20:250.doi:10.1186/s12884-020-02910-2 pmid:http://www.ncbi.nlm.nih.gov/pubmed/32345240
      OpenUrlPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • SK and JVE contributed equally.

    • Contributors MAF, SK and JVE developed the study design in collaboration with RMR, MR, SR and TB. RMR, MR and MAF collected the data. MAF, SK and JVE contributed to the analysis. MAF wrote the first draft of the manuscript. All authors contributed to the manuscript. All authors read and approved the final manuscript. JVE is responsible for the overall content as the guarantor.

    • Funding This work was supported by the German Federal Ministry of Education and Research [grant number 01KA1922] and the BIH-Charité Digital Clinician Scientist Program funded by the Charité – Universitätsmedizin Berlin and the Berlin Institute of Health.

    • Map disclaimer The inclusion of any map (including the depiction of any boundaries therein), or of any geographic or locational reference, does not imply the expression of any opinion whatsoever on the part of BMJ concerning the legal status of any country, territory, jurisdiction or area or of its authorities. Any such expression remains solely that of the relevant source and is not endorsed by BMJ. Maps are provided without any warranty of any kind, either express or implied.

    • Competing interests None declared.

    • Patient and public involvement Neither patients nor the public were involved in study development, choice of outcome measures and patient recruitment or any other aspect of this study.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.