Irrational use of medicines is a global problem that results in the mismanagement of patients, and the wastage of critical resources, and has contributed to the growing burden of antibiotic resistance.^1,2 As an effort to address irrational medicine use, indicators (core and complementary) for the evaluation of medicine use in health facilities were developed and validated by the World Health Organization (WHO).³ The core indicators are divided into prescribing, hospital, and patient care;^4,5 and they have been reported to be less likely to change over time and context, more informative, and easier to perform.^6,7

Various settings have used the WHO indicators to evaluate medicine use and prescribing practices. As there are no international standards that have been developed to assess prescribing, these WHO indicators are used as proxies.⁶ The core World Health Organization/International Network for Rational Use of Drugs (WHO/INRUD) prescribing indicators are average number of medicines (WHO standard = 2–3), average percentage of medicines prescribed by generic name (WHO standard = 100%), average percentage of prescriptions with one or more antibiotics (WHO standard ≤ 30%), average percentage of prescriptions with one or more injections (WHO standard ≤ 10%) and percentage of medicines prescribed from the essential medicines list (EML) (WHO standard = 100%).²

Although literature suggests that awareness on inappropriate prescribing is relatively high, medicines, particularly antibiotics, continue to be used irrationally. In low- and middle-income countries (LMICs), a 10-year review of empirical evidence showed that prescribing practices did not improve over time; the average number of medicines prescribed per patient increased from 2.1 to 2.8, and prescribing of antibiotics increased from 45% to 54%.⁸ A systematic review of prescribing practices conducted in 11 African countries showed that prescribing in the African region was not in line with WHO recommendations, more so in private than in public sector facilities.⁶ In Ethiopia, a study found that while the prescribing and dispensing practices were good, irrational use of antibiotics and the average number of medicines per prescription remained problematic.⁹ Irrational use of medicines has also been reported in other LMICs such as Sudan, Yemen, Thailand, Mali and Uganda.^{10,11,12,13,14,15}

The success of interventions against irrational prescribing is debatable. Prescribing practices are influenced by educational, regulatory, economic and managerial factors.² Managerial interventions are diverse and include limiting medicines available for prescribing in a setting, hence influencing the decisions that those prescribers would make,¹⁶ monitoring the use of guidelines and providing supervision, and auditing prescriptions and giving feedback to prescribers and pharmacists.¹⁷

There is evidence to support the use of multifaceted interventions in addressing irrational medicine use.¹⁸ A systematic review showed that a prescription audit and feedback intervention, a managerial strategy, improved healthcare professionals’ practice, although the effects were small to moderate.¹⁹ Other studies echo that interventions may have positive impacts on the rational use of medicines but those impacts tend to drop post-intervention.^20,21

Minimal work on the rational use of medicines has been conducted in The Kingdom of Eswatini (previously Swaziland and hereafter referred to as Eswatini). In Eswatini, the WHO/INRUD indicators have not been integrated into the health system among the different levels of care, and hence, no prescribing targets and rates have been set for the five prescribing indicators. To provide initial data and the authors assessed prescribing indicators in the public sector and faith-based facilities across Eswatini. The authors then evaluated the effects of a short intervention, designed to improve WHO/INRUD prescribing indicators, immediately after the intervention and assessed whether the effects were sustained 6 months later. Findings from this study could be used to evaluate the applicability of the WHO/INRUD prescribing indicators in assessing prescribing in public sector facilities in Eswatini. Furthermore, findings could inform the need to validate the WHO/INRUD indicators for Eswatini and to allow for the development of different indicators to guide future prescribing in this setting, including prescribers’ compliance to the standard treatment guidelines.²²

Figure 1 shows the sampling of facilities, their allocation to intervention and control arms and the total number of facilities analysed at the end of the study.

A total of 3,200 prescriptions were analysed at baseline. The intervention arm had 16 facilities: 6 (38%) in the Hhohho region, 1 (6%) in Shiselweni, 3 (19%) in Lubombo and 6 (38%) in the Manzini region. The control arm also had 16 facilities: 4 (25%) in the Hhohho region, 7 (44%) in Shiselweni, 2 (13%) in Lubombo and 3 (19%) in the Manzini region as shown in Table 1. Table 1 also reports baseline findings of the WHO/INRUD prescribing indicators by intervention and control facilities. Rational prescribing standards as recommended by WHO were only met by the average number of medicines per prescription (though the average was slightly higher than the standard 2–3 medicines per prescription) and the percentage of prescriptions with injections of 10.7% (though slightly higher than the WHO standard of ≤ 10%), in both intervention and control facilities (Table 1). The other prescribing indicators did not reach the recommended standards in both intervention and control facilities (Table 1). Of note at baseline was the high use of antibiotics, above 50%, in both intervention and control facilities (the WHO recommendation is for the average percentage of prescriptions with one or more antibiotics to be ≤ 30%) (Table 1).

A total of 3,200 post-intervention and 3,200 post-follow-up prescriptions were analysed at the end of the study. There were no deviations from the intervention protocol, that is, facilities were analysed in the arms to which they were allocated.

At baseline, we found a high prevalence of irrational prescribing practices, particularly overuse of antibiotics. The percentage of prescriptions with one or more antibiotics prescribed were higher than 50%, in both intervention and control facilities. A prescription audit, feedback and small-group education intervention had no effect on the five WHO/INRUD prescribing indicator standards.

Findings showed that WHO/INRUD prescribing indicators were outside the recommended WHO standards immediately after the intervention and 6 months after the intervention. The high use of antibiotics in this LMIC is of concern as this could promote antimicrobial resistance (AMR). This high use of antibiotics in Eswatini could be because of the high burden of infectious diseases such as diarrhoea and sexually transmitted diseases. In this setting, the availability of microbiology testing services is poor, and this may result in the prescribing of antibiotics even for conditions that do not require them. Interventions targeted at increasing microbiology testing and human resources for this testing need to be prioritised to minimise the management of non-bacterial conditions with antibiotics.

With growing global evidence on the spread of multidrug-resistant bacteria, the country could soon run out of treatment options for bacterial infections. O’Neill²⁶ states that if proactive solutions to curb the spread of AMR are not provided, an estimated 10 million lives a year globally are at risk of suffering drug-resistant infections.^1,26 Apart from irrational antibiotic prescribing, the high number of medicines per prescription shows a high prevalence of polypharmacy in the country, which may be because of increasing co-morbidities in Eswatini. This study’s findings showed irrational prescribing as generic prescribing, prescribing of medicines from the EML and prescribing of injections were also out of recommended WHO standards. Therefore, there is a need for Eswatini to strengthen policies and interventions that promote rational use of medicines.

A small reduction, though not statistically significant, in the use of antibiotics in the intervention arm was observed at the end of the follow-up period, showing that, though not evident immediately post-intervention, the intervention might be beneficial in reducing the use of antibiotics. More research over a longer period needs to investigate whether this can be sustained in Eswatini. Regional pharmacists who were appointed in 2018,²⁷ after the presentation of baseline study findings to senior officials under pharmaceutical services, could incorporate the prescription audit and feedback sessions into their supervisory and mentorship visits. There is a possibility that if the intervention is sustained over a longer time, there would be an effect in improving rational prescribing. The effect of system and policy changes could be evaluated using interrupted time series. Furthermore, other interventions that have been shown to reduce antibiotic use in literature still need to be explored in Eswatini. These include the availability of policies that will advocate for the implementation of a national AMR containment strategy, availability of an entity that will provide information on medicines to healthcare professionals and the community, and availability of a functional department dedicated to promoting rational use of medicines within the Ministry of Health which will ensure functional pharmacy and therapeutics committees (PTCs) and antimicrobial stewardship (AMS) committees in health centres, hospitals and regions.²⁸ A good intervention to promote rational use of antibiotics is the establishment and use of AMS committees that can be constituted at all levels of care.^29,30

Antimicrobial stewardship committees improve the use of antibiotics through facilitating timely microbiology and antimicrobial susceptibility testing (AST) and proper patient management.³¹ Antimicrobial stewardship committees could also establish surveillance mechanisms to monitor antibiotics use to generate information that can be used to produce facility antibiograms that can influence antibiotics use in specific facilities; in turn, national AMR data could be used to influence policies and treatment guidelines.³² Studies to further assess antibiotics that Eswatini uses according to the WHO access, watch reserve (AWaRe) classification are needed.

Although the study combined an educational and managerial intervention to influence prescribing in line with recommendations from literature to use multifaceted interventions, the intervention had no effect on prescribing practices. This has also been noted in other settings, where irrational prescribing of antibiotics did not go down significantly after intervention.^33,34,35 In this study, this could have been because of several factors. Prescribing behaviour is a complex behaviour that is influenced by both intrinsic factors such as education, attitudes, experience, complacency and sociodemographic factors and extrinsic factors such as patient-related factors, health system and costs.³⁶ A qualitative study conducted in Eswatini showed that system (stockouts of essential medicines), patient (pressure from patients to be prescribed many medicines) and provider (prescribers prescribing as many medicines as possible to reduce visits to the facility and reduce patient volumes) factors impact prescribing practices.²⁵ Therefore, it may require more time and more intensive interventions to change prescribing behaviours. Further, the feedback meeting held with all facilities at the end of the baseline survey provided both control and intervention personnel with educational aspects of the intervention and may have influenced behaviour in both arms, as can be seen in the reduction in the percentages of prescriptions with antibiotics and injections in both intervention and control facilities at post-intervention and post-follow-up time points. Lastly, prescribers were aware of future assessments across both arms and could have changed their prescribing behaviour in anticipation.

Although we observed no effect of the intervention in this study, this study has shown that a small-group education intervention coupled with prescription audit and feedback is feasible in a LMIC. Future studies need to assess the use of other interventions. A limitation of the WHO/INRUD prescribing indicators is that they may not be sensitive in detecting changes in particular LMIC settings, as was shown in a previous study in Namibia.³⁷ The Namibian study tested the sensitivity of each of the WHO/INRUD indicators and found that sensitivity rates of these indicators ranged from 11.6% to 72.3%.³⁷ Other studies may be required to assess the applicability of the WHO/INRUD prescribing standards in Eswatini. Also, additional indicators, such as prescribers’ compliance to the standard treatment guidelines (STGs), to guide and assess the quality of future prescribing based on the WHO definition of rational medicine use are needed in the country. There may be a need for Eswatini to develop thresholds for the WHO/INRUD prescribing indicators so that future studies may compare against these. This study may provide initial data thresholds for these indicators. If the assumption would be that after the intervention prescribers were prescribing rationally, data from this study could be used as a basis for the development of thresholds that regional pharmacists can use during their mentorship and supportive supervision visits. However, more studies are required.

Another limitation of our study is that we did not assess the long-term durability of the intervention.

Lastly, staff rotations may have affected the effect of the intervention, and the introduction of an electronic system, CMIS, may have affected the performance of the intervention. The electronic system is preloaded with medicines both in trade and generic names. Also, prescribers can choose any treatment for any diagnosis, that is, after selecting the diagnosis, CMIS does not limit the medicines that one can select to manage that condition. This allows prescribers to select antibiotics to manage NCDs. Assessments on whether the information on CMIS is aligned to information in the Standard Treatment Guidelines and EML of Common Medical Conditions in the Kingdom of Eswatini need to be conducted.

In a lower-middle-income setting context with a high prevalence of irrational prescribing practices, a prescription audit, feedback and small-group education intervention did not have benefits in improving rational antibiotic prescribing. System changes such as the introduction of an electronic prescribing and dispensing system, rotation of staff members between intervention and control facilities and the appointment of regional pharmacists during the study could have influenced the performance of the intervention. The implementation of the intervention by the researcher and not as a service integrated into facility processes could have also affected the performance of the intervention.²⁵