Practise defensibly within the medico-legal and litigation environment of South African O&G

In one line

South African obstetrics carries the country's largest single medico-legal liability — cerebral-palsy and birth-asphyxia claims dominate a contingent liability running to roughly R99 billion — and defensible practice turns less on avoiding error than on a contemporaneous, honest, well-documented account of care that can withstand a negligence claim built on the four elements of duty, breach, causation and damages.

The statutory groundwork — consent, the CTOP Act, the Children's Act, mandatory reporting — sits in SA O&G law at Intermediate level. The consultant layer is the law of negligence itself: how a claim is constructed and defended, why this country's obstetric liability is uniquely large, and how documentation, disclosure and the expert process either protect or sink you.

Mechanism & pathophysiology

Litigation is not a clinical disease, but it has a pathogenesis, and treating it as random misfortune is the first error. The South African crisis is a system in positive feedback, and each driver is something a consultant either feeds or interrupts.

The arithmetic is stark. By 2018/19 the contingent liability of provincial health departments for medico-legal claims had reached roughly R99.2 billion, with around R2 billion paid out that year, on a trajectory growing near 30% a year for contingent liability and 23% a year for payments since 2014. Obstetrics is the epicentre. Of state claims, roughly half are cerebral-palsy-type, and in a representative year (2017) O&G accounted for about 52% of claims against the state, of which some three-quarters were cerebral palsy. A single such claim now settles in the order of millions to tens of millions of rand, because the quantum is dominated by the cost of a lifetime of care for a severely disabled child.

Three things make cerebral palsy the engine of the crisis specifically. First, quantum: future care for a child with spastic quadriplegic cerebral palsy, costed over a normal life expectancy, dwarfs almost any other award. Second, the long limitation tail: a child injured at birth can litigate years later — and a minor's prescription period runs from majority, so a claim can surface two decades after the delivery, when the only evidence of what happened is the contemporaneous record. Third, the causal narrative is contestable but emotionally compelling: an intrapartum hypoxic-ischaemic insult is one cause of cerebral palsy, but the majority of cerebral palsy is antenatal or unascertained in origin, so the litigation hinges on whether this child's injury was intrapartum and avoidable — exactly the question the four elements are designed to test, and exactly where good and bad records diverge in their outcomes. The clinical substance of that causal argument — the timing and pattern of injury, the templates used to attribute an intrapartum cause — sits in neonatal-encephalopathy-cerebral-palsy, and the population-level appraisal that frames the crisis in maternity-statistics-critical-appraisal; the work here is the legal architecture those facts are litigated within.

Two structural features of the SA system amplify all of this. First, most of the liability sits with the state, and it is vicarious: the Member of the Executive Council (MEC) for Health of a province is sued for the negligence of employees acting in the course of their duties, so the named defendant in the great obstetric claims is "the MEC for Health", not the individual registrar — the state carries the quantum, but the individual's record and conduct are still on trial and the practitioner can face parallel HPCSA proceedings. Second, the injury is frequently generated at the district–regional interface, where the reasonable-practitioner standard must be read against the resources actually available: a delayed caesarean because the single theatre was occupied, or a missed deterioration because one midwife was covering a full labour ward, is judged by what was reasonable in that post with those resources — but the system failure that produced the under-resourcing is exactly what a root-cause analysis (and a plaintiff) will expose. The consultant who understands that the claim is institutional and the standard is context-bound argues both more accurately.

The downstream pathology is defensive medicine, and the clearest fingerprint is the caesarean rate. SA private-sector caesarean rates have run above 70% for years — far beyond any rate the maternal-fetal evidence supports — and fear of being "sued for not doing a caesarean" is a recognised, if multifactorial, driver. This is a controversy worth naming honestly: defensive caesarean section trades a contested medico-legal risk for real, measurable maternal morbidity (haemorrhage, sepsis, the accreta spectrum, future uterine rupture) and offers no proven reduction in cerebral palsy, because most cerebral palsy is not intrapartum and is not prevented by abdominal delivery. A consultant who reaches for the knife to manage their own anxiety has converted a legal fear into an obstetric harm — and that decision is itself defensible only if the indication is documented and genuine.

Assessment

"Assessment" in this setting means appraising your own exposure and the integrity of the record before a complaint ever lands — and, when one does, reading the claim against the legal elements rather than panicking.

• Anatomy of a negligence claim (the four elements — the plaintiff must prove all four). A claim in delict succeeds only if each link holds:
  • Duty of care — a legal duty owed to the patient. Rarely contested in a clinical relationship; it exists the moment you accept the patient.
  • Breach of the standard of care — that you fell below the standard of the reasonable practitioner in the same field. A specialist is held to the specialist standard, a registrar in a referral setting to what is reasonable for that post, judged by the knowledge available at the time, not with hindsight.
  • Causation — both factual ("but for" the breach, would the harm have occurred?) and legal (was the harm a sufficiently close, non-remote consequence?). This is where most obstetric defences are won or lost: even a proven breach does not pay out if the injury would have happened anyway.
  • Damages — actual, quantifiable harm (here, the lifetime cost of disability plus general damages).
• The standard of care is the reasonable practitioner, not perfection. A bad outcome is not negligence; an error is not automatically negligence; the question is whether a reasonable practitioner in your position would have acted as you did. Guidelines (NDoH Maternity Care Guidelines, the partogram, fetal-surveillance standards) are the de facto yardstick of "reasonable" — a documented, reasoned departure from a guideline can be defensible; an undocumented one usually is not.
• Read the record as a court will. The single most important assessment a consultant makes after an adverse event is whether the notes tell a coherent, contemporaneous story: times, decisions, who was called and when they came, the CTG interpretation as it was read at the time, and the decision-to-delivery interval. A record that is legible, timed, signed and internally consistent makes a claim defensible; one that is sparse, retrospectively altered or contradictory makes even good care indefensible.
• Identify when the burden may shift. Ordinarily the plaintiff proves negligence, but in the retained-swab / retained-instrument type of case the facts may "speak for themselves" — res ipsa loquitur — and an inference of negligence arises that the defendant must answer (see the evidence section). Recognising a res ipsa situation early changes the entire defensive posture: the swab count and its documentation become the case.

Management

Defensible practice is built, not improvised, and it runs on the same immediate → ongoing → long-term structure as any clinical plan: what you do in the encounter, what you do around the adverse event, and what the system does to reduce recurrence and liability over time.

Immediate — the contemporaneous record and the consent conversation. The note made during or immediately after care is the foundation of every later defence.

• A defensible clinical record is contemporaneous, accurate, complete, legible, signed, timed and attributable to a named, identifiable practitioner. In obstetrics the load-bearing documents are the partogram and the CTG: both are legal documents, and both must record interpretation, not just data. A partogram with a crossed alert line and no documented action, or a pathological CTG with no recorded decision, is a record that convicts.
• Document the times that matter: time of the decision, time the consultant/theatre was called, time of arrival, decision-to-delivery interval, time of birth, cord gases. In a hypoxic-injury claim years later, these times are the case.
• Never alter a record after the fact. Corrections are made by a dated, signed addendum that leaves the original legible — backdating or obliterating an entry is the act that turns a defensible case into an indefensible one and exposes you to a finding of dishonesty (the court in Michael v Linksfield marked its disapproval of a practitioner's dishonesty with a punitive costs order).
• Consent is documented and patient-centred. South African law has, since Castell v De Greef, required disclosure of the risks a reasonable patient in this patient's position would consider material — the same patient-centred standard the UK reached later in Montgomery. Document the material risks discussed, the alternatives offered, and that the patient understood, in a language she understands (record the interpreter). For caesarean, trial of labour after caesarean, instrumental delivery and shoulder-dystocia-prone deliveries, the consent conversation is itself a defensible-practice intervention.

Ongoing — open disclosure after an adverse event. When harm has occurred, the response shapes both the patient's experience and the litigation.

• Open disclosure (the duty of candour) is honest, prompt, and includes an apology. Tell the patient and family clearly what happened, what it means in the short and long term, and what will be done; express genuine sympathy. An apology is not an admission of legal liability, and the evidence is that open, early disclosure reduces — not increases — the likelihood of litigation and complaint, because much obstetric litigation is driven by a failure of explanation as much as by the injury itself.
• Note the SA-specific gap: South Africa has no standalone statutory duty of candour (unlike England, Scotland, Ireland and Victoria, where it is legislated). The HPCSA ethical rules impose a duty of honesty and to act in the patient's best interests, but there is no specific error-disclosure statute — so candour here is an ethical and risk-management discipline you must own, not a box a law forces you to tick.
• Report and analyse the incident. Enter the event into the facility incident/adverse-event system, escalate per protocol, and ensure it enters morbidity-and-mortality and root-cause analysis review. Root-cause analysis looks for the system failures behind the proximate error (the missing machine, the absent senior, the communication breakdown) rather than a person to blame — and a credible RCA is both a quality intervention and, later, evidence that the institution responded responsibly.
• Preserve the evidence. Secure the CTG trace, the partogram, the swab-count record, the placenta if relevant, and the full record — do not let the very documents that may exonerate you go missing.

Long-term — records governance, systems and the litigation process.

• Retention is regulated. HPCSA guidance requires a patient health record to be kept for at least 6 years from the date it became dormant (the patient's last treatment); for a minor, until the 21st birthday (because a child can litigate up to three years past majority); for a mentally incapacitated patient, for the patient's lifetime. Obstetric records therefore have a long mandatory life precisely because the injured-at-birth claim arrives late — discarding a delivery record on the ordinary adult schedule destroys your own defence.
• Records are protected personal information. Under the National Health Act (confidentiality, s14) and POPIA (health data is "special personal information"), records must be securely stored, lawfully processed, disclosed only with consent or lawful authority, and securely destroyed at end of life — with the Information Regulator now an enforcement reality.
• System-level liability reduction. Skills drills (shoulder dystocia, eclampsia, PPH, CTG interpretation), credentialling, MEOWS-driven escalation, and adherence to the NDoH guidelines reduce both harm and exposure. The National Core Standards / Office of Health Standards Compliance and structured NCCEMD (Saving Mothers) and perinatal mortality audit are the governance scaffolding into which individual defensible practice fits.

The expert witness and the legal process

When a claim proceeds, the consultant meets the legal system in two roles — as a factual witness to care they gave, and potentially as an expert witness opining on the standard of care — and confusing the two is a classic error.

• The record drives the process. The plaintiff's attorneys obtain the full medical record by subpoena or in discovery; the case is then built on, and contested through, what that record contains. A complete, contemporaneous, legible record is the single greatest determinant of how the litigation runs — which is why records governance is litigation strategy, not administration.
• The factual witness testifies to what they did and observed, from the record and memory. The discipline is to stay within fact, not to reconstruct or defend with hindsight, and never to be caught contradicting a contemporaneous note — the reason an honest, complete record made at the time is your protection years later when memory has faded.
• The expert witness opines on whether the care met the standard of the reasonable practitioner. The court's approach since Michael v Linksfield is decisive here: an expert's opinion carries weight only if it is logically and reasonably defensible — the court is not bound to accept it merely because it is sincerely held or supported by colleagues. An expert who overlooks a risk that should have been guarded against, or whose view cannot withstand logical scrutiny, will not persuade. The expert's duty is to the court, not to the party who instructed them; partisanship destroys credibility.
• Giving evidence is a skill: answer the question asked, concede what is properly conceded, do not speculate beyond the record or your expertise, and do not let advocacy substitute for honesty. The practitioner whose record is sound and whose testimony is measured and truthful is in the strongest position; the one caught having altered a note, or overstating, loses the case on credibility regardless of the clinical merits.
• Subpoenas and confidentiality. A subpoena or court order is one of the lawful gateways through which confidential records are released without the patient's consent — the National Health Act and POPIA permit disclosure compelled by law. Respond to a valid subpoena; do not unilaterally withhold or, conversely, hand over records on an informal request without lawful basis.

Guidelines compared

The "guidelines" here are partly clinical and partly legal-professional, and they diverge in instructive ways.

The substantive divergence to hold is candour: the UK and several other jurisdictions have made open disclosure a legal duty with regulatory teeth, whereas SA leaves it to professional ethics and risk-management prudence. The practical consequence is that a South African consultant cannot point to a statute to justify (or compel) disclosure — the discipline has to be internal, and it remains the right thing to do clinically, ethically and defensively.

The evidence & the controversy

The defining SA controversy is what to do about a liability that the public health system cannot sustain, and the consultant should be able to argue it because every reform changes the environment they practise in.

The South African Law Reform Commission's Project 141 (Discussion Paper 154, 2021) is the centre of this debate. It canvasses moving from lump-sum awards to structured/periodic payments and to payment in kind (the state providing the future medical care it would otherwise have to fund in cash), capping of certain heads of damage, mandatory mediation and alternative dispute resolution, specialist medico-legal courts, and tighter scrutiny of claims to separate genuine from fraudulent or inflated ones. Each has a defensible logic and a real objection: periodic payments protect against a windfall lost to inflation or misappropriation but assume a functioning state service to deliver the care; capping controls quantum but risks under-compensating the most catastrophically injured. The honest position in a viva is that the crisis is real, the reform direction is contested, and no single lever fixes it — and that the most reliable individual contribution remains the prevention of avoidable intrapartum injury and the keeping of records that let genuinely good care be recognised as such.

A second live controversy is the defensive caesarean. The evidence does not support abdominal delivery as a cerebral-palsy-prevention strategy — most cerebral palsy is not intrapartum hypoxia — yet litigation pressure pushes the rate up, importing maternal harm. The defensible stance is to operate on indication and to document it, not to let an unprovable legal fear drive a provable surgical morbidity; the same logic that makes the four elements protective (no causation, no liability) is the logic that should reassure the clinician practising good, evidence-based, well-documented obstetrics.

A third thread is records as digital data. As facilities move to electronic records and as POPIA enforcement matures, the medico-legal centre of gravity is shifting toward data integrity, audit trails (which make covert alteration detectable — a double-edged feature) and breach liability. The consultant who understands that an electronic record's metadata can both convict a backdated entry and exonerate a properly contemporaneous one is reading the direction of travel correctly.

Landmark trials & key evidence

The "evidence" in a medico-legal chapter is the case law and the policy analysis a candidate must be able to cite by name. Each below was verified against its source.

A worked figure for the viva: contingent liability around R99.2 billion (2018/19) with ~R2 billion paid that year means the liability on the books was roughly fifty times the cash actually leaving in a year — the gap between provision and payment is itself the crisis, because the liability accrues faster (~30%/yr) than the system can pay or reform it.

Exam traps & red flags

• Treating a bad outcome as proof of negligence. Negligence requires breach of the reasonable-practitioner standard and causation; a cerebral-palsy child is not, by itself, evidence that anyone fell short.
• Forgetting causation. The commonest defensible obstetric case is the proven breach with no causation — the injury would have occurred regardless. A candidate who concedes liability the moment a breach is shown has lost the case the law would have won.
• Hindsight standard. The standard is what a reasonable practitioner knew and could do at the time, in that setting (district vs tertiary) — not what is obvious after the MRI.
• Altering or backdating the record. The single most destructive act after an adverse event; corrections are dated, signed addenda that preserve the original. Dishonesty about records draws punitive costs and destroys credibility (Michael v Linksfield).
• A thin partogram or an un-interpreted CTG. These are legal documents; data without documented interpretation and action is the record that convicts. The decision-to-delivery interval and the times of calls/arrivals are the heart of a hypoxia claim.
• Defensive caesarean as risk management. Operating to manage your own legal anxiety imports real maternal morbidity and does not prevent the cerebral palsy that is mostly not intrapartum — and an undocumented "soft" indication is itself indefensible.
• Silence after harm. Failing to disclose, explain and apologise both wrongs the patient and increases litigation; candour is protective even though SA does not legislate it.
• Discarding obstetric records early. A minor litigates from majority; the record must be kept until the patient's 21st birthday (lifetime if incapacitated). Routine adult-schedule destruction erases your own defence.
• Missing the res ipsa situation. A retained swab or instrument shifts the evidential burden to you; the swab count and its documentation become the whole case (Goliath).
• Conflating ethics, consent statute and negligence law. The reasonable-patient consent standard (Castell), the four elements of delict, and the HPCSA ethical rules are distinct frames; answering a negligence question with a consent rule, or vice versa, reads as unstructured.

Worked viva

A typical stem: "A 31-year-old delivers at a regional hospital after a prolonged second stage; the CTG was pathological for the last 90 minutes, cord arterial pH is 6.9, and the neonate develops hypoxic-ischaemic encephalopathy. Two years later you receive a letter of demand alleging negligence and claiming R18 million. The registrar asks how the hospital should respond." A high-scoring answer runs:

1. Frame it by the four elements. "Liability requires the plaintiff to prove all four: a duty of care — which existed; a breach of the reasonable-practitioner standard; factual and legal causation; and damages. The contest will be breach and causation, and both turn on the contemporaneous record."
2. Interrogate the record, honestly. "I would secure and read the CTG, the partogram and the notes as they stand — times of CTG interpretation, when the registrar recognised the pathological trace, when the consultant and theatre were called and when they arrived, the decision-to-delivery interval, and the cord gases. I would not alter anything; any clarification is a dated, signed addendum."
3. Test causation, not just breach. "Even if there was a delay, the defensible question is whether earlier delivery would have prevented the injury — the cord pH, the pattern and timing of the CTG abnormality, and the placental histology all speak to whether this was an avoidable intrapartum insult or an established antenatal process. Most cerebral palsy is not intrapartum; this analysis is the heart of the defence."
4. Disclose and support the family. "Independently of the claim, the family is owed open disclosure — an honest account and a genuine apology, which is not an admission of liability and which reduces, not increases, litigation. SA has no statutory duty of candour, but the HPCSA honesty duty and good practice require it."
5. Run the institutional process. "The event goes through incident reporting, M&M and root-cause analysis looking for system failures — staffing, escalation, CTG-interpretation competence — and the file, the trace and the partogram are preserved. The medical-defence organisation or state legal services is notified early."
6. Place it in the SA frame. "This is the archetypal claim driving the country's ~R99 billion contingent liability, which is why the SALRC reforms — periodic payments, payment in kind, ADR — are being debated, and why the durable individual contribution is preventing avoidable intrapartum injury and keeping records that let good care be recognised as such."

Evidence anchors

• Castell v De Greef 1994 (4) SA 408 (C) — reasonable-patient standard for disclosure of material risk in consent (Cape Provincial Division).
• Montgomery v Lanarkshire Health Board [2015] UKSC 11 — UK reasonable-patient materiality test; the comparator to Castell.
• Michael v Linksfield Park Clinic (Pty) Ltd 2001 (3) SA 1188 (SCA); [2001] ZASCA 12 — expert evidence must be logically/reasonably defensible (the Bolitho qualification).
• Goliath v MEC for Health, Eastern Cape 2015 (2) SA 97 (SCA); [2014] ZASCA 182 — res ipsa loquitur / inference of negligence; retained surgical swab.
• Lee v Minister of Correctional Services 2013 (2) SA 144 (CC); [2012] ZACC 30 — flexible "but-for" factual causation.
• South African Law Reform Commission — Project 141, Discussion Paper 154: Medico-legal Claims (2021) — the reform reference; contingent-liability figures and proposed remedies.
• Prinsen L. The leading causes of medicolegal claims and possible solutions. SAMJ 2023;113(4) — drivers and three-tier prevention; ~half of claims cerebral-palsy-type.
• Taylor B, Cleary S. Medicolegal cases against obstetricians and gynaecologists in South Africa's private sector. SAMJ 2021;111(7) — the private-sector claim landscape.
• HPCSA guidelines on patient record-keeping — retention at least 6 years from dormancy; minors until the 21st birthday; mentally incapacitated patients for life; records contemporaneous, accurate, legible, attributable.
• National Health Act 61 of 2003 (consent s7, confidentiality s14, records) and POPIA (Act 4 of 2013) — health records as special personal information; lawful processing, security and secure destruction; Information Regulator enforcement.
• NDoH National Core Standards / Office of Health Standards Compliance and the NCCEMD (Saving Mothers) and perinatal mortality audit — the governance framework into which defensible individual practice fits.