Professor Affette McCaw-Binns, a Reproductive Health Epidemiologist at the University of the West Indies (Mona) and Dr Simone Spence, Director of Family Health Services at the Ministry of Health in Jamaica explain how legislation and policy strengthened the reporting of maternal deaths in Jamaica. This case study describes how the policy framework was amended to improve the reporting of maternal deaths and how other interventions implemented simultaneously together strengthen the maternal death surveillance and response (MDSR) system.
In the early 1980s1,2, maternal deaths in Jamaica were significantly under-reported in vital registration records by as much as 75%. With over 80% of all live births occurring in public hospitals2 it was suggested that establishing a surveillance system at public hospitals could capture needed information about the number of maternal deaths in the country. Given the findings3, the government agreed to implement an active (as opposed to the pre-existing passive) surveillance system to monitor maternal deaths.
This case study will describe the approaches that the government adopted, including how the legal framework was used in support of strengthening the MDSR system and reversing under-reporting.
Approaches to improving reporting of maternal deaths
Maternal mortality as a notifiable event
In Jamaica, maternal mortality was first classified as a Class I notifiable event in 1998 and integrated into the existing Class I disease surveillance system that monitors infectious diseases under the 1985 Public Health Act, revised in 20044,5,6. This Act allows the Minister of Health to legally “make regulations [….] in relation to a notifiable and communicable disease, the treatment and prevention thereof […]”6.
By classifying a maternal death as a notifiable event within the existing disease surveillance framework, there was no need for separate legislation to enable the surveillance of maternal deaths. It was, therefore, more efficient to include maternal mortality surveillance under the umbrella of existing laws and regulations as legislative change can be a lengthy process.
Defining a maternal death for surveillance purposes
At the start, the definition of a “maternal death” was simplified for surveillance purposes to “any death in a woman of reproductive age (described as 10-50 years) with evidence of a pregnancy in the last six weeks.”8 Each week, surveillance nurses would check hospital death registers for women of reproductive age who died. Once identified, their medical records were examined for any report of a pregnancy in the past six weeks. Where evidence of a pregnancy was found, the death was reported “on suspicion” to the National Epidemiology Surveillance Unit and the Parish (local) Health Department, which would conduct the investigation. This simplification made it easier for non-clinicians to identify potential maternal deaths and reduced the number of deaths that may have been overlooked.
An evaluation of the maternal mortality surveillance system was conducted to assess the acceptability of the process and develop strategies to ensure improvements in maternal health9,10. Participants reported that the process had improved awareness of the problem however they expressed concern that the WHO definition11 of a maternal death may limit understanding of factors contributing to pregnancy-related deaths, especially if lives had been prolonged by Intensive Care Unit admission or if women died just outside that six-week period from factors arising from pregnancy (e.g. cardiovascular complications). They also felt it necessary to make an effort to identify deaths in the private sector and the community (e.g. ruptured ectopic pregnancies, abortions and external causes such as suicide, accidents or violence)10.
As a result of this evaluation and to improve surveillance, the case definition of maternal mortality was expanded to include late maternal and coincidental deaths and is now: “Death in any woman of reproductive age (10-50 years) in whom there is evidence of pregnancy or termination of pregnancy within the past one year”5 (p.4). The Maternal Mortality Surveillance Guidelines (revised 2012)5 set out clear criteria for identifying cases as:
- “Gender: Female
- Age: 10-50 years
- Evidence of Pregnancy: Pregnant and died undelivered, history of abortion, miscarriage, stillbirth or live birth
- Period of observation for Pregnancy: From date of death to one year preceding death” (p.4).
Active versus passive surveillance
Because maternal deaths are notifiable events, all clinicians including those in the private sector are expected to report deaths suspected of being pregnancy related within 24 hours of occurrence to the relevant authorities. This process is known as passive surveillance, but can be unreliable and prone to reporting bias.
In order to strengthen the MDSR system, the public sector adopted an active surveillance system to improve the number of maternal deaths that were identified and investigated. Active surveillance involves Class I surveillance officers conducting weekly record reviews to identify potential cases occurring in public hospitals. All deaths among women of reproductive age are reviewed for evidence of pregnancy in the last year; those meeting the case definition criteria are then investigated.7
In contrast to public facilities, the private sector attends less than 5% of all births. Cases with severe complications requiring tertiary care (e.g. intensive care) are usually transferred to the public sector. Aside from precipitate events (e.g. obstetric haemorrhage) that can lead to a rapid death with little opportunity to arrange for a transfer to intensive care in the public sector, and given the small number of women attended, maternal deaths in the private sector are rare10,12. Thus, the vast majority of maternal deaths due to known complications should be identified through the public sector.
The system has also strengthened the reporting of deaths outside public hospitals (e.g. in the community and private sector). Community health teams will actively follow up on information (e.g. “rumours”) about a recent death of a mother. Community health workers in primary care assigned to work in specific geographical areas regularly know the mothers in their catchment population and would be aware of these events. Other sources of information about possible maternal deaths include press and police reports as well as communications from community members and healthcare providers in both the public and private sectors. Incomplete information on confirmed cases will be noted and at minimum documented as a death due to undetermined causes.
Since the system began in 1998, MDSR records have never been used for litigation against a health worker or facility. Cases are reviewed as confidential enquiries where the names of attending physicians are not included in the case-investigation and documentation process. The data are summarised on regional and national levels, and no individual case reports are shared outside the case review process.
Focus on systemic failures
In Jamaica, review meetings were established at the start of the process and are led by trained reviewers to focus on systemic failures (i.e. failures of the health system), rather than individual circumstances that may have contributed to a death. In some instances, a death is found to be “unavoidable”. However, where “avoidable features” have been identified, the review team tries to determine how the death could have been prevented (e.g. personnel, equipment, skills and other resources). If, during a review, more than one death occurred from a similar condition, then the team reviews the cases together, again focusing on the systemic issues and the challenges in managing these conditions.
To help reinforce the opportunity to learn from the case review process and the overall principles of the system, members of the national committee make an effort to attend the regional reviews across the four regions of the country. Regional reviews are attended by multi-disciplinary teams of health providers including community health aides, midwives, public health nurses, obstetricians, public health specialists, pathologists, epidemiologists and statisticians. Where specific skill gaps are identified, local or national training programmes are developed. The response component is most effective when specific individuals are assigned to lead the process to address particular deficiencies at the local level. Problems common to multiple health teams are discussed at national meetings and are shared with the national policy directorate for resolution by the Director of Family Health Services.
Challenges and lessons learnt
Deaths in early pregnancy
A 2008 study12 found that the surveillance process missed precipitate deaths in the community especially if the patient did not interact with the public sector team at any time. These include deaths in the first-trimester of pregnancy where the family or the community may have been unaware that the deceased was pregnant (e.g. ectopic pregnancy, terminations of pregnancy). Later in pregnancy, obstetric emergencies such as eclampsia or placental abruption can result in a rapid death. When sudden deaths such as these occur, they are usually investigated by pathologists from the Ministry of Justice. As these may become medico-legal cases, delays occur while the case is investigated by the justice system, limiting the timely reporting and registration of these deaths. As a result, sudden deaths are not routinely reported to the surveillance system. Health officials, however, are seeking to address this gap by having the Ministry of Health participate in an inter-agency committee aimed to improve inter-sectoral data sharing.
The Registration (Births and Deaths) Act13 requires that all stillbirths of 28-weeks of gestation or more be registered2. Any live birth and subsequent death regardless of age must be registered under this Act, including early and late neonatal deaths.
A review of Jamaica’s 2008 vital registration system for foetal and under-five deaths14 documented challenges in reporting stillbirths (69%), early neonatal deaths (89%) and late neonatal deaths (83%). Most under-reporting of deaths among children under-five years of age was due to delays in registering Coroners’ cases, as was similarly noted above for maternal deaths.
International guidelines recommend that the target for completeness of reporting should be at least 90%15 of all known deaths and registered within sufficient time to be included in routine demographic statistics. This review14 found that whilst only 26% of sudden deaths among children under-five were registered within a year of the event, 91% of all under-five deaths certified by an attending physician were registered within the recommended time frame.
Due to the challenge in registering stillbirths, routinely generated mortality statistics (based on in-patient data from the Ministry of Health) will help monitor and summarise the number and registration of stillbirth occurrences in each facility on a monthly basis.
In the health sector, the mortality register is used to monitor the event of the death of a woman of reproductive age across all hospital wards. The screening of these deaths is, therefore, not confined to cases on the obstetric block alone.
For deaths occurring outside the health sector, the Ministry of Health is awaiting the formalisation of the data-sharing agreement (mentioned above) with the Ministry of Justice. In the interim, the Ministry of Health continues to work with community sources (e.g. “rumours”, media reports, etc.) to identify cases of women of reproductive age who were recently pregnant and died.
In response to the persistent under-reporting of maternal deaths through vital registration12, the Registrar General’s Department added a comprehensive pregnancy check box to the medical certificate of cause of death. Early returns suggest that reporting of maternal deaths through the civil registration process has improved significantly16,17.
Managing indirect causes of maternal death
Jamaica has improved its management of direct pregnancy-related complications and has had some successes with indirect conditions such as HIV and Type II diabetes. However, challenges persist due to the changing epidemiological profile of the antenatal population (i.e. delayed childbearing and an increase in essential morbidity e.g. obesity, essential hypertension, heart disease, diabetes, sickle cell disease) as well as a health and training system still focused on attending young, healthy, low-risk mothers.
Today, the care of a woman of reproductive age is complex and requires access to a range of specialists. In Jamaica, working across disciplines (e.g. obstetrics and internal medicine) has resulted in new approaches, such as improving the skill set of the obstetric team by training specialists in maternal-foetal medicine. The first group of fellows will begin practicing in 2017.
In response to the lack of adequate infrastructure, the Ministry of Health is also seeking to improve equipment provision at referral hospitals. It is expected that this process will be lengthy as it requires retrofitting the health facilities (e.g. new construction). This development is being supported by an external grant and loan financing.
Furthermore, there is a need to review and adjust the current policy framework for service provision, including the development of and training in using the revised treatment guidelines. These interventions are challenged by staff turnover in a relatively under-financed public health sector.
Multidisciplinary teams and the public
The national Maternal Death Surveillance Guidelines5 provide the policy framework that supports multidisciplinary reviews. These reviews include clinicians, epidemiologists, pathologists, community nurses, midwives and community health workers, to name a few, which enable the triangulation of information across the health team and levels of care. The development of new strategies that focus on effective responses, will help inform how to involve managerial officers in the review and response process.
There is a long-term need to improve communication channels amongst and across members of the health team and the wider public. Investing in locally appropriate solutions that are effective and evidence-based can also help strengthen relations between patients and the health team as well as develop innovative approaches in engaging the community to address barriers to accessing timely care.
Two decades of MDSR experience in Jamaica has shown that a more robust process, with an enabling and pragmatic legal and policy framework, can be established and grow when based on the collective support of key stakeholders. However, for the system to be effective it must respond to the stories of each mother’s death to identify the needs of both the community and health sector. The social, cultural and quality of care aspects related to a mother’s death are all important to consider to determine the contributing factors of each event. The local setting must be taken into account as solutions to outcomes (e.g. eclampsia) can vary across settings. For example, the underlying determinant in a remote rural area may be transportation while elsewhere it may be limited access to evidence-based methods, such as magnesium sulphate.
1 Walker, G.J., Ashley, D.E., McCaw, A.M., Bernard, G.W. (1986). Maternal mortality in Jamaica. The Lancet, 1(8479): 486-8.
2 McCaw-Binns, A.M., Fox, K., Foster-Williams, K.E., Ashley, D.E., Irons, B. (1996). Registration of births, stillbirths and infant deaths in Jamaica. International Journal of Epidemiology, 25(4):807-13).
3 McCaw-Binns, A., Standard-Goldson, A., Ashley, D., Walker, G., MacGillivray, I. (2001). Access to care and maternal mortality in Jamaican hospitals: 1993-95. International Journal of Epidemiology, 30(4): 796-81.
4 GTR (McCaw-Binns, A. & Mullings, J.). (n.d.). Epidemiological Surveillance of Maternal Mortality (1981-2012): Country: Jamaica. Retrieved 1 December, 2016, from https://www.msh.org/sites/msh.org/files/gtr_casestudy_jamaica_eng.pdf
5 Ministry of Health [Jamaica]: Family Health Services. (2012). Maternal Death Surveillance Guidelines. Kingston: Government of Jamaica.
6 Ministry of Justice [Jamaica]. (1985). Public Health Act. Retrieved 9 January, 2017, from http://moj.gov.jm/laws/public-health-act
7 McCaw-Binns. A. & Lewis-Bell. K. (2009). Small victories, new challenges: two decades of maternal mortality surveillance in Jamaica. West Indian Medical Journal, 58(6): 518-32
8 Spence, S. Ministry of Health in Jamaica, Director of Family Health Services Jamaica. Personal Communication. 19 December 2016.
9 McCaw-Binns, A., Alexander, S.F., Lindo, J.L., Escoffrey, C., Spence, K., Lewis-Bell, K., Lewis, G. (2007). Epidemiologic transition in maternal mortality and morbidity: new challenges for Jamaica. International Journal of Gynaecology and Obstetrics, 96(3): 226-32.
10 McCaw-Binns, A.M., Lindo J.L.M., Lewis-Bell, K.N., Ashley, D.E. (2008). Maternal mortality surveillance in Jamaica. International Journal of Gynecology and Obstetrics, 100(1): 31-36.
11 World Health Organization (WHO). (2012). The WHO Application of ICD-10 to deaths during pregnancy, childbirth and the puerperium: ICD-MM. Geneva: WHO.
12 McCaw-Binns, A.M., Mullings, J.A., Holder, Y. (2015). Vital registration and under-reporting of maternal mortality in Jamaica. International Journal of Gynaecology and Obstetrics, 128(1):62-7.
13 Ministry of Justice [Jamaica]. (1889, amended 1982). Registration (births and deaths) Act. Retrieved 12 February 2017, from http://moj.gov.jm/laws/registration-births-and-deaths-act
14 McCaw-Binns, A., Mullings, J., Holder, Y. (2015b). The quality and completeness of 2008 perinatal and under-five mortality data from vital registration, Jamaica. West Indian Medical Journal, 64(1):3-16.
15 United Nations (UN). (2014). Principles and recommendations for a vital statistics system: Revision 3. New York: UN.
16 Statistical Institute of Jamaica. (2016). Demographic Statistics: 2015 (Table 27). Retrieved on 12 February 2017, from http://statinja.gov.jm/PublicationReleases.aspx. Accessed 12 February 2017.
17 Statistical Institute of Jamaica. (2015). Demographic Statistics: 2014 (Table 27). Retrieved on 12 February 2017, from http://statinja.gov.jm/PublicationReleases.aspx
18 Asnani, M.R., McCaw-Binns, A.M., Reid, M.E. (2011). Excess risk of maternal death from sickle cell disease in Jamaica: 1998-2007. Plos One.