Understanding and improving malaria diagnosis in health facilities in Dar es Salaam, Tanzania

In Tanzania, as in most settings of sub-Saharan Africa, malaria is the first reported cause of
attendance in health facilities. The National Bureau of Statistics estimates that a total of 16 million
cases and 100,000 deaths (mainly in children) are due to malaria each year. In Dar es Salaam, the
main city, approximately 3 million attendances are recorded, of which about one third are due to
fever, mostly considered as presumptive malaria. Recent data show that transmission intensity is
much lower in urban settings than in rural lowland areas. This is especially true for Dar es Salaam
where only a small fraction of all fever episodes in children and adults are actually associated with
Plasmodium parasitaemia.
Clinical presentation of malaria is largely unspecific. No reliable clinical predictor that allows including
or excluding the diagnosis of malaria has been identified. In this context, and in the absence of
diagnostic test, WHO recommended in the past all fever episodes to be treated with antimalarials.
Such blanket treatment leads first to substantial over-treatment with malaria drugs (in Dar es Salaam
up to 95% of all treatments are unnecessary) and second to increased risk of missing alternative
diagnoses with potentially fatal outcome. To address this issue of high public health relevance, we
undertook a project called IMALDIA (Improving Malaria Diagnosis) aimed at improving the
management of febrile patients in health facilities in Dar es Salaam, mainly through the
implementation of Rapid Diagnostic Tests for malaria (mRDT). The project had 3 major components:
(1) Evaluating the safety of withholding antimalarials in febrile children with a negative mRDT living in
a moderate and a highly endemic area
(2) Introducing laboratory diagnosis for malaria in the routine management of fever cases, using
mRDT. The focus of this operational research was to document how feasible and effective the
introduction of these tests is in the context of the routine management of fever cases.
(3) Understanding the aetiologies of fever cases in children by screening a group of 1000 children
with detailed clinical assessments and a range of laboratory tests in order to better identify the
diversity of the causes of fever in small children living in an urban and a rural area.
The overall aim of the IMALDIA project was to improve the diagnostic approach and management of
fever cases in health facilities in Dar es Salaam, contribute to a more efficient and effective health
sector, and help Tanzania on its way to reducing infant and child mortality.In a first step, we
assessed the diagnostic performance of mRDT when used by health workers in routine practice. For
this purpose, a quality assurance system both at central and peripheral level was set up. This system
did not detect major problem and showed that the final result of mRDT by health workers was
reliable.
Summary
X
The purpose of the second step was to better estimate the pre-test probability of malaria in
populations targeted by mRDT (febrile patients of all age groups attending a health facility of any
type). To this end we undertook a systematic review of the studies giving the proportion of patients
with associated P. falciparum parasitemia (PFPf) in Sub-Saharan Africa. We found that the median
PFPf was 35%, and that it had decreased by half when comparing the period before with the period
after the year 2000 (44% versus 22%). This relatively low pre-test probability nowadays is another
reason to implement mRDT in Africa. In Dar es Salaam the PFPf was very low (below 10%) hence it
was even more urgent to start using a reliable malaria test. Microscopy was available in almost all
public health facilities of the city but its performance was extremely low, with an overall sensitivity of
71% and a specificity of only 47%.
On the request of several Tanzanian stake-holders, in particular clinicians working routinely with
patients, we assessed the safety of withholding antimalarials in children under five years with a
negative malaria test. We did not observe any complication or death due to a missed diagnosis of
malaria in our cohort of 1000 children, of which 60% were negative by mRDT. We concluded that the
strategy of withholding antimalarials in negative children is safe and does not expose the child to an
increased risk.
The results of the systematic review coupled with the findings of the safety study led us to question
the appropriateness of the previous WHO recommendation of treating all fevers with antimalarials in
children less than five years living in highly endemic areas. WHO has now changed its policy,
confirming that the IMALDIA findings were very relevant to the changed situation of many African
countries, including Tanzania.
The core of this thesis, and the main objective of the IMALDIA project, was to investigate the
feasibility and value of implementing mRDT in the management of fever episodes in an urban
malaria setting. Using 2 different designs and 2 independent data sources, we found a three quarter
reduction in antimalarial consumption following RDT implementation. This massive reduction was
due to the higher accuracy of routine mRDT compared to routine microscopy (that led to a dramatic
reduction in the number of positive patients) and to the confidence of health workers in mRDT results
(the proportion of negative patients treated with antimalarials dropped from 53% to 7%). The impact
was maintained up to the end of the observation period (18 months). Not surprisingly, mRDT
implementation increased the prescription of antibiotics by 50% and unfortunately did not have a
major impact on the quality of the medical consultation.
We took the opportunity of our near-to-program implementation of mRDT to perform a cost-saving
analysis in a real situation and in a setting representative of many moderate endemic places in
Africa. The conclusion was that costs can be saved on drugs, from both the provider and from the
client’s perspective. For this reason, the overall expenditure for the patient was lower in health
facilities using mRDT (by 0.31 USD per patient). However, the overall expenditure for the health
Summary
XI
system was higher (by 1.31 USD per patient) when using mRDT instead of routine microscopy,
mainly because of the relatively high price of the device.
The aim of the last study was to explore the other causes of fever (beside malaria), in order to
generate evidence for a revision of the existing clinical decision-charts for the management of
patients, in particular the Integrated Management of Childhood Illness (IMCI). Half of the fever
episodes in children were due to acute respiratory infections (ARI), of which 2/3 were probably of
viral origin. Only 5% of all ARI were documented pneumonia. Gastroenteritis contributed to 9% of all
fevers, of which at least 1/3 were due to a virus. In 1/5 of the children, no aetiology of high probability
could be found but most of them recovered without treatment. Most of the children with acute fever
thus do not need to receive an antibiotic. Based on these findings, we proposed a limited series of
modifications to the IMCI chart and concluded that new point-of-care laboratory tests for the main
infectious diseases are urgently needed.
In conclusion, the IMALDIA project provided a deep insight into many aspects of the implementation
of mRDT in near-to-programme conditions in Tanzania. Our findings show that the introduction of
mRDT is safe, feasible and useful for the routine management of fever cases in all age groups and at
all levels of the health system. Implementation at large scale will require flexibility on the part of the
health care provider in order to be able to change his/her behaviour and a strong commitment of all
persons involved. As malaria diagnosis is only one aspect of the management of patients presenting
with fever, this will not solve all obstacles for making a proper differential diagnosis and prescribing
the appropriate treatment for fever episodes. To really improve the quality of care it will be essential
to develop new improved guidelines for clinicians. These decision charts should be based on the new
available evidence and could include novel point-of-care tests for the key diseases, once these
become available.