Global S&T Development Trend Analysis Platform of Resources and Environment
| DOI | 10.1126/science.abg4374 |
| Insulin imports fail to meet many countries' needs | |
| Abhishek Sharma; Warren A. Kaplan | |
| 2021-07-30 | |
| 发表期刊 | Science
 |
| 出版年 | 2021 |
| 英文摘要 | Limited physical availability of various forms of insulin, and its concomitant high price, is a global health concern, particularly in low- and middle-income countries (LMICs) that have experienced an unprecedented increase in the prevalence of diabetes over the past two decades ([ 1 ][1], [ 2 ][2]). Global insulin need is expected to increase by 20% over the next decade, making insulin supply, delivery, and availability a public health imperative ([ 3 ][3]). But limited insulin production has led to a reliance on the few exporting countries that house corporate and manufacturing facilities of major multinational corporations (MNCs) ([ 4 ][4]). We use an innovative linkage of two datasets in an attempt to answer two questions: From where do countries with no domestic production source their insulin? And of these countries, are their present levels of insulin imports sufficient to treat the in-country number of persons with diabetes?
For various economic, political, and geographic reasons, a disproportionately large amount of insulin is traded among certain high-income countries themselves rather than with others ([ 5 ][5]). Some analyses have been done on global imports and exports of insulin in an attempt to understand the global insulin trading structure and the unit costs of traded insulin ([ 4 ][4], [ 5 ][5]). But the relationship between this prior work on trade supply of insulin and estimates of the insulin need of patients has been unexplored. We analyzed both epidemiological data and global trade data for retail insulin [using the United Nations Commodity Trade (UN COMTRADE) ([ 5 ][5]) data of 194 trading countries] to, in effect, compare insulin imports and in-country need for persons with diabetes as an estimated “supply-need” ratio.
Insulin is manufactured in only 20 countries ([ 4 ][4]), so in principle there are 174 import-dependent countries in the COMTRADE database. However, many of those 174 import-dependent countries also exported varying volumes of insulin at various times during the study period (2000 to 2018). We limited our study sample to those countries lacking domestic insulin production that (a) exported either no insulin at all or just trivial amounts (annual average of 104 kg insulin or less, which approximates insulin for less than 500 persons with diabetes) and (b) had import and diabetes prevalence data for at least 7 years between 2000 and 2018. We applied criterion (a) to ensure that our analysis is reflective of those countries whose insulin importation was intended for domestic patients rather than driven by the foreign-exchange, business, or economic incentives of reexportation. These criteria led to our final count of 82 import-dependent “study countries,” which represent ∼35% of bilateral insulin trade transactions globally. As a sensitivity test, we relaxed criterion (a) to increase the export limit to 600 kg, resulting in 22 “test countries.” In total, these 104 countries represent 83 million persons with diabetes. See detailed methods in supplementary materials (SM). We used trade data reported by importing countries, as imports reported by the receiving country are usually recorded more accurately than corresponding exports from the originating country, probably because of taxing and tariff requirements ([ 5 ][5]). Whereas insulin exports are driven primarily by MNCs and government regulations and incentives, there are multiple, interacting stakeholders driving insulin imports in most countries considered here. Typically, there is a public sector model whereby the government procures medicines, including insulin, and distributes them to health facilities using a publicly run Central Medical Store and a government-owned transport fleet. Insulin and related products funded by some bilateral and international development agencies are procured directly by such agency or a procurement agent acting on its behalf. The private retail sector also plays a role in provision of insulin.
![Figure][6]
Flow of retail insulin into study-country regions from 2012 to 2018
The lower half of this diagram represents insulin exporters; the upper half represents importing regions with the number of study countries in parentheses. Flow of traded insulin [colored arrows proportional to millions of kilograms (M kg) net weight] is shown for each country or region. For this study, Caribbean and Oceania countries were combined to ensure adequate sample size for summary statistics. See supplementary materials for further details on methods and country-level data.
GRAPHIC: K. FRANKLIN/ SCIENCE BASED ON A. SHARMA AND W. A. KAPLAN
Of the 82 study countries, 32 were from Africa, followed by 13 each in Central and South America and in the Caribbean and Oceania regions. The remaining 24 countries were from Asia, Eastern Europe, and the Middle East (table S1). These import-dependent study countries received insulin from various export partners by year and by region. However, between 2012 and 2018, Denmark, France, and Germany exported the majority of insulin—by net weight—to all our study country regions (see the first figure). Study countries in the African region received 27.2% of their insulin imports from countries that themselves do not make it and may be “passing through” insulin sourced from elsewhere (see tables S3 and S4 for these exporting countries). The remainder was imported from France (27.1%), Germany (19.1%), Denmark (17.5%), India (4.1%), China (3.7%), and other manufacturers (1.0%). Study countries in Asia received almost half (46.4%) their insulin from India. Most of the insulin (92.1%) exported by Mexico and Brazil to the study countries was for those located in the Americas, and most insulin exported from China (93.7%) was bound for study countries in Africa.
For each study country, we converted annual net weight of retail insulin imports [kilograms (kg)] during 2000.2018 to insulin international units (IUs) as the supply (1 kg of insulin = 70,000 IUs), using assumptions and formulae as shown in the SM. Then, to estimate the insulin IUs needed for treating persons with diabetes annually, we used data from the NCD Risk Factor Collaboration (NCD-RisC) on country-specific diabetes prevalence—defined as fasting plasma glucose of 7·E0 mmol/liter or higher, or history of diagnosis with diabetes, or use of insulin or oral hypoglycemic medicines ([ 6 ][7]). These NCD-RisC diabetes prevalence data are based on pooled analysis of data from 751 individual population surveys and studies that measured at least one of the markers listed above. We assumed that a fixed percentage (10%) of all such patients in a given country require a daily defined insulin dose of 40 IUs. This insulin-use percentage of 10% is a relatively conservative estimate based on review of published literature on reported use of insulin by persons with diabetes in countries of varying development status (table S2). With the estimated retail insulin supply in IUs (numerator) and estimated need for insulin among those with diabetes based on lab markers and/or actual diagnosis (denominator), we determined the annual “supply-need ratio” for all 82 study countries during 2000.2018. An annual ratio ≥1.00 would mean that insulin supply is enough to meet estimated need of the diabetes-prevalent population (as defined above) for at least that 1 year. See SM for detailed methodology.
![Figure][6]
Regional median insulin supply-need ratios from 2000 to 2018
For clarity, the regional medians are horizontally displaced in ascending order. The vertical lines reflect the 25th to 75th percentile range. For presentation purposes, the 75th percentile for Eastern Europe is truncated at 5.50 in the last two time segments. See supplementary materials for study country–level data and statistical comparisons
GRAPHIC: K. FRANKLIN/ SCIENCE BASED ON A. SHARMA AND W. A. KAPLAN
We estimated median regional supply-need ratio (using retail insulin for study countries) over three time segments from 2000 to 2018 (see the second figure). These ratios increased over time in all regions except Africa and Asia and, except for these two regions, show a mismatch toward oversupply (apparent surplus) rather than undersupply of insulin and its estimated regional need (median ratio >1.00) in later years. Although—except for Africa and Asia—the overall situation with respect to the mismatch trends toward apparent oversupply over time, there are study countries in almost all regions that had median ratios <1.00 during 2012–2018 (see the third figure and table S6). For example, in the Middle East, the ratios for Yemen and Palestine during 2012–2018 were <1.00, far less than those for Qatar, Oman, Kuwait, and Bahrain. Similarly, in the Caribbean and Oceania regions, countries with ratios <1.00 included Dominican Republic, Fiji, Samoa, and Papua New Guinea, and those in the Americas included Bolivia, Guatemala, Ecuador, and Guyana. Three of the eight countries in Eastern Europe (Montenegro, Bosnia and Herzegovina, and Macedonia) had median ratios far greater than 1.00.
The 22 “test countries” had no statistically significant impact on the 2012–2018 regional median ratios (tables S7 and S13). The following “test” countries had ratios <1.00: Tanzania, Uganda, Malaysia, Peru, Croatia, Cyprus, Georgia, and Jordan.
The COMTRADE database also provides information on “bulk” insulin (insulin in form not ready for retail sale or patient use). Nonetheless, by net weight, the importation of bulk insulin is trivial relative to retail insulin imports. Including bulk insulin in the imported “supply” (numerator) has no statistically significant impact on supply-need ratios (fig. S2 and table S10). Some study countries in Africa and the Middle East are importing bulk insulin (table S3), but it is unclear whether these countries can perform conversion of insulin crystals to retail insulin, although they may be repackaging bulk liquid insulin.
Aside from its public health value as an essential medicine, we chose insulin for this analysis because it is the only specifically identified, single pharmaceutical in the entire COMTRADE database, as other pharmaceutical commodities are more aggregated (e.g., medicines “containing penicillins, streptomycins, or their derivatives”). In addition, we chose insulin because it is a specific treatment for a specific indication. The various over- and undermismatches identified herein may well extend to other pharmaceuticals in global trade, suggesting an area for further study.
### Insulin imports and its unmet need
Insulin supply depends on multiple stakeholders interacting with respect to insulin sources, costs, need forecasting and procurement, supply chain management, and health care delivery ([ 1 ][1], [ 2 ][2]). It is well known that only a few companies—with corporate headquarters in Denmark, France, Germany, and the United States—control most of the value and volume of the global insulin market, largely owing to the relatively complex manufacturing process for insulin ([ 2 ][2], [ 4 ][4], [ 5 ][5]).
![Figure][6]
Global distribution of median insulin supply-need ratios, 2012 to 2018
Eighty-two study countries (solid) and 22 “test” countries (striped) are shown, as well as 20 countries with domestic insulin manufacturing capacity (dark gray). Teal color marks countries with ratios ≥1, and red colors mark those with ratios <1. See supplementary materials for details.
GRAPHIC: K. FRANKLIN/ SCIENCE BASED ON A. SHARMA AND W. A. KAPLAN
Moreover, our import-dependent study countries have relatively few import “connections” with Brazil, India, or China (see the first figure), notwithstanding that these countries are major producers and suppliers of medicines and also host insulin-producing plants (table S5) ([ 4 ][4]). Although some countries in the Americas have domestic insulin production capacity, their exports are primarily to markets within a limited geographic outreach, i.e., to countries in the Western Hemisphere that lack such capacity. Even US insulin exports to study countries are primarily to those in the Americas.
African countries appear to have no such regional “champion,” and their insulin imports come primarily from Denmark [Novo Nordisk (NN)], France (Sanofi, NN, Eli Lilly), and Germany (Sanofi), as well as a robust supply from countries that are not obviously making insulin domestically. This suggests that African countries are getting much of their insulin from intermediary trading partners who themselves are passing it through. This is an interesting finding, but the global trade data itself are currently insufficient to learn more about this. Furthermore, only if the study countries in the Middle East and Eastern Europe were to improbably reexport all their 2018 insulin import surplus to those in Africa, could most of the insulin need in Africa be satisfied (table S12).
Our supply-need ratios show that there is a disconnect between the quantity of insulin being imported and the quantity needed to treat the population with diabetes at any one time. The mismatches that we observe are a symptom of many economic, social, political, health care delivery, and access-related factors that determine the journey of insulin in-country from the customs officials who record it to the patient who requires it ([ 1 ][1], [ 2 ][2]). Lack of access to health care may result in missed opportunities to detect diabetes—a process that requires specific diagnostic tests, e.g., the fasting plasma glucose and/or the oral glucose tolerance test. These tests are commonly arranged within a comprehensive health care setting that recognizes the risk profiles of its patients. Without access to such care, patients would be less likely to receive screening and diagnostic services, and their diabetes may be missed by the health care system ([ 7 ][8], [ 8 ][9]).
“Underimportation” (low ratios) compared to our estimated number of patients that require insulin appears consistently in many study countries. Several explanations are possible. Many people with measured high blood sugar may not actually be diagnosed as having diabetes by health providers, and/or those with diagnosed diabetes may not be prescribed insulin. Further, insulin import forecasts may be artificially lowered because access barriers may be preventing patients from getting insulin, and/or more of those on oral hypoglycemics should really be getting insulin but are not ([ 1 ][1], [ 7 ][8], [ 8 ][9]). We suspect all these factors may be contributing to the mismatches in Africa and Asia.
As some persons with diabetes may remain undiagnosed and/or remain untreated with insulin even if they require it, we explored whether the actual need for insulin among diagnosed persons is being met by imported quantities. For that, we calculated country-specific “supply-need” ratios using an alternate denominator (available only for 2018)—i.e., insulin needed for persons with diabetes who are actually being treated with insulin ([ 3 ][3]). We found that the resulting regional median ratios were mostly higher than our original values but were not statistically different with the sole exception being Africa (table S11). This highlights that although countries may work to serve the patients who are prescribed and physically able to avail themselves of insulin treatment, there are many with diabetes—particularly in Africa and Asia—who are simply not getting the needed insulin. If the amount of insulin imported does not take into account the “missed” patients who have high blood sugar but are not formally diagnosed and/or are untreated, the amount of insulin imported may only match the need for those who are lucky enough to be both clinically diagnosed and have access. Many countries in Africa and Asia appear to have gaps between assumed insulin “need” (our 10% assumption) and the actual number of patients who are treated.
Notwithstanding our emphasis on health care systems, a country could actually have an accurate estimate of the domestic need but simply be unwilling or unable to import enough to meet the estimated need. This might reflect factors such as limited health system financing, weak and confused national governance, and lack of accountability regarding noncommunicable diseases in general. Indeed, some study countries do not report any insulin imports for multiple years in a row (table S6).
In this context, “overimportation” (high ratios) may be similarly problematic. Some countries outside of Africa and Asia consistently show high and increasing ratios (fig. S3 and table S6), i.e., more insulin is being imported than is required for current use (∼10% assumption). This mismatch may also be hiding several challenges, i.e., more persons have diabetes than are being reported and/or insulin is being overprescribed and used. This latter situation would be exacerbated by marketing strategies of the pharmaceutical industry ([ 9 ][10]).
### Limitations of UN COMTRADE data
We assumed that the insulin trade reported by countries is accurate but must acknowledge the ways in which it may not be. Submitting COMTRADE trade data (obtained mostly through customs records) is voluntary ([ 10 ][11]). Despite the UN instructions that countries must report the net weight of insulin only (i.e., liquid weight minus the packaging), the reported data may include the weight of packaging material and thus overestimate the insulin supply by an unknown amount—leading to even lower supply-need ratios. Notably, a 2008 survey found that 87% of importing countries complied with the COMTRADE recommendation to measure weight excluding the packaging ([ 11 ][12]).
We cannot say a priori whether the quantities of imported insulin reflect clerical mistakes in transcribing customs data. In this regard, a paper suggested that global trade data quality “has been deteriorating in the past three decades…” in part because “…low-quality reporters have recently increased their share in global trade” ([ 10 ][11]). These authors suggested that “over-invoicing the value of a shipment to take advantage of certain export support schemes or under-invoicing for duty evasion…” is one reason for poor data quality.
To improve insulin availability for patients who need it, coordinated on-the-ground efforts are required from multiple stakeholders. First, national governments must strengthen mechanisms to identify their patient populations requiring insulin. In the absence of robust health insurance plans and electronic medical records, it may be that many LMICs do not have sufficient information about their own diabetes populations, treatment patterns, and utilization rates of medicines and insulin ([ 8 ][9]). There should be allocation of sufficient funding and resources (e.g., inclusion in national benefit packages) for adequate insulin sourcing and in-country distribution.
To optimize insulin sourcing, governments and insulin buyers could benefit from a global or regional pooled procurement mechanism to achieve the best value for money by encouraging monopsony, whereby a single buyer dominates a market ([ 1 ][1]). Advanced market commitment mechanisms that promise large purchase orders over time at established prices can create healthier market dynamics. The GAVI Vaccine Alliance may be a useful model ([ 12 ][13]) to incentivize manufacturers to develop and produce insulin and guarantee a sustainable price for broader coverage.
Currently, mechanisms exist to incentivize smaller manufacturers (e.g., those headquartered outside of Western Europe and the United States) to expand production and supply of their approved, quality-assured insulin ([ 2 ][2]). The World Health Organization (WHO), in November 2019, included insulin in its “prequalification” program. This allows insulin manufacturers to submit their products to the WHO for regulatory review of quality, safety, and efficacy, which may facilitate approval and launch of more products in the international markets and ensure a steadier supply, something that smaller companies in Asia, South America, and the United Arab Emirates could find mutually beneficial with regard to Africa. The resultant increased market competition could lead to favorable prices for governments, procurement agencies, and patients ([ 1 ][1], [ 2 ][2]).
To improve in-country supply of affordable insulin for patients through evidence-based policy, governments and/or donors should ensure adequate funding for repeated or continuous surveillance monitoring of availability, prices, rational use, and affordability of insulin—including the type of insulin being used and clinical outcomes. In this context, national electronic health records would be useful, but they are often non-existent in many LMICs. Information from commercial health care market databases on insulin use and/or prescriptions filled is very valuable. However, such data are often expensive, are typically collected only from the private—not the public—sector, and are often not available in lower-income countries and small countries with no domestic insulin production, as these may not have a relationship with the commercial databases. Evidence regarding major global public health issues—such as access to insulin—should be informed by open, quality data available as needed.
On that note, a conspicuous lack of granularity in the COMTRADE coding system limits the ability to characterize medicine trade patterns from a public health perspective. This impedes evidence-based policy-making. Although the COMTRADE coding system may be adequate to capture traded quantities of other commodities (such as mineral ores, oils, or cotton), specific codes for more individual medicines and vaccines (instead of group-level codes) along with variables that are more relevant to the pharmaceuticals—such as dosage form and strength—would be needed. With regard to insulin, updating the coding system to capture the type of traded insulin (human or analog insulin), its strength (40 or 100 IU/ml), and dosage form (vials or pens) would go a long way toward informing global and national policies to optimize insulin procurement, supply security, and availability.
Although it has been 100 years since its discovery, insulin remains inaccessible to millions of people with diabetes, and multi-stakeholder initiatives are required to improve access ([ 1 ][1], [ 2 ][2]). In LMICs, where the majority of individuals with diabetes live ([ 6 ][7]), the inability to access high-quality health care services and diabetes care may simply not allow many to obtain whatever insulin has entered the country, and it may well be insufficient to meet the actual requirements of all those who need it. Access to insulin needs to be framed within the wider health system, and all the elements—the public sector, private sector, manufacturers, regulatory bodies, and academics and clinicians—must align for appropriate diabetes care for patients.
[science.sciencemag.org/content/373/6554/494/suppl/DC1][14]
1. [↵][15]1. D. Beran,
2. M. Ewen,
3. R. Laing
, Lancet Diabetes Endocrinol. 4, 275 (2016).
[OpenUrl][16]
2. [↵][17]1. D. Beran,
2. M. Lazo-Porras,
3. C. M. Mba,
4. J. C. Mbanya
, Diabetologia 64, 954 (2021).
[OpenUrl][18]
3. [↵][19]1. S. Basu et al
., Lancet Diabetes Endocrinol. 7, 25 (2019).
[OpenUrl][20]
4. [↵][21]1. D. Beran et al
., Diabet. Med. 36, 726 (2019).
[OpenUrl][22]
5. [↵][23]1. W. A. Kaplan et al
., “Insulin trade profile: April 2016” (Health Action International, Amsterdam, 2016); |
| 领域 | 气候变化 ; 资源环境 |
| URL | 查看原文 |
| 引用统计 | |
| 文献类型 | 期刊论文 |
| 条目标识符 | http://119.78.100.173/C666/handle/2XK7JSWQ/335526 |
| 专题 | 气候变化 资源环境科学 |
| 推荐引用方式 GB/T 7714 | Abhishek Sharma,Warren A. Kaplan. Insulin imports fail to meet many countries' needs[J]. Science,2021. |
| APA | Abhishek Sharma,&Warren A. Kaplan.(2021).Insulin imports fail to meet many countries' needs.Science. |
| MLA | Abhishek Sharma,et al."Insulin imports fail to meet many countries' needs".Science (2021). |
| 条目包含的文件 | 条目无相关文件。 | |||||
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论