SA and Brazil: the C-section Champions League 

Caesarean section (C-section) rates are higher in SA private hospitals than in American “celebrity hospitals”. But Brazil’s are even higher. Why so high and what can be done?

The caesarean section (C-section) rate just published in Discovery Health’s 2017-18 Healthcare Claims Tracker has prompted media comparisons with US hospitals where celebrity moms like Gwyneth Paltrow, Kim Kardashian and Victoria Beckham deliver their celebrity offspring. Cedars-Sinai (Los Angeles) (24.3%) and Mount Sinai Hospital in New York (21.5%) lag behind Discovery’s reported rate of 74.5%.

Medicolegal concerns are cited as the prime driver of the overwhelming local preference for C-sections over vaginal birth. Practitioners now pay enormous annual premiums for medical malpractice insurance to cover themselves against multi-million rand legal awards that are being made for obstetrical adverse events. The perception is that operative delivery is the safer option and protective against lawsuits; the vast majority of such suits are apparently associated with labour and vaginal birth. This is despite the reality that even elective C-sections in healthy women are associated with a host of potential complications (see below).

Brazil is an interesting and appropriate comparator, with a mix of public and private health care delivery not dissimilar to South Africa’s. Surprisingly, Brazil’s overall C-section rate – 52% – is the highest in the world. More surprising still: Brazil’s private hospitals brought 88% of newly minted citizens into the world by the abdominal route.

What’s the problem?

With C-sections the norm in the private sector even in low risk pregnancies it may be difficult to convince many local practitioners, as well as mothers-to-be, that there is a problem. It is true that C-sections save lives in high-risk deliveries and the safety of operative delivery has improved over the last few decades but C-sections are not minor interventions.

Potential maternal complications include postpartum haemorrhage (linked to a higher risk for hysterectomy), wound infections, venous thrombosis, pelvic and abdominal trauma/injury, not to mention the risks of anaesthesia. Subsequent fertility may be reduced and future pregnancies complicated by abnormal placentation, uterine rupture, or difficult repeat surgery. Bowel obstruction may occur as a late complications secondary to adhesions. Risks for the baby include accidental lacerations, respiratory complications (especially in elective C-sections before term) leading to NICU admission, and a lower success rate for breastfeeding. Babies delivered abdominally develop a different gut microbiome, with negative implications for subsequent health and development.

What’s the “right” cesarean section rate?

Across European countries, only 16% of Dutch babies are delivered by C-section. Sweden is at 18%, France 21%, Spain 25%, the UK 26%, and Germany 30%. Excellent maternal and neonatal outcomes are recorded in countries with C-section rates a third of Discovery’s; it seems reasonable then to conclude that rates in the SA private sector are too high.

The “optimal” overall C-section rate is not obvious however and lower is not necessarily better. These operations are life-saving in certain situations. Rates below 10% are a major public health problem, linked to major maternal and neonatal mortality and morbidity. A recent Lancet study across 22 African countries revealed alarming perinatal outcomes including a C-section-related maternal mortality rate 50x higher than in high income countries, catastrophic not only for mothers but for children left behind. These severe outcomes reflect operative deliveries that occur too late in patients with severe underlying risk factors and hospitals with inadequate resources and systems for safe delivery.

Without minimizing the “defensive medicine” issue it seems clear that solo, private, obstetrical practice has built-in structural factors that are driving our C-section rates and making it the default choice even in low risk pregnancy. Not working in a team – which could include skilled midwives and birth assistants, as well as fellow obstetricians and anaesthesiologists – obstetrical specialists are alone facing the stress and inconvenience of unplanned elective surgery, after-hours work and intrusions into daily appointment schedules that are difficult to deal with.

Maternal preferences are no doubt also a factor. Patients are said to desire an exact delivery date that is convenient for them and their families. A certain local celebrity is on record as being disappointed that the birth of his first child could not be arranged to occur under a favorable astrological sign. Less frivolously, many women have a natural fear of labour pain and a realistic concern about long lasting pelvic or perineal injury and dysfunction. It is perceived that “patients do not value a natural birth anymore and want to be in control.”

South African responses 

Tinkering with financial incentives has not moved the dial; obstetricians are now, on average, reimbursed by medical schemes slightly less for a C-section than for a vaginal birth but rates have not changed.

In response to these issues, the BetterObs programme was initiated by the South African Society of Obstetricians and Gynaecologists. This program promotes “safer deliveries, healthier babies, better outcomes” and therefore, it is assumed, fewer opportunities for litigation.

Participating obstetricians must adhere to a published protocol, attend morbidity and mortality (M&M) meetings, complete a report on each delivery (“to be seen as a medico legal document”), comply with peer review, and inform the patient about the program, the SASOG mediation committee and its function. Each treating paediatrician is required to complete a discharge summary on each neonate, notify the delivering obstetrician if a neonate is re admitted, and attend M&M meetings if an adverse event occurred. The hospital is required to make the patient aware of the BetterObs program, ensure M&M meetings take place, and withdraw the delivery rights of an obstetrician not attending the minimum number of meetings. Patients must acknowledge they have been informed of the available complaint channels. An obstetrician leads the program in each hospital and is responsible for teaching sessions for the nursing staff in each hospital. Mediation addresses potential litigation cases (if adverse events occur) in order to prevent legal action, and to ensure a fair outcome for all involved.

Details on the impact of BetterObs have not yet been published.

The Brazilian response

In Brazil published guidelines and voluntary adherence programs were used to reduce the nearly universal use of C-section for childbirth in their private sector, but achieved very little. However increased awareness elicited further efforts to address the issue.

In 2012 an 85-bed private, not-for-profit hospital (Hospital and Maternity Santa Isabel – HMSI) in Sao Paulo partnered with the US-based Institute for Healthcare Improvement and applied quality improvement methodology (“improvement science”) to increase the rate of vaginal birth.

Goals were set – to achieve a vaginal birth rate of at least 40% (80% in primiparous women) with low (“European”) levels of perinatal mortality, reduced NICU admission rates, and a lower cost of care.

Seven months into the intervention the target vaginal birth rate of 40% (66% among publicly insured patients) was reached (Figure 1). Over 2 years the perinatal mortality rate decreased 25%, and 95% of women receiving 6 or more prenatal consultations, another goal of the initiative. Per capita hospital costs related to childbirth dropped 27%, with a 61% decrease in costs relating to the neonatal intensive care unit (NICU) and a 55% reduction in the admission rate. Obstetricians’ remuneration increased by 72%. The proportion of pregnant women satisfied and very satisfied with their obstetrical care reached 86%.

Figure 1. Control chart showing the increase in vaginal birth rate at HMSI. Image source: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0100-72032015001000446&lng=en&nrm=iso&tlng=en

How was this achieved?

Four primary drivers were defined that described what would need to be done to achieve the planned goals:

1. Changing the attitudes of pregnant women, families, physicians and the hospital community, with specific actions for each target group involved in childbirth, such as educational actions and health engagement
2. Redesign of the birthing care model, hiring nurse specialists in obstetrics and reorganising physician attendance
3. Focusing on prenatal care
4. Changing remuneration for obstetrical care, with physicians being paid per shift rather than by procedure (if the percentage of vaginal deliveries exceeded 50% of total deliveries,

An important intervention was the inclusion of pregnant women and their families in designing the new care model. A questionnaire administered 3 months after birth used questions about how they felt at each point of contact with the childbirth care system (first consultation, ultrasound, pregnancy course, prenatal consultations, hospitalization, childbirth, discharge, etc). Survey participants attributed an emotional state (joy, security, irritation, insecurity, fear, etc.) to the interaction at each point of contact.

More than 70 Plan-Do-Study-Act (PDSA) cycles were run that tested the changes on a small scale and were then rolled out.

Projeto Parto Adequado

A few years later similar concepts were scaled up and applied to a much larger initiative – Projeto Parto Adequado (PPA) – involving 35 hospitals in an 18 month collaborative. The four primary drivers of PPA were similar to those defined in the HMSI project:

1. Governance: forming a coalition between leadership in the health sector, aligning quality and safety in labour and childbirth;
2. Participation of women and families: empowering women and families so they actively participate in the entire process of pregnancy, birth and postpartum care.
3. Reorganisation of care: changing the model of perinatal care to favour the physiological evolution of labour and ensuring that the decision to do a C-section is based on clinical criteria;
4. Monitoring: structuring information systems that allow lifelong learning.

At the beginning of the PPA project (April 2015), the rate of vaginal deliveries in the pilot group of 12 hospitals was about 20%; 16 months later it had nearly doubled, to 37.5% (Figure 2). Hospitals also reduced their NICU admission rate from 63 to  48 per 1000 live births (Figure 3). Almost 90% of hospitals were able to increase the percentage of vaginal deliveries and more than half reached or exceeded the target rate of 40% of normal births five months before the end of the pilot phase of the project.

Figure 2. Increasing vaginal births in the PPA project. Image source: http://www.ans.gov.br/images/stories/Materiais_para_pesquisa/Materiais_por_assunto/web_total_parto_adequado.pdf

Figure 3. Decreasing NICU admissions in the PPA project. Image source: http://www.ans.gov.br/images/stories/Materiais_para_pesquisa/Materiais_por_assunto/web_total_parto_adequado.pdf

Observations on the South African and Brazilian experience

1. Although the SASOG guidelines and BetterObs program articulate many evidence-based aspects of good obstetrical care, the structural aspects of solo obstetrical private practice that drive high C-section rates are not being explicitly addressed; it will be interesting to see what impact the program has on outcomes, including on safety, adverse events and resulting litigation. The new peer review process is an important positive element but it is not clear whether this review is to occur only in response to adverse events (M&M) or is being applied to ongoing learning and improvement at a hospital and program level.
2. The Brazilian approach has produced positive results, scaling up from one hospital to many, includes explicit program theory about what is driving operative delivery rates, addressing each of the main presumed drivers, and uses well-tested quality improvement methodologies, including the Breakthrough Series model.
3. In complex health care delivery systems attention to a single factor or single outcomes is likely to be insufficient to guide intervention. In terms of interventions, education and guidelines though essential tend to have weak effects on their own. HMSI and PPA both emphasised care model redesign. They also changed their approach to reimbursement, as have at least some SA private insurers, in an attempt to remove financial incentives for delivery by C-section. Local experiments with new care models, such as PPOServe’s Birthing Team – do exist and worth watching.
4. Appropriate target outcomes should be patient-centred – and do not necessarily involve a specific C-section rate. They should include defined maternal and neonatal clinical outcomes, including rates of late preterm birth, and NICU admission rates, and must occur at an affordable cost. Incorporating informed maternal preferences and measuring patient-reported outcomes and experience is key. A process of co-design should be welcomed but will involve significant effort and a new paradigm of patient involvement. If successful, maternal preferences, when well informed, may shift toward vaginal birth.
5. The local approach should also address the enormously high medical malpractice costs, borne initially by doctors but ultimately transferred to society at large. Solutions may lie in part in the legal system itself, as well as in better, more humane responses to the needs of injured patients, such as Communication and Resolution Programs.

Conclusion

The SA private sector can learn from the successful experience with quality improvement methods that are being applied to childbirth in Brazil. However we need our own design. Driving factors in SA are not necessarily identical. A forthcoming qualitative study will help us understand whether and how the interventions in Brazil achieved fidelity, and how they were sustained and spread, but cannot replace our own investigation into context and local adaptation of interventions.  These questions can only be answered, as the Brazilians have done, through tests of change and developing a “structured system for lifelong learning”.

Figure 4. The PPA driver diagram, capturing program theory and the major drivers of the C-section rates targeted for reduction. Image source: https://reproductive-health-journal.biomedcentral.com/articles/10.1186/s12978-018-0636-y