Anti-malarial prescription practices among children admitted to six public hospitals in Uganda from 2011 to 2013

Malaria is reported by the Ministry of Health’s Health Management Information System to be the leading cause of morbidity and mortality in Uganda, accounting for 30–50 % of outpatient visits, 35 % of hospital admissions, and 9–14 % of hospital deaths [1]. More recent estimates from the World Health Organization (WHO) listed Uganda as having one of the highest malaria burdens in the world with an estimated 10,338,093 probable and confirmed cases, 592,264 inpatient malaria cases, and 6585 inpatient deaths in 2012 [2]. For children admitted to the hospital with malaria, case management including early diagnosis and prompt, effective treatment is a key strategy for minimizing deaths [3].

In 2010, WHO introduced two major changes to treatment guidelines for malaria case management. First was the replacement of presumptive treatment with parasitological diagnosis whenever possible, and second was the replacement of intravenous quinine with intravenous artesunate as first-line treatment for severe malaria [4]. The Uganda Ministry of Health adopted both changes in 2011 [5]. The new emphasis on parasitological diagnosis does not distinguish between patients with uncomplicated and severe malaria. However, due to the life-threatening nature of severe malaria, the WHO emphasizes treatment should not be delayed in the interest of obtaining a test result [6]. Parasitological confirmation of malaria is essential for targeting anti-malarial treatment to appropriate patients [7, 8]. Advantages of targeted treatment include minimizing unnecessary use of drugs, avoiding selection of drug resistance, and reducing exposure of patients to adverse events associated with unnecessary anti-malarial drug treatments [9‐11]. In the case of severe malaria, parasitological confirmation of disease minimizes misclassification of other severe febrile illnesses as malaria [6].

Despite changes in policy recommendations and reported benefits, reports indicate that anti-malarial therapy continues to be prescribed to children even following a negative malaria test result [12]. Utilizing data collected from an enhanced inpatient malaria surveillance system from July 2011 through December 2013, presented are results on anti-malarial prescription practices among children with test results admitted to six government hospitals in Uganda from areas of varying malaria endemicity. To improve the understanding of prescription practices, presented to are factors associated with the prescription of anti-malarials to children with negative malaria test results.

Demonstration of the presence of malaria parasites prior to treatment with appropriate anti-malarial medicine has recently been adopted as a global standard of care for malaria case management [15]. Proper diagnosis and treatment of malaria is especially crucial in Ugandan hospitals as a majority of admitted children have fever as well as other symptoms and signs suggestive of complicated malaria. Presented is data on malaria prescription practices from a sentinel site programme, where malaria diagnostic capacity is supported, clinicians receive malaria case management training, and the testing of each admitted child for malaria is encouraged and subject to external quality assurance. Presented data are notable for high rates of malaria diagnostic testing, a recent shift from the prescription of intravenous quinine to intravenous artesunate for malaria, and frequent prescription of anti-malarials to children with a negative malaria test result. The analysis also reveals factors associated with the prescription of anti-malarials to children with only negative malaria test results, findings which will be critical to designing programmes to improve provider compliance with malaria case management guidelines.

Through sustained supervision, adopting a goal of universal testing, and buffering of the supply chain for essential materials required for malaria diagnostic testing, testing rates over 95 % among children admitted to the six public hospitals were achieved. Among children with positive tests results, intravenous quinine was the most commonly prescribed anti-malarial. At the start of the programme, intravenous artesunate, recommended by WHO and the Ministry of Health as first line treatment for complicated malaria in children, was rarely prescribed. However, in the latter half of 2013, a dramatic increase in prescription rates of intravenous artesunate was noted across all sites, though this was not sustained at some sites in subsequent months. The rise in prescription rates of intravenous artesunate corresponded to the actual implementation of the policy change adopted 2 years earlier in Uganda. The rapid decline in prescription rates observed at two sites following the initial surge was attributed to failed reacquisition of intravenous artesunate. These difficulties highlight the challenges of implementing and sustaining new policies in resource-limited settlings like Uganda.

Among children with only negative malaria test results, data showed high rates of prescription of anti-malarials, results comparable to those reported in a similar setting in Tanzania prior to the global recommendation for universal diagnostic testing for malaria [12]. Despite training efforts, only one hospital, located in the lowest transmission setting, demonstrated a decline in this practice. A case for the presumptive treatment in endemic areas in the absence of laboratory confirmation has been argued (especially in the setting of unreliable laboratory diagnosis) [16‐20], however this argument becomes less acceptable in low transmission settings where the positive predictive value of clinical diagnosis is low [8]. Furthermore, the practice of withholding anti-malarials from children after a negative malaria test has been shown to be safe and effective in a variety of transmission settings, including hospitalized children [20‐23]. Indeed, in our study over 96 % of patients were tested at admission and only 1 % of those who tested negative at admission went on to have a positive test. Additionally, prescription of anti-malarials to hospitalized children with negative test results has been associated with increased risk of mortality, a fact attributed to mismanagement of the true underlying (non-malarial) cause of illness [24‐26]. Contrary to expectations, these results show that prescription rates to children with negative test results at the six hospitals did not correlate with reported endemicity levels at respective sites, implying that other factors, such as qualifications, beliefs, and attitudes of health workers were likely in play [18]. Regardless of variation in practice, and for as long as laboratory test results are credible, as was the case in the settings where this study was conducted, prescribing anti-malarials to children with negative test results is wasteful, puts patients at risk of harmful effects of medicines, and may promote drug resistance [12, 16, 27, 28]. Furthermore, amidst procurement and supply challenges, sustaining supply of intravenous artesunate cannot be guaranteed amidst the indiscriminate prescription of anti-malarials [24, 25, 29, 30]. Finally, diagnosing and treating children with negative malaria test results can erode health worker confidence in otherwise effective medicines, miss other serious causes of febrile illness, and contribute to poor outcomes, including a possible increase in mortality [12, 19, 26, 31]. It should be noted that while the rate of anti-malarial prescription to children with a negative malaria test was high at the hospitals involved in this study, it does compare favorably to rates published at other Ugandan hospitals [32, 33], likely reflecting the influence of UMSP training efforts.

Key to lowering the rate of prescription of anti-malarials to children with negative test results will be a better understanding of factors associated with this case management decision, and the results suggest possible reasons for this practice. With limited clinical skills and diagnostic capacity to evaluate patients, clinicians are unable to confirm alternate diagnoses and often disregard negative test results. Presented data support this hypothesis—children presenting with signs and symptoms similar to those of complicated malaria, such as pallor, and histories of fever, convulsions, and passing tea coloured urine were more likely to be prescribed anti-malarials after a negative test than children who did not have these findings. Additional evidence in support of this hypothesis is that the magnitude of association between study sites and anti-malarial prescription among children with negative test results doubled upon regression analysis limited to children with clinical findings suggestive of complicated malaria, as compared to analysis limited to those without clinical findings suggestive of complicated malaria. In addition, qualitative research work has shown that health worker beliefs towards perceived benefits of anti-malarial treatment are often overrated superseding rational clinical judgment. This behaviour is likely influenced by promotion of malaria as an important disease [34], peer pressure and patient demand for anti-malarial drugs which are often available at most health facilities [31, 35, 36], encouraging clinicians to treat malaria even in situations where malaria is unlikely [37]. In as much as prescription of anti-malarial to children with negative results is irrational, circumstances may exist when the practice is justified. For example, it may be reasonable to prescribe anti-malarials to someone who tested negative but was admitted after therapy had already been initiated at another health facility. Additionally, long turnaround times for test results warrant starting therapy before the test results are available. Importantly, the findings demonstrate the importance of repeating a negative test result when in doubt, as even in the study setting a small proportion of children who were initially negative became positive upon repeat testing.

This study had some limitations. First, it is possible that some treatments documented on the MRF were never administered and others were administered without documentation. Additionally, study data came from a malaria sentinel surveillance site programme and may not reflect what is happening at other public hospitals in Uganda. Furthermore, classification of those with malaria was limited to diagnostic testing results and it was not possible to ascertain where patients fell on the spectrum of disease from asymptomatic parasitaemia to complicated malaria. Lastly, data on treatment prior to hospitalization was not captured, which may have influenced some clinicians to prescribe anti-malarials even when test results were negative.

Demonstrated is that the practice of irrational prescription of anti-malarials remained high at six hospitals serving as malaria sentinel sites. These findings are of concern considering that Uganda changed policy to laboratory based confirmation, and that testing rates were high in the hospitals where the study was done, implying that the situation could be more serious in government hospitals where opportunities for testing, training and supervision are not comparable. Given this, it will continue to be extremely challenging to implement the WHO test and treat policy in Uganda and similar settings [37]. Sustained interventions targeted towards strengthening diagnostic and treatment capacity for malaria and non-malaria febrile disease, and changing health worker beliefs and attitudes towards malaria diagnosis are required. Behaviour change can be achieved through implementation of standard training programmes integrated into vigorous evaluation, supervision, and mentorship systems.