Response to Pardo et al, Malaria Journal 2006; 5:10
Immo Kleinschmidt, Medical Research Council of South Africa
29 March 2006
On behalf of the partners involved in implementing the Bioko Island Malaria Control Project, we commend Pardo et al [1] for their paper in Malaria Journal which attempts to evaluate strategies for the control of Plasmodium falciparum infection on the island of Bioko (Equatorial Guinea). Their findings confirm our own data showing that indoor residual spraying (IRS) “has proven an effective control strategy” on Bioko Island. However, we were surprised that the authors did not acknowledge the role of the Bioko Island Malaria Control Project and its implementing partners namely Medical Development International, One World Development Group, the Medical Research Council of South Africa, Harvard University, and the Malaria Control Program of the Equatorial Guinea Ministry of Health. Instead, the reader of this article could easily conclude, mistakenly, that the authors were reporting an evaluation of their own intervention. Similarly, the article’s reference to “a private initiative by oil companies” warrants at least a footnote that specifically cites the private sector partners i.e. Marathon Oil Company, Noble Energy, and Atlantic Methanol. Without their support the reductions in the prevalence of infection reported by Pardo et al would be dramatically different.
Unfortunately the authors misreport many important details of the BIMCP intervention. For example, IRS is described as space fumigation rather than treatment of surfaces, there are inaccuracies in reporting the frequency of spraying and the type of insecticide used. The authors make no mention of the comprehensive programme of case management and intermittent preventive treatment of pregnant women which is an important part of this project. There are other major errors in the paper, such as reporting the size of the population of Bioko as 60,000, a gross underestimate that would have been unlikely to occur if they had known that around 100,000 structures have been sprayed on the island in each spray round. However, the most disappointing aspect of this article is the very poor design of the evaluation that is described. Essentially this consists of two surveys carried out in 2004 and in 2005 respectively. Very little detail is given about the sampling method, but one can hardly take a survey of 162 children at a handful of locations around the island as a representative reflection of prevalence of infection in children on Bioko, and then compare this to prevalence measured in the second survey of a bigger but still small sample of 433 children, taken at a different set of locations. Apart from the inadequate sample size, particularly in the baseline survey, we know from our own data that there is substantial spatial heterogeneity in prevalence of infection on Bioko, and any pre- versus post intervention comparison should make comparisons at the same sentinel sites, to avoid any apparent reduction in prevalence being confounded by spatial variation in prevalence. Given the overlapping confidence intervals of the two point estimates of prevalence, one suspects that what is claimed to be significant with p<0.05 is only very marginally significant. It is poor practice to dichotomise p-values as either significant or non-significant, given the arbitrary nature of the 0.05 limit. Instead the actual value would be more informative, citing p<0.001 if the value is very small.
Furthermore, in table 2 the variation by location, age and mosquito net is unclear: in the text we are told there were significant differences depending on location, age and use of nets: does the significance test refer to differences in intervention effect or to differences in prevalence within each year. If prevalence does vary significantly by these factors then it would have been important to adjust the results for these factors. It appears that the 2005 results are at higher altitude than villages surveyed in 2004, which could explain in part the sharp reduction in parasitemia, which we regard as overestimated.
For a more comprehensive description of this large scale malaria control intervention using IRS, the first of its kind in an equatorial setting in recent times, and for a more rigorous evaluation of its impact than Pardo et al were able to present, we would refer readers to our forthcoming publication on this project [2].
Signed:
Joseph Carter, Medical Care Development International, 8401 Colesville Rd Suite 425, Silver Spring MD 20910;
Immo Kleinschmidt(corresponding author), Medical Research Council, 491 Ridge Rd, Overport, Durban 4091, South Africa;
Andy Spielman, Harvard School of Public Health, 677 Huntington Ave. Boston, MA 02115, U.S.A.;
Jaime Kuklinski, One World Development Group, 5745 SW 75 St., #219 Gainesville, FL 32608, U.S.A.
References
1. Pardo G, Descalzo MA, Molina L, Custodio E, Lwanga M, Mangue C, Obono J, Nchama A, Roche J, Benito A, Cano J. Impact of different strtaegies to control Plasmodium infection and anaemia on the island of Bioko (Equatorial Guinea). Malaria Journal 2006, 5:10.
2. Kleinschmidt I, Sharp B, Benavente L, Schwabe C, Torrez M, Kuklinski J, Morris N, Raman J, Carter J. Reduction in infection with p.falciparum one year after the introduction of malaria control interventions on Bioko Island, Equatorial Guinea. Am. J Trop Med Hyg 2006 (in Press).
Response to Pardo et al, Malaria Journal 2006; 5:10
29 March 2006
On behalf of the partners involved in implementing the Bioko Island Malaria Control Project, we commend Pardo et al [1] for their paper in Malaria Journal which attempts to evaluate strategies for the control of Plasmodium falciparum infection on the island of Bioko (Equatorial Guinea). Their findings confirm our own data showing that indoor residual spraying (IRS) “has proven an effective control strategy” on Bioko Island. However, we were surprised that the authors did not acknowledge the role of the Bioko Island Malaria Control Project and its implementing partners namely Medical Development International, One World Development Group, the Medical Research Council of South Africa, Harvard University, and the Malaria Control Program of the Equatorial Guinea Ministry of Health. Instead, the reader of this article could easily conclude, mistakenly, that the authors were reporting an evaluation of their own intervention. Similarly, the article’s reference to “a private initiative by oil companies” warrants at least a footnote that specifically cites the private sector partners i.e. Marathon Oil Company, Noble Energy, and Atlantic Methanol. Without their support the reductions in the prevalence of infection reported by Pardo et al would be dramatically different.
Unfortunately the authors misreport many important details of the BIMCP intervention. For example, IRS is described as space fumigation rather than treatment of surfaces, there are inaccuracies in reporting the frequency of spraying and the type of insecticide used. The authors make no mention of the comprehensive programme of case management and intermittent preventive treatment of pregnant women which is an important part of this project. There are other major errors in the paper, such as reporting the size of the population of Bioko as 60,000, a gross underestimate that would have been unlikely to occur if they had known that around 100,000 structures have been sprayed on the island in each spray round. However, the most disappointing aspect of this article is the very poor design of the evaluation that is described. Essentially this consists of two surveys carried out in 2004 and in 2005 respectively. Very little detail is given about the sampling method, but one can hardly take a survey of 162 children at a handful of locations around the island as a representative reflection of prevalence of infection in children on Bioko, and then compare this to prevalence measured in the second survey of a bigger but still small sample of 433 children, taken at a different set of locations. Apart from the inadequate sample size, particularly in the baseline survey, we know from our own data that there is substantial spatial heterogeneity in prevalence of infection on Bioko, and any pre- versus post intervention comparison should make comparisons at the same sentinel sites, to avoid any apparent reduction in prevalence being confounded by spatial variation in prevalence. Given the overlapping confidence intervals of the two point estimates of prevalence, one suspects that what is claimed to be significant with p<0.05 is only very marginally significant. It is poor practice to dichotomise p-values as either significant or non-significant, given the arbitrary nature of the 0.05 limit. Instead the actual value would be more informative, citing p<0.001 if the value is very small.
Furthermore, in table 2 the variation by location, age and mosquito net is unclear: in the text we are told there were significant differences depending on location, age and use of nets: does the significance test refer to differences in intervention effect or to differences in prevalence within each year. If prevalence does vary significantly by these factors then it would have been important to adjust the results for these factors. It appears that the 2005 results are at higher altitude than villages surveyed in 2004, which could explain in part the sharp reduction in parasitemia, which we regard as overestimated.
For a more comprehensive description of this large scale malaria control intervention using IRS, the first of its kind in an equatorial setting in recent times, and for a more rigorous evaluation of its impact than Pardo et al were able to present, we would refer readers to our forthcoming publication on this project [2].
Signed:
Joseph Carter, Medical Care Development International, 8401 Colesville Rd Suite 425, Silver Spring MD 20910;
Immo Kleinschmidt(corresponding author), Medical Research Council, 491 Ridge Rd, Overport, Durban 4091, South Africa;
Andy Spielman, Harvard School of Public Health, 677 Huntington Ave. Boston, MA 02115, U.S.A.;
Jaime Kuklinski, One World Development Group, 5745 SW 75 St., #219 Gainesville, FL 32608, U.S.A.
References
1. Pardo G, Descalzo MA, Molina L, Custodio E, Lwanga M, Mangue C, Obono J, Nchama A, Roche J, Benito A, Cano J. Impact of different strtaegies to control Plasmodium infection and anaemia on the island of Bioko (Equatorial Guinea). Malaria Journal 2006, 5:10.
2. Kleinschmidt I, Sharp B, Benavente L, Schwabe C, Torrez M, Kuklinski J, Morris N, Raman J, Carter J. Reduction in infection with p.falciparum one year after the introduction of malaria control interventions on Bioko Island, Equatorial Guinea. Am. J Trop Med Hyg 2006 (in Press).
Competing interests
None