Prescribed Minimum Benefits (PMB) are a legally guaranteed minimum level of cover that every medical scheme in South Africa must pay for in full, regardless of the plan selected. The purpose of this system is to protect medical scheme members from being unable to access necessary treatment once their cover has been exhausted. This guarantee makes PMB both a key safeguard for members and one of the areas most vulnerable to fraud in the industry.

The logic is straightforward: when a payer cannot cap the amount, apply a benefit limit or refuse payment under the terms of a plan, an unscrupulous party has an incentive to “repackage” an ordinary bill as a PMB claim. At the same time, schemes seeking to protect their reserves tend to tighten their controls and sometimes reject legitimate PMB claims. MAINS LAB experts examine both sides of this conflict: the legal framework for PMB, common fraud schemes, responses by medical schemes and the regulator, and the impact of reforms, including NHI.

What PMB Are and How They Work

PMB are enshrined in the Medical Schemes Act 131 of 1998 and its Regulations. All schemes must cover the diagnosis, treatment and care of three groups of conditions:

Component

Scope

Emergency conditions

Any sudden condition that threatens life or an organ

DTP (Diagnosis Treatment Pairs)

A list of 271 conditions defined by a combination of diagnosis and treatment

CDL (Chronic Disease List)

26 chronic conditions

Key Legal Principles

Payment in full

Regulation 8 requires a scheme to pay for PMB treatment without limits, co-payments or deductions from the medical savings account (MSA), provided that the member uses a designated service provider (DSP) or seeks emergency care.

No exclusions

The cause of the condition is irrelevant: if septicaemia develops after cosmetic surgery, treatment for sepsis must be assessed for PMB cover.

Identification through ICD-10

The ICD-10 code entered by the doctor on the bill helps identify whether a claim may qualify for PMB cover; the final assessment requires the condition to meet the clinical criteria for the applicable basis of cover.

DSPs and tariffs

If a member voluntarily chooses not to use a DSP, the scheme may pay only part of the cost, for example 80% of the scheme tariff, with the member paying the difference.

The Scale of FWA in the Private Sector

According to the latest official figures, South Africa’s private healthcare sector lost approximately USD 1.80 billion* to fraud, waste and abuse (FWA) in 2025**. At the BHF conference in July 2026, the acting head of the SIU estimated these losses at around 15% of healthcare expenditure***.

The minister listed common violations including billing for services not provided, substituting branded medicines for generics, code manipulation, duplicate claims, excessive testing and claims submitted in the names of people who are not scheme members. Many of these mechanisms directly exploit the structure of PMB.

PMB Vulnerabilities and Common Fraud Schemes

PMB create three structural incentives for abuse: the absence of a price ceiling through the requirement for payment “in full”, payment from the risk pool rather than the savings account, and the dependence of PMB status on an ICD-10 code assigned by the provider. Each incentive gives rise to its own category of schemes.

The opportunity for arbitrage arises at the coding stage: the same consultation, when billed under an ordinary code, is paid subject to plan limits, from the MSA and with co-payments; when billed under a PMB code, it is paid from the risk pool and in full. This difference creates an incentive to substitute one code for another.

Risk

Potential violation

Unjustified PMB coding

The diagnosis or severity of the condition is misrepresented to obtain cover

False emergency status

Planned treatment is presented as requiring immediate intervention

Services not provided or billed more than once

The bill includes services that were not provided or have already been paid for

Improper unbundling

Components of a single service are billed separately in breach of tariff rules

Unsubstantiated registration of a chronic condition

Cover is requested without sufficient diagnostic evidence

Misrepresentation of the beneficiary

A service is attributed to a member who did not receive it

Unjustified volume of services

Investigations or procedures are not consistent with clinical need

The first two schemes are particularly difficult to detect. The line between legitimate clinical judgement and manipulation is blurred: for many DTPs, an ICD-10 code provides only an indication of potential PMB status, while the final decision requires clinical information.

Experts observe that inflated claims often involve cases in which the claimed PMB code is poorly supported by clinical documentation. An illustrative example is PMB DTP 901E, which relates to hypertension. Here, insufficient evidence means a lack of support for classifying the particular episode as a PMB case, rather than weak scientific evidence about the disease itself.

Code 901E covers hypertension with acute life-threatening complications, malignant hypertension, renal artery stenosis and other curable forms of hypertension. A bill entry stating “hypertension” is therefore not sufficient to justify this particular DTP: the clinical records must substantiate the relevant condition. Hypertension is also included in the Chronic Disease List (CDL), however; a dispute over the justification for code 901E does not in itself mean that there is no entitlement to PMB cover on another basis.

According to the expert estimates cited, only around 30% of cases submitted to a medical scheme as PMB actually meet the criteria for such cover. This figure should be treated as an observation from practice: without information on the sample, period and review methodology, it cannot be extrapolated to the South African market as a whole. Nor does it imply that the remaining 70% constitute fraud: unsubstantiated PMB status must be distinguished from coding errors, missing documentation and deliberate misrepresentation.

Epilepsy as a Basis for Payment

Epilepsy is one of the 26 chronic conditions on the Chronic Disease List. Cover includes medicines, consultations and investigations related to the condition, subject to the applicable PMB protocols and terms. This guarantee protects patients who need long-term treatment. At the same time, it creates a financial incentive to record a chronic diagnosis whose validity has yet to be established.

Claims analysis indicates that up to 30% of patients admitted with an epilepsy code had no treatment for the condition recorded in the data either before admission or during the six months following it.

The examples follow a recurring pattern: a short hospital stay, charges for accommodation and investigations, followed by no recorded dispensing of antiepileptic medication.

A 41-year-old man was admitted for one day after an episode of loss of consciousness. He underwent a contrast-enhanced brain MRI, carotid ultrasound, ambulatory ECG monitoring and one consultation. The bill carried an epilepsy code. The total was USD 1,570, of which USD 350 was charged for ward accommodation.

1 A 26-year-old woman also spent one day in hospital. The bill listed EEG monitoring with interpretation and one consultation. Of the total of USD 1,680, ward accommodation accounted for USD 1,120, or around two-thirds of the bill.

2 A 69-year-old woman was admitted for one day. The supplied description of the bill lists a ward accommodation charge of USD 2,000 and an age modifier; the total was USD 2,460.

3

Each episode may have a clinical explanation in its own right. Loss of consciousness requires its cause to be established, a suspected seizure warrants investigation, and an older patient’s condition may require observation. However, the need for investigation does not in itself substantiate the chronic diagnosis on which payment is being claimed.

The key question in a review is therefore not simply why the patient was not prescribed medication. It is necessary to establish what evidence supports the epilepsy diagnosis, what the investigations showed, which diagnosis was recorded at discharge and what follow-up plan was offered to the patient. Reviewers should also establish whether medicines could have been supplied through another channel and recorded in a different system.

It is the combination of circumstances that raises concern: a chronic disease code, a short admission, substantial accommodation costs and no visible continuation of treatment. This warrants a request for medical records, but is not yet proof of a fabricated diagnosis or an unnecessary admission.

Emergency Status as a Basis for Payment

The second vulnerability concerns the classification of an admission as an emergency. PMB protect patients who need immediate care: a definitive diagnosis cannot always be made before treatment begins. A medical scheme is nevertheless entitled to request supporting information within a reasonable period. The hospital’s initial assessment does not make the claimed emergency status indisputable.

A financial incentive arises when a planned investigation or intervention is presented as an emergency episode. Instead of seeking payment under the ordinary terms of the plan, the hospital requests PMB cover.

Almost Nine out of Ten Admissions Are Classified as Emergencies

At one private hospital described in the supplied analysis, almost nine out of ten admissions were classified as emergencies, compared with around half in the comparison group. Planned admissions at this hospital were rare.

In addition, around four in ten of the bills analysed contain no recorded medicines, whether tablets, injections or infusions. This does not automatically mean that no medicines were administered: the completeness of the data and the billing arrangements need to be checked. Combined with the unusually high proportion of emergency admissions, however, this finding calls for an explanation.
The comparison must take account of the hospital’s clinical profile, patient mix and case severity. Individual episodes should then be reviewed to establish what required immediate admission, what treatment was provided and why it could not have been delivered on a planned or outpatient basis.

Pneumonia with No Antibiotic on the Bill

A 25-year-old woman spent nine nights in hospital with a diagnosis of pneumonia. The bill included both a bed charge and a separate room charge for the same day, an abdominal CT scan, an ultrasound of the heart, three COVID tests, blood cultures and daily doctor visits. No antibiotic appears on the bill supplied.

The review must establish what treatment the patient actually received and what explained the length of her stay. Further questions concern the indications for the investigations and the simultaneous charging of two accommodation items: these may be separate, permissible tariff components or duplicate billing. Only the itemised bill and the tariff rules can distinguish between them.

The absence of an antibiotic from a single bill does not in itself prove that the admission was unnecessary. However, the mismatch between the recorded diagnosis, length of stay and visible treatment requires comparison with the medication orders and clinical records.

An Emergency Admission with Investigations for Chronic Conditions

A 58-year-old woman presented to the emergency department with elevated blood pressure and spent two nights in hospital. Laboratory investigations accounted for around one-third of the bill: ten tests for autoimmune diseases, thyroid function testing, glycated haemoglobin and cholesterol measurements, and a glucose tolerance test. No blood pressure-lowering medicines appear on the bill supplied.

This mix of services raises questions about how each investigation related to the reason for the emergency admission. It is necessary to establish which signs of an acute condition were documented on presentation, what justified two nights in hospital, and which investigations formed part of emergency care rather than the assessment of coexisting conditions.

For this article, the case is particularly illustrative in relation to DTP 901E: a record of elevated blood pressure does not in itself establish a condition that qualifies for this basis of PMB cover. Entitlement to cover for chronic hypertension and the justification for a particular emergency admission require separate assessments.

An Intestinal Infection and ENT Surgery on the Same Day

A one-year-old boy was admitted with a diagnosis of an acute intestinal infection. On the same day, an ENT surgeon removed his tonsils and adenoids and inserted ventilation tubes in his ears under anaesthesia. The following day, the hospital billed for ward accommodation and an emergency department visit.

The relationship between the admission diagnosis and the surgery performed requires an explanation. Reviewers should check when the procedure was scheduled, its indications, the preoperative documentation and the basis for emergency status. A separate question is what the emergency department visit charge on the following day covered, given that the child was already an inpatient.
If the records show that the operation had been planned in advance and the acute diagnosis was used to change the basis of payment, this would provide substantial grounds for concluding that the claim had been submitted improperly.

Who Pays for Unjustified Claims

PMB costs are paid from the medical scheme’s common risk pool. The same pool funds necessary care for other members, including treatment for severe chronic diseases and emergency conditions. An unjustified claim therefore affects everyone who pays contributions.

Systematic overpayments increase scheme expenditure and may put pressure on future contributions and the scope of additional benefits. However, contribution increases cannot automatically be attributed to fraud, as other factors also play a role.
The practical purpose of claims controls is to verify the basis for payment without delaying necessary care. These cases show why a diagnosis code or an “emergency” designation alone is insufficient for a full assessment. They must be considered alongside the patient’s condition, the services actually provided, treatment orders and subsequent follow-up.

Abuse by Medical Schemes

Controls can also fail in the opposite direction, leaving patients without the funding to which they are entitled. A study published in the South African Medical Journal in 2024 analysed 1,124 PMB complaints received by the CMS between 2014 and 2018. Medicines and surgical treatment featured particularly prominently among the services for which complainants had been refused payment in full.****

These data demonstrate persistent disputes over cover, but do not measure the current proportion of wrongful refusals. A sample of complaints does not represent all claims, and a complaint does not in itself establish a violation by a scheme.
The assessment of a payer’s performance should therefore measure both sides of decision quality: how many unjustified payments were prevented and how many legitimate claims were wrongly delayed or rejected. Otherwise, growing “savings” may conceal a shift in costs to patients and providers.

Section 59 and the Question of Racial Bias

The Section 59 investigation became an important part of the debate over fairness in anti-fraud procedures. According to a parliamentary statement dated 8 July 2025, the independent panel identified systemic racial discrimination in the application of FWA mechanisms. Parliament’s account states that Black providers in the data examined were 1.5–3.5 times as likely to receive an adverse FWA classification.*****

The panel’s findings are disputed. In its own reporting, Discovery states that no overt bias has been established in its algorithms and processes, criticises the data and methodology used, and emphasises that the report’s findings are not legally binding.****** The company’s position must be distinguished from the independent panel’s findings; neither should be presented as a final judicial resolution of all the disputes.

It is important here to distinguish between two positions: the independent panel’s findings and the company’s response. Neither constitutes a final court judgment, so the dispute cannot be regarded as settled. The practical lesson for the industry is that an anti-fraud system must both identify violations and be able to demonstrate that its selection criteria are transparent, verifiable and do not discriminate against particular groups of providers.

How to Organise PMB Controls in Practice

The proposed control model centres on the episode of care, from the reason for seeking care to the services actually provided and the payment decision. It provides a structured way to answer five questions.

Has entitlement to cover been confirmed?

The review checks scheme membership, the applicable terms and whether the condition meets PMB criteria. Missing information should prompt a targeted request for documents explaining what needs to be substantiated.

1 Do the services match the clinical episode?

The diagnosis is compared with the investigations, procedures, medicines and treatment timeline. Statistical deviations are considered in the context of case severity and coexisting conditions.

2 Have the correct payment terms been established?

DSP use, the reasons for seeking care outside the network, the applicable tariffs, the bill’s components and the permissibility of a co-payment are assessed separately. Whether a service is clinically justified and whether its price is correct are separate checks.

3 Is there evidence of a violation?

Automated analysis helps prioritise cases for review. A final determination requires checking the records, assessing the provider’s explanations and distinguishing between an error, a disputed interpretation and deliberate misrepresentation.

4 Can the decision be explained and reconsidered?

For each disputed claim, the grounds for the decision, the information used and the outcome of any review of objections should be recorded. Feedback should be incorporated into the rules and models so that the same cause of false positives does not recur.
The quality of this process should be assessed using several measures together: confirmed overpayments and actual recoveries, the proportion of reviews in which no violation is established, decisions overturned on appeal, the time taken to pay justified claims, and the administrative burden on those involved. An increase in the number of suspicious bills is not, in itself, a measure of success.

5

____________________________________

* CurrencyBeacon exchange rate as at 26 October 2026: USD 1 = ZAR 16.6358. Amounts are rounded to two decimal places. Source: https://currencybeacon.com/rates/USD/ZAR
** https://newsday.co.za/south-africa/27812/south-africas-private-health-sector-loses-r30-billion-annually-to-fraud-waste-and-abuse/
*** https://www.theherald.co.za/news/2026-07-07-r30bn-stolen-a-year-in-healthcare-fraud-while-millions-drop-medical-aid/
**** https://samajournals.co.za/index.php/samj/article/view/1007/980
***** https://www.parliament.gov.za/press-releases/media-statement-committee-health-welcomes-findings-section-59-investigation-panel
****** https://www.discovery.co.za/wcm/discoverycoza/assets/investor-relations/2025/discovery-sustainability-report.pdf

About Mains Lab

Mains Lab is a global provider of artificial intelligence and machine learning solutions focused on improving integrity and efficiency within health insurance systems. The company operates across the Middle East, Europe, Latin America, and Africa. To date, Mains Lab’s technology has enabled insurers and TPAs to achieve over USD 1 billion in total savings, helping organisations detect fraud, reduce waste, and optimise claims management through advanced real-time analytics.

www.mainslab.com

BACK TO NEWS