(Posted as supplied by author)
Table A: Countries analysed in each African region
AfrD 
AfrE 
Algeria 
Botswana 
Angola 
Burundi 
Benin 
Central African Republic 
Burkina Faso 
Congo 
Cameroon 
Cote d’Ivoire 
Cape Verde 
Democratic Republic of the Congo 
Chad 
Eritrea 
Comoros 
Ethiopia 
Equatorial Guinea 
Kenya 
Gabon 
Lesotho 
Gambia 
Malawi 
Ghana 
Mozambique 
Guinea 
Namibia 
Guinea Bissau 
Rwanda 
Liberia 
South Africa 
Madagascar 
Swaziland 
Mali 
Uganda 
Mauritania 
United Republic of Tanzania 
Mauritius 
Zambia 
Niger 
Zimbabwe 
Nigeria  
Sao Tome and Principe  
Senegal  
Seychelles  
Sierra Leone  
Togo 
(Posted as supplied by author)
Table B: Estimation of effectiveness
Intervention/ Parameter 
Value 
Source/Notes 
ITN  
Adherence 
65% 
Estimate within range suggested to achieve impact^{w8} 
Probability of success when not fully compliant 
0% 
Estimate based on binary response of whether or not an individual has slept under the net 

50% 
^{w8} 

20% 
^{w9} 
IRS  
Adherence 
100% 
Estimate based on household level to achieve impact 
Probability of success when not fully compliant 
0% 
Estimate based on dosing schedule and pharmacokinetic properties 

50% 
Assumed to be similar to ITNs^{w3,w10,w11} 

20% 
Assumed to be similar to ITNs^{w3,w10,w11} 
CQ  
Adherence 
40% 
Estimate based on relative length and complexity of the regimen 
Probability of success when not fully compliant 
20% 
Estimate based on dosing schedule and pharmacokinetic properties 
Initial resistance 
0.3 
^{w2} 
Growth rate of resistance 
0.1 
^{w2} 

27% 
See appendix (drugs) for details of calculation 
SP  
Adherence 
90% 
Estimate based on relative length and complexity of the regimen 
Probability of success when not fully compliant 
0% 
Estimate based on dosing schedule and pharmacokinetic properties 
Initial resistance 
0.1 
^{w2} 
Growth rate of resistance 
0.4 
^{w2} 

44% 
See appendix (drugs) for details of calculation 
CQSP 
Assumed to be coadministered, not coformulated  
Adherence 
35% 
Estimate based on relative length and complexity of the regimen 
Probability of success when not fully compliant 
45% 
Estimate based on dosing schedule and pharmacokinetic properties 
Initial resistance 
0.1 
^{w2} 
Growth rate of resistance 
0.1 
^{w2} 

48%  
ACT 
Assumed to be blisterpacked and/or coformulated  
Adherence 
40% 
Estimate based on relative length and complexity of the regimen 
Probability of success when not fully compliant 
45% 
Estimate based on dosing schedule and pharmacokinetic properties 
Initial resistance 
0.001 
^{w2} 
Growth rate of resistance 
0.05 
^{w2} 

63% 
See appendix (drugs) for details of calculation 
IPTp (SP)  
Adherence 
80% 
Estimate based on relative length and complexity of the regimen 
Probability of success when not fully compliant 
10% 
Estimate based on dosing schedule and pharmacokinetic properties 
Initial resistance 
0.1 
^{w2} 
Growth rate of resistance 
0.1 
^{w2} 

3.4% 
Assumes baseline reduction of 5% based on model developed by Goodman et al 2000^{w12}; See methods section for details of calculation of net effect 
(Posted as supplied by author)
Table C: Detailed costeffectiveness results
Description of intervention, Region 
Average yearly costs, $int 
Average yearly effectiveness (DALYs averted) 
Average cost effectiveness ($int per DALY averted) 
Incremental cost effectiveness ($int per DALY averted) 
AfrD  
ITN  95% cov. 
254,277,140 
8,872,378 
29 
Dominated 
IPTp  95% cov. 
22,782,339 
92,104 
247 
Dominated 
IRS  95% cov. 
282,976,169 
8,872,378 
32 
Dominated 
CQ  95% cov. 
93,804,789 
3,921,597 
24 
Dominated 
SP  95% cov. 
93,838,374 
6,448,511 
15 
Dominated 
Comb  95% cov. 
93,960,837 
7,034,184 
13 
Dominated 
ACT  95% cov. 
95,609,717 
9,254,473 
10 
15.65 
ITN & CQ  95% cov. 
313,765,847 
10,611,816 
30 
Dominated 
ITN & SP  95% cov. 
313,893,918 
11,725,990 
27 
Dominated 
ITN & Combo  95% cov. 
313,985,895 
11,983,490 
26 
Dominated 
ITN & ACT  95% cov. 
314,977,037 
12,957,175 
24 
Dominated 
IRS & CQ  95% cov. 
337,452,801 
10,611,816 
32 
Dominated 
IRS & SP  95% cov. 
337,611,872 
11,725,990 
29 
Dominated 
IRS & Combo  95% cov. 
337,711,011 
11,983,490 
28 
Dominated 
IRS & ACT  95% cov. 
338,729,226 
12,957,175 
26 
Dominated 
IRS & ACT & IPTp  95% cov. 
339,298,766 
12,972,791 
26 
Dominated 
ITN & ACT & IPTp  95% cov. 
315,546,119 
12,972,791 
24 
59 
IRS & ITN  95% cov. 
444,540,358 
12,675,957 
35 
Dominated 
IRS & ITN & ACT  95% cov. 
467,328,380 
14,554,556 
32 
Dominated 
IRS & ITN & ACT & IPTp  95% cov. 
467,673,321 
14,561,792 
32 
96 
ITN  80% cov. 
224,730,197 
7,629,171 
29 
Dominated 
IPTp  80% cov. 
18,404,318 
77,556 
237 
Dominated 
IRS  80% cov. 
239,471,420 
7,629,171 
31 
Dominated 
CQ  80% cov. 
70,895,223 
3,298,390 
21 
Dominated 
SP  80% cov. 
70,913,875 
5,419,512 
13 
Dominated 
Comb  80% cov. 
71,014,283 
5,910,663 
12 
Dominated 
ACT  80% cov. 
72,386,626 
7,771,018 
9 
9 
ITN & CQ  80% cov. 
260,865,735 
9,284,603 
28 
Dominated 
ITN & SP  80% cov. 
293,776,655 
10,344,348 
28 
Dominated 
ITN & Combo  80% cov. 
261,049,692 
10,589,200 
25 
Dominated 
ITN & ACT  80% cov. 
294,978,860 
11,514,826 
26 
Dominated 
IRS & CQ  80% cov. 
276,362,347 
9,284,603 
30 
Dominated 
IRS & SP  80% cov. 
276,488,434 
10,344,348 
27 
Dominated 
IRS & Combo  80% cov. 
276,576,876 
10,589,200 
26 
Dominated 
IRS & ACT  80% cov. 
277,522,810 
11,514,826 
24 
Dominated 
IRS & ACT & IPTp  80% cov. 
277,868,610 
11,529,641 
24 
Dominated 
ITN & ACT & IPTp  80% cov. 
262,319,382 
11,529,641 
23 
Dominated 
IRS & ITN  80% cov. 
385,160,639 
11,123,636 
35 
Dominated 
IRS & ITN & ACT  80% cov. 
407,116,404 
13,261,860 
31 
Dominated 
IRS & ITN & ACT & IPTp  80% cov. 
407,407,031 
13,270,038 
31 
Dominated 
ITN  50% cov. 
166,705,327 
4,961,812 
34 
Dominated 
IPTp  50% cov. 
17,190,375 
48,466 
355 
Dominated 
IRS  50% cov. 
148,997,834 
4,961,812 
30 
Dominated 
CQ  50% cov. 
64,076,020 
2,056,500 
31 
Dominated 
SP  50% cov. 
64,083,603 
3,373,747 
19 
Dominated 
Comb  50% cov. 
64,144,985 
3,678,178 
17 
Dominated 
ACT  50% cov. 
64,992,004 
4,829,313 
13 
Dominated 
ITN & CQ  50% cov. 
193,184,118 
6,229,432 
31 
Dominated 
ITN & SP  50% cov. 
256,065,911 
7,039,542 
36 
Dominated 
ITN & Combo  50% cov. 
193,290,021 
7,226,565 
27 
Dominated 
ITN & ACT  50% cov. 
257,254,619 
7,933,061 
32 
Dominated 
IRS & CQ  50% cov. 
192,344,080 
6,229,432 
31 
Dominated 
IRS & SP  50% cov. 
192,405,053 
7,039,542 
27 
Dominated 
IRS & Combo  50% cov. 
192,464,601 
7,226,565 
27 
Dominated 
IRS & ACT  50% cov. 
193,158,204 
7,933,061 
24 
Dominated 
IRS & ACT & IPTp  50% cov. 
193,347,670 
7,944,332 
24 
Dominated 
ITN & ACT & IPTp  50% cov. 
194,162,562 
7,944,332 
24 
Dominated 
IRS & ITN  50% cov. 
267,866,814 
7,499,565 
36 
Dominated 
IRS & ITN & ACT  50% cov. 
289,831,107 
9,510,715 
30 
Dominated 
IRS & ITN & ACT & IPTp  50% cov. 
289,990,658 
9,518,346 
30 
Dominated 
AfrE  
ITN  95% cov. 
243,893,889 
5,908,330 
41 
Dominated 
IPTp  95% cov. 
20,606,834 
58,589 
352 
Dominated 
IRS  95% cov. 
244,519,970 
5,908,330 
41 
Dominated 
CQ  95% cov. 
71,620,401 
2,503,983 
29 
Dominated 
SP  95% cov. 
71,633,984 
4,109,851 
17 
Dominated 
Comb  95% cov. 
71,726,621 
4,481,207 
16 
Dominated 
ACT  95% cov. 
73,000,256 
5,886,159 
12 
12.40 
ITN & CQ  95% cov. 
283,198,057 
7,285,807 
39 
Dominated 
ITN & SP  95% cov. 
253,901,457 
8,166,581 
31 
Dominated 
ITN & Combo  95% cov. 
283,367,274 
8,369,968 
34 
Dominated 
ITN & ACT  95% cov. 
254,755,715 
9,138,452 
28 
55.89 
IRS & CQ  95% cov. 
294,706,173 
7,285,807 
40 
Dominated 
IRS & SP  95% cov. 
294,813,050 
8,166,581 
36 
Dominated 
IRS & Combo  95% cov. 
294,898,541 
8,369,968 
35 
Dominated 
IRS & ACT  95% cov. 
295,848,380 
9,138,452 
32 
Dominated 
IRS & ACT & IPTp  95% cov. 
297,508,799 
9,150,654 
33 
Dominated 
ITN & ACT & IPTp  95% cov. 
285,960,854 
9,150,654 
31 
Dominated 
IRS & ITN  95% cov. 
417,786,013 
8,745,929 
48 
Dominated 
IRS & ITN & ACT  95% cov. 
441,216,954 
10,721,678 
41 
118 
IRS & ITN & ACT & IPTp  95% cov. 
442,342,075 
10,729,154 
41 
151 
ITN  80% cov. 
217,147,049 
5,156,555 
42 
Dominated 
IPTp  80% cov. 
19,689,795 
50,388 
391 
Dominated 
IRS  80% cov. 
207,739,476 
5,156,555 
40 
Dominated 
CQ  80% cov. 
68,956,191 
2,152,560 
32 
Dominated 
SP  80% cov. 
68,965,211 
3,531,517 
20 
Dominated 
Comb  80% cov. 
69,043,378 
3,850,229 
18 
Dominated 
ACT  80% cov. 
70,123,532 
5,055,437 
14 
Dominated 
ITN & CQ  80% cov. 
252,277,432 
6,428,029 
39 
Dominated 
ITN & SP  80% cov. 
281,775,820 
7,240,619 
39 
Dominated 
ITN & Combo  80% cov. 
252,417,222 
7,428,216 
34 
Dominated 
ITN & ACT  80% cov. 
282,861,375 
8,136,888 
35 
Dominated 
IRS & CQ  80% cov. 
257,113,360 
6,428,029 
40 
Dominated 
IRS & SP  80% cov. 
257,196,381 
7,240,619 
36 
Dominated 
IRS & Combo  80% cov. 
257,271,146 
7,428,216 
35 
Dominated 
IRS & ACT  80% cov. 
258,126,588 
8,136,888 
32 
Dominated 
IRS & ACT & IPTp  80% cov. 
259,253,805 
8,148,125 
32 
Dominated 
ITN & ACT & IPTp  80% cov. 
254,386,921 
8,148,125 
31 
Dominated 
IRS & ITN  80% cov. 
363,023,981 
7,730,361 
47 
Dominated 
IRS & ITN & ACT  80% cov. 
384,638,944 
9,678,725 
40 
Dominated 
IRS & ITN & ACT & IPTp  80% cov. 
385,586,428 
9,686,077 
40 
Dominated 
ITN  50% cov. 
163,783,165 
3,312,998 
49 
Dominated 
IPTp  50% cov. 
18,770,275 
31,489 
596 
Dominated 
IRS  50% cov. 
133,035,146 
3,312,998 
40 
Dominated 
CQ  50% cov. 
65,718,603 
1,343,275 
49 
Dominated 
SP  50% cov. 
65,722,956 
2,201,621 
30 
Dominated 
Comb  50% cov. 
65,771,428 
2,399,767 
27 
Dominated 
ACT  50% cov. 
66,443,687 
3,148,238 
21 
Dominated 
ITN & CQ  50% cov. 
191,879,461 
4,213,508 
46 
Dominated 
ITN & SP  50% cov. 
232,573,857 
4,788,216 
49 
Dominated 
ITN & Combo  50% cov. 
191,957,235 
4,920,806 
39 
Dominated 
ITN & ACT  50% cov. 
233,468,213 
5,421,383 
43 
Dominated 
IRS & CQ  50% cov. 
183,392,851 
4,213,508 
44 
Dominated 
IRS & SP  50% cov. 
183,430,404 
4,788,216 
38 
Dominated 
IRS & Combo  50% cov. 
183,478,586 
4,920,806 
37 
Dominated 
IRS & ACT  50% cov. 
184,067,612 
5,421,383 
34 
Dominated 
IRS & ACT & IPTp  50% cov. 
184,771,948 
5,429,305 
34 
Dominated 
ITN & ACT & IPTp  50% cov. 
193,244,869 
5,429,305 
36 
Dominated 
IRS & ITN  50% cov. 
253,426,000 
5,082,819 
50 
Dominated 
IRS & ITN & ACT  50% cov. 
270,276,675 
6,603,300 
41 
Dominated 
IRS & ITN & ACT & IPTp  50% cov. 
270,868,802 
6,609,013 
41 
Dominated 
(Posted as supplied by author)
Figure A: Cost effectiveness plane showing 60 analysed interventions (20 individual and combination interventions at 3 assumed coverage levels) and expansion path (see text), AfrD.

(Posted as supplied by author)
Figure B: Cost effectiveness plane showing 60 interventions (20 individual and combination interventions at 3 assumed coverage levels) and expansion path (see text), AfrE.

Appendix (posted as supplied by author)
Individual and combination interventions
In addition to the seven individual interventions discussed in the main text (reproduced here for convenience):
 Insecticidetreated bed nets (ITN);
 Indoor residual spraying (IRS);
 Case management with chloroquine (CQ);
 Case management with sulfadoxinepyrimethamine (SP);
 Case management with nonartemisininbased (CQSP) combination therapy (Comb);
 Case management with artemisininbased combination therapy (ACT);
 Intermittent presumptive treatment with SP in pregnancy (IPTp);
the following combinations were analysed:
 ITN & CQ;
 ITN & SP;
 ITN & Comb;
 ITN & ACT;
 IRS & CQ;
 IRS & SP;
 IRS & Comb;
 IRS & ACT;
 IRS & ACT & IPTp;
 ITN & ACT & IPTp;
 IRS & ITN;
 IRS & ITN & ACT;
 IRS & ITN & ACT & IPTp.
Each of these 20 interventions was analysed at each of the three coverage levels (50%, 80% and 95%) and in each of the two regions (AfrD and AfrE).
Interventionspecific assumptions
Drugs
The current subregional coverage of each casemanagement intervention was estimated as a populationweighted average on the basis of the country population and countrylevel estimates of the proportion of children who receive CQ or other antimalarials.^{w1} If no data were available on the proportion of sick children receiving drugs, rather than assume zero coverage for the country, the country was excluded from both the numerator and denominator of the coverage calculation. Estimates for the use of ‘any antimalarial’ that were greater than those estimates quoted specifically for CQ were assumed to refer to that country’s firstline drug (if other than CQ, otherwise that country’s secondline drug). If specific coverage estimates for CQ were not available, it was assumed that 80% of ‘any antimalarial’ consisted of that country’s firstline drug, with the remaining 20% consisting of the secondline drug.
Estimating drug resistance
Resistance was estimated by means of a standard logistic growth function (eq. 1). Projected effectiveness estimates for case management interventions and IPTp were thus adjusted for changing resistance profiles with increasing duration of use. Projected drug resistance was estimated as a function of the initial level of resistance prior to intervention implementation, the growth rate of resistance, and the maximum level of resistance (carrying capacity). As the estimated growth in resistance over the 10year implementation horizon was in all cases approximately linear, annual numbers of treatment failures were averaged over the 10year implementation period to obtain an average failure rate.
Resistance was measured as:^{w1}
where R_{i}_{, t} is the proportion of parasites that resistant to drug i at time t, R_{0} is the initial level of resistance, r is the growth rate in resistance, k is the maximum level of resistance and t is time (see Table 1, main text). The growth rate of resistance was estimated in conjunction with biogeneticists on the basis of drugspecific pharmacokinetic properties and experience in the field.^{w2}
Drug treatment failure
Treatment failure was calculated as:^{w1}
where F is treatment failure, m is the probability of full adherence, and p is the probability of treatment success without full compliance (see Table 2, main text).
Net drug effectiveness
Net effectiveness, E, was calculated as:
where B is baseline efficacy, and F_{i}_{,t} is treatment failure.
Preventive interventions
No vector resistance to insecticide was accounted for. Potential benefits of prevention on unborn children (i.e. in terms of higher birth weight) were not included.
IRS and ITN were assumed to benefit from a 9% "knockon" effect (i.e. transmission externality) at modelled coverage levels: that is, the costs of a given level of effective coverage were reduced by about 8% (= 1  1/1.09), while keeping the effects unchanged. This is likely a conservative estimate.^{w3}
For ITNs, coverage estimates refer to individuals, while, for IRS, coverage estimates refer to the household (e.g. a 50% coverage level for IRS indicates that 91% of houses within 50% of transmission areas are assumed to be effectively sprayed.
ITN
ITNs were assumed to be long lasting and thus not requiring reimpregnation. ITNs were assumed to last on average 5 years.^{w4} One ITN was attributed to each 1.5 persons.^{w5} Estimates of ITN effectiveness assumed 65% adherence, defined as sleeping under an untorn net.
IRS
Each spray team was assumed to cover 1170 houses per year.
Case Management
Case management refers to the firstline antimalarial drug administered to symptomatic patients presenting at outpatient clinics. One health center visit per case was assumed, as well as that a patient given an antimalarial drug does indeed have malaria (and therefore receives the drug benefit).
IPTp
Women of childbearing age were assumed to be those aged 15 to 49. Sixty percent of firsttime mothers (estimated to account for 30% of births) were assumed to receive two doses of IPTp within the context of routine antenatal care (ANC) services. Sixty percent of pregnancies considered "at risk" due to HIV (estimated to account for an additional 20% of births) were assumed to receive three doses. A total of 15% of pregnant mothers (split evenly between first and multiparous highrisk pregnancies) treated with IPTp were assumed to have babies that would not survive the neonatal period. A potential reduction in incidence of placental malaria from IPTp was excluded from consideration, as its associated morbidity is difficult to quantify due to the asymptomatic nature of the disease in endemic areas.^{w6}
Resistance to SP within the IPTp context was based on estimates of resistance to SP when used as a firstline treatment.^{w7} However, it was assumed that when SP is used for IPTp it is no longer used in standard case management, since if SP is concurrently used as the first line treatment for malaria, it’s usefulness as a preventive treatment will be diminished. Resistance to SP was assumed to grow more slowly when used only for IPTp.
References
w1. Goodman C, Coleman PG, Mills AJ. Changing the first line drug for malaria treatment  costeffectiveness analysis with highly uncertain intertemporal tradeoffs. Health Economics 2001;10:731749.
w2. Roper C. personal communication, 2004.
w3. Curtis CF, Maxwell CA, Finch RJ, Njunwa KJ. A comparison of use of a pyrethroid either for house spraying or for bednet treatment against malaria vectors. Tropical Medicine and International Health 1998;3(8):619631.
w4. WHO, RBM, UNICEF, PSI, MSH. Sources and Prices of Selected Products for the Prevention, Diagnosis and Treatment of Malaria, 2004.
w5. Lines J. personal communication, 2005.
w6. WHO. Global burden of disease estimates 2002. Geneva: http://www3.who.int/whosis/menu.cfm?path=evidence,burden , 2002.
w7. CostIt [program]. 1 version. Geneva: WHO, 2002.
w8. WHO. The Africa Malaria Report 2003. Geneva: UNICEF and World Health Organization, 2003:120.
w9. Lengeler C. Insecticidetreated bednets and curtains for preventing malaria. Cochrane Library, Issue 1. Oxford, 2001.
w10. Curtis CF, Mnzava AEP. Comparison of house spraying and insecticidetreated nets for malaria control. Bulletin of the World Health Organization 2000;78(12):13891400.
w11. Lengeler C. Comparison of malaria control interventions. Bull World Health Organ 2001;79(1).
w12. Goodman CA, Coleman PG, Mills AJ. Economic analysis of malaria control in subSaharan Africa. Geneva: Global Forum for Health Research, 2000:185.