Fever, vomiting, cough and convulsions were the commonest symptoms. (Table I) Symptoms like headache and muscle aches commonly described in adults were rarely reported. Reported cases were from older children. On the other hand, diarrhoea was reported in a quarter of the children. In almost all cases, the duration of fever approximated that of the illness and this was about two and half days. Although fever, cough and vomiting were the commonest symptoms, mothers acted fast and brought the children to the unit if they had convulsions or mental confusion (Table I).

Using a modification of the WHO 1990 criteria for severe malaria by Marsh at el³, the children were categorized as either having defining or supporting criteria for severe malaria. The commonest defining manifestations were severe anaemia, respiratory distress, multiple generalized convulsions, hypoglycaemia and cerebral malaria. (Table II)

The total inpatient mortality was 3.1%. Of the 23 cases of death, 5 (21.7%) occurred within 1 hour of arrival to hospital. A total of 17 deaths (73.9%) occurred in the emergency room and the remaining soon after transfer to the main Paediatric wards but all within 24 hours. Cerebral malaria accounted for 47% of these deaths. Different degrees of impairment of consciousness, respiratory distress, multiple generalized convulsions and jaundice were significantly associated with the occurrence of death. In almost all cases, the presence of any of the manifestations of severe malaria relatively increased the resulting mortality.

Although provided for in the hospitals’ guidelines for the work up of children with severe malaria, the emergency unit does not routinely offer blood glucose level determination. The study therefore provided this service and was able to demonstrate hypoglycaemia in 11.4% of the children. Similarly, renal function is not routinely assessed nor urine input/output recorded for children with severe malaria. We were therefore unable to determine the prevalence of renal failure. However, 3 children definitely had renal problems. All 3 were anuric for more than 24 hours despite adequate rehydration and fluid challenge. They all regained normal urine outputs over the next 72 hours.

During the study period, a particular problem was noted with tests for haemoglobin. The reagent for the test was often lacking. In as many as 30% of patients, the unit was unable to perform haemoglobin tests resulting in some severely anaemic patients being transfused without determining their haemoglobin levels.

Quinine, oxygen, normal saline, 5% dextrose, 25 or 50% dextrose, blood for transfusion, blood transfusion sets, intravenous cannulae, intravenous fluid giving sets, and syringes were considered the minimum drugs and supplies needed to resuscitate children presenting with severe malaria. If any of these items was needed to resuscitate a child and was not available, it was regarded as missing. Overall, 54.8% of the children had at least one item needed for their resuscitation missing in the emergency room. The most frequently missing items were intravenous cannulae (53.1%), syringes (23.3%) and blood transfusion sets (15%).1 There were significant differences in the percentages of children who presented during day with resuscitation items missing and those who came at night. (Table III) Figure II illustrates missing essential drugs and supplies.

In almost all cases, oxygen, normal saline and 5% dextrose were available when needed. With the exception of blood, the caretakers bought the missing essential drugs/supplies from nearby pharmacies. There was no significant association between lack of any essential supply and the occurrence of death; 2.8% of the children who lacked essential resuscitation supplies died compared to 3.5% of those who did not (OR 0.79 95% CI 0.35– 1.82).

Table III: A comparison of care received during day and at night

Selected measures of quality of care and outcome of severe malaria

A child was considered to have received quality care if a doctor attended to her/him within 1 hour of arrival, received the first dose of quinine within 2 hours, all essential drugs and supplies needed for his/her resuscitation were available, and the caretaker felt the service rendered was good or very good. Over 90% of the caretakers were satisfied with the services but less than half of the children had all the essential items available.

The relationship between these measures of quality and outcome of severe malaria is shown in table IV. Cases where services were considered good or very good had a significantly low mortality; 0.9 vs 16.9%, RR 0.05 (0.02 – 0.12). The level of education did not influence the Aspects of the clinical assessment of patients by clinicians were poor. Important measurements like the blood pressure, though provided for in the management protocol, were not performed in the majority of children. Some children were transfused before determining the haemoglobin level. Renal function was not assessed in almost all children. For purposes of care and documentation, these areas need improvement.

Severe malaria develops rapidly from uncomplicated disease. The mean duration of illness at admission of only two and half days is similar to findings in previous studies.⁸ The strategy of giving children anti malarial drugs within 24 hours to avert disease progression should therefore be emphasized.⁹ It was however sad to note that severe malaria developed despite pre-admission Chloroquine in > 50% of cases! The high Chloroquine resistance described in areas around Kampala may explain this¹⁰and supports the view that Chloroquine alone may not be useful for uncomplicated malaria.

As expected in a malaria endemic area, severe anaemia was the commonest defining feature.¹¹ The percentage of those with cerebral malaria was rather low³ although the distribution of other manifestations was within expected limits.³Of particular importance was the 11.4% prevalence of hypoglycaemia since this unit does not routinely perform tests for blood sugar though it is provided for in the protocol. Similarly, blood pressure measurements were often not performed making it difficult to assess circulatory failure adequately. Children with this severe manifestation of malaria may die without any attempts to correct possible shock. It is time blood glucose assessment and blood pressure measurement in children with severe malaria is enforced in the unit.

Different degrees of impairment of consciousness, respiratory distress, multiple generalized convulsions and jaundice were significantly associated with death. The prominence of the associations between impairment of consciousness and respiratory distress on one hand and death on another confirms the high risk these two clinical features subject children to.^3,12 The most interesting results however were those that demonstrated that patients appear to receive better services at night. This was the exact opposite of our initial hypothesis. The relatively small numbers of patients at night are seen quickly and started on specific malaria treatment resulting in significant reductions in mortality. This finding strengthens the view that with more hands in the emergency room especially during day, mortality in the unit can be reduced.

The unit fails the test of attending to children within 1 hour of arrival. Less than 25% of the children were seen within 1 hour and less than half of this received the first dose of quinine in 2 hours! This is an area that needs to be addressed urgently. The time taken to receive the first dose of quinine was further delayed by another 30 minutes among those who had to go and buy resuscitation materials. A steady supply of these resuscitation materials will cut down the time to onset of specific treatment with quinine. Firm action and commitment are needed to achieve these concerns. There was a slightly higher mortality among patients who passed the quality tests of being seen within 1 hour or receiving quinine within 2 hours. The most likely reason for this was that they were very sick children with poor prognosis in who resuscitation attempts failed despite early intervention by attending physicians. Haemoglobin level determination brought to light another problem of supplies. This basic test could not be done in over 30% of those who needed it because of lack of reagents. Several severely anaemic children were transfused on the basis of clinical observation. A mechanism should be put in place to plan for and ensure a steady supply of such important yet basic reagents.

Despite the gleam picture painted by the ratings of the measures of quality set in this study, the good news was the assessment by caretakers. Over 90% were satisfied with the care given to their children rating it either as good or very good and this was not influenced by the level of education a of caretaker. There was a significant correlation between the occurrence of death and services being considered poor. It is possible that caretakers who lost children would feel that not enough was done to save their children and rate the service poor. This assumption may be valid in cases where children died within a few hours of admission and questions on caretaker satisfaction with care were asked after the death of their children.

The caretakers also offered several suggestions for quality improvement. Most centred round increased availability and steady supply of essential materials, more staff, faster services, reduction in delays, enforcement of discipline among the staff, decongesting the unit and offering explanations to patients. These should be areas the unit should address as quality improvement processes are initiated as they bear directly to what clients feel.

We thank all caretakers of our patients and the patients themselves. May the souls of the 23 who died rest in peace. We thank the doctors and nurses in the emergency unit for their tireless effort in filling the case record forms, the laboratory staff and data entry staff.

The study received financial support from Makerere University Faculty of Medicine/Sida Sarec Research Grant and the Child Health and Development Centre/TORCH Project Novice Research grant.