A survey of availability, price and affordability of essential medicines from 2011 to 2016 in Chinese secondary and tertiary hospitals

Information pertaining to the use of essential drugs, including dosage form, strength, purchase time, specification, manufacturer, and price information, was extracted from the China Medicine Economic Information database (CMEI). The CMEI, which was constructed in 1993, is a large observational database of drug procurement records that cover sample institutes of 396 secondary hospitals (accounting for 6.5% secondary hospitals in China) and 763 tertiary hospitals (accounting for 38.7% tertiary hospitals in China). All of these hospitals are public and are located in 28 provinces across mainland China (excluding Qinghai, Tibet and Hainan). All member hospitals submitted ample drug procurement records to the CMEI monthly and then the records were aggregated and standardized for researchers and policy-makers.

Meanwhile, Management Sciences for Health (MSH) provided the reference price from 2011 to 2013 of the generic medicines for calculation of median price ratio (MPR) [11]. That is to say, a MPR of 2 would mean that the local medicine price is twice the international reference price. MSH international reference prices, which are generally offered by not-for-profit suppliers to developing countries, are recommended as the most useful standard. Generally, a MPR of one or less indicates an efficient public sector procurement system.

China’s population income distribution was obtained from the China Statistical Yearbook from 2011 to 2014 [12‐15]. (Additional file 1: Table S4) The China Statistical Yearbook provided information about income of urban and rural residents per capita for calculation of affordability.

Thirty medicines were surveyed from January 2011 to November 2016 in the database: 13 from the WHO/HAI core global and regional lists (representing medicines for common acute and chronic disorders) and 17 locally selected supplementary medicines chosen for their local importance and disease burden in China, with input from an advisory committee of practicing pharmacists, academics and experts. Of the 30 medicines surveyed, 27 were listed in the NEML database [16] and three were on medicines procurement supplementary list of several provinces. Eighteen medicines treat acute disorders, whereas 12 treat chronic disorders. Out of 10 therapeutic classes we surveyed, hypertension, rheumatoid arthritis and diabetes have the highest prevalence [17] among Chinese patients and so we chose four medicines of treatment for these three diseases to assess affordability (Additional file 1: Table S5).

The availability of each medicine was reported as the percent availability of an individual medicine at the surveyed hospitals. As mentioned above, we included all of the strengths of sample medicines through measuring the chemical entities by smallest unit. Median availability of selected medicines was used in statistical analysis. We compared the availability for adjacent years, different regions, and that between innovator brands and generics.

The median price ratio (MPR), which represented the ratio of one medicine’s median unit price to the international reference price (IRP), was used for price evaluation. The median MPR of selected medicines was used in statistical analysis. To facilitate comparisons of reported figures, all reference prices of MSH from 2011 to 2013 were converted into Chinese Renminbi (CNY) over the same period by purchasing power parity (PPP) conversion rates [18]. For the absence of reference prices of MSH from 2014 to 2016, we used discount factor (DF) to discount all drug price from 2014 to 2016 into price in 2013 and then compared them with reference price in 2013 to calculate MPR(see calculation details in Additional file 1: Table S3) [19]. Finally, we analyzed differences between annual and regional MPR and between MPR of innovator brands, and MPR of generics.

Due to a lack of data about the average daily wage of the lowest-paid unskilled government workers (LPGWs), we used incidence of catastrophic drug expenditure (CDE) to assess affordability instead of LPGWs. CDE is a concept borrowed from catastrophic health expenditure (CHE) [20], which is widely used to describe all types of health expenditures that threaten the financial capacity of a household to maintain its subsistence needs. Different thresholds are used to define CHE in different researches. Generally, out-of-pocket healthcare payments (OOP) that comprise ≥10% of total household expenditures and out-of-pocket healthcare payments that comprise ≥40% of a household’s non-subsistence income are widely used [21, 22]. When only examining households with catastrophic out-of-pocket drug expenditures, a drug budget share equal to or greater than 10% is universally accepted [23]. Finally, we estimated the percentage of households with catastrophic drug expenditures (defined as a drug budget share of 10% or more) and used 7.5% and 12.5% to do sensitivity analysis. In this study we used drug expenditure instead of OOP to calculate CDE on the basis that in China copayment of hypertensive and diabetic outpatients is up to 85% and that outpatients almost fully cover drug expenditure for chronic diseases [24]. Moreover, reimbursement rate is slightly different among different provinces, therefore, drug expenditure before reimbursement is a better indicator for regional comparison. Given that we only included 3 chronic diseases which required lifelong treatment with medications, we calculated the daily average cost of medicine to estimate drug expenditure.

Due to the fact that urban residents earn an income exceeding that of rural residents, we have the figures of income per capita calculated separately for both residents.

Among the 30 essential medicines, their median availability of different region from 2011 to 2016 were shown in Table 1. The nationwide availability remained steady and exceeded 50% in 2012 and 2015. The trend of availability in each area was consistent with the national trend and the availability of drugs in the eastern region was significantly higher than the central and western regions. Fifteen medicines had > 50% availability (Additional file 1: Table S5). Azithromycin ranked the highest and glyburide had the lowest availability. Another finding captured differences among drugs under the same category. The availability of Melformin, for example, which is a common diabetes drugs, was about 80% while another diabetes drugs such as glyburide had the lowest availability of 3%. But as for brand name and generic drug (Table 2), we found no significant difference.

The nationwide median MPR of 30 drugs kept steady around 5 from 2011 to 2015 but increased significantly in 2016 (Table 3). Additionally, we captured a different MPR of drugs in different areas. The MPR of drugs in the eastern region was significantly higher than the central and western regions. As for specific essential medicines, beclometasone received the highest MPR of 793.79 while ranitidine had the lowest MPR of 0.87 in 2011. The MPR of innovator brand medicines were significantly higher than the MPR of generic medicines all the time (Table 4). The ratio of MPR of innovator brands to that of generics increased from 3.66 to 6.32 during the study period. The result of median MPR of specific essential medicines is shown in Additional file 1: Table S5.

Overall, the incidence of essential medicines causing CDE have decreased over time. It is clear that brand name medicines were more likely to cause CDE than generics and rural patients have a greater tendency to fall into CDE. The sensitivity analysis indicated that the results were robust for varying percent of income (Table 5).

As mentioned above, we improved the WHO/HAI survey manual and exceeded the specific strengths list so that the nationwide availability outclassed that of other contemporaneous research in China. Yet the study has three limitations. First, we found significant differences among drugs under the same category for MPR and availability. That is to say, we may exaggerate or underestimate the severity of the condition if we only assess the access to one drug in our survey list neglecting other therapeutic alternatives. What is more the results will be worse if we take in account other expenses such as consultation fees and diagnostic tests. Therefore, further studies can focus on assessing the access to essential medicines from the perspective of the curing process of disease instead of limited sample drug and include the alternative drugs. Second, although China’s population income curve is a good proxy for estimating income, there is still a narrow gap between the income curve from the China Statistical Yearbook and the real income distribution in China. Third, we used three parameters –– availability, prices and affordability — to reflect access to essential medicines. Other barriers, not mentioned in this work, may impair or diminish the population’s access to medicines and also need concerns for further study.