STRENGTHS AND LIMITATIONS OF THIS STUDY

This study represents the largest multicentre assessment of antibiotic use over 5 years in Vietnam, a country with high burden of antibiotic resistance.

WHO AWaRE classification, Essential Medicine List (EML) and The Anatomical Therapeutic Chemical (ATC) classification system were used together to provide a more comprehensive analysis of the pattern of antibiotic consumption.

The major limitation of this study is the use of convenience sampling, which may limit the generalisability of the results.

Another limitation is the use of procurement data as the proxy of antibiotic use could lead to a bias in the estimation of the actual antibiotic consumption.

Introduction

Antibiotic consumption is one of the most important drivers of the emergence of antibiotic resistance. Global antibiotic use has increased by 46% from in 9.8 defined daily doses (DDD) per 1000 population per day in 2000 to 14.3 DDDs per 1000 population per day in 2018.1 The AWaRe classification of antibiotics was developed by WHO in 2017 and revised in 2019 and 2021, as an interactive tool to support countries in planning and monitoring antibiotic use and the effect of antibiotic stewardship policy.2In the revision in 2021, the AWaRe Classification includes details of the 258 antibiotics classified as Access, Watch and Reserve, their pharmacological classes, Anatomical Therapeutic Chemical (ATC) codes and WHO Essential Medicines List status. The WHO 13th General Programme of Work 2019–2023 has set the country-level target of at least 60% of total antibiotic consumption being Access group antibiotics by 2023.3

The WHO launched the Global Antimicrobial Resistance and Use Surveillance System (GLASS) in 2015 and piloted the surveillance of antimicrobial consumption in 2016.4 The surveillance protocol and tools were standardised to support countries in implementation, especially in low- and middle-income countries. At the national level, countries collect and analyse antimicrobial consumption on a routine basis, from information of value chain of medicines, including procurement and supply, distribution, prescribing, dispensing and patient use.4 As of 2020, only 22/126 countries, territories and areas participated in GLASS contributed to the most recent GLASS report published in 2022.5

Vietnam is a lower middle-income country that has a high burden of drug resistance. It has reported antibiotic resistance data to GLASS but has not participated in the report of antibiotic consumptions by 2023.6 In a study between 2000 and 2015 in 76 countries using the antibiotic sales data, Vietnam ranked the second among countries with the highest increase in Watch antibiotic consumption (10.6 DDDs per 1000 population per day), after Turkey (16.2 DDDs per 1000 population per day).7 At the community level, the proportions of Access and Watch group antibiotic encounters were 59.0% and 39.3%,8 while in the hospital settings, they were 47.2% and 52.4, respectively.9

With the health expenditure per capita of US $166 (in 2020)10 and a high frequency of antibiotic use in hospitals,11 monitoring the antibiotic consumption and expenditure of essential antibiotics is important for Vietnam to develop antibiotic stewardship policy and promote more equitable and cost-effective treatment and maintain national budget sustainability. This study aims to describe the pattern of antibiotic use and cost in public hospitals in Vietnam from 2018 to 2022 using the AWaRE classification.

MethodsStudy setting

Vietnam is divided 63 provinces and in 2020, the country had a total of 350 068 hospital beds in 34 central-level hospitals and 13 university hospitals under the Ministry of Health, 471 provincial hospitals, 952 district hospitals/primary healthcare centres under provincial department of health and 231 private hospitals.12 The public sector accounted for 96% and private sector accounted for remaining 4% of total number of hospital beds nationalwide.12 Of the public hospitals beds, central-level hospitals comprised 9.8% and provincial hospitals and district hospitals/primary healthcare centres compromised 43.3% and 46.9% of hospitals beds, respectively.13

In Vietnam, the bidding process for the supply of drugs for public hospitals and healthcare centres is guided by the government regulations. The procurement and supply of drugs can be performed by a centralised bidding process in which a healthcare authority organise the bidding for large amounts of commonly used drugs for many healthcare facilities in their catchment areas or a decentralised bidding process, where individual hospitals or healthcare centres independently conduct their own bidding processes to address their actual needs.14 15 Procurement units and suppliers are obligated to complete at least 80% of the total contract value during the contract validity period.14 15 Therefore, amount of drug procurement can be used as a proxy for measuring the drug consumption.

This was a cross-sectional study of antibiotic use in healthcare facilities in Vietnam. This report analysed and classified the drug consumptions by the following groups: (1) national or provincial centralised drug procurement and (2) decentralised drug procurement with the subgroups of the central, provincial and district hospitals.

Data resources

The Law on the Procurement stipulates that the bidding opportunities, contract award (purpose, contractor and value) must be publicly announced, and it is encouraged to publish on the local government agencies’, ministries’ websites or mass media.16 This analysis is based on the data of the tender-winning bids that is published in the website of the Drug Administration of Vietnam (https://dav.gov.vn/), which is the Ministry of Health regulatory authority for the period January 2018 and December 2022.17

In the database, each drug record included the name of the active ingredient, trademarks, strength, dosage form, package, route of administration, registration identification, pharmaceutical manufacturer, country of origin of manufacturer, unit of measurement, quantity, unit price and total purchase expenditures.14 15 The public disclosure of procurement data was checked to ensure data transparency and accessibility. Institutional data were de-identified before analysis.

Statistical analysis

Antibiotics were classified as Access, Watch, Reserve and non-recommended by the 2021 AWaRE classification, 2021 WHO Essential Medicine List (EML)2 and analysed using the Anatomical Therapeutic Chemical (ATC) classification and DDD methodology developed by the WHO Collaborating Centre for Drug Statistics Methodology (ATC/DDD system).18 All antibiotics were selected by the list of antibacterials for systemic use (ATC group J01) for this analysis.

The DDD is a measurement unit of the average maintenance dose of a drug per day for a 70 kg adult for its main indication. The number of DDDs was calculated for each antibiotic by dividing the total amount of purchased antibiotics by the DDD conversion of the corresponding antibiotic in the ACT/DDD 2021 system. The percentage of Access, Watch and Reserve group antibiotics were calculated as the DDDs of antibiotics in each group divided by the total DDDs. The average expenditure per DDD for an antibiotic group was calculated by dividing the bidding price for that antibiotic group by the number of DDDs and weighted by the actual amount of purchased antibiotics. All prices were adjusted for inflation using the consumer price index to 2022 price and converted to US$ with the exchange rate of 2022 according to the annual average official exchange rates of the World Bank (US $1 = 22,602.05 VND in 2018 and = 23,271.21 VND in 2022).19 This allows comparison between antibiotic price between years.

Datasets were cleaned and joined using Microsoft Excel (Office 365, Microsoft Corporation, Redmond, Washington, USA) with Power Query Editor. Data were analysed using IBM SPSS Statistics for Windows (version 27.0. Armonk, NY: IBM Corp). Standard descriptive statistics were calculated for categorical (in percentage) and continuous variables (in median and IQR). Differences were considered statistically significant at p values ≤0.05.

Patient and public involvement

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Results

Between January 2018 and December 2022, there were 390 published winning bids for drug procurement, including 65 national or provincial centralised procurement units and 325 individual hospitals totalling approximately US $12.79 billions (table 1). There were 210/390 procurement units represented for at least 3 of 5 years. The overall antibiotic expenditures accounted for 24.7% (US $3.16 billions) of total drug expenditures. A total 2.54 million antibiotic DDDs was included in this analysis. There were 86 antibiotics available in the country corresponding to 35.1% (86/245) of antibiotics defined by ATC classification. Of these 86 antibiotics, 25 (29.1%) substances belong to the Access category, 45 (52.34%) substances belong to Watch category (fosfomycin belongs both Access and Watch category), 7 (8.1%) substances belong to the Reserve category and 10 (11.6%) substances belong to non-recommended group of antibiotics by AWaRe classification (single dosage form of ornidazole is not defined by AWaRe classification).

Table 1

Characteristics of datasets

There were 10 products (combinations of ampicillin, benzathine benzylpenicillin, benzylpenicillin, ciprofloxacin and tinidazole combinations, combinations of penicillins, norfloxacin and tinidazole, spiramycin, spiramycin and metronidazole and sulfamethoxazole and trimethoprim) that were excluded from analysis of numbers of DDDs as without DDD assignment. Parenteral antibiotics accounted for 11.9% (303.4 million DDDs) of the total antibiotic consumption. The largest proportion of parenteral antibiotic use was in tertiary hospitals (57.8%, 21.4 million DDDs), followed by provincial hospitals (17.5%, 36 million DDDs) and district hospitals (9.9%, 5.1 million DDDs). In overall, the top three antibiotic classes were second-generation cephalosporins (21.4%, 541.6 million DDDs), beta-lactam/beta-lactamase-inhibitor (15.2%, 386 million DDDs) and penicillins (14.6%, 371.1 million DDDs).

Excluding the non-recommended group, the Access group antibiotic accounted for 43.9% (1.11 billion DDDs) of total antibiotic procurement for 5 years, ranging from 40.9% to 53.8%. The majority of antibiotic consumption was from the Watch group, which accounted for 56% (1.41 billion DDDs) of total antibiotic procurement (ranging from 46.0%–59%). Consumption of Watch antibiotics was highest in national hospitals (ranging from 72.4% to 76.9%), followed by provincial hospitals (ranging from 56% to 58.1%)) and was lowest in district hospitals (range from 41.6–53.9%) with a consistent pattern over the study period. Proportions of procured antibiotics according to their AWaRe categories are presented in figure 1.

Figure 1

Proportions of antibiotic procurement in DDD (%) and expenditures by the level of hospitals.

In the Access group, the three most commonly used antibiotics were amoxicillin and beta-lactamase inhibitor (33.5%, 372.5 million DDDs), amoxicillin (30.8%, 342.4 million DDDs), cefalexin (13.7%, 152 million DDDs) and cefadroxil (6.9%, 76.5 million DDDs), which together accounted for 84.8% (83.5–88.8%) of DDDs and for 58.2% (51.9–65.4%) of funds spent on this category (figure 1). Cefuroxime and linezolid were the most commonly used antibiotics in the Watch and Reserve groups, which accounted for 32.4% (460.3 million DDDs) and 50.8% (1.4 million DDDs) of the number of DDDs in the corresponding groups, respectively.

Reserve group antibiotics represented only 0.1% (2.8 million DDDs) of the number of DDD during the study period but accounted for 5% (US $149.9 millions) of the total of expenditure. The same disproportion between the share of quantities and expenditure of Reserve group antibiotics was seen in all levels of hospitals, with the ratios of shares of expenditure to the share of DDDs ranging 7.6:1 in national levels to 26:1 in provincial hospitals and 42:1 in district hospitals in figure 1.

The yearly average cost of antibiotics per DDD was $0.69 for the Access group (ranging from US $0.62 in 2019 to US $0.73 in 2018), US $1.45 for the Watch group (ranging from US $1.19 in 2018 to US $1.82 in 2019) and US $53.6 for the Reserve group (ranging from US $48.2 in 2018 to US $67.1 in 2022). The most expensive was ceftazidime/avibactam and ceftolozane/beta-lactamase inhibitor (Reserve groups), which were first available from 2020 at US $357.35–372.72 per DDD and US $210.26–222.85 per DDD, respectively. The least expensive antibiotic was doxycycline (US $0.02–0.04). The cost ratio per DDD between Watch- and Access group antibiotics was 2.1 (1.6–2.9) and between Reserve- and Access groups were 77.9 (66.3–92.4).

Of these 86 antibiotics, 26 (30.2%) were essential, 44 (51.2%) were non-essential and 16 (18.6%) was not defined by the EML. The proportions of essential antibiotic substances were 12/25 (48%) in the Access group, 11/45 (24.4%) in the Watch group and 3/7 (42.9%) in the Reserve groups.

When different level of care was taken into account, a difference was observed in the purchase of non-essential antibiotics in national, provincial and district hospitals. The proportion of essential antibiotics in the numbers of DDD and expenditures by the AWaRe classification was showen in figure 2. Non-essential medicines consumed an average of 20.2% (510.8 million DDDs) of purchased antibiotic DDD and 44.4% (US $1.31 billions) of antibiotic expenditure. The most commonly purchased non-essential antibiotic was levofloxacin (4.6%, 116.3 million DDDs) in terms of total number of DDDs and was cefoxitin (5.3%, US $168.6 millions) in terms of total expenditure. In the Reserve group, linezolid was the most commonly purchased antibiotic, which accounted for 50.8% (1.4 million DDDs) (ranged from 49.4%–54.3%) of DDDs and 22.1% (US $33.1 millions) (20.3%–26.6%) of the expenditure (table 2 and online supplemental table 1). The non-essential antibiotics were more expensive than essential antibiotics in the corresponding AWaRe categories. The yearly weighted average costs of non-essential antibiotics per DDD varied from US$ 1.77 (Access antibiotic), US$ 2.85 (Watch group) to US$ 71.82 (Reserve group), while the average costs of these three essential antibiotic groups by AWaRe categories were US $0.53, US $0.95 and US $59.34, respectively.

Figure 2

Proportion of non-essential antibiotics in DDD (%) and expenditures by EML.

Table 2

Proportions of antibiotic procurement in DDD (%) and expenditures by AWaRe classification for 5 years

Discussion

To the current knowledge, this was the largest study in Vietnam to date evaluating the consumption of antibiotics using the AWaRe classification in public hospitals. With 5 year data from the procurement of drugs from 2018 to 2022 showing that the yearly average consumption of the Access group antibiotic ranged from 40.9% to 53.8%, the public hospital sector in Vietnam was under the WHO target of at least 60% of total antibiotic consumption being the Access group antibiotic by 2023.

In 2020, data on antibiotic consumption from 26 countries and territories in the GLASS report showed that the median consumption was 16.6 (range, 12.3–31.2) in DDDs per 1000 population per day globally and 15.3 DDDs per 1000 population per day in South-East Asian and West Pacific Regions (including five countries, namely, Bhutan, Nepal, Maldives Mongolia, Lao People’s Democratic Republic and Brunei Darussalam).5 Amoxicillin with/without beta-lactamase inhibitors was the most consumed antibiotics, and ceftriaxone was the most often used parenteral antibiotics.5 The Access group antibiotic accounted for >60% of the total antibiotic consumption in 17/26 enrolled settings with a median of 67% and range from 26% in Nepal to 88% in Bhutan. Reserve group antibiotics accounted for a median of 0.1% (ranged 0–3%). In a point prevalence survey in 69 countries between 2015 and 2018, West and Central Asia was the region with the highest proportion of Watch group antibiotic use (66.1%) and the lowest proportion of Access group use (28.4%).20

Bidding procurement data can be used as a proxy for monitoring antibiotic consumption. In Asia, this approach was used in some countries, such as China, Malaysia and Myanmar. In China, the survey showed that antibiotic consumption increased from 2012 to 2014, then decreased by 2016, and the top three antibiotics consumed were penicillins, quinolones and macrolides.21 In Myanmar, the study included systemic antibiotic procurement data from almost a third of all public hospitals. During 2014 and 2017, the proportion of the Access group antibiotics increased from 42% to 54%, while the proportion of Watch group antibiotics decreased from 46% to 38%.22 In Malaysia, the retrospective study used the nationwide pharmaceutical procurement and sales data from all public health facilities between 2018 and 2021. The Access group antibiotics consistently accounted for at least 90% in the public sector, while it was lower in the private sector (64.2 to 68.3%).23

Vietnam is the country with the highest prevalence of antibiotic usage in Southeast Asia as measured in children aged <5 years with symptoms of lower respiratory tract infections (74% vs a median usage of 51% in the region).1 In a study of 1747 patients admitted to the critical care units in Vietnam, 63.6% received empirical antibiotic treatment. Of those patients with antibiotic initiation, 24.1% received the Access group antibiotics, 87.3% received the Watch group antibiotics, 0.6% received the Reserve group antibiotic and 5% received non-recommended antibiotics.11 In a study of patients in surgical and internal medicine wards in provincial hospitals, cephalosporins were the most commonly used antibiotics (53.9%), followed by fluoroquinolone (25.4%), penicillin/β-lactamase inhibitor (10.4%)24 and carbapenems were prescribed in 2.4% of patients. In Vietnam, the insurance policy restricts the reimbursement for selected Watch group antibiotics (eg, carbapenems, levofloxacin) and all Reserve group antibiotics in tertiary and secondary hospitals.25 This policy can partly explain the less use and the lower expenditure for these group antibiotics in primary hospitals.

The high consumption of Watch group antibiotics in Vietnamese hospitals may be understandable in the context of the high prevalence of antibiotic resistance in the country. The study of 42 553 bacterial isolates from 13 hospitals between 2016 and 2017 in Vietnam showed that 58% (663/1136) of Staphylococcus pneumoniae strains were resistant to penicillin, 59% (4,085/6,953 isolates) of Escherichia coli strains and 40% (1,186/2,958 isolates) of Klebsiella pneumoniae produced extended-spectrum β-lactamase and 73% (3,302/4,515 isolates) of Staphylococcus aureus strains were methicillin resistant.26

The details of the price of individual antibiotics in Vietnam were reported in the previous report.9 In this analysis, the median prices of the Reserve group antibiotics showed a downward trend throughout 5 years. However, it was still 28 times more expensive than the Watch group antibiotics and 46 times more expensive than the Access group antibiotics. At the current national monthly income of $182 per capita,27 the cost of the last resource antibiotics can cause the financial burden to patients who are currently infected with drug-resistant pathogens. Therefore, the national-level target of at least 60% of total antibiotic consumption being the Access group will promote the responsible use of antibiotics and slow the spread of antibiotic resistance while making health services affordable and accessible for the population.

The WHO’s Model List of Essential Medicines (MLEM) included selected quality-assured drugs that are the most effective, safest and comparative cost-effectiveness for the affordability of the health system.28 These findings of the availability of essential antibiotics were consistent with a previous study in 23 countries with 1130 medicines in 2012, which showed that the median availability of essential antibiotics for systemic uses was 78.4%.29 However, financial data of essential versus non-essential medicines are not well established. In this study, there was a suboptimal expenditure of non-essential antibiotics (44.5% of total antibiotic expenditure) that needs particular attention to increase the efficiency of medicine expenditures.

This study has several limitations. The major limitation was the selection bias due to the incomplete published data on medicines procurement. There were some uncertainties in the representativeness of levels of hospitals in the national or provincial centralised drug procurement. Additionally, the 390 procurement datasets came from a diversity of procurement units and the complete datasets were not available for all facilities for the whole 5 years (table 2), so this analysis was insufficient to describe the overall trend in antibiotic consumption over years. The pharmaceutical market in Vietnam has experienced significant growth with the value increasing from 2.7 billion USD in 2015 to 6.2 billion USD in 2020 and is forecasted to reach a value of 13 billion USD in 2026.30 However, with a large geographical coverage and based on the estimation of national drug expenditure (6.2 billion USD in 2020), this study offered an annual snapshot of antibiotic use and estimated that this analysis represented approximately 20–30% of the national funds spend on medication. Another limitation is the use of procurement data as the proxy of antibiotic use could lead to a bias in the estimation of the actual antibiotic consumption as not all antibiotics were put into use. Understanding about the links between different stages of the procurement cycle is important for interpretation and effective decision-making.

Conclusion

Consumption of Access group antibiotics in Vietnam by 2022 represented less than the target of ≥60% set by the WHO, and there was a disproportionate share of expenditures for the AWaRe antibiotic categories. Vietnam should urgently implement the monitoring of the national-level antibiotic consumption in hospitals that could be used for national stewardship interventions and consider redirecting resources to essential drugs. Further works need to be done to investigate the feasibility, effectiveness and acceptability of using The WHO AWaRe antibiotic book31 as a guidefor clinical management of common infections in children and adults in the setting of high prevalence of antibiotic resistance.

Data availability statement

Data are available upon reasonable request.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

Not applicable.