Discussion

The key pharmaceutical processes of registration, selection and procurement of medicines for five major childhood cancers seem to be aligned in SA’s public healthcare system, indicating good operationalisation of SA’s policies and processes. The bottleneck seems to lie in the procurement of essential medicines through national tenders. Private sector formularies listed a limited selection of priority chemotherapeutics, indicating potential restrictions in what may be reimbursed to their beneficiaries.

While our findings also indicate alignment with international processes, the few gaps in comparison to the WHO EMLc may have a big impact. In fact, in a 2022 cross-sectional survey to determine priority essential childhood cancer medicines, dactinomycin was in the top 10 of most frequently selected drugs by paediatric oncologists when asked what medicines would achieve greatest benefit in children.28 Thus, the lack of market authorisation for dactinomycin in SA indicates that deficiencies in therapeutic care exist. Although certain legislative loophole arrangements—in SA’s case in the form of Section 21 access—can still allow the use of unregistered drugs after named-patient approval, this access pathways is associated with a range of challenges.29 These include the obtaining of hospital and/or provincial approval and the associated administrative burden on clinicians, the universally limited budgets to buy products outside of the NEML, and considerable delays in supply when products need to be imported. In SA’s private sector, medical schemes are under no obligation to reimburse section 21 medicines.16

Procurement issues potentially constraint access to some of the key chemotherapeutics in the management of childhood cancers such as cytarabine and etoposide.25 From the evidence obtained in this study, it cannot be deduced whether submitted bids were not awarded by the DoH or whether companies are not submitting any bids, but anecdotal evidence suggests that the DoH’s price expectations are too low to make bidding profitable.29 Additionally, even if some products were eventually successfully awarded in an additional tender for 2020–2022, this supplementary round brings a considerable delay of about 4 months based on the tender documents. These delays also affected core chemotherapeutics such as doxorubicin and vincristine.28 Noteworthy, DoH’s price expectations were unchanged for the additional tender round. In the meantime products must be bought out by provincial governments or individual hospitals to meet the demand, putting considerable strain on hospital pharmacists and continuous supply cannot be guaranteed during this time.30

Although our findings indicate potential difficulties in procurement, actual accessibility remains hard to predict based on these tender documents alone: medicines can be procured through buy-outs if contracts were not awarded, or medicines may be in short supply despite a contract. With regimens generally consisting of four or five active ingredients, even intermittent supply issues for one drug can negatively impact care for these aggressive cancers; omitting or switching of drugs is undesirable and could have detrimental effects.2 7 Surveys on the ground are required to get a more complete picture of supply and availability issues and how these impact patient outcomes.

Notwithstanding the considerable number of red crosses for access to supportive medicines, we have identified no major issues in accessibility of these active ingredients based on the registration and selection step alone. Not only are the gaps in registration status and NEML selection largely in line with international guidelines,20 but also not all within-class alternatives are required to be accessible if another from the same therapeutic class is (also stipulated in guideline9). It is, however, relevant that at least one alternative can be accessed if the medicine of first choice is not well tolerated. In this South African case study, we find that at least one active ingredient per class should be accessible—except for weak opioids. This gap does not seem problematic, since there is no international consensus on their use due to a lack of evidence.31 In the other antiemetics group, the inclusion of aprepitant on the NEML could be an important future addition, since aprepitant or analogues may be used as a further escalation in care if other antiemetics are insufficient.7 10

Although minimum coverage in the private sector (eg, PMB level) is supposed to be similar to the care as provided in the state sector and across all medical schemes,15 major gaps are visible as compared with the WHO EMLc and the public sector’s NEML. This misalignment already starts in the ICON formulary and is further exacerbated for the two medical scheme formularies. Although the large PSOMS is reported to take guidance from ICON,16 the inconsistencies between both formularies rather imply that other resources and factors also play a role in the establishment of their formularies. With that, the role of ICON in the private healthcare system is unclear.16

The rather large number of red crosses for the ICON formulary and two private sector medical schemes must, however, be interpreted with caution. Reimbursement of cancer therapy for many medical aid plans depends on monetary benefit limits: a predetermined amount from which consultation fees, various investigative scans and treatments including medicines are initially funded.16 It is only after this limit has been reached that patients may be restricted to formularies to avoid co-payments, especially if their diagnosis is not one of the 270 PMB covered indications (including some paediatric cancers). With that, the gaps identified may be of particular relevance to those without additional oncology benefits and PMB-level insured members. In addition, schemes may opt to use specific oncology protocols to define processes in care and access to medicines, but these are not publicly available. Nonetheless, it is difficult to predict what these results mean for individual medical schemes and insured members. This lack of transparency in what will be—or will not be—covered and when these formularies apply, creates challenges for members when having to navigate the system.16

In the interpretation of all of our findings, we acknowledge that even when no major barriers seem to exist on the active ingredients level, access may be more constrained for specific finished pharmaceutical products (FPPs). This is of particular importance for oral dosage forms in paediatrics, since different dosage strengths are required for children of different ages and manipulation of products could introduce quality issues and errors.32 An exploratory assessment of solid oral dosage forms was conducted with this in mind, but the lack of data in the NEML and ICON’s formulary on specific (required) FPPs prevented more comprehensive comparisons. The absence of accurate guidance on required FPPs in these sources constitutes a significant gap in itself, particularly as these documents guide subsequent procurement. A more detailed NEML, potentially complemented by STGs, could address this deficiency. Alternatively, making SAOC’s and ICON’s treatment guidelines publicly available could play an important role in addressing this gap.

Nevertheless, the fact that some products were not found in either the tender bid packs nor the MPR in the exploratory assessment implies that these common products may no longer be marketed in SA. With that, this exploratory assessment confirmed anecdotal reports that products are disappearing from the market.29 Similarly, ICON referred to bleomycin access through a section 21 exemption, despite a bleomycin product having market authorisation in SA. This again implies that the registered product is not widely available in SA, and alternative, equivalent products may be accessed via this loophole arrangement.

A limitation of this study is that an Africa-wide treatment guideline from 2017 was used to inform our basket, due to a lack of a South African equivalent. This may have resulted in active ingredients of local importance being missed, particularly some of the innovative medicines that may not be available in most of the other countries on the African continent. Additionally, treatment protocols, clinical insights and available therapies may have changed since then. For example, chlormethine, tramadol and niflumic acid do not seem to be medicines of first choice anymore, also explaining their absence from the WHO EMLc and national sources. Nonetheless, most of the priority medicines were also identified as such in a recent international survey among paediatric oncologists and paediatricians in LMICs,28 showing general representativeness of our sample. Additionally, details on FPPs were not provided in this resource—nor in some of the other data sources—limiting our analyses to active ingredient level and hence limiting the accuracy of our findings. Despite its limitations, the present basket allowed us to study the alignment of pharmaceutical processes, as was the primary aim of this study. Furthermore, we were limited to the use of publicly available data for this study. This also restricted us to the use of data on tenders as a proxy for procurement as a whole. The SAHPRA database was used to assess registration status, but it is unclear as to how often this database is updated. To mitigate the risk of incomplete data, the MPR was used to verify whether medicines were on the private market meaning they must have been registered. Finally, in the interpretation of these findings we must stress that even though we have not identified any barriers in access to the majority of these active ingredients via database evaluation, this does not guarantee that a medicine is indeed available on the shelf. The conduct of longitudinal availability surveys would be of particular complementary value to our findings.

The novelty and significance of this study lie in the scope of pharmaceutical processes studied. Where previous studies in other countries have compared national EMLs with the WHO model list33–35 or with national drug registries,11 this study is the first to include data on procurement and potential reimbursement in addition to (international) selection and registration. By studying these steps together, a more comprehensive picture of potential gaps was obtained and specific bottlenecks could be identified. With that, this study also emphasises the need for making information publicly available, including treatment guidelines, procurement documents and outcomes. Finally, the exploratory assessments performed highlight the importance of checking multiple sources to validate findings and enable the placing of results in the often complex context of a health system.