To analyse life-threatening obstetric complications that occurred in public hospitals in Argentina.
To analyse life-threatening obstetric complications that occurred in public hospitals in Argentina.
Multicentre collaborative cross-sectional study.
Twenty-five hospitals included in the Perinatal Network of Buenos Aires Metropolitan Area.
Women giving birth in participating hospitals during a 1-year period.
All cases of severe maternal morbidity (SMM) and maternal mortality (MM) during pregnancy (including miscarriage and induced abortion), labour and puerperium were included. Data were collected prospectively.
Identification criteria, main causes and incidence of SMM; case-fatality rates, morbidity–mortality index and effective intervention's use rate.
A total of 552 women with life-threatening conditions were identified: 518 with SMM, 34 with MM. Identification criteria for SMM were case-management (48.9%), organ dysfunction (15.2%) and mixed criteria (35.9%). Incidence of SMM was 0.8% (95% confidence interval [95% CI] 0.73–0.87%) and hospital maternal death ratio was 52.3 per 100 000 live births (95% CI 35.5–69.1). Main causes of MM were abortion complications and puerperal sepsis; main causes of SMM were postpartum haemorrhage and hypertension. Overall case-fatality rate was 6.2% (95% CI 4.4–8.6): the highest due to sepsis (14.8%) and abortion complications (13.3%). Morbidity-mortality index was 15:1 (95% CI 7.5–30.8). Use rate of known effective interventions to prevent or treat main causes of MM and SMM was 52.3% (95% CI 46.9–57.7).
This study describes the importance of life-threatening obstetric complications that took place in public hospitals with comprehensive obstetric care and the low utilisation of known effective interventions that may decrease rates of SMM and MM. It also provides arguments that justify the need to develop a surveillance system for SMM.
Maternal mortality (MM) is among the most important indicators of global health inequities. Preventable MM results from deficits in access to and quality of health services, and from the failure of public policies to promote women's rights.[1, 2] Maternal Mortality Ratio has been selected to evaluate progress towards Millennium Development Goal 5. Although some countries have significantly decreased Maternal Mortality Ratio, many others have not reached their goals.[3, 4]
In 2010, there were an estimated 287 000 maternal deaths worldwide, yielding a Maternal Mortality Ratio of 260 maternal deaths per 100 000 live births (LB). In Argentina, the Maternal Mortality Ratio in 2010 was 44 per 100 000 LB, with significant differences between provinces, and main causes being miscarriage and induced abortion complications (20.5%), infections/sepsis (10.9%), hypertensive pregnancy disorders (11.2%), haemorrhages (12.7%), and other direct (15.7%) and indirect (29.0%) obstetric causes. These events are influenced by structural flaws in health systems, quality of care, accessibility to health services and management of obstetric emergencies. The risk for maternal death doubles with every increase in age of 10 years, and is ten times higher in small hospitals (<1500 deliveries per year).
Reviews of MM are necessary to understand the epidemiology and causes of the life-threatening complications that precede maternal deaths, including cases of severe maternal morbidity (SMM).[8-16] Nearly half of maternal deaths could be prevented with improvements in the quality of care that may result from these reviews.[17-21]
Even when knowledge exists, health providers often fail to translate this knowledge into beneficial practices to reduce MM and poor obstetric outcomes.[22-25]
The aim of this study is to present a situation analysis of the incidence and main causes of MM and SMM and use of known effective interventions in a network of public hospitals in Argentina.
This is a multicentre collaborative cross-sectional study conducted in 25 hospitals in the Perinatal Network of Buenos Aires Metropolitan Area (AMBA). The network was established in 2004 to improve the quality of care of women in maternity hospitals. This network (30 hospitals) conducts nearly 90 000 deliveries per year (1100 to 7000 per hospital) and provides comprehensive emergency obstetric care and perinatal training programmes.
Case identification was prospective during 1 year (1 June 2008 to 31 May 2009). All cases of life-threatening obstetric complications including SMM (identified using management criteria and organ dysfunction, see Supplementary material, Figure S1) and MM associated with complications during pregnancy (including miscarriage and induced abortions), labour and the puerperium were included through an obligatory consecutive sampling.
Exclusion criteria included: accidents, homicide or suicide, cases that took place more than 42 days after the end of pregnancy. Twenty-five out of 30 hospitals were included; five were excluded because of lack of research capacities.
A pilot study was carried out to develop methods used for data collection, train hospital research teams, generate forms for case notification, and create an operations manual (see Figures S2 and S3). Cases were identified through daily record reviews of women admitted to: Obstetrics and Gynaecology, Intensive Care Unit (ICU), Clinical Care, Surgery and Coronary Care Unit (CCU). After case identification, the coordinating centre was notified by e-mail within 48 hours. Following case resolution (death, discharge or referral), the data collection form was sent to the research coordinating centre to determine the main causes of the event and to evaluate the use of known effective interventions for case management. Clinical audits were carried out to evaluate the quality of case notification by hospital. To avoid selection bias and under-reporting, lists of discharges, admissions to ICU and blood transfusions of all women between 10 and 49 years were reviewed. Clinical records of those cases likely to be MM or SMM were analysed for final confirmation.
Reported outcomes included hospital maternal death ratio, incidence rate of SMM, identification criteria for SMM, proportional distribution of principal causes of MM and SMM, overall mortality and fatality rates, cause-specific fatality rate, overall morbidity–mortality index, cause-specific morbidity–mortality index, and the specific and overall utilisation rates of known effective interventions to reduce MM and SMM.
The information available for all MM and SMM cases was entered by two independent data-entries using the Epi-Info 2000 database (Centers for Disease Control, Atlanta, GA, USA). Descriptive statistics (rates and ratios) were calculated with their respective 95% confidence intervals (95% CI) and frequency distributions of MM and SMM causes. To estimate the 95% CI for Maternal Mortality Ratio and incidence rate of SMM, we assumed a Poisson distribution. STATA version 8.0 (Stata Corp, College Station, TX, USA) statistical software was used for analysis.
The World Health Organization International Classification of Diseases, 10th revision, was used to classify the principal causes of MM and SMM. Six known effective interventions to reduce MM or SMM were selected for analysis (see, Figure S4).[29-35] World Health Organization classifications were used for outcomes and operational definitions (see Figure S5).[6, 28, 36-38]
For comparative purposes, a comprehensive literature search was conducted using Medline and Lilacs (from January 1989 to February 2011). Studies were excluded when there was no mention of criteria to identify cases of SMM; full texts could not be found, and incidences of SMM could not be estimated.
The study protocol was approved by the institutional review board of the Durand Hospital with the written agreement of participating hospitals. Confidentiality for data on participating women and hospitals was preserved.
During the study period, a total of 65 033 live births were registered and 996 women with potentially life-threatening conditions were identified (1.5% of the total population); of which 552 (0.85%) were women with life-threatening conditions, 518 (0.8%) resulted in SMM and 34 (0.05%) ended in MM.
SMM and MM represented 93.8% and 6.2% of all women with life-threatening conditions respectively. Figure 1 summarises the sequence of events associated with the reproductive process.
Table 1 describes the sociodemographic characteristics of the studied population. The mean age was 27.4 years (range 13–45 years), 26.6% were nulliparous and 37.5% had three or more previous gestations. 35% had complications in previous pregnancies; 17.2% of pregnancies ended before 22 weeks of gestation: 91.5% were miscarriages and induced abortions and 8.5% were ectopic pregnancies. Among pregnancies of more than 22 weeks of gestation, 80.1% ended in caesarean delivery. Finally, 4.3% of women remained pregnant after the event of SMM. The highest proportion of events took place during the puerperium (41.8%) followed by antepartum (38.2%). Seventy-three percent of cases were identified in the emergency room, while nearly a quarter were identified in the ICU/CCU (Table 2).
|N = 552 n (%)|
|Foreigners, n (%)||128 (23)|
|Age, median (range)||27.4 (13–45)|
|<20 years||89 (16.1)|
|20–35 years||343 (62.1)|
|>35 years||120 (21.8)|
|Parity, n (%)|
|Multiparous (>3 previous deliveries)||207 (37.5)|
|Residence, n (%)|
|Buenos Aires||70 (12.6)|
|Buenos Aires Province||467 (84.6)|
|Educational level, n (%)|
|No formal education||8 (1.4)|
|Civil status, n (%)|
|Stable relationship||396 (71.7)|
|n (%) (n = 552)|
|Type of delivery|
|Gestational age <22 weeks||95 (17.2)|
|Laparotomy for ectopic pregnancy||8 (8.5)|
|Gestational age >22 weeks||433 (78.5)|
|Vaginal delivery||82 (18.9)|
|Caesarean delivery||347 (80.1)|
|Instrumental delivery||4 (1.0)|
|Still pregnant||24 (4.3)|
|Time of the reproductive process where cases were identified a|
|Intrapartum (labour and delivery)||87 (20.0)|
|Sector of the healthcare facility where cases were identified|
|Obstetrics and Gynaecology||402 (73.0)|
|Intensive Care Unit/Coronary Unit||126 (22.8)|
|Clinical Medicine||19 (3.4)|
The incidence of SMM was 0.8% (95% CI 0.73–0.87%) and hospital maternal death ratio was 52.3 per 100 000 LB (95% CI 35.5–69.1 per 100 000 LB). The overall case-fatality rate was 6.2% (95% CI 4.4–8.6%) and the overall morbidity–mortality index was 15:1 (95% CI 7.5–30.8%).
Table 3 shows the frequency in the use of criteria (management-based or organ dysfunction) for SMM case identification. Blood transfusion was the most frequent organ dysfunction criterion used and admission to ICU was the most frequently used management criterion.
|Type of criterion ( n = 518)|
|Case management criteria||254 (48.9)|
|Organ dysfunction criteria||79 (15.2)|
|Mixed criteria (both case management and organ dysfunction criteria)||187 (35.9)|
|Case management criteria ( n = 441) a|
|Admission to the ICU||284 (64.4)|
|Emergency hysterectomy||56 (12.7)|
|Both criteria combined||101 (22.9)|
|Organ dysfunction criteria ( n = 266) b|
|Pulmonary oedema (clinical diagnosis associated with intravenous furosemide requirement and/or intubation)||8|
|Hypovolaemia requiring five or more units of packed red blood cells||140|
|Admission to the ICU for sepsis||46|
|Emergency hysterectomy for sepsis||26|
|Intubation and ventilation for more than 60 minutes except for general anaesthesia||43|
|Peripheral O2 saturation <90% for more than 60 minutes||14|
|Oliguria, defined as diuresis <400 ml/24 hours, refractory to careful hydration or to furosemide or dopamine||20|
|Acute urea deterioration to >270 mg/dl or creatinine >7, 2 g/dl||5|
|Jaundice with pre-eclampsia||22|
|Diabetic ketoacidosis or thyrotoxic crisis||2|
|Acute thrombocytopenia requiring transfusion of platelets||32|
|Coma for more than 12 hours||20|
|Subarachnoid or intraparenchymatous haemorrhage||1|
Table 4 shows the distribution of main causes of MM and SMM. The first three causes of all cases of MM and SMM were abortion complications, haemorrhage and hypertensive disorders. Abortion complications represented the main cause of MM (29.4%), followed by puerperal sepsis (11.8%) and hypertensive disorders (8.8%).
|Main causes||Specific causes||All cases||MM||SMM|
|(n = 552), n (%)||(n = 34), n (%)||(n = 518), n (%)|
|Abortion||Miscarriage and induced abortion complications||75 (13.5)||10 (29.4)||65 (12.5)|
|Haemorrhagic disorders||192 (34.8)||2 ( 5.9)||190 (36.7)|
|Antepartum||Placenta praevia||7 (1.3)||7 (1.4)|
|Unknown||1 (0.2)||1 (0.2)|
|Intrapartum||Uterine rupture||13 (2.4)||1 (2.9)||12 (2.3)|
|Abruptio placentae||26 (4.7)||26 (5.0)|
|Unknown||16 (2.9)||16 (3.1)|
|Postpartum||Placenta accreta||51 (9.2)||51 (9.8)|
|Uterine atony||46 (8.3)||1 (2.9)||45 (8.7)|
|Unknown||32 (5.8)||32 (6.2)|
|Hypertensive disorders||164 (29.7)||3 (8.8)||161 (31.1)|
|Pregnancy-induced hypertension||6 (1.1)||6 ( 1.2)|
|Chronic hypertension||4 (0.7)||4 ( 0.8)|
|Pre-eclampsia||62 (11.2)||1 (2.9)||61 (11.8)|
|Eclampsia||69 (12.5)||69 (13.3)|
|HELLP (haemolysis, elevated liver enzymes, low platelet count)||23 (4.2)||2 (5.9)||21 (4.1)|
|Sepsis||Puerperal sepsis||27 (4.6)||4 (11.8)||23 (4.4)|
|Other causes||94 (17.0)||15 (44.1)||79 (15.3)|
|Ectopic pregnancy||11 (2.0)||11 (2.1)|
|Molar pregnancy||3 (0.5)||1 (2.9)||2 (0.4)|
|Other infections/sepsis unrelated to pregnancy||21 (3.8)||4 (11.8)||17 (3.3)|
|Cardiovascular or pulmonary complications||15 (2.7)||3 (8.8)||12 (2.3)|
|Pulmonary embolism||6 (1.1)||3 (8.8)||3 (0.6)|
|Anaemia||6 (1.1)||6 (1.2)|
|HIV||3 ( 0.5)||2 (5.9)||1 (0.2)|
|Others||29 (5.3)||2 (5.9)||27 (5.2)|
Table 5 shows the morbidity–mortality indices and specific mortality rates for each of the main causes of MM and SMM. Sepsis and abortion complications were the principal causes of mortality. The morbidity–mortality indices were higher for haemorrhage management and hypertensive disorders (95:1 and 54:1, respectively). The specific mortality rate was eight times higher in post-abortion sepsis compared with post-abortion haemorrhage.
|Causes||MM||SMM||Morbidity–Mortality Index (95% CI)||Case fatality rate (%) (95% CI)|
|Abortion complications a||10||65||6.5:1 (2.7–15.6)||13.3 (6.9–23.6)|
|Post abortion sepsis||9||34||3.8:1 (1.5–9.3)||20.9 (10.6–36.5)|
|Postabortion haemorrhageb||1||31||31:1 (4.0–242.2)||3.1 (0.2–18.0)|
|Puerperal sepsis||4||23||5.8:1 (1.8–18.3)||14.8 (4.9–34.6)|
|All sepsis (puerperal sepsis + postabortion sepsis)||13||57||4.4:1 (1.9–10.0)||18.6 (10.6–30.0)|
|Hypertensive disorders||3||161||53.7:1 (14.8–195.0)||1.8 (0.5–5.7)|
|Haemorrhagic disorders||2||190||95.0:1 (20.8–433.6)||1.0 (0.2–4.1)|
|Others||15||79||5.3:1 (2.4–11.7)||16.0 (9.5–25.3)|
|All causes||34||518||15.2:1 (7.5–30.8)||6.2 (4.4–8.6)|
The specific case-fatality rate for sepsis was 18.6%, being 41% higher in post-abortion sepsis compared with puerperal sepsis.
The utilisation rate of known effective interventions to reduce MM and SMM showed that active management of the third stage of labour to prevent postpartum haemorrhage was used in 58.1% of women with vaginal delivery who presented with haemorrhage. Magnesium sulphate was administered to 62% of women with pre-eclampsia, and to 86.8% of women with eclampsia. Prophylactic antibiotics were used in 58.1% of all cases with endometritis or surgical wound infection after caesarean delivery, and in 64.7% of all cases of endometritis due to premature rupture of membranes. The rate of use of manual vacuum aspiration for management of incomplete miscarriage and induced abortion was 4.6% (Table 6).
|Complications||Effective interventions||No. of times used/No. of casesa||Specific use rate (%)||95% CI|
|Hypertensive disorders||Magnesium sulphate in pre-eclampsia||62/100||62.0||51.7–71.4|
|Magnesium sulphate in eclampsia||66/76||86.8||76.7–93.2|
|Active management of third stage of labour||18/31||58.1||39.3–74.9|
|Sepsis/Infections||Antibiotic prophylaxis in preterm premature rupture of membranes||11/17||64.7||38.6–84.7|
|Antibiotic prophylaxis in caesarean delivery||18/31||58.1||39.3–74.9|
|Abortion complications||Manual endouterine aspiration for treatment of incomplete abortion||4/87||4.6||1.5–12.0|
|Overall use rateb||179/342||52.3||46.9–57.7|
This study estimated an incidence of 0.8% for SMM, differential aetiological patterns for SMM and for MM, an overall case-fatality rate for MM and SMM causes of 6.2% (being higher for sepsis and abortion complications), and finally an overall use rate of known effective interventions to reduce MM and SMM of 52.3%.
According to our bibliographic review, unjustifiable variation in the incidence of SMM (0.01–8.23%) may be the result of the lack of uniform criteria to identify cases and standardised reporting systems[8, 13, 39-41] (Table S1).
We reached a consensus for case definition of SMM based on multiple criteria to achieve more sensitivity. All cases of MM and SMM related to pregnancy loss before 22 weeks of gestation (including abortion complications, miscarriage and induced abortions) were included. For case notification, management criteria were the most used (48.9%), while organ dysfunction criteria had a lower but still significant contribution (15.2%), facilitating the identification of performance problems in the hospitals. This methodology could be helpful for the development of active surveillance systems for SMM and for case management with a potential impact on the reduction of MM. Differential aetiological patterns for SMM (mainly postpartum haemorrhages and hypertensive disorders) and for MM (abortion complications and puerperal sepsis) require cause-specific approaches. Abortion complications and septic shock showed a mortality rate eight times higher than hypertension and nearly 13–15 times greater than haemorrhage, while post-abortion sepsis was seven times higher compared with postabortion haemorrhage. Nearly 1 in 5 women with septic shock died, mostly in those women with sepsis due to abortion complications compared with other causes of sepsis. The analysis of the morbidity–mortality index showed a better quality of care for hypertensive and haemorrhagic disorders compared with management of abortion complications and sepsis. Most of the complications associated with illegal abortions in Argentina affect poor women with lack of access to sexual and reproductive health programmes, highlighting an area ripe for public health policy reform.
The overall use rate of known effective interventions to prevent or treat the main causes of MM and SMM implied that implementation opportunities were lost half of the time.[22-25] However, a different pattern was observed between the use of manual vacuum aspiration (MVA) (<5%) and other known effective obstetric practices (almost 70%), despite the known comparative benefits of MVA and its support by health authorities. This would suggest that the low rate of use of MVA can be explained by the lack of skilled health providers. The use of active management of the third stage of labour in patients with SMM was much lower than what was observed in a previous local study (58% versus 93%, respectively). However, for preventive management of eclampsia and sepsis due to preterm premature rupture of membranes, we showed a nearly two-thirds increase in the use rate of magnesium sulphate and prophylactic antibiotics in comparison to the use rates reported in 2009. Approximately 40% of women with caesarean delivery complicated by puerperal sepsis did not receive prophylactic antibiotics. Finally, the highest use rate was observed with magnesium sulphate for eclampsia (90%). This study showed two different professional behaviour profiles. One with high use rates for those obstetric emergencies that require acute care, and the other with low rates for preventive interventions that often do not have immediate effects. It is necessary to reconsider implementation strategies of known beneficial practices at hospital level.[43, 44]
Main strengths lie in the value of collaborative research in the public health sector and its population representativeness (nearly 50% of the births in the public sector of AMBA and a quarter of total deliveries in public sector hospitals in Argentina occurred in our network, 2009). This situation allows for the extrapolation of our results for public hospitals at the national level. Considering that 99.4% of births nationally take place in hospitals, and, based on an overall morbidity–mortality index of 15:1 and 410 maternal deaths for 2009, 6150 cases of SMM can be expected yearly in Argentina. This study reconstructed a theoretical model of outcomes associated with the reproductive process that contributes to understanding the magnitude of the problem and the underutilisation of appropriate interventions.
Limitations of this study lie in the fact that information bias could be present considering that private hospitals and small maternity units (<1500 deliveries/year) were not included. Second, although a strict methodology was used to assure case identification and quality of data, it is possible that some cases were not captured. If this was the case, we consider that the overall impact of this situation would be negligible based on the number of cases analysed.
This study demonstrates the importance of understanding local health realities when conducting population-based research. The opportunity to provide policy makers with reliable information can lead to the development of specific strategies aimed at reducing MM and SMM such as perinatal regionalisation, active surveillance systems for SMM and implementation of known effective practices for case management, including the use of MVA, antibiotics and magnesium sulphate where appropriate.
The ultimate goal of this study is to contribute to the reduction of inequities in women's health and sexual and reproductive rights by addressing the influence of social determinants through inter-sectoral collaboration with education, social affairs, media and nongovernmental organisations. This approach to maternal health cannot be restricted to MM. With the knowledge and experience accrued through this and other studies, it is now an ethical and professional duty to consider strategies and policies for the prevention and treatment of SMM at the national and international levels as a critical component to improve maternal and perinatal health.
All the authors declare that they do not have any conflict of interest.
AK helped to conceive the project, designed the study and formulated the methodology, conducted the project and was project leader, analysed and interpreted the data, wrote the first draft of the manuscript, provided background knowledge to the data analysis and interpretation, and will act as study guarantor. RM helped to conceive the project, designed the study and formulated the methodology, analysed and interpreted the data, assisted in the writing and editing of the manuscript, and provided background knowledge to the data analysis and interpretation. PM coordinated the project, helped to supervise enrolment, analysed and interpreted the data, assisted in the writing and editing of the manuscript, performed the literature search and reviewed all abstracts and full text articles with assistance from AK and RM. CO helped to supervise enrolment, analysed and interpreted the data, assisted in the writing and editing of the manuscript, performed the literature search and reviewed all abstracts and full text articles with assistance from AK and RM. AMa and OF helped to supervise enrolment, analysed and interpreted the data, performed the literature search and reviewed all abstracts and full text articles with assistance from AK and RM. AMe, RW, MCF, JC, HCE, AL, CT, MD, RL, CM, BP, SN, DN, ÁB, GB, LDB, DMV, AD, EM, DF, LM, MH, CL, RS and CG helped to supervise enrolment, and were involved in the design of the study as members of the expert panel and provided feedback on earlier drafts of the manuscript. All authors critically revised earlier versions of the paper and gave approval of the final version to be published.
The protocol was approved by the Durand Hospital Ethics Committee (Authorizing Disposition Number No. 00418/2006).
Funding was received from the Ministry of Health of Ciudad Autónoma de Buenos Aires: grant for A. Karolinski (Decree No. 634/2009—Investigator's Career) and scholarships for A. Mazzoni and O. Fontana 2007–09; the National Ministry of Health: scholarship ‘Ramón Carrillo—Arturo Oñativia' for P. Micone 2007–09; and through Italian cooperation (MAE): through the CESTAS NGO (8249/CESTAS/ARG Project).
The authors are grateful to the local hospital research teams for their commitment to ensure the quality of data; Hospital Directors and Chiefs of Obstetrics and Gynaecology, Haematology, Intensive Care services and information technology personnel who facilitated access to hospital information systems and the exhaustive search for cases of MM and SMM; to Angeles Sala and Jeffrey Goldhagen for editing; to Joao Paulo Souza Department of Reproductive Health and Research (RHR), World Health Organization for his critical comments and technical support; to the authorities of the Ministries of Health of the Nation, City of Buenos Aires and PBA, Municipal Health Departments, Directors of Maternal–Child Health Programs and Networks; and the Italian cooperation (MAE) through the CESTAS NGO (8249/CESTAS/ARG Project) for their research support. Additional funding was provided by the National Ministry of Health (scholarship ‘Ramón Carrillo–Arturo Oñativia' for P. Micone 2007–09) and the Ministry of Health of Ciudad Autónoma de Buenos Aires (grant for A. Karolinski—Decree No. 634/2009—Investigator's Career and scholarships for A. Mazzoni and O. Fontana 2007–09).
Many studies to date reporting on maternal mortality statistics describe preventable maternal deaths in areas afflicted with poverty, decreased access to education, and fewer skilled clinicians. This study is of major significance because it not only identifies rates of preventable maternal mortality, but also reports the usage of known acceptable and successful interventions to reduce maternal mortality. It is therefore important to carefully examine this study and the implications of its findings for public health policy.
This study is set in Argentina, a country with the 84th highest maternal mortality rate in the world (http://data.worldbank.org/indicator/SH.STA.MMRT). Preventable maternal mortality was attributed mainly to complications of miscarriage and induced abortions (20.5%), infections (10.9%), hypertensive disorders in pregnancy (11.2%) and haemorrhage (12.7%). In addition to identifying and describing the incidence of these events, this study also describes the utilisation rates of known effective interventions to reduce maternal mortality. Although it was encouraging to see an 86.8% utilisation rate of magnesium sulphate in cases of eclampsia, it was astonishing to find only a 4.6% utilisation rate of manual vacuum aspiration for the management of incomplete miscarriage and induced abortion. These findings have broad public health implications.
How should we interpret these results given the study's cross-sectional design? Although it is accurate to say that the design is suboptimal in determining a causal relationship between underuse of accepted interventions and the risk of maternal mortality, the authors should be commended for essentially establishing a surveillance system for maternal mortality in these hospitals. Although only including a network of 30 hospitals that provide comprehensive emergency obstetric care and perinatal training programmes, the exclusion of many other smaller and more rurally located hospitals results in an underestimation of maternal morbidity and utilisation rates, and therefore provides us with useful information.
The authors suggest the low rate of manual vacuum aspiration use was probably explained by a lack of skilled healthcare providers. Whether it is a lack of skilled providers, or an unwillingness of providers to use this skill, it is clear that there are political and religious barriers contributing to the underutilisation of this intervention, which has a profound impact on maternal mortality rates in Argentina. Women who would probably benefit from this intervention often obtain an abortion illegally from an unskilled provider. So, despite the methodological weaknesses of the study design, this research identifies significant modifiable maternal mortality indicators that should be published and highlighted as salient points in need of public health policy reform.
KB and MD have no commercial interests to disclose.
K Brookfield & M Druzin
Division of Maternal Fetal Medicine, Stanford University School of Medicine/Lucile Packard Children's Hospital, Stanford, CA, USA