Malaria Control During Mass Population Movements and Natural Disasters (2003)

Effective case management is the minimum requirement for any malaria control program. Because falciparum malaria can become life threatening within just 48 hours, first-line therapy needs to be offered at the most peripheral facilities. A system of referral needs to be established so that patients requiring more aggressive therapy for severe disease can get to a central facility where appropriate care can be given, such as parenteral antimalarials and supportive care. Health care providers in central locations need to be trained and able to deliver both specific and supportive care to unconscious malaria patients.

As mentioned previously, the choice of drugs used for uncomplicated severe malaria at different levels of the health care system must be based on relevant data regarding drug resistance patterns and expected efficacy. Occasionally, sufficient information can be obtained from the host country or existing literature; however, specific drug efficacy studies conducted with standard methods in the population being served are the best way to determine the most appropriate drugs to use. Simple methods for assessing the efficacy of first-line antimalarial drugs are presented in Appendix B. Currently, standardized in vivo efficacy studies are rarely conducted during the acute phase of an emergency; however, these data are necessary in order to make decisions regarding effective malaria treatment. Often nongovernmental organizations (NGOs) are overburdened by the demands of providing essential services or lack staff members with experience in conducting such studies. The Roll Back Malaria Technical Resource Network for Ma-

laria Control in Complex Emergencies is an example of a resource for technical assistance in this area (see Appendix D for more information on resources).

A number of options exist or are in development for diagnosing malaria (see Table 6-1). The most common methods of diagnosis are either those based on clinical signs and symptoms alone or those based on microscopy. Newer technologies, especially rapid diagnostic tests, are becoming more widely used. Cost and prevailing conditions (availability of electricity, equipment, supplies, adequate training, and supervision) typically determine the method used. In Africa, for instance, clinical diagnosis of malaria is most commonly used, both in complex emergency situations and stable populations. Rapid diagnostic tests have rarely been used operationally in complex emergencies. When they have been used, it is usually in conjunction with rapid surveys rather than routine diagnosis (Kassankogno et al., 2000).

Although reliable diagnosis cannot be made on the basis of signs and symptoms alone because of the nonspecific nature of clinical malaria, clinical diagnosis of malaria is common in many malarious areas. In much of the malaria-endemic world, resources and trained health care personnel are so scarce that presumptive clinical diagnosis is the only realistic option and is a common approach in the context of complex emergencies, especially in sub-Saharan Africa. Clinical diagnosis offers the advantages of ease, speed, and low cost. In areas where malaria is prevalent, clinical diagnosis usually results in all patients with fever and no other apparent cause being treated for malaria. This approach can identify most patients who truly need antimalarial treatment but is also likely to misclassify many who do not (Olivar et al., 1991). Overdiagnosis can be considerable and contributes to misuse of antimalarial drugs, places individuals at unnecessary risk of side effects, and can contribute to the development of drug resistance. Considerable overlap exists between the signs and symptoms of malaria and other common diseases, especially acute lower respiratory tract

infections and acute viral syndromes, and can greatly increase the frequency of misdiagnosis and mistreatment (Redd et al., 1992; Rey et al., 1996).

Attempts to improve the specificity of clinical diagnosis for malaria by including signs and symptoms other than fever or history of fever have met with only minimal success (Smith et al., 1994). The Integrated Management of Childhood Illnesses (IMCI) is a strategy developed to improve diagnosis and treatment of the most common childhood illnesses in areas that must rely on relatively unskilled health care workers without access to laboratories or special equipment. With this strategy, every febrile child living in a “high-risk” area for malaria is considered to have, and is treated for, malaria (although many children will likely be treated for other illnesses as well). “High risk” is defined in the IMCI Adaptation Guides as being any situation where as little as 5 percent of febrile children between the ages of 2 and 59 months are parasitemic (World Health Organization, 1997a), a definition that will likely lead to significant overdiagnosis of malaria in areas with low-to-moderate malaria transmission. Although this strategy may improve diagnosis of childhood illnesses, it requires substantial investment in training and supervision. Its usefulness in a complex emergency will likely be contingent on the relief agency’s commitment to sustain this level of effort.

Wherever possible, laboratory-based diagnosis should be used. It offers the advantage of identifying not only patients truly in need of malaria treatment but also patients who do not and who should therefore be reassessed for another cause of their fever. By reducing unnecessary malaria treatment, overall drug pressure on the parasite, an important contributing factor in the development of drug resistance, can also be reduced.

Although laboratory-based diagnosis is typically more expensive and labor intensive than clinical diagnosis, it has been shown to be cost effective in some settings and can potentially lower actual malaria treatment costs by reducing the amounts of antimalarial drugs that are used. This is especially true in settings where inexpensive antimalarials, such as chloroquine and sulfadoxine/pyrimethamine, can no longer be used because of high levels of drug resistance.

Species diagnosis is another important advantage to laboratory-based diagnosis. In most areas of the world, nonfalciparum malaria can be reliably

treated with chloroquine alone or chloroquine and primaquine, reducing overall malaria treatment costs while ensuring appropriate treatment of falciparum malaria.

Simple light microscopic examination of (preferably) Giemsa-stained blood films is the most widely practiced and useful method for definitive malaria diagnosis. Advantages include differentiation between species, quantification of parasite density, and the ability to distinguish clinically important asexual parasite stages from gametocytes, which may persist without causing symptoms. These advantages can also be critical for proper case management. For example, the ability to quantify parasite density offers a method to evaluate parasitological response to treatment, which is needed for early identification of treatment failure. The disadvantages are

that slide collection, staining, and reading can be time consuming, and microscopists need to be trained and supervised to ensure consistent reliability. While the availability of microscopic diagnosis has been shown to reduce drug use in some trial settings (Jonkman et al., 1995), in practice the results are often unreliable or disregarded by clinicians (Barat et al., 1999). Any program aimed at improving the availability of reliable microscopy should also retrain clinicians in the use and interpretation of microscopic diagnosis.

A second method is a modification of light microscopy called the quantitative buffy coat method (QBC, by Becton-Dickenson). Originally developed to screen large numbers of specimens for complete blood cell counts, this method has been adapted for malaria diagnosis (Levine et al., 1989). The technique uses microhematocrit tubes precoated with flourescent acridine orange stain to highlight malaria parasites. With centrifugation, parasites are concentrated at a predictable location. Advantages to QBC

are that less training is required to operate the system than for reading Giemsa-stained blood films, and the test is typically quicker to perform than normal light microscopy. Field trials have shown that the QBC system may be marginally more sensitive than conventional microscopy under ideal conditions (Rickman et al., 1989; Tharavanij, 1990). Disadvantages are that electricity is always required, special equipment and supplies are needed, the per-test cost is higher than simple light microscopy, and species-specific diagnosis is not reliable. This methodology is currently used infrequently and supplies are increasingly difficult to locate.

A third diagnostic approach involves the rapid detection of parasite antigens using rapid immunochromatographic techniques (also known as “dipstick” tests). Multiple experimental tests have been developed targeting a variety of parasite antigens (Mackey et al., 1982; Fortier et al., 1987; Khusmith et al., 1987). A number of commercially available kits are based on the detection of the histidine-rich protein 2 (HRP-II) of P. falciparum. Compared with light microscopy and QBC, this test yields rapid and highly sensitive diagnosis of P. falciparum infection (World Health Organization, 1996b; Craig and Sharp, 1997). Advantages to this technology are that no special equipment is required, only minimal training is necessary and no electricity is needed. The principal disadvantages are the currently high per-test cost and an inability to quantify the density of infection. Furthermore, for tests based on HRP-II, detectable antigen can persist for days after adequate treatment and cure; therefore, the test cannot adequately distinguish a resolving infection from treatment failure due to drug resistance, especially early after treatment (World Health Organization, 1996b). Additionally, a test based on detection of a specific parasite enzyme (lactate dehydrogenase or pLDH) has been developed (OptiMAL, by Flow, Inc., Portland, Oregon). It reportedly detects only viable parasites, which, if true, eliminates prolonged periods of false positivity posttreatment (Makler et al., 1998; Piper et al., 1999; Palmer et al., 1999). Newer-generation antigen detection tests are able to distinguish between falciparum and nonfalciparum infections, greatly expanding their usefulness in areas where nonfalciparum malaria is transmitted frequently. Finally, there is growing concern about the stability, shelf life, and reliability of these tests under field conditions. Exposure to heat and humidity have been raised as possible factors that could compromise test function (L. Causer, Centers for Disease

Control and Prevention, personal communication, 2002). For a more complete review of rapid diagnostic tests for malaria, see Moody (2002).

Detection of parasite genetic material or resistance-conferring mutations through polymerase chain reaction techniques is becoming a more frequently used tool in the diagnosis of malaria, as well as the diagnosis and surveillance of drug resistance in malaria (Plowe et al., 1995). At present, however, this method of diagnosis is well beyond the capacity of most NGOs or relief agencies. Since sample collection is exceedingly easy (a dried blood spot on filter paper), collaboration with an outside agency or ministry of health that does have this capacity would allow its use for monitoring patterns of resistance to chloroquine or sulfadoxine/pyrimethamine in an emergency setting, particularly in areas where in vivo studies would be difficult to conduct. This approach is being used to supplement in vivo drug efficacy testing in the Democratic Republic of Congo (P. Bloland, Centers for Disease Control and Prevention, unpublished data, 2001). This methodology has limitations, however. Because samples must be analyzed in specialized laboratories, it may take months to obtain results. Mutations related to resistance have been identified only for chloroquine, sulfadoxine/ pyrimethamine, and atovaquone. Finally, although the relationship between the presence of resistance-conferring mutations and clinical or parasitological response to treatment has been investigated (Djimdé et al., 2001), it probably differs from location to location depending on underlying levels of immunity in the population and the specific drug being tested.

The method to be used to diagnose malaria is an important issue for consideration when setting up malaria control programs. Not all situations will allow for definitive diagnosis, such as facilities seeing very high volumes of patients, community outreach programs, or during epidemics. At a minimum, some form of laboratory-based diagnosis (most likely microscopy) should be available in all situations as an aid in the diagnosis and management of severe malaria. However, the decision to use laboratory tests (as well as the decision of which diagnostic tests to use) for routine diagnosis of all suspected malaria cases needs to be based on:

• patient load;

• time required to do the test;

• cost of the test;

• cost savings that might be expected from avoiding unnecessary malaria treatment;

• level of training and supervision of laboratory and clinical staff and number of trained staff members available; and

• availability of electricity, appropriate equipment, and supplies.

In typical relief efforts in Africa, patient load alone makes the use of laboratory-based diagnosis difficult. In western Tanzania, for instance, although microscopy was available, it was used only for diagnosis of malaria among inpatients, staff, or particularly confusing cases (e.g., to help decide between malaria and an acute lower respiratory tract infection; H. Williams and P. Bloland, Centers for Disease Control and Prevention, unpublished data, 1998). Because the microscopes relied on sunlight, no microscopy was available on particularly cloudy days or during evening hours. The majority of patients were diagnosed using clinical impressions alone. While rapid diagnostic tests could have been used in this setting, the current cost of these tests would have been prohibitive for the budgets of most NGOs.

Prompt provision of effective therapy that is capable of preventing the progression of illness from mild to severe has been estimated to provide as much as a 50-fold reduction in the risk of mortality. In contrast, effective treatment of malaria once it has progressed to severe illness has only a five-fold reduction in the probability of dying (White, 1999). Mortality among patients being treated for severe or cerebral malaria in a hospital is 10 to 40 percent, even when the best treatment is given (Greenberg et al., 1989; Molyneux et al., 1989). Progression of death can be rapid; the mean time between onset of illness and death among Gambian children was 2.8 days (Greenwood et al., 1987).

The choice of an optimal treatment regimen for uncomplicated malaria depends on local drug resistance patterns, acceptability (in terms of safety, side effects, ability to use during pregnancy, ability to treat young children), and cost. Other considerations in choosing a treatment regimen for a given situation include the need to train or retrain staff, ease of administration,

probable adherence to recommended dosing regimen, and an assured supply of drugs.

Combinations of artemisinin derivatives (such as 3 days of artesunate and another common antimalarial drug, such as mefloquine, sulfadoxine/ pyrimethamine, or amodiaquine) are currently being evaluated as options for treating uncomplicated P. falciparum malaria. Artemisinin (ART) derivatives produce rapid killing of parasites, greatly reducing the parasite load. Because ART derivatives must be used for 7 days when used alone, combination with a second drug allows fewer days of treatment (3 days) while maintaining high efficacy, thus improving compliance. In specific areas of low transmission in Southeast Asia, ART combination therapy appears to have slowed the development of drug resistance and possibly decreased malaria transmission (Price et al., 1997; Nosten et al., 2000).

Although the combination of ART and mefloquine has been used extensively and successfully in refugee camps in Thailand (as well as in stable populations in parts of Southeast Asia), there has been very limited experience with the programmatic use (for either displaced or stable populations) of this or other ART combinations in Africa or the Americas, although preliminary studies and evaluations are being conducted (Doherty et al., 1999; von Seidlein et al., 2000; Bloland et al., 2000; Kachur et al., 2001).

Two often-cited reasons to use an ART combination treatment strategy—inhibition of development of resistance and reduction in malaria transmission rates—are unproven in areas of moderate-to-intense malaria transmission, such as is found in most of tropical Africa (Price et al., 1996; White et al., 1999; Bloland et al., 2000). Nonetheless, ART in combination with a second drug appropriate to the region would probably offer optimum efficacy. If these additional properties of artemisinin-containing combination therapy (ACT) are proven, a strong argument could be made to preferentially use such combinations to protect antimalarial drugs and to reduce malaria transmission, an especially desirable effect during an epidemic. Presently, however, the decision to utilize ACT, especially in areas of moderate-to-high transmission, should be based on comparison between ACT and other available malaria drugs in terms of efficacy, availability, cost, effectiveness, and other operational factors and not on unproven characteristics. Other available therapeutic options are presented in Appendix A and Table 3-3. Not all of the options presented would be appropriate, however, for any given setting or situation. Additionally, some drug choices that would appear optimal in the context of a complex emergency may

conflict with the national malaria treatment policy of the host country. Many national malaria treatment policies lag behind current status of drug resistance, especially in Africa. Nonetheless, ministries of health may require evidence that the prevailing national treatment policy is inadequate to treat refugee populations before allowing a deviation from national policy. For example, prior to 1999 in refugee camps in western Tanzania, the first-line treatment for uncomplicated malaria was chloroquine in accordance with the national treatment guidelines. Faced with increasing reports of chloroquine failures, standardized in vivo studies were conducted in the camps and surrounding communities. Based on these data, a decision was made to allow sulfadoxine/pyrimethamine to be used for first-line therapy in the camps, even though chloroquine continued to be recommended for the national population (H. Williams, Centers for Disease Control and Prevention, unpublished data, 1999).

As mentioned, drug resistance patterns can be highly variable within relatively small geographic areas, and predictions of likely therapeutic efficacy based on the geographic location of the area in which a displaced population settles may prove to be inaccurate. Although some generalizations can be made (see Table 3-4), more accurate information should be obtained through operational research or from reliable local or international sources.

Chloroquine remains a highly effective treatment for nonfalciparum infections in most areas of the world. In areas where nonfalciparum malaria infections are relatively common, distinguishing nonfalciparum infections from falciparum infections could result in cost savings, especially if treatment of falciparum infections is expensive. Primaquine reduces the probability of relapse due to P. vivax and P. ovale; in situations where rapid reinfection with P. vivax or P. ovale or poor compliance with a 14-day regimen is likely, however, use of primaquine as a routine antirelapse treatment is often not a priority. Nonetheless, data from areas of relatively low transmission suggest that routine use of 14 days of primaquine (but not less) could potentially prevent a substantial amount of febrile illness and anemia associated with preventable relapses of P. vivax (Luxemberger et al., 1999; Rowland and Durrani, 1999).

Chloroquine-resistant P. vivax has been reported to occur at low levels

in some regions. In some locations it can occur frequently enough to raise questions about the wisdom of continued use of the drug, such as Irian Jaya, Indonesia, and Papua New Guinea (Murphy et al., 1993). In areas where chloroquine-resistant vivax is common, such infections should be treated as for chloroquine-resistant falciparum. However, some drugs that can be used for P. falciparum are not optimally effective for P. vivax, especially sulfadoxine/pyrimethamine (Pukrittayakamee et al., 2000a).

Some geographic areas have P. vivax parasites with a reduced sensitivity to primaquine. This is particularly true of vivax infections from Southeast Asia and Oceania (Collins and Jeffery, 1996). These infections can sometimes be successfully treated using 30 mg daily for 14 days, rather than the standard 15-mg daily regimen adult dose. A new drug related to primaquine (tafenoquine; see Table 3-3) is being developed that offers increased efficacy and may replace primaquine in the treatment of P. vivax (Walsh et al., 1999).

Primaquine can cause severe or fatal hemolysis in patients with G6PD deficiencies. Populations with certain ethnic backgrounds are at increased risk of severe G6PD deficiency and therefore increased risk of hemolysis. Severe deficiency (<10 percent residual enzyme activity) is seen in some people of Mediterranean and Asian decent; moderate deficiency (10 to 60 percent residual activity) is common among populations of African descent. Ideally, all patients needing primaquine treatment should be tested for G6PD deficiency before treatment begins, but in a complex emergency this rarely occurs. Patients with severe deficiency (and any pregnant women) should not receive primaquine; most patients with moderate deficiency (African A variant) can be treated with a larger dose of primaquine

less often (45 mg primaquine base orally once a week for 8 weeks for adults) without producing life-threatening hemolysis.

Treatment of severe P. falciparum malaria should be with immediate administration of parental quinine or quinidine, beginning with a loading dose and followed by 8 to 12 hourly maintenance doses (World Health Organization, 2000a, 2000b). This regimen should be continued until the patient can take oral medications. Quality nursing care and close clinical monitoring are essential, especially for comatose patients.

Quinine remains the standard treatment for severe malaria illness because of its ready availability, efficacy, and rapid action. While artemisinins have been used for the treatment of severe malaria, there is no clear indication that they are superior to quinine in terms of patient survival (World Health Organization, 2000a). However, in areas where quinine resistance is known to occur, artemisinins offer obvious advantages.

Quinine use is often restricted to facilities that have the capacity to carefully monitor patients for potential cardiac side effects from the quinine therapy and to properly administer and monitor intravenous fluids and medicines. Another consideration, therefore, that might favor the use of artemisinins in the context of complex emergencies is a lack of capacity of facilities to administer parenteral quinine safely and effectively. Patients with severe falciparum malaria being managed at facilities with minimal capacity could potentially be treated with an artemisinin suppository before being transferred to a better-equipped and better-staffed facility for complete management (Krishna et al., 2001). Alternatively, if a facility has the capacity to provide injections but not intravenous therapy, patients could be treated with either artemether or artesunate intramuscularly. Dosing with artemisinin or artesunate intramuscularly can be done on a once-daily basis, greatly increasing the ease of treatment. (See Appendix C for recommendations for possible alternative regimens.) Of note, some large studies of artemisinin use for treating patients with severe falciparum malaria have noted delayed resolution of coma; the reasons and implications of this delayed recovery are unknown (World Health Organization, 2000a).

Severely ill malaria patients, especially those who are comatose, require

careful and attentive nursing to reduce case fatality rates. This care should be available 24 hours a day, 7 days a week. Some general considerations for nursing care are as follows (World Health Organization, 2000a):

• Comatose patients are at risk of aspiration of fluids and must be kept on their sides.

• Patients should be turned frequently to avoid bed sores.

• Hyperthermia (rectal temperature >39ºC) should be identified quickly and managed with antipyretics, tepid sponging, and fanning.

Careful monitoring of patients is needed, including their mental status, fluid administration rates, fluid intake and output, glucose, temperature, pulse, respiration, parasite density, and blood pressure.

In populations with little or no immunity, malaria can be a very serious infection, with a high risk of severe maternal morbidity or mortality and fetal loss (Nosten et al., 1991; World Health Organization, 2000a). In populations with high levels of acquired immunity, the primary complications of malaria during pregnancy are maternal anemia and low-birthweight babies, with increased risk primarily seen among women with first or second pregnancies (McGregor, 1984). In either setting, the impact of malaria during pregnancy can be great. Drug resistance, however, has greatly complicated the provision of safe and effective malaria treatment during pregnancy.

The choice of an appropriate first-line treatment for malaria in pregnant women should be based on local drug resistance patterns and the underlying level of immunity of the population. Treatment of otherwise uncomplicated malaria during pregnancy generally follows the recommendations for treatment of uncomplicated malaria in nonpregnant individuals. There are, however, a few drugs that are contraindicated for use during pregnancy and should be avoided (e.g., primaquine, tetracycline, and related drugs).

In many settings, quinine is often the drug of choice for treating symptomatic malaria during pregnancy. While artemisinins have been used successfully and safely during pregnancy (primarily in combination with mefloquine) and may in fact be preferable to quinine (because of a combination of efficacy, ease of administration, and low incidence of side effects,

especially hypoglycemia) for the treatment of malaria during pregnancy, their safety has not been fully established (McGready et al., 1998). This is especially so for use during the first trimester of pregnancy. Until more experience has been gained with the use of artemisinins (alone or in combination with other drugs), a definitive statement on their safety in pregnancy cannot be made. Nonetheless, artemisinin derivatives, alone or in combination with another safe drug, offer one of the few highly effective treatment options during pregnancy in areas where multidrug-resistant malaria occurs (McGready et al., 1998).

Anemia is a frequent and clinically important complication of malaria, as it can significantly add to morbidity and mortality in displaced populations. Nevertheless, it is frequently overlooked, especially during complex emergencies. Implementing a systematic assessment for anemia (see Table 6-2), as part of an initial patient evaluation, will likely identify many more patients needing treatment for anemia than would otherwise be identified. Preparations for this increase in patient load associated with anemia

should be made in advance. Given the staff and equipment limitations in most emergency situations, many of the options listed in Table 6-2 are not being implemented. Training staff to include pallor as part of their initial examination is an easy first step in addressing this complication. However, it is also important to remember that anemia can be the result of a wide range of infectious and noninfectious causes and that often more than one cause can be identified in a given patient or population.

The single most important component of the management of anemia in areas where malaria poses a significant risk is effective antimalarial treatment. In these areas all anemic patients would likely benefit from effective treatment for malaria, regardless of blood smear status, in addition to whatever anthelminthic, micronutrient, or other ancillary treatment might be appropriate to the setting.

A blood transfusion can be a life-saving intervention, but it can also be unnecessary or even harmful, especially if screening for bloodborne

pathogens is not reliable or universally applied (Lackritz, 1998; Obonyo et al., 1998; Kinde-Gazard et al., 2000). Studies have shown that the survival benefit of transfusion is greatest for severely anemic patients with signs of respiratory distress (Lackritz et al., 1992, 1997). This respiratory distress is due to an underlying severe metabolic acidosis (English et al., 1997; World Health Organization, 2000a; Day et al., 2000). Furthermore, severely anemic patients with or without respiratory distress had improved survival only when transfusions were given within the first 24 to 48 hours of admission. Additionally, in one study the presence of malaria parasitemia after transfusion negated the hematological benefits of transfusion, suggesting the need for providing severely anemic children with some form of malaria prevention in addition to blood transfusion (Lackritz et al., 1997). Development of guidelines (that are appropriate to the local situation) for the proper use of blood transfusions and training in the recognition of respira-

tory distress need to be emphasized to ensure optimal management of severe malarial anemia (World Health Organization, 2000a; English et al., 1997).

Nutritional supplementation, especially when combined with effective malaria treatment, can greatly enhance hematological recovery among anemic patients. A strategy worth further investigation is adding post-hospitalization malaria prevention (via drugs or insecticide-treated bed nets) in combination with micronutrient supplementation as a way to further improve hematological status (see also Preventive Use of Antimalarial Drugs, Chapter 7). Some data, however, suggest that concurrent folate supplementation may increase treatment failures of sulfadoxine/pyrimethamine for malaria treatment (van Hensbroek et al., 1995; Shankar, 2000; Bayley and Macreadie, 2002). The appropriate response to these observations is unclear and is being studied. Until more information is available, it would probably be best to withhold folate supplementation, at least temporarily, after treatment with an antifol antimalarial in order to gain maximum benefit from the malaria treatment.

• Identify the differences in levels of transmission and immunity for both the environment from which the displaced population comes and the environment in which it settles to help guide the development of curative and preventive strategies.

• Involve the host government (if functioning) immediately and constructively in discussions about the health needs of the displaced population. Unilateral decisions on the part of relief organizations are unlikely to be welcomed by the host government. The host government may well have pertinent information, advice, and expertise that could be brought to the situation. Discussions should include the consideration of treatment policies, local drug resistance patterns, acceptability, availability, cost, and other operational factors when establishing malaria treatment guidelines.

• Obtain specific local information regarding current drug resistance patterns from standardized drug efficacy trials, the host country’s data, and other sources of technical information (such as malaria country profiles from Roll Back Malaria in order to choose the optimal first-line treatment guidelines.

• Use laboratory-based (microscopy) diagnosis whenever feasible.

• Supplement clinical diagnosis (if used) with rapid blood smear surveys to evaluate the reliability of clinical diagnosis on a regular basis.

• Train clinical staff in the recognition and proper treatment of severe malarial disease and important malaria-related complications, such as anemia.

• Establish a referral system to facilitate the transfer of severely ill patients from peripheral health posts to central facilities capable of providing proper management.

• Institute laboratory-based diagnosis for all referral sites to aid in the management of severe disease.