Reforming healthcare coding in South Africa

Healthcare coding is for more than getting claims paid. It allows clinical information to be recorded consistently, services to be described accurately, medicines and devices to be identified reliably and health system activity to be analysed in different settings. In South Africa, those functions have developed in separate, overlapping and often poorly governed layers.

The result is a coding environment that works just well enough to keep payment systems moving, but not well enough to support a modern health system. Providers, schemes, administrators, hospitals, software vendors and regulators all depend on coding, but they do not operate within a coherent national architecture. Clinical documentation, diagnosis classification, procedure coding, medicine and device identification, tariff rules, claims edits and reimbursement decisions are blurred together.

This creates avoidable effort and waste. Clinicians face rejected or delayed claims, inconsistent interpretation of codes and uncertainty about new procedures. Funders live with a lack of standardisation, limited transparency and difficulty managing affordability. Patients are exposed to shortfalls, disputes and confusing benefit rules. At a system level, the fragmentation limits costing, quality measurement, public-private comparison, strategic purchasing and planning.

There is an opportunity to move beyond incremental fixes. The adoption of SNOMED CT, the gradual international shift towards ICD-11 and the growing need for better public and private sector data create a moment to redesign coding as health-system infrastructure. The task is to separate the layers, define their roles, govern the mappings between them and build a system that can support care, payment, quality improvement and policy.

Image: Piet Mondrian. Did he paint this to represent reorganizing the healthcare coding landscape into a governed national architecture?

How the system emerged

The private-sector billing system grew from the Doctors Billing Manual, first published in 1944 and developed under the auspices of the South African Medical Association. Over time, these billing codes became embedded in the relationships between practitioners, medical schemes, administrators and software vendors. For many years, the same structure was used both to describe services and support payment.

The system came under pressure as provider charges, scheme reimbursement rates and the cost of practice moved apart. The National Health Reference Price List of 2006 was intended to provide a national reference point, but it did not survive a legal challenge in 2010. South Africa was left without a legitimate, enforceable national reference tariff. Later attempts, including HPCSA guideline tariffs, did not create a binding payment system for medical schemes.

Since then, the private sector has operated through professional coding manuals, scheme reimbursement rates, proprietary rules, modifiers and bilateral arrangements. Practitioners may charge according to their own practice economics, while schemes reimburse according to benefit rules and rates. Patients carry the difference through out-of-pocket payments or gap cover. A tariff problem is layered on top of a coding problem.

The current coding landscape

Today’s private-sector coding environment is complex. The SAMA Medical Doctors’ Coding Manual is the main procedural billing reference for most practitioners. The Complete CPT for South Africa localises the American Medical Association’s Current Procedural Terminology for some hospitals and specialist disciplines. Pharmaceutical and device claims depend on NAPPI codes together with ATC from the World Health Organisation for medicine classification. Diagnosis coding is based on ICD-10. The public sector uses the Uniform Patient Fee Schedule for billing purposes and lacks consistent patient-level clinical coding.

New procedures and technologies may enter professional coding manuals before schemes and administrators have updated their systems. Where no suitable code exists, clinicians use unlisted or generic codes, which often trigger manual review and inconsistent payment. Schemes rely on rules, modifiers and edits to manage risk and affordability. Providers experience this as delay, underpayment or arbitrary rejection. Funders experience it as necessary claims control in a poorly standardised system.

The Private Healthcare Information Standards Committee has contributed to clinical coding, messaging standards, diagnosis-related groups and electronic health standards. This shows that the industry can reach technical consensus. But PHISC is private-sector, technical and voluntary. It cannot by itself resolve national terminology governance, tariff methodology, product coding or statutory reimbursement reform.

The system is therefore not wholly unregulated or static. It is fragmented and under-governed. Professional bodies, schemes, administrators, hospitals, software vendors and regulators each hold part of it, but no trusted national process integrates coding updates, tariff methodology, digital standards, product identifiers and claims rules.

Proprietary dependence and innovation costs

CPT is a proprietary terminology owned by the American Medical Association. Its use in software, claims platforms and commercial products generally requires licensing. In South Africa, SAMA distributes localised coding products, including CCSA and electronic coding tools. These products provide structure, updates and professional input.

The problem is not that these products have no value but that reliance on expensive and proprietary procedural content creates avoidable barriers for local software developers, administrators and digital health platforms.

Public-private fragmentation

The public-private divide is an equally important issue. Private-sector claims data are detailed and granular. Public-sector data are much weaker at patient level, including diagnosis coding, procedure coding and episode costing. This limits case-mix analysis, hospital costing, quality measurement and any future move towards strategic purchasing under National Health Insurance.

A national system cannot be built by forcing private billing codes into the public sector, or by treating the public-sector fee schedule as a national clinical language. South Africa needs a shared coding architecture that can support records, referrals, discharge summaries, surveillance, quality measurement, costing and reimbursement in both sectors.

The SNOMED CT and ICD-11 opportunity

South Africa’s adoption of SNOMED CT is an important milestone. But SNOMED CT should be understood as a clinical terminology layer, not as a replacement for classifications, procedure codes or payment systems. Countries such as England, Australia, Canada and the United States use SNOMED CT in electronic records or clinical data standards while still maintaining separate classifications, procedure systems, medicines terminologies or reimbursement schedules.

ICD-11 is also important. It is better suited to modern reporting than ICD-10 and supports more detailed coding through post-coordination. But mapping between SNOMED CT, ICD-11, procedure classifications and reimbursement categories is complex. It requires governance, tooling, validation, version control and local rules. Automatic mapping may help in selected domains, but it is not a universal solution.

An architecture for the future

The target state should be a governed family of code sets, each with a clear role.

SNOMED CT should become the preferred clinical reference terminology for electronic records, referrals, discharge summaries and clinical data exchange.

ICD-11 should become the national statistical classification for morbidity reporting, introduced gradually with training, tools, quality assurance and continuity with ICD-10 during the transition.

Procedure and intervention coding should be rationalised over time. South Africa needs a governance layer that maintains mappings between existing procedural systems, including MDCM, CCSA/CPT, UPFS, SNOMED CT procedure concepts and relevant international classifications. It should also define the attributes needed for costing and payment, such as laterality, setting, anaesthesia, assistants, equipment and multiple procedure rules. SNOMED CT can describe clinical procedures, but reimbursement needs separate costing and payment logic.

Medicine and device coding should move towards a more open national product master-data architecture. NAPPI is already embedded in claims, formularies, pharmacy benefit management and supply chains. A practical first step would be an open mapping and governance layer around NAPPI, SAHPRA data, ATC and relevant SNOMED product concepts.

Reimbursement should be governed separately. Clinical terminology can describe what happened and why. It cannot decide what should be paid. Tariffs require costing, actuarial modelling, affordability decisions, benefit design, professional input and legal protection for transparent collective processes.

A Pricing and Coding Commission

A statutory Pricing and Coding Commission is a sensible idea, but its role must be carefully defined. It should govern coding, publish mappings and version histories, oversee costing studies, define principles for relative value, support public-sector case-mix development and coordinate with bodies such as PHISC.

Such a body would need representation from clinicians, hospitals, schemes, administrators, patients, regulators, health economists, actuaries, software vendors and public-sector managers. It would also need competition-law clarity. Without that, collective tariff work will remain vulnerable to legal challenge.

A possible transition pathway

The transition should be evolutionary. In the first one to two years, South Africa should establish governance arrangements, terminology services, licensing arrangements, priority mappings and pilot sites. The priority should be to avoid disrupting claims payment while building the foundation.

In years three to five, dual coding could be piloted in selected domains such as hospital episodes, high-volume procedures, pathology, radiology, chronic disease reporting and public-sector discharge summaries. The aim would be to test data quality, coding burden, reimbursement impact and dispute rates before wider implementation.

Over five to ten years, South Africa could move towards broader ICD-11 implementation, SNOMED-enabled records, better public-private comparability, open product mapping and more credible tariff methodology. Legacy code sets may remain in use during this period, but they should be mapped to a national architecture rather than operating as isolated systems.

Conclusion

South Africa’s coding problem has no single villain and no quick fix. The country has inherited a fragmented system shaped by outdated tariff history, proprietary dependence, scheme-specific payment rules, weak public-sector clinical data and voluntary private-sector standards.

SNOMED CT and ICD-11 create an opportunity for reform, but only if they are used as part of a governed national architecture. The goal should be a system in which clinical data are captured once, defined consistently, mapped transparently and used for care, payment, quality improvement and planning.

Coding reform should be treated as health-system infrastructure improvement, not a tariff dispute

Sources

1. Council for Medical Schemes. CMScript 10 of 2024: Coding and funding of claims. 2024. https://www.medicalschemes.co.za/cmscript-10-of-2024-coding-and-funding-of-claims/
2. Hospital Association of South Africa Ltd v Minister of Health and Others. High Court of South Africa, 2010. https://www.globalhealthrights.org/wp-content/uploads/2013/01/HC-2010-Hospital-Association-of-South-Africa-v.-Ministry-of-Health-and-Ors..pdf
3. Private Healthcare Information Standards Committee. About PHISC. https://www.phisc.net/about-us
4. Private Healthcare Information Standards Committee. PHISC standards. https://www.phisc.net/standards/phisc-standards
5. SNOMED International. Major milestone for Africa: South Africa adopts SNOMED CT to drive digital health transformation. 2026. https://www.snomed.org/news/major-milestone-for-africa%3A-south-africa-adopts-snomed-ct-to-drive-digital-health-transformation
6. NHS England. Clinical coding – SNOMED CT. https://www.england.nhs.uk/long-read/clinical-coding-snomed-ct/
7. NHS England. Terminology and classifications. https://digital.nhs.uk/services/terminology-and-classifications
8. Australian Digital Health Agency. SNOMED CT-AU with Australian Medicines Terminology. 2026. https://www.digitalhealth.gov.au/healthcare-providers/product-releases/snomed-ct-au-with-australian-medicines-terminology-amt-february-2026-release
9. Canadian Institute for Health Information. Canadian Coding Standards for ICD-10-CA and CCI. 2022. https://secure.cihi.ca/free_products/canadian-coding-standards-2022-en.pdf
10. US National Library of Medicine. SNOMED CT. https://www.nlm.nih.gov/healthit/snomedct/index.html
11. World Health Organization. ICD-11 implementation frequently asked questions. https://www.who.int/standards/classifications/frequently-asked-questions/icd-11-implementation