Potential Cost Savings from Generic Medicines - Protecting the Prescribed Minimum Benefits

Abstract

Background: South Africa has followed a pro-generic policy since the introduction of the National Drug Policy in 1996. The selection processes in the public and private sectors have, however, remained largely disconnected, and at times contradictory. Medicines provided outside of hospitals accounted for 17% of medical aid spend in 2006, up 8.8% from the previous year. Of particular concern to funders has been the expenditure on the 27 chronic conditions listed as Prescribed Minimum Benefits. The Medical Schemes Act (Act 131 of 1998) provides for the definition of Prescribed Minimum Benefits, which stipulate a package of services or care a medical scheme must provide for in its benefit design. There is pressure to reconsider these requirements in order to increase the affordability of medical scheme coverage. This study assessed the potential savings that would be achievable by substituting generics for brand name (originator) medicines listed in the chronic disease algorithms set out by the CMS. It set out to determine the number of generics available for the treatment of each chronic condition and to establish the extent of any price differentials between the generic and the brand name medicines for each condition. Methods: All medicines listed in the 25 chronic diseases algorithms made available by the Council for Medical Schemes were identified. Brand and generic versions were identified in the Monthly Index of Medical Specialties (MIMS, May 2006). Single exit prices inclusive of value added tax, were obtained from the web site of the Pharmaceutical Blue Book and the cost per defined daily dose for one month was then calculated. Cost differentials, where available, were then identified for each medicine listed in the algorithms. Cost differentials for medicines within each algorithm were presented as the median of the difference between brand and generic medicines listed for that algorithm, and also as the median of differences between generic medicines for the same condition. Results: Three of the algorithms (diabetes insipidus, haemophilia and hypothyroidism) list medicines for which no generic equivalent was available at the time of the study. The median cost differential between brand and generic equivalents for the remaining 22 chronic conditions ranged from 19.5% (for type 1 diabetes mellitus) to 97% (for Addison’s disease). Across the entire chronic disease algorithm set, 80 medicines with generic equivalents were listed for 22 conditions. The median cost differential between brand and generic versions of these 80 medicines was 49.9% (interquartile range 32.0 to 78.5%). Of all generic medicines identified, 67.5% were more than 40% cheaper, per DDD per month, than the branded version. In 16 medicines the cost differentials between generic versions were 1% or less. Some correlation between the number of generics and the size of the cost differential was apparent (correlation coefficient 0.49). There were examples of high cost differentials in highly competitive areas of the market. Conclusions: An argument could be made for more closely aligning the process of developing the National Essential Drugs List) and the development of the Council for Medical Schemes algorithms. By being more specific about which medicines should be covered, needless expenditure on “me-too agents of doubtful additional benefit could be avoided. Where clinically warranted, appropriate choices could be provided. Finality in respect of the pricing of medicines needs to be achieved. This applies not only to the dispensing fee but also to the proposed benchmarking process and the proposed differential between brand and generic medicines.