It has been suggested  that because 53 % of United States patents on insulin were linked to the delivery devices and not the insulin itself , intellectual property is not a barrier for earlier versions of insulin entering the market. Patentable innovations in insulin delivery devices are designed to extend the overall patent protection of medicine/device product combinations. Such innovations are incremental but very common .
The statement that insulin IP is not a barrier to market entry is accurate only for the presently marketed insulins not linked to devices (Fig. 1: black symbols), and the main insulin producers are continually filing for patents on analog insulins in their R&D pipelines so their market exclusivity (assuming that these patent applications mature into issued patents) are likely to continue for years to come (Fig. 1: red symbols). In short, analysis of publicly-available data on global insulin patents and manufacturers indicates that the vast majority of the world’s insulin markets are dominated by brand-name manufacturers long after the original product and process patents have expired.
The North American insulin market is dominated by the small number of companies who are the sole suppliers of one or more of six insulin analogs, which are available exclusively as brand name products. There is no US or Canadian human, non-analog insulin. Although third parties are likely free to exploit technology claimed by expiring OB/HC patents, it is possible that existing (i.e., non-expired) IP portfolios of Lilly, Novo, Sanofi and Pfizer in the U.S. and Canada (Fig. 1: red symbols) would prevent or hinder such exploitation. Given that the IP for recombinant human insulin, including DNA sequences and vectors is long off-patent, the existing insulin portfolios are unlikely to be sufficient to block production of human, recombinant insulin. Patent barriers are not the main reason for a lack of a generic version of recombinant human insulin in the U.S. marketplace or indeed, anywhere else in the world.
Moreover, insulin markets have evolved towards containing the newest, most expensive analog products not only in the US and Europe but in every measured insulin market in the world. These shifts greatly complicate access to medicines for the 2.8 billion people living on less than $2 a day, and for many living on higher incomes as well. Stimulating markets for acceptable, yet older products is critical for changing insulin market dynamics; otherwise, brand name companies will continue to introduce upgraded and patented products, deeming their older offerings as obsolete and pulling them from the market. We do not know what fraction of the domestic production of insulin in areas outside the US and Canada is based on producing insulin under license for Novo Nordisk, Lilly, Sanofi and possibly for Pfizer. The positive relationship between INPADOC patent threads for these four large multinational companies and diabetes prevalence (Additional file 3) we infer as manifestation of the scaling effect of market size.
We observed that only 10 % of the 40 putative insulin manufacturers identified in low- and middle-income countries were filing patent applications related to insulin. From this, we infer that they have intentions to market their own insulin in these countries and/or are already marketing their own insulin. For example, there are many companies making insulins for the Indian market and these products include, among others, purified bovine insulin (Bovine Longact® from USV), recombinant human insulin (Wosulin®: rDNA human monocomponent isophane Insulin from Wockhardt; Insugen®, human insulin from Biocon) and various insulin analogs (Lantus®- insulin glargine from Sanofi Aventis; Novomix-30®, Soluble insulin aspart 30 %, insulin aspart protamine 70 % from Novo Nordisk; Glaritus®, Insulin glargine from Wockhardt; Basalog®, insulin glargine from Biocon;) and combinations (e.g., Mixulin®, Porcine Insulin 30 %, Isophane Insulin 70 % from Cadila) [34–36].
Consider the following thought experiment: Assume Company X is producing both human analog insulin and human non-analog insulin in Ethiopia and wants to export both (i.e., respectively, a Lantus® and Humulin® equivalent) into the United States, Europe and a low income country (LIC). At the outset, we reiterate that within a few years patents in all these destinations (U.S., Europe and the LIC), if they exist at all, are unlikely to be a barrier to commercialization of the analog and there are no IP barriers to production of recombinant human insulin.
What regulatory options exist to stimulate more competitive insulin markets? First, if imported into the US or made in the US under contract with Company X, both insulins will be regulated as a “drug” not as a biologic  and the regulatory dossier would be under the ANDA (“Abbreviated New Drug Application” pathway of US FDA Section 505(b)2. Indeed, this pathway was already used in 2006 for approval of a generic recombinant growth hormone product, Omnitrope® by Sandoz relying in-part on the FDA’s prior approval of Pfizer’s pioneer rhGH product, Genotropin® .
In August 2014, the US FDA granted tentative approval for Eli Lilly’s Basaglar®, a recombinantly produced insulin glargine analog for treating diabetes. As a 505(b)(2) product, approval relied in part on clinical studies carried out for the originator, Sanofi's Lantus® (insulin glargine). Basaglar® does not have final approval due to patent litigation involving Sanofi's patents. Time to tentative approval was rapid, however. It was exactly ten months . The same product was approved as a “biosimilar” in 2014 in Europe. In the US and Europe, a recombinant version of non-analog human insulin would follow the same respective pathways . Analog insulin glargine has recently been approved in Mexico  according to the biocomparable approvals pathway defined in 2012 (i.e., Galactus®, under license to PiSA Pharmaceuticals). A key issue, at least for the United States FDA, is whether a biosimilar insulin can be freely substituted at the pharmacy level . The interchangeability of different small-molecule generics leads to substantially reduced drug pricing. When there is no interchangeability, it is not clear whether or not price competition will have an impact unless there is coherence with other policy interventions .
It is an open question as to whether or not the LICs could rely on the regulatory authorities in the US, India or Mexico and allow marketing of a version of glargine or human insulin. Notwithstanding the relative ease of US and European approval of Basaglar®, different manufacturing processes may result in subtly different insulin products. Such differences between versions of all insulins and their respective reference products could be expected .
Regulatory solutions can only partly address the structural problems contributing to uncompetitive off-patent insulin markets, if they do not address the broader problems of physician and patient preference. One of the biggest barriers to widespread access is the fact that doctors may be influenced by claims that insulin analogs are superior to human insulin when the evidence is equivocal. According to the WHO, no clear advantage (with lack of clinically important benefits) of analog insulin over recombinant human insulin has been established . To be sure, if there are clinical complications associated with human insulin use, patients may indeed not want to switch from analog products to a human generic. In markets dominated by analogs, when a patient gets diagnosed (and needs insulin), he/she will likely be given the (multinational) analog insulin. If the patient feels better, they would want to continue with the same (analog) insulin and not switch to other (human) products/brands. Switching to another insulin would mean that a patient will have to regularly visit the doctor for tests/readings, and the patient would likely prefer to remain stable with one insulin. Simply put, the multinational companies have a wide physician network which reinforces their brand perceptions. In low- and middle-income countries where human insulin is still the predominant market share  this behavioral situation may well be less onerous yet, irrespective of insulin type, we suspect physician acceptance is a critical access barrier to overcome.
Finally, once approved for market, the buyers of, as well as the payers for, these generic human insulins will need to negotiate for price, although in the US this opportunity is limited . At present, the major sellers of insulin are well organized and their buyers are not. As pointed out recently , by contrast with antiretrovirals, which were paid for by donors such as the Global Fund, insulin is not purchased by donors, but rather directly from country budgets. In situations where pooled procurement of essential medicines is ongoing [47, 48] or proposed , its implementation may have a great influence on procurement prices for insulins of all types. Pooled procurement, in principle, avoids the costs of sustaining local production facilities that may not be viable in any case. However, it is difficult to investigate the extent to which such pooled procurement is effective in significantly increasing medicine penetration at the national level. But if the end result is that lower prices are being offered and more patients have access to medicines, the health system still benefits. One lesson from the ARV situation is that a possible barrier to pooled procurement is a lack of regulatory and procurement capacity at the country level .
Another option that has been used is a restricted tender system (in contrast to open tenders) for purchasing from well-known pre-qualified suppliers whose products have been previously authorised and with whom the procurement authority has had satisfactory results. However, a potential concern is that restricted tendering rounds may increase the likelihood of market concentration if the same suppliers win contracts, so that competitors let their product market authorisations expire. This is challenge for buyers to be mindful of. Some level of competition is naturally critical for tendering to work effectively, bearing in mind that quality and the continuity of supply are also important considerations .
Some arrangements allowing for tenders might be set up in a way that several manufacturers are selected for supplying the medicine at the same price. If this can be done so that competition is still suppressing prices, this might, in principle, prevent excessive concentration and its negative effects on future prices . Further, the time period for which tenders are awarded could be limited to encourage more diversity in the market. Other criteria besides price can be included in a request for tender, such as quality of the product, quality of the delivery system (e.g., insulin vials versus insulin pens) and security of supply. The tendering system could be structured to ensure patients and their doctors retain adequate choice of subsidised treatments.
A limitation of our method is that, in order for our study to be feasible and replicable, we confined our international patent search to the only international patent databases freely available (i.e., the EPO’s INPADOC via Espacenet, WIPO PatentScope) and India’s national patent database where many major generic pharmaceutical companies are based. However, there are other premium international patent databases (e.g., Derwent) and all other national patent databases [20, 53] which may yield additional records. Nonetheless, the EPO and the WIPO facilitate procedures and communications on a global or regional level. These organizations have the most official and complete information on global applications as well as adjunct information. They should always be used for any serious research that has legal and financial ramifications and for verifying information found in other sources.