International Journal of Gynecology and Obstetrics 124 (2014) 123–127
Contents lists available at ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
CLINICAL ARTICLE
Insights into maternal mortality in Madang Province, Papua New Guinea John W. Bolnga a,⁎, Nancy N. Hamura a, Alexandra J. Umbers b, Stephen J. Rogerson b, Holger W. Unger b a b
Department of Obstetrics and Gynecology, Modilon General Hospital, Madang, Papua New Guinea Department of Medicine, Royal Melbourne Hospital, The University of Melbourne, Melbourne, Australia
a r t i c l e
i n f o
Article history: Received 29 May 2013 Received in revised form 11 August 2013 Accepted 28 October 2013 Keywords: Cause Maternal death Maternal mortality ratio Papua New Guinea Reporting Three-delays model
a b s t r a c t Objective: To assess the frequency, causes, and reporting of maternal deaths at a provincial referral hospital in coastal Papua New Guinea (PNG), and to describe delays in care. Methods: In a structured retrospective review of maternal deaths at Modilon General Hospital, Madang, PNG, registers and case notes for the period January 2008 to July 2012 were analyzed to determine causes, characteristics, and management of maternal death cases. Public databases were assessed for underreporting. Results: During the review period, there were 64 maternal deaths (institutional maternal mortality ratio, 588 deaths per 100 000 live births). Fifty-two cases were analyzed in detail: 71.2% (n = 37) were direct maternal deaths, and hemorrhage (n = 24, 46.2%) and infection (n = 16, 30.8%) were the leading causes of mortality overall. Women frequently did not attend prenatal clinics (n = 34, 65.4%), resided in rural areas (n = 45, 86.5%), and experienced delays in care (n = 45, 86.5%). Maternal deaths were underreported in public databases. Conclusion: The burden of maternal mortality was found to be high at a provincial hospital in PNG. Most women died of direct causes and experienced delays in care. Strategies to complement current hospital and national policy to reduce maternal mortality and to improve reporting of deaths are needed. © 2013 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction Reducing maternal deaths remains a global health priority, particularly as maternal death rates are disproportionately higher in low- and middle-income countries [1,2]. The disparity between the frequency of maternal deaths in low- and middle-income countries and that in high-income nations indicates that most deaths are preventable. Millennium Development Goal 5 (MDG5) was established to reduce maternal mortality by three-quarters by 2015 [3], particularly in lowincome countries. In the South Pacific region, maternal deaths are frequent in Papua New Guinea (PNG), a low-income country with a population of 7 million [4,5]. The national maternal mortality ratio (MMR) is high (773 deaths per 100 000 live births in 2007, 312 in 2008, and 230 in 2010 [1,6,7]), as are fertility rates (4.4 births per woman) [4]; by contrast, contraception use and prenatal clinic attendance are low, and most deliveries are unsupervised [8]. There is evidence that for some of these indicators significant rural–urban differences exist [7]. It remains a matter of debate whether the downward trend in recent mortality figures is due to an actual reduction in maternal deaths, or simply a result of variations in the methodology and data sets used to derive them [9]. Up-to-date regional information on maternal mortality is scarce [9,10]. In a hospital in the PNG Eastern Highlands, half of maternal ⁎ Corresponding author at: Department of Obstetrics and Gynecology, Modilon General Hospital, Modilon Road, 511 Madang, Papua New Guinea. Tel.: +675 73340292; fax: +675 4223719. E-mail address: [email protected] (J.W. Bolnga).
deaths were attributed to overwhelming sepsis (puerperal or after abortion) [9]. By contrast, hemorrhage accounted for 65% of deaths in Milne Bay Province, coastal PNG, and delays in seeking and obtaining appropriate pregnancy care were found to be common [10]. Modilon General Hospital (MGH) is situated in Madang on the north coast of PNG. It is the principal referral center for the Madang provincial population (which is estimated at 490 000 and largely resides in rural areas) and manages approximately 2500 deliveries annually. MGH provides family planning, prenatal care, and comprehensive emergency obstetric care (EmOC). Institutional reviews of maternal death can assist with identifying targets for intervention to reduce maternal mortality at an institutional, provincial, and national level, and can provide data to determine the burden and potential underreporting of maternal deaths [11]. The aim of the present study was to assess the frequency, causes, and reporting of maternal deaths at MGH, and to describe delays in pregnancy care, thereby generating data to inform policy to reduce maternal mortality. 2. Materials and methods In a descriptive data review, registers from the labor ward, obstetrics and gynecology ward, and emergency department of MGH, Madang, PNG, were searched for maternal deaths that occurred during the period January 1, 2008, to July 31, 2012. Ethical approval for the study was obtained from the MGH directorate. Informed consent was not required. At MGH, prenatal care and comprehensive EmOC are provided. Blood transfusion services rely on patients’ relatives as donors. HIV testing has formed part of routine prenatal care at MGH since 2007:
0020-7292/$ – see front matter © 2013 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijgo.2013.08.012
124
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127
approximately 1.1% of women presenting for prenatal care are HIVpositive. Malaria is endemic, and pregnant women are provided with intermittent preventive treatment for malaria (single dose of sulfadoxine/ pyrimethamine and weekly chloroquine before 2012, 3-monthly doses of sulfadoxine/pyrimethamine since 2012). Maternal deaths at MGH are reviewed by clinical staff and discussed. Mortality and delivery figures are reported to provincial and national health information offices, and detailed maternal death reports are also sent to the lead obstetrician in Port Moresby, PNG. Of note, the hospital did not have a specialist obstetrician from mid-2008 to December 2010. In the present study, each case of maternal death was allocated a unique identification number, ensuring anonymity. Charts were obtained when possible, and relevant information was extracted and analyzed. Medical causes of death were categorized by 2 clinicians (J.W.B. and H.W.U.) using established definitions [11–14]. Hindrances to women accessing and receiving the care that they needed were evaluated via the “3-delays model” [15]. Phase-1 delays were defined as delays in seeking medical help by the women. Phase-2 delays were defined as delays in reaching the health center and/or hospital after a decision to seek care had been made (including delayed transfers from health centers to MGH). Phase-3 delays were defined as delays in receiving timely and appropriate care at aid posts and at hospital. Institutional delivery data were obtained from the labor room register. The numbers of maternal deaths and deliveries at MGH were subsequently compared with data registered with the Provincial Health Information Office (PHIO) in order to assess the adequacy of reporting. Because patient identifiers were unavailable for the PHIO data, a capture–recapture analysis was not possible, and the assessment of potential underreporting was restricted to a simple comparison of summative statistics. Data are reported as number (percentage), mean ± SD, or median (interquartile range). Statistical analysis was performed using Stata version 12.0 (StataCorp, College Station, TX, USA).
Table 2 Background characteristics of maternal death cases (n = 52).
Age, y Mean ± SD Median (interquartile range) Range Marital status Married Single missing data Employment Housewife Subsistence farmer Formal employment Missing data Religious affiliation Catholic Lutheran Other Missing data Region of origin Madang/Morobe Sepik New Britain Highlands Rural/urban Rural Urban Residence district Bogia Madang Middle Ramu Rai Coast Sumkar Usino Bundi Traveling time to hospitalb ≤1 h 1–3 h 3–6 h 6–12 h 12–24 h ≥24 h Referred from health center Yes No
3. Results During the study period, there were 64 maternal deaths and 10 891 live births at MGH, resulting in an institutional MMR of 588 per 100 000 live births (Table 1). Twelve women were excluded from detailed analysis owing to a lack of case notes. Limited information in registers indicated that these maternal deaths were caused by obstetric hemorrhage (n = 8), sepsis (n = 3), and ruptured ectopic pregnancy (n = 1). As a result, data from 52 women were analyzed in detail. The mean ± SD age of the women who died was 28 ± 6.4 years (Table 2). Most were married and lived in rural areas. Two-thirds of patients (n = 33) lived 3 hours or more in travel time distance from MGH. Twenty-five percent of women (n = 13) were nulliparous, and 17.3% were grand multiparous (Table 3). Among parous women (71.2%), the median parity was 3 (IQR, 1–4.5, range 1–10). Twelve (23.1%) had a previous obstetric complication. Only 18 women (34.6%) had attended prenatal care in the index pregnancy.
Valuesa
Characteristic
28 ± 6.4 28 (19–38) 17–42 46 (88.5) 1 (1.9) 5 (9.6) 15 (28.9) 14 (26.9) 3 (5.8) 20 (38.4 15 (28.9) 6 (11.5) 4 (7.7) 27 (51.9) 42 (80.8) 5 (9.6) 3 (5.8) 2 (3.8) 45 (86.5) 7 (13.5) 11 (21.2) 17 (32.7) 5 (9.6) 4 (7.7) 5 (9.6) 10 (19.2) 13 (25.0) 6 (11.5) 11 (21.2) 14 (26.9) 1 (1.9) 7 (13.5) 40 (76.9) 12 (23.1)
a
Values are given as number (percentage) of women unless stated otherwise. Estimates were made on the basis of geographic distance and transport available for travel from residence to hospital. b
Most cases were direct maternal deaths (n = 37, 71.2%) (Table 4, Supplementary Material S1), and hemorrhage (owing to a range of underlying causes) was a prominent factor in these cases. Uterine rupture secondary to prolonged obstructed labor was common. Of the indirect maternal deaths (n = 15, 28.8%), non-pregnancy-related infection was the most common cause (n = 11, 21.2%). There was 1 death in early pregnancy owing to septic abortion.
Table 1 Comparison of maternal deaths, live births, and MMR between Modilon General Hospital data and Madang PHIO data 2008–2012. Year
2008 2009 2010 2011 Total 2008–2011 2012b Total 2008–2012
Modilon General Hospital (review)
Madang Province (PHIO)a
Modilon General Hospital (PHIO)
Maternal deaths
Live births
MMR
Maternal deaths
Live births
MMR
Maternal deaths
Live births
MMR
14 8 23 15 60 4 64
1633 2043 2669 2753 9098 1793 10891
857 392 862 545 660 223 588
4 4 6 5 19
501 569 526 552 2148
798 703 1141 906 885
13 19 34 28 94
17543 18098 18513 19010 73164
74 105 184 147 128
Abbreviations: MMR, maternal mortality ratio; PHIO, Provincial Health Information Office. a Includes data from Modilon General Hospital. b January 1 to July 31, 2012; PHIO data for 2012 unavailable.
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127 Table 3 Clinical characteristics of maternal death cases (n = 52). Characteristic Parity Nulliparous 1–2 3–4 Grand multipara (≥5) Missing data Gestational age First trimester Second trimester Third trimester Prenatal clinic attendance Yes No Multiple pregnancy Twin Delivered singleton Undelivered singleton Undelivered and unknown Missing data Previous obstetric history Stillbirth Retained placenta Multiple pregnancy Postpartum hemorrhage No documented history Mode of delivery Singleton pregnancies SVD Breech delivery EMLUSC EMLUSC & TAH Vacuum extraction Twin pregnancies (twins 1 & 2) SVD & SVD SVD & forceps SVD & vaginal delivery after internal podalic version Vacuum extraction EMLUSC Not applicable (undelivered) Fetal outcome Singleton pregnancies Live birth Stillbirth Twin pregnancies Both live births Twin 1 live birth, twin 2 stillbirth Not applicable (undelivered) Time between hospital admission and death Dead on arrival ≤1 h 1–6 h 6–24 h 24–48 h 2 d to 1 wk 1–2 wk ≥2 wk Frequency of delay ≥1 delay No delay Type of delay Phase 1 Phase 2 Phase 3 None
125
Table 4 Principal underlying causes of maternal death (n = 52). Number (%) 13 (25.0) 15 (28.9) 13 (25.0) 9 (17.3) 2 (3.8) 1 (1.9) 5 (9.6) 46 (88.5) 18 (34.6) 34 (65.4) 8 (15.4) 34 (65.4) 4 (7.7) 5 (9.6) 1 (1.9) 5 (9.6) 4 (7.7) 1 (1.9) 2 (3.9) 40 (76.9)
18 (34.6) 5 (9.6) 7 (13.5) 2 (3.6) 2 (3.6) 2 (3.6) 1 (1.9) 1 (1.9) 2 (3.9) 2 (3.9) 10 (19.2)
19 (36.5) 15 (28.9) 5 (9.6) 3 (5.8) 10 (19.2) 2 (3.9) 2 (3.9) 5 (9.6) 8 (15.4) 11 (21.2) 15 (28.9) 6 (11.5) 3 (5.8) 45 (86.5) 7 (13.5) 21 (40.4) 23 (44.2) 35 (67.3) 7 (13.5)
Abbreviations: EMLUSC, emergency lower uterine segment cesarean; SVD, spontaneous vertex delivery; TAH, total abdominal hysterectomy.
Most women were in the third trimester of pregnancy (Table 3). Among women in the second and third trimester of pregnancy (n = 51), 9 (17.7%) died before, 2 (3.9%) died during, and 40 (78.4%) died after delivery (8 of whom were unsupervised). Multiple pregnancy, breech presentation, and grand multiparity were common features in these cases (Table 3).
Underlying cause of death
Number (%)
Direct maternal deaths Hemorrhage Uterine rupture Third-stage hemorrhagea Postpartum hemorrhage (uterine atony) Prepartum hemorrhage High vaginal laceration Placenta accreta Pregnancy-related infection Hypertensive disorder Other Obstetric embolism Retained placenta (with sepsis) Complication of anesthesia Septic abortion Indirect maternal deaths Non-pregnancy-related infection Likely bacterial sepsis Malaria Tuberculosis Other Cardiac disease Malignancy
37 (71.2) 24 (46.2) 7 (13.5) 6 (11.5) 6 (11.5) 5 (9.6) 1 (1.9) 1 (1.9) 5 (9.6) 4 (7.7) 4 (7.7) 1 (1.9) 1 (1.9) 1 (1.9) 1 (1.9) 15 (28.8) 11 (21.2) 9 (17.3) 1 (1.9) 1 (1.9) 4 (7.7) 2 (3.9) 2 (3.9)
a
Hemorrhage owing to retained placenta.
Common causes of death among women with twin pregnancies were pre-eclampsia or eclampsia (n = 3) and hemorrhage (atony, n = 2; uterine rupture, n = 1). Six grand multiparous women died of hemorrhage. Among women with a history of retained placenta (n = 4), 2 had a recurrent retained placenta, 1 had placenta accreta, and 1 had uterine rupture. In 15 of 34 (41.2%) singleton pregnancies, the neonate was stillborn (Table 3). Among women with a documented history of stillbirth (n = 5), 4 had a recurrent stillbirth. Most women (n = 40, 76.9%) had been referred from rural aid posts (Table 3). On hospital admission, midwives commonly assessed the patient first (n = 24, 46.2%), followed by junior medical staff (n = 23, 44.2%). Forty-seven women (90.4%) had indications of physiologic compromise on arrival (≥1 of the following: blood pressure b90/60 or N140/90 mm Hg; heart rate N 110 beats per minute; temperature ≥37.5 °C; respiratory rate N24; anuria or oliguria). Two-thirds of the women (35/52) died 24 hours or more after admission (Table 3). Sixteen women (30.8%) required instrumental or operative assistance (Table 3). Fifty-two percent (n = 27) of women were transfused with a median of 2 units of red cell concentrate (range, 1–7 units; mean, 2.7 units). Among the women who died because of postpartum hemorrhage (after delivery of both neonate and placenta either owing to atony alone or associated with uterine rupture or prolonged third stage), only half (9/20) had documented use of uterotonic drugs. Most women who died from pregnancy- or non-pregnancy-related sepsis (n = 12/14, 85.7%) had been referred: only 3 had received broad-spectrum antibiotics before arrival, and 9 received them at MGH. Malaria was suspected for 23 women (44.2%), and blood slides were examined for 4 women: 2 were parasitemic, and 1 woman died as a consequence (Table 4). Nineteen received artemether as per national guideline [16]. Data on uptake of intermittent preventive treatment for malaria and HIV status were unavailable. Forty-five women (86.5%) experienced 1 or more delays, 27 (51.9%) had 2 or more delays, and phase-3 delays were the most common (Table 3). Phase-3 delays occurred at MGH (n = 18), the referring health center (n = 12), or both (n = 5). Delays at health centers included lack of adequate treatment (e.g. antibiotics, uterotonic drugs, magnesium sulfate, or fluids) (n = 17) and delayed referral (n = 8). At hospital, 14 women experienced delays in receiving blood, 10 had delays in their review by senior medical staff, and 10 had delays in crucial treatment (antibiotics, magnesium sulfate, fluids). Others delays
126
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127
included review by junior medical staff (n = 3) and necessary surgical intervention (n = 3). There were discrepancies between the number of maternal deaths and live births at MGH identified in the hospital records and those published in PHIO summary statistics. In the period 2008–2011, there was underreporting to PHIO of maternal deaths (n = 19/60, 31.7%) and live births (n = 2148/9098, 23.6%) at MGH (Table 1). During this period, the number of deliveries at MGH almost doubled, but PHIO figures did not reflect this trend (Table 1). 4. Discussion Maternal death was found to be common at a provincial referral hospital in PNG, and most women died as a result of hemorrhage or infection. These causes of maternal death have been previously recorded as predominant in PNG, although regional heterogeneity exists [9,10,17]. Only 2 indirect deaths (due to malignancy) could not have been prevented through adequate provision and uptake of family planning, prenatal care, supervised delivery, and basic or comprehensive EmOC. Most women lived rurally at some distance from MGH, and many had been referred from aid posts. The recent drive to improve capacity for maternity care at the community level, including basic EmOC, will assist in improving case management and reduce the high number of direct maternal deaths [18–20]. Many women did not seek or receive prenatal care, although the review was unable to identify reasons for this low uptake. Obstetric risk factors were frequently present, and it is important that women with high-risk pregnancies seek prenatal care early, are recognized as such, and are referred to hospital before delivery. Effective mechanisms for transport to facilities that can provide comprehensive EmOC will help [18–20]. Waiting houses for obvious high-risk pregnancies may further reduce the risk of maternal death [10], but require evaluation in the context of PNG. Most women experienced at least 1 delay—that is, a missed “window of opportunity” that was potentially critical to the outcome. Research indicates that delays are ultimately related to women’s poverty, lack of education, information, and awareness, adverse experiences with healthcare staff, and gender inequality (phase 1); a lack of safe, accessible, and low-cost transport mechanisms (phase 2); and a shortage of medical supplies, and trained and motivated staff at health facilities (phase 3) [10,15,21]. Many women did not receive appropriate fluid resuscitation and treatment (antibiotics, uterotonic drugs, or magnesium sulfate) at the health center level, for which a lack of supplies and basic EmOC providers are possible explanations. At the hospital level, there were delays in blood transfusion: a failure to recognize the need for transfusion, and a lack of staff, blood bank reserves, and reagents are likely contributors to this. Laboratory services undoubtedly need strengthening, given the high number of women requiring blood and low numbers of women who had blood smear microscopy for malaria. The lack of a specialist obstetrician at the institution during parts of the study period, and problems with prompt recognition and treatment of ill women by junior staff might have prevented some women from receiving optimal care, including operative delivery. Delays will have contributed to the irreversible physiologic compromise that most women were experiencing by the time of arrival at hospital: women frequently succumbed to renal failure, disseminated intravascular coagulation, and overwhelming sepsis within 48 hours of admission (data not shown). By contrast, patients who died after having been admitted to hospital for some time commonly had chronic conditions such as malignancy, tuberculosis, and rheumatic heart disease. The present study has captured only a proportion of maternal deaths occurring in Madang province. Women die at rural health centers (Table 1), and some do not reach a health facility [10] and their deaths remain undetected. It is likely that proportionally more deaths “occur” at hospital than in rural areas [22]; hence, institutional MMRs need to
be interpreted with caution. Nevertheless, the observed underreporting of deaths to PHIO might also occur at the health center level. The use of PHIO data to estimate the burden of, and progress with reducing, maternal deaths may lead to imprecise and therefore potentially misleading data. Both maternal deaths and live births were found to be disproportionately underreported, further decreasing the precision of MMR estimates. Suboptimal documentation and data flow between the hospital and PHIO are possible explanations for this. Real-time surveillance mechanisms for maternal deaths are required [14], and clear and irrefutable post-MDG5 targets, such as reducing the MMR to below 30 per 100 000 by 2030, are more useful [23]. Establishing a national maternal death surveillance and response system that expands and optimizes the current use of maternal death proformas complies with aims of the PNG National Health Plan [20]. Technical support from the WHO to implement this activity is underway [14], and will facilitate a swift, efficient, and specific response to monitoring maternal deaths, at district and provincial levels. The use of current PHIO figures for health budgeting purposes will result in suboptimal resource allocation: the number of deliveries at MGH almost doubled over the study period, but this trend was not reflected in PHIO statistics (Table 1). The National Health Plan aims to increase the number of health facilities able to provide comprehensive EmOC [20]. The present findings suggest that this should be preceded by strengthening the capacities of existing facilities in terms of staffing, training, ambulance services, and laboratory and blood transfusion services. Future research might include a prospective evaluation of “near misses” [24]—that is, women whose lives were saved but only just. Furthermore, studies might attempt to evaluate the frequency of late maternal deaths in the area [25], and to review each maternal death soon after it occurred, with the aim of clearly establishing the exact nature of possible delays (by interviewing relatives and staff). Research evaluating underreporting at the aid post level and estimating deaths in villages will broaden our understanding of this hidden burden of maternal death in PNG. The present study has limitations. First, not all cases had notes available for in-depth review. Second, late maternal deaths (those occurring N42 days but b 1 year postpartum) were not reviewed because the documentation in registers was inadequate to attempt their evaluation. Third, it is possible that the present review missed maternal deaths in early pregnancy—regional differences may only partly explain this [9]— and women may have died of ruptured ectopic pregnancy and abortions (including septic abortions) prior to reaching a formal healthcare provider. Last, documentation bias may have resulted in underestimating delays. When a delay was evident, its underlying cause could not always be established from case notes alone and qualitative research is needed in this area [10]. Current government strategies need to be complemented by efforts to improve the capacities of existing providers of comprehensive EmOC. Establishing real-time maternal death surveillance systems that can monitor progress and provide precise data for resource allocation planning are needed. Further research is required to identify the burden of village deaths and the barriers preventing women in PNG from receiving the pregnancy care that they need. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijgo.2013.08.012. Acknowledgments H.W.U. and A.J.U. received salary support from the Malaria in Pregnancy Consortium, which receives funding from the Bill & Melinda Gates Foundation. Conflict of interest The authors have no conflicts of interest.
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127
References [1] Hogan MC, Foreman KJ, Naghavi M, Ahn SY, Wang M, Makela SM, et al. Maternal mortality for 181 countries, 1980–2008: a systematic analysis of progress towards Millennium Development Goal 5. Lancet 2010;375(9726):1609–23. [2] Hill K, Thomas K, AbouZahr C, Walker N, Say L, Inoue M, et al. Estimates of maternal mortality worldwide between 1990 and 2005: an assessment of available data. Lancet 2007;370(9595):1311–9. [3] United Nations General Assembly. United Nations Millennium Declaration. A/Res/55/ 2. http://www.un.org/millennium/declaration/ares552e.pdf. Published September 18, 2000. Accessed February 18, 2013. [4] The World Bank. Papua New Guinea - Reproductive health at a glance. http:// documents.worldbank.org/curated/en/2011/04/14695221/papua-new-guineareproductive-health-glance. Published April 2011. Accessed February 21, 2013. [5] United Nations Development Programme. Papua New Guinea Country Profile: Human Development Indicators. http://hdrstats.undp.org/en/countries/profiles/PNG.html. Published 2011. Accessed February 18, 2013. [6] World Health Organization, UNICEF, UNFPA, The World Bank. Trends in maternal mortality: 1990 to 2010. http://whqlibdoc.who.int/publications/2012/9789241503631_ eng.pdf. Published 2012. Accessed February 16, 2013. [7] Papua New Guinea National Statistical Office. Papua New Guinea Demographic and Health Survey 2006: National Report. Port Morseby: Papua New Guinea Government Printer; 2009. [8] United Nations Population Fund. The State of the World’s Midwifery 2011. Country data and profiles: Papua New Guinea. http://www.unfpa.org/sowmy/resources/en/ country_info.htm. Published 2011. Accessed February 12, 2013. [9] Sanga K, de Costa C, Mola G. A review of maternal deaths at Goroka General Hospital, Papua New Guinea 2005–2008. Aust N Z J Obstet Gynaecol 2010;50(1): 21–4. [10] Kirby B. Too Far to Travel—Maternal Mortality in Milne Bay Province—Beyond the Numbers. THOR – Hands of Rescue Foundation: Alotau; 2010. [11] World Health Organization. The WHO application of ICD-10 to deaths during pregnancy, childbirth and puerperium: ICD MM. http://apps.who.int/iris/bitstream/ 10665/70929/1/9789241548458_eng.pdf. Published 2012. Accessed January 17, 2013. [12] World Health Organization. Health statistics and health information systems: Maternal mortality ratio (per 100 000 live births). http://www.who.int/healthinfo/
[13]
[14]
[15] [16] [17] [18]
[19]
[20]
[21]
[22] [23] [24]
[25]
127
statistics/indmaternalmortality/en/index.html. Published 2010. Accessed January 12, 2013. Cross S, Bell JS, Graham WJ. What you count is what you target: the implications of maternal death classification for tracking progress towards reducing maternal mortality in developing countries. Bull World Health Organ 2010;88(2):147–53. Hounton S, De Bernis L, Hussein J, Graham WJ, Danel I, Byass P, et al. Towards elimination of maternal deaths: maternal deaths surveillance and response. Reprod Health 2013;10:1. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med 1994;38(8):1091–110. Papua New Guinea National Department of Health. Malaria Treatment Protocol. Port Moresby: Papua New Guinea National Department of Health; 2009. Garner P, Lai D, Baea M. Childbirth in rural areas: maternal deaths, village deliveries and obstetric service use. P N G Med J 1994;37(3):166–72. Morgan C, Byrne A. Improving maternal, newborn and child health in PNG through family and community care. http://pacificinstitute.anu.edu.au/outrigger/2011/12/ 11/improving-maternal-newborn-and-child-health-in-png-through-family-andcommunity-care/. Published December 12, 2011. Accessed February 23, 2013. Papua New Guinea National Department of Health. Ministerial Taskforce on Maternal Health in Papua New Guinea. http://www.unfpa.org/sowmy/resources/docs/ library/R149_DOH_PNGUINEA_2009_Ministerial_Taskforce_report_final_version_3. pdf. Accessed January 19, 2013. Government of Papua New Guinea. National Health Plan 2011–2020: Volume 1 Policies and Strategies. http://www.wpro.who.int/countries/png/PNGNHP_Part1.pdf. Published June 2010. Hunt P, Bueno de Mesquita J. Reducing Maternal Mortality: The contribution of the right to the highest attainable standard of health. http://www.unfpa.org/webdav/ site/global/shared/documents/publications/reducing_mm.pdf. Published 2010. Ronsmans C, Graham WJ. Lancet Maternal Survival Series steering group. Maternal mortality: who, when, where, and why. Lancet 2006;368(9542):1189–200. Gilmore K, Gebreyesus TA. What will it take to eliminate preventable maternal deaths? Lancet 2012;380(9837):87–8. Cecatti JG, Souza JP, Oliveira Neto AF, Parpinelli MA, Sousa MH, Say L, et al. Pre-validation of the WHO organ dysfunction based criteria for identification of maternal near miss. Reprod Health 2011;8:22. Høj L, da Silva D, Hedegaard K, Sandström A, Aaby P. Maternal mortality: only 42 days? BJOG 2003;110(11):995–1000.
Contents lists available at ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
CLINICAL ARTICLE
Insights into maternal mortality in Madang Province, Papua New Guinea John W. Bolnga a,⁎, Nancy N. Hamura a, Alexandra J. Umbers b, Stephen J. Rogerson b, Holger W. Unger b a b
Department of Obstetrics and Gynecology, Modilon General Hospital, Madang, Papua New Guinea Department of Medicine, Royal Melbourne Hospital, The University of Melbourne, Melbourne, Australia
a r t i c l e
i n f o
Article history: Received 29 May 2013 Received in revised form 11 August 2013 Accepted 28 October 2013 Keywords: Cause Maternal death Maternal mortality ratio Papua New Guinea Reporting Three-delays model
a b s t r a c t Objective: To assess the frequency, causes, and reporting of maternal deaths at a provincial referral hospital in coastal Papua New Guinea (PNG), and to describe delays in care. Methods: In a structured retrospective review of maternal deaths at Modilon General Hospital, Madang, PNG, registers and case notes for the period January 2008 to July 2012 were analyzed to determine causes, characteristics, and management of maternal death cases. Public databases were assessed for underreporting. Results: During the review period, there were 64 maternal deaths (institutional maternal mortality ratio, 588 deaths per 100 000 live births). Fifty-two cases were analyzed in detail: 71.2% (n = 37) were direct maternal deaths, and hemorrhage (n = 24, 46.2%) and infection (n = 16, 30.8%) were the leading causes of mortality overall. Women frequently did not attend prenatal clinics (n = 34, 65.4%), resided in rural areas (n = 45, 86.5%), and experienced delays in care (n = 45, 86.5%). Maternal deaths were underreported in public databases. Conclusion: The burden of maternal mortality was found to be high at a provincial hospital in PNG. Most women died of direct causes and experienced delays in care. Strategies to complement current hospital and national policy to reduce maternal mortality and to improve reporting of deaths are needed. © 2013 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction Reducing maternal deaths remains a global health priority, particularly as maternal death rates are disproportionately higher in low- and middle-income countries [1,2]. The disparity between the frequency of maternal deaths in low- and middle-income countries and that in high-income nations indicates that most deaths are preventable. Millennium Development Goal 5 (MDG5) was established to reduce maternal mortality by three-quarters by 2015 [3], particularly in lowincome countries. In the South Pacific region, maternal deaths are frequent in Papua New Guinea (PNG), a low-income country with a population of 7 million [4,5]. The national maternal mortality ratio (MMR) is high (773 deaths per 100 000 live births in 2007, 312 in 2008, and 230 in 2010 [1,6,7]), as are fertility rates (4.4 births per woman) [4]; by contrast, contraception use and prenatal clinic attendance are low, and most deliveries are unsupervised [8]. There is evidence that for some of these indicators significant rural–urban differences exist [7]. It remains a matter of debate whether the downward trend in recent mortality figures is due to an actual reduction in maternal deaths, or simply a result of variations in the methodology and data sets used to derive them [9]. Up-to-date regional information on maternal mortality is scarce [9,10]. In a hospital in the PNG Eastern Highlands, half of maternal ⁎ Corresponding author at: Department of Obstetrics and Gynecology, Modilon General Hospital, Modilon Road, 511 Madang, Papua New Guinea. Tel.: +675 73340292; fax: +675 4223719. E-mail address: [email protected] (J.W. Bolnga).
deaths were attributed to overwhelming sepsis (puerperal or after abortion) [9]. By contrast, hemorrhage accounted for 65% of deaths in Milne Bay Province, coastal PNG, and delays in seeking and obtaining appropriate pregnancy care were found to be common [10]. Modilon General Hospital (MGH) is situated in Madang on the north coast of PNG. It is the principal referral center for the Madang provincial population (which is estimated at 490 000 and largely resides in rural areas) and manages approximately 2500 deliveries annually. MGH provides family planning, prenatal care, and comprehensive emergency obstetric care (EmOC). Institutional reviews of maternal death can assist with identifying targets for intervention to reduce maternal mortality at an institutional, provincial, and national level, and can provide data to determine the burden and potential underreporting of maternal deaths [11]. The aim of the present study was to assess the frequency, causes, and reporting of maternal deaths at MGH, and to describe delays in pregnancy care, thereby generating data to inform policy to reduce maternal mortality. 2. Materials and methods In a descriptive data review, registers from the labor ward, obstetrics and gynecology ward, and emergency department of MGH, Madang, PNG, were searched for maternal deaths that occurred during the period January 1, 2008, to July 31, 2012. Ethical approval for the study was obtained from the MGH directorate. Informed consent was not required. At MGH, prenatal care and comprehensive EmOC are provided. Blood transfusion services rely on patients’ relatives as donors. HIV testing has formed part of routine prenatal care at MGH since 2007:
0020-7292/$ – see front matter © 2013 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijgo.2013.08.012
124
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127
approximately 1.1% of women presenting for prenatal care are HIVpositive. Malaria is endemic, and pregnant women are provided with intermittent preventive treatment for malaria (single dose of sulfadoxine/ pyrimethamine and weekly chloroquine before 2012, 3-monthly doses of sulfadoxine/pyrimethamine since 2012). Maternal deaths at MGH are reviewed by clinical staff and discussed. Mortality and delivery figures are reported to provincial and national health information offices, and detailed maternal death reports are also sent to the lead obstetrician in Port Moresby, PNG. Of note, the hospital did not have a specialist obstetrician from mid-2008 to December 2010. In the present study, each case of maternal death was allocated a unique identification number, ensuring anonymity. Charts were obtained when possible, and relevant information was extracted and analyzed. Medical causes of death were categorized by 2 clinicians (J.W.B. and H.W.U.) using established definitions [11–14]. Hindrances to women accessing and receiving the care that they needed were evaluated via the “3-delays model” [15]. Phase-1 delays were defined as delays in seeking medical help by the women. Phase-2 delays were defined as delays in reaching the health center and/or hospital after a decision to seek care had been made (including delayed transfers from health centers to MGH). Phase-3 delays were defined as delays in receiving timely and appropriate care at aid posts and at hospital. Institutional delivery data were obtained from the labor room register. The numbers of maternal deaths and deliveries at MGH were subsequently compared with data registered with the Provincial Health Information Office (PHIO) in order to assess the adequacy of reporting. Because patient identifiers were unavailable for the PHIO data, a capture–recapture analysis was not possible, and the assessment of potential underreporting was restricted to a simple comparison of summative statistics. Data are reported as number (percentage), mean ± SD, or median (interquartile range). Statistical analysis was performed using Stata version 12.0 (StataCorp, College Station, TX, USA).
Table 2 Background characteristics of maternal death cases (n = 52).
Age, y Mean ± SD Median (interquartile range) Range Marital status Married Single missing data Employment Housewife Subsistence farmer Formal employment Missing data Religious affiliation Catholic Lutheran Other Missing data Region of origin Madang/Morobe Sepik New Britain Highlands Rural/urban Rural Urban Residence district Bogia Madang Middle Ramu Rai Coast Sumkar Usino Bundi Traveling time to hospitalb ≤1 h 1–3 h 3–6 h 6–12 h 12–24 h ≥24 h Referred from health center Yes No
3. Results During the study period, there were 64 maternal deaths and 10 891 live births at MGH, resulting in an institutional MMR of 588 per 100 000 live births (Table 1). Twelve women were excluded from detailed analysis owing to a lack of case notes. Limited information in registers indicated that these maternal deaths were caused by obstetric hemorrhage (n = 8), sepsis (n = 3), and ruptured ectopic pregnancy (n = 1). As a result, data from 52 women were analyzed in detail. The mean ± SD age of the women who died was 28 ± 6.4 years (Table 2). Most were married and lived in rural areas. Two-thirds of patients (n = 33) lived 3 hours or more in travel time distance from MGH. Twenty-five percent of women (n = 13) were nulliparous, and 17.3% were grand multiparous (Table 3). Among parous women (71.2%), the median parity was 3 (IQR, 1–4.5, range 1–10). Twelve (23.1%) had a previous obstetric complication. Only 18 women (34.6%) had attended prenatal care in the index pregnancy.
Valuesa
Characteristic
28 ± 6.4 28 (19–38) 17–42 46 (88.5) 1 (1.9) 5 (9.6) 15 (28.9) 14 (26.9) 3 (5.8) 20 (38.4 15 (28.9) 6 (11.5) 4 (7.7) 27 (51.9) 42 (80.8) 5 (9.6) 3 (5.8) 2 (3.8) 45 (86.5) 7 (13.5) 11 (21.2) 17 (32.7) 5 (9.6) 4 (7.7) 5 (9.6) 10 (19.2) 13 (25.0) 6 (11.5) 11 (21.2) 14 (26.9) 1 (1.9) 7 (13.5) 40 (76.9) 12 (23.1)
a
Values are given as number (percentage) of women unless stated otherwise. Estimates were made on the basis of geographic distance and transport available for travel from residence to hospital. b
Most cases were direct maternal deaths (n = 37, 71.2%) (Table 4, Supplementary Material S1), and hemorrhage (owing to a range of underlying causes) was a prominent factor in these cases. Uterine rupture secondary to prolonged obstructed labor was common. Of the indirect maternal deaths (n = 15, 28.8%), non-pregnancy-related infection was the most common cause (n = 11, 21.2%). There was 1 death in early pregnancy owing to septic abortion.
Table 1 Comparison of maternal deaths, live births, and MMR between Modilon General Hospital data and Madang PHIO data 2008–2012. Year
2008 2009 2010 2011 Total 2008–2011 2012b Total 2008–2012
Modilon General Hospital (review)
Madang Province (PHIO)a
Modilon General Hospital (PHIO)
Maternal deaths
Live births
MMR
Maternal deaths
Live births
MMR
Maternal deaths
Live births
MMR
14 8 23 15 60 4 64
1633 2043 2669 2753 9098 1793 10891
857 392 862 545 660 223 588
4 4 6 5 19
501 569 526 552 2148
798 703 1141 906 885
13 19 34 28 94
17543 18098 18513 19010 73164
74 105 184 147 128
Abbreviations: MMR, maternal mortality ratio; PHIO, Provincial Health Information Office. a Includes data from Modilon General Hospital. b January 1 to July 31, 2012; PHIO data for 2012 unavailable.
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127 Table 3 Clinical characteristics of maternal death cases (n = 52). Characteristic Parity Nulliparous 1–2 3–4 Grand multipara (≥5) Missing data Gestational age First trimester Second trimester Third trimester Prenatal clinic attendance Yes No Multiple pregnancy Twin Delivered singleton Undelivered singleton Undelivered and unknown Missing data Previous obstetric history Stillbirth Retained placenta Multiple pregnancy Postpartum hemorrhage No documented history Mode of delivery Singleton pregnancies SVD Breech delivery EMLUSC EMLUSC & TAH Vacuum extraction Twin pregnancies (twins 1 & 2) SVD & SVD SVD & forceps SVD & vaginal delivery after internal podalic version Vacuum extraction EMLUSC Not applicable (undelivered) Fetal outcome Singleton pregnancies Live birth Stillbirth Twin pregnancies Both live births Twin 1 live birth, twin 2 stillbirth Not applicable (undelivered) Time between hospital admission and death Dead on arrival ≤1 h 1–6 h 6–24 h 24–48 h 2 d to 1 wk 1–2 wk ≥2 wk Frequency of delay ≥1 delay No delay Type of delay Phase 1 Phase 2 Phase 3 None
125
Table 4 Principal underlying causes of maternal death (n = 52). Number (%) 13 (25.0) 15 (28.9) 13 (25.0) 9 (17.3) 2 (3.8) 1 (1.9) 5 (9.6) 46 (88.5) 18 (34.6) 34 (65.4) 8 (15.4) 34 (65.4) 4 (7.7) 5 (9.6) 1 (1.9) 5 (9.6) 4 (7.7) 1 (1.9) 2 (3.9) 40 (76.9)
18 (34.6) 5 (9.6) 7 (13.5) 2 (3.6) 2 (3.6) 2 (3.6) 1 (1.9) 1 (1.9) 2 (3.9) 2 (3.9) 10 (19.2)
19 (36.5) 15 (28.9) 5 (9.6) 3 (5.8) 10 (19.2) 2 (3.9) 2 (3.9) 5 (9.6) 8 (15.4) 11 (21.2) 15 (28.9) 6 (11.5) 3 (5.8) 45 (86.5) 7 (13.5) 21 (40.4) 23 (44.2) 35 (67.3) 7 (13.5)
Abbreviations: EMLUSC, emergency lower uterine segment cesarean; SVD, spontaneous vertex delivery; TAH, total abdominal hysterectomy.
Most women were in the third trimester of pregnancy (Table 3). Among women in the second and third trimester of pregnancy (n = 51), 9 (17.7%) died before, 2 (3.9%) died during, and 40 (78.4%) died after delivery (8 of whom were unsupervised). Multiple pregnancy, breech presentation, and grand multiparity were common features in these cases (Table 3).
Underlying cause of death
Number (%)
Direct maternal deaths Hemorrhage Uterine rupture Third-stage hemorrhagea Postpartum hemorrhage (uterine atony) Prepartum hemorrhage High vaginal laceration Placenta accreta Pregnancy-related infection Hypertensive disorder Other Obstetric embolism Retained placenta (with sepsis) Complication of anesthesia Septic abortion Indirect maternal deaths Non-pregnancy-related infection Likely bacterial sepsis Malaria Tuberculosis Other Cardiac disease Malignancy
37 (71.2) 24 (46.2) 7 (13.5) 6 (11.5) 6 (11.5) 5 (9.6) 1 (1.9) 1 (1.9) 5 (9.6) 4 (7.7) 4 (7.7) 1 (1.9) 1 (1.9) 1 (1.9) 1 (1.9) 15 (28.8) 11 (21.2) 9 (17.3) 1 (1.9) 1 (1.9) 4 (7.7) 2 (3.9) 2 (3.9)
a
Hemorrhage owing to retained placenta.
Common causes of death among women with twin pregnancies were pre-eclampsia or eclampsia (n = 3) and hemorrhage (atony, n = 2; uterine rupture, n = 1). Six grand multiparous women died of hemorrhage. Among women with a history of retained placenta (n = 4), 2 had a recurrent retained placenta, 1 had placenta accreta, and 1 had uterine rupture. In 15 of 34 (41.2%) singleton pregnancies, the neonate was stillborn (Table 3). Among women with a documented history of stillbirth (n = 5), 4 had a recurrent stillbirth. Most women (n = 40, 76.9%) had been referred from rural aid posts (Table 3). On hospital admission, midwives commonly assessed the patient first (n = 24, 46.2%), followed by junior medical staff (n = 23, 44.2%). Forty-seven women (90.4%) had indications of physiologic compromise on arrival (≥1 of the following: blood pressure b90/60 or N140/90 mm Hg; heart rate N 110 beats per minute; temperature ≥37.5 °C; respiratory rate N24; anuria or oliguria). Two-thirds of the women (35/52) died 24 hours or more after admission (Table 3). Sixteen women (30.8%) required instrumental or operative assistance (Table 3). Fifty-two percent (n = 27) of women were transfused with a median of 2 units of red cell concentrate (range, 1–7 units; mean, 2.7 units). Among the women who died because of postpartum hemorrhage (after delivery of both neonate and placenta either owing to atony alone or associated with uterine rupture or prolonged third stage), only half (9/20) had documented use of uterotonic drugs. Most women who died from pregnancy- or non-pregnancy-related sepsis (n = 12/14, 85.7%) had been referred: only 3 had received broad-spectrum antibiotics before arrival, and 9 received them at MGH. Malaria was suspected for 23 women (44.2%), and blood slides were examined for 4 women: 2 were parasitemic, and 1 woman died as a consequence (Table 4). Nineteen received artemether as per national guideline [16]. Data on uptake of intermittent preventive treatment for malaria and HIV status were unavailable. Forty-five women (86.5%) experienced 1 or more delays, 27 (51.9%) had 2 or more delays, and phase-3 delays were the most common (Table 3). Phase-3 delays occurred at MGH (n = 18), the referring health center (n = 12), or both (n = 5). Delays at health centers included lack of adequate treatment (e.g. antibiotics, uterotonic drugs, magnesium sulfate, or fluids) (n = 17) and delayed referral (n = 8). At hospital, 14 women experienced delays in receiving blood, 10 had delays in their review by senior medical staff, and 10 had delays in crucial treatment (antibiotics, magnesium sulfate, fluids). Others delays
126
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127
included review by junior medical staff (n = 3) and necessary surgical intervention (n = 3). There were discrepancies between the number of maternal deaths and live births at MGH identified in the hospital records and those published in PHIO summary statistics. In the period 2008–2011, there was underreporting to PHIO of maternal deaths (n = 19/60, 31.7%) and live births (n = 2148/9098, 23.6%) at MGH (Table 1). During this period, the number of deliveries at MGH almost doubled, but PHIO figures did not reflect this trend (Table 1). 4. Discussion Maternal death was found to be common at a provincial referral hospital in PNG, and most women died as a result of hemorrhage or infection. These causes of maternal death have been previously recorded as predominant in PNG, although regional heterogeneity exists [9,10,17]. Only 2 indirect deaths (due to malignancy) could not have been prevented through adequate provision and uptake of family planning, prenatal care, supervised delivery, and basic or comprehensive EmOC. Most women lived rurally at some distance from MGH, and many had been referred from aid posts. The recent drive to improve capacity for maternity care at the community level, including basic EmOC, will assist in improving case management and reduce the high number of direct maternal deaths [18–20]. Many women did not seek or receive prenatal care, although the review was unable to identify reasons for this low uptake. Obstetric risk factors were frequently present, and it is important that women with high-risk pregnancies seek prenatal care early, are recognized as such, and are referred to hospital before delivery. Effective mechanisms for transport to facilities that can provide comprehensive EmOC will help [18–20]. Waiting houses for obvious high-risk pregnancies may further reduce the risk of maternal death [10], but require evaluation in the context of PNG. Most women experienced at least 1 delay—that is, a missed “window of opportunity” that was potentially critical to the outcome. Research indicates that delays are ultimately related to women’s poverty, lack of education, information, and awareness, adverse experiences with healthcare staff, and gender inequality (phase 1); a lack of safe, accessible, and low-cost transport mechanisms (phase 2); and a shortage of medical supplies, and trained and motivated staff at health facilities (phase 3) [10,15,21]. Many women did not receive appropriate fluid resuscitation and treatment (antibiotics, uterotonic drugs, or magnesium sulfate) at the health center level, for which a lack of supplies and basic EmOC providers are possible explanations. At the hospital level, there were delays in blood transfusion: a failure to recognize the need for transfusion, and a lack of staff, blood bank reserves, and reagents are likely contributors to this. Laboratory services undoubtedly need strengthening, given the high number of women requiring blood and low numbers of women who had blood smear microscopy for malaria. The lack of a specialist obstetrician at the institution during parts of the study period, and problems with prompt recognition and treatment of ill women by junior staff might have prevented some women from receiving optimal care, including operative delivery. Delays will have contributed to the irreversible physiologic compromise that most women were experiencing by the time of arrival at hospital: women frequently succumbed to renal failure, disseminated intravascular coagulation, and overwhelming sepsis within 48 hours of admission (data not shown). By contrast, patients who died after having been admitted to hospital for some time commonly had chronic conditions such as malignancy, tuberculosis, and rheumatic heart disease. The present study has captured only a proportion of maternal deaths occurring in Madang province. Women die at rural health centers (Table 1), and some do not reach a health facility [10] and their deaths remain undetected. It is likely that proportionally more deaths “occur” at hospital than in rural areas [22]; hence, institutional MMRs need to
be interpreted with caution. Nevertheless, the observed underreporting of deaths to PHIO might also occur at the health center level. The use of PHIO data to estimate the burden of, and progress with reducing, maternal deaths may lead to imprecise and therefore potentially misleading data. Both maternal deaths and live births were found to be disproportionately underreported, further decreasing the precision of MMR estimates. Suboptimal documentation and data flow between the hospital and PHIO are possible explanations for this. Real-time surveillance mechanisms for maternal deaths are required [14], and clear and irrefutable post-MDG5 targets, such as reducing the MMR to below 30 per 100 000 by 2030, are more useful [23]. Establishing a national maternal death surveillance and response system that expands and optimizes the current use of maternal death proformas complies with aims of the PNG National Health Plan [20]. Technical support from the WHO to implement this activity is underway [14], and will facilitate a swift, efficient, and specific response to monitoring maternal deaths, at district and provincial levels. The use of current PHIO figures for health budgeting purposes will result in suboptimal resource allocation: the number of deliveries at MGH almost doubled over the study period, but this trend was not reflected in PHIO statistics (Table 1). The National Health Plan aims to increase the number of health facilities able to provide comprehensive EmOC [20]. The present findings suggest that this should be preceded by strengthening the capacities of existing facilities in terms of staffing, training, ambulance services, and laboratory and blood transfusion services. Future research might include a prospective evaluation of “near misses” [24]—that is, women whose lives were saved but only just. Furthermore, studies might attempt to evaluate the frequency of late maternal deaths in the area [25], and to review each maternal death soon after it occurred, with the aim of clearly establishing the exact nature of possible delays (by interviewing relatives and staff). Research evaluating underreporting at the aid post level and estimating deaths in villages will broaden our understanding of this hidden burden of maternal death in PNG. The present study has limitations. First, not all cases had notes available for in-depth review. Second, late maternal deaths (those occurring N42 days but b 1 year postpartum) were not reviewed because the documentation in registers was inadequate to attempt their evaluation. Third, it is possible that the present review missed maternal deaths in early pregnancy—regional differences may only partly explain this [9]— and women may have died of ruptured ectopic pregnancy and abortions (including septic abortions) prior to reaching a formal healthcare provider. Last, documentation bias may have resulted in underestimating delays. When a delay was evident, its underlying cause could not always be established from case notes alone and qualitative research is needed in this area [10]. Current government strategies need to be complemented by efforts to improve the capacities of existing providers of comprehensive EmOC. Establishing real-time maternal death surveillance systems that can monitor progress and provide precise data for resource allocation planning are needed. Further research is required to identify the burden of village deaths and the barriers preventing women in PNG from receiving the pregnancy care that they need. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijgo.2013.08.012. Acknowledgments H.W.U. and A.J.U. received salary support from the Malaria in Pregnancy Consortium, which receives funding from the Bill & Melinda Gates Foundation. Conflict of interest The authors have no conflicts of interest.
J.W. Bolnga et al. / International Journal of Gynecology and Obstetrics 124 (2014) 123–127
References [1] Hogan MC, Foreman KJ, Naghavi M, Ahn SY, Wang M, Makela SM, et al. Maternal mortality for 181 countries, 1980–2008: a systematic analysis of progress towards Millennium Development Goal 5. Lancet 2010;375(9726):1609–23. [2] Hill K, Thomas K, AbouZahr C, Walker N, Say L, Inoue M, et al. Estimates of maternal mortality worldwide between 1990 and 2005: an assessment of available data. Lancet 2007;370(9595):1311–9. [3] United Nations General Assembly. United Nations Millennium Declaration. A/Res/55/ 2. http://www.un.org/millennium/declaration/ares552e.pdf. Published September 18, 2000. Accessed February 18, 2013. [4] The World Bank. Papua New Guinea - Reproductive health at a glance. http:// documents.worldbank.org/curated/en/2011/04/14695221/papua-new-guineareproductive-health-glance. Published April 2011. Accessed February 21, 2013. [5] United Nations Development Programme. Papua New Guinea Country Profile: Human Development Indicators. http://hdrstats.undp.org/en/countries/profiles/PNG.html. Published 2011. Accessed February 18, 2013. [6] World Health Organization, UNICEF, UNFPA, The World Bank. Trends in maternal mortality: 1990 to 2010. http://whqlibdoc.who.int/publications/2012/9789241503631_ eng.pdf. Published 2012. Accessed February 16, 2013. [7] Papua New Guinea National Statistical Office. Papua New Guinea Demographic and Health Survey 2006: National Report. Port Morseby: Papua New Guinea Government Printer; 2009. [8] United Nations Population Fund. The State of the World’s Midwifery 2011. Country data and profiles: Papua New Guinea. http://www.unfpa.org/sowmy/resources/en/ country_info.htm. Published 2011. Accessed February 12, 2013. [9] Sanga K, de Costa C, Mola G. A review of maternal deaths at Goroka General Hospital, Papua New Guinea 2005–2008. Aust N Z J Obstet Gynaecol 2010;50(1): 21–4. [10] Kirby B. Too Far to Travel—Maternal Mortality in Milne Bay Province—Beyond the Numbers. THOR – Hands of Rescue Foundation: Alotau; 2010. [11] World Health Organization. The WHO application of ICD-10 to deaths during pregnancy, childbirth and puerperium: ICD MM. http://apps.who.int/iris/bitstream/ 10665/70929/1/9789241548458_eng.pdf. Published 2012. Accessed January 17, 2013. [12] World Health Organization. Health statistics and health information systems: Maternal mortality ratio (per 100 000 live births). http://www.who.int/healthinfo/
[13]
[14]
[15] [16] [17] [18]
[19]
[20]
[21]
[22] [23] [24]
[25]
127
statistics/indmaternalmortality/en/index.html. Published 2010. Accessed January 12, 2013. Cross S, Bell JS, Graham WJ. What you count is what you target: the implications of maternal death classification for tracking progress towards reducing maternal mortality in developing countries. Bull World Health Organ 2010;88(2):147–53. Hounton S, De Bernis L, Hussein J, Graham WJ, Danel I, Byass P, et al. Towards elimination of maternal deaths: maternal deaths surveillance and response. Reprod Health 2013;10:1. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med 1994;38(8):1091–110. Papua New Guinea National Department of Health. Malaria Treatment Protocol. Port Moresby: Papua New Guinea National Department of Health; 2009. Garner P, Lai D, Baea M. Childbirth in rural areas: maternal deaths, village deliveries and obstetric service use. P N G Med J 1994;37(3):166–72. Morgan C, Byrne A. Improving maternal, newborn and child health in PNG through family and community care. http://pacificinstitute.anu.edu.au/outrigger/2011/12/ 11/improving-maternal-newborn-and-child-health-in-png-through-family-andcommunity-care/. Published December 12, 2011. Accessed February 23, 2013. Papua New Guinea National Department of Health. Ministerial Taskforce on Maternal Health in Papua New Guinea. http://www.unfpa.org/sowmy/resources/docs/ library/R149_DOH_PNGUINEA_2009_Ministerial_Taskforce_report_final_version_3. pdf. Accessed January 19, 2013. Government of Papua New Guinea. National Health Plan 2011–2020: Volume 1 Policies and Strategies. http://www.wpro.who.int/countries/png/PNGNHP_Part1.pdf. Published June 2010. Hunt P, Bueno de Mesquita J. Reducing Maternal Mortality: The contribution of the right to the highest attainable standard of health. http://www.unfpa.org/webdav/ site/global/shared/documents/publications/reducing_mm.pdf. Published 2010. Ronsmans C, Graham WJ. Lancet Maternal Survival Series steering group. Maternal mortality: who, when, where, and why. Lancet 2006;368(9542):1189–200. Gilmore K, Gebreyesus TA. What will it take to eliminate preventable maternal deaths? Lancet 2012;380(9837):87–8. Cecatti JG, Souza JP, Oliveira Neto AF, Parpinelli MA, Sousa MH, Say L, et al. Pre-validation of the WHO organ dysfunction based criteria for identification of maternal near miss. Reprod Health 2011;8:22. Høj L, da Silva D, Hedegaard K, Sandström A, Aaby P. Maternal mortality: only 42 days? BJOG 2003;110(11):995–1000.
