|
|
||||||||
| ABSTRACT |
|
|
|---|
| INTRODUCTION |
|
|
|---|
Most morbidity studies associated with ITNs have relied on active surveillance. However, routine active surveillance to monitor efficacy is problematic for large-scale control programs.10,11 In contrast, routine passive surveillance by national programs, while simple and cheap, is generally considered less reliable, though efforts are underway to strengthen district level surveillance.12 During the course of a randomized controlled trial of ITNs in western Kenya, we analyzed data from peripheral health facilities to determine the impact of ITNs on all-cause and malaria-associated sick child visits (SCVs). We also explore the potential role of these facilities for passive morbidity surveillance in this rural setting.
| MATERIALS AND METHODS |
|
|
|---|
Passive surveillance in peripheral health facilities. A passive surveillance system was established in 13 of 15 peripheral health facilities in Asembo. Three-fourths of these facilities were government funded. Drugs were equally available to peripheral health facilities serving control and ITN villages and all facilities served both control and ITN-treated communities. Two facilities that did not routinely examine at least five sick children each day were excluded. Monitoring began in June 1996 with the assignment of one project staff per facility who interviewed the caregiver accompanying each sick child using a pre-coded questionnaire. In addition to demographic information and data on ITN use, details of the childs illness were recorded using local disease terminology based upon baseline ethnographic studies. At the end of each day, the staff member transcribed to the questionnaire data from facility records on symptoms, diagnosis, and treatment of each child. Project staff did not participate in diagnosis or provide treatment or advice. Facility personnel were unaware of the randomization status of children. Retrospective review of SCVs prior to intervention was conducted in each facility. Project recorders transcribed all visits listed in health facility registers by date of visit, childs name, age, village of residence, diagnosis, and treatment.16 Of 13 health facilities surveyed, seven that monitored SCVs continuously from January 1995 through March 1999 were selected for this analysis. The period January 1995 to December 1996 is the pre-intervention period; January 1997 to March 1999 is the intervention period, and for simplicity of reporting, January 1998March 1999 is referred to as 1998.
Statistical analysis. Surveillance forms were collected weekly, checked, coded, and forwarded to the Kenya Medical Research Institute/Centers for Diseases Control and Prevention main office for data entry. Analysis was performed using Statistical Analysis System (SAS System for Windows, Release 8.01; SAS Institute, Cary, NC). The incidence of SCVs was calculated as the number of visits per 1,000 child-years. Each child could contribute multiple times to the numerator. The denominator was based on census data for the ITN study for the first quarter of 1997. The analysis assumed that the population remained constant over the intervention period regardless of intervention status.
All analysis was based on intention to treat. Data from children residing in four villages that were split into two for purposes of randomization were therefore excluded from the analysis, since their randomization status could not be accurately determined from village names. Preliminary analysis of pre-intervention data showed that the incidence of SCVs was higher in villages randomized to receive ITNs, compared with subsequent control villages. This complicated the interpretation of post-intervention comparisons between ITN and control villages since analysis had to take into account historical differences: the change in incidence was expressed as a percentage reduction; 100 x (1 - incidence rate ratio [IRR])%, where the ratio represents incidence during the intervention divided by pre-intervention rate. The difference in the percentage reduction in ITN and control villages was defined as the ITN-associated reduction: 100 x (IRRcontrol villages - IRRITN villages)%. The statistical significance of the differences over time between ITN and control villages was assessed using the Wald chi-square statistic of the interaction term between randomization status (ITN or control village) and time effect (pre- or during intervention). A significant interaction term indicates that the change over time is different between the ITN and control villages. The impact of ITNs by age group and by intervention study year (1997 versus 1998) was assessed using analogous interaction terms.
We estimated the potential number of SCVs prevented by ITNs per 1,000 child-years by multiplying the estimate of the bed net-associated reduction described earlier by the crude overall incidence rate in the entire pre-intervention study population. The potential rather than the actual number of visits prevented was calculated because use of the latter would have overestimated the effect of ITNs due to the unequal pre-intervention distribution of SCVs between subsequent ITN and control villages.
Poisson regression using GEE methods in the GENMOD procedure in SAS software was used. Because the trial used village as the unit of randomization, all P values and 95% confidence intervals were corrected for clustering at this level, assuming an exchangeable correlation structure for measures from the same village. Two sided P values < 0.05 were considered statistically significant.
Ethical clearance. The ITN trial was reviewed and approved by the institutional review boards of the Kenya Medical Research Institute (Nairobi, Kenya) and the Centers for Disease Control and Prevention (Atlanta, GA).
| RESULTS |
|
|
|---|
|
|
|
|
The incidence of prescriptions was reduced 30% more in ITN than control villages in parallel with the same overall reduction in the incidence of sick child visits. Thus, sick children from both from ITN and control villages who attended a health facility were each issued an average of 2.1 drugs per illness episode. Similarly, of children diagnosed with malaria, an average of 2.4 drugs were issued per malaria episode diagnosed, again regardless of their randomization status.
Impact of ITNs by diagnosis.
Similar to the overall number of SCVs, a 30% greater reduction in the incidence of SCVs diagnosed as malaria was observed among visits from children living in ITN villages (30% versus 5%; P = 0.04) (Table 3
). Diagnoses of severe malnutrition and anemia were rare and constituted only 1% and 2% of all SCVs, respectively, among children from the entire study population. The incidence of anemia, diarrhea, malnutrition, respiratory tract infections, and other illnesses in any age group also decreased with time, and this decrease was greater in intervention villages. However, none of these differences were statistically significant. Comparison by diagnosis of the potential number of SCVs prevented shows that the greatest reduction was in children diagnosed with malaria (Figure 3
).
|
|
| DISCUSSION |
|
|
|---|
The reduction in SCVs from bed net villages was higher in the first year of the study compared with the second year, but it is possible, for a variety of reasons, that this does not reflect an actual decrease in efficacy. In MaySeptember 1998, the number of SCVs decreased markedly in both ITN and control villages, possibly due to reduced rainfall during this part of the year,13 resulting in lower malaria transmission. Lower malaria transmission may also have resulted from a strong community effect by ITNs observed in this study, reducing malaria associated morbidity in children living within 300 meters of intervention villages (25% of the control population).17 A stronger community effect in the second study year resulting from the cumulative effects of insecticide on reducing mosquito densities over time may also partly explain the reduced morbidity in the control villages in the second year.17 In addition, because this analysis was based on data generated from routine statistics, we used a fixed denominator over time. However, the ITNs were found to have a marked impact on mortality in the first year in Asembo (reduction of 34%).18 Thus, the overall number of children, particularly infants, in ITN villages, but not in control villages, increased over time, as more children survived the first year to contribute to the second year. The use of a fixed, instead of time-dependent, denominator will thus underestimate the true denominator, leading to an overestimate of the incidence of SCVs in ITN villages in the second year, and thus an underestimate of the efficacy measurement. Other possible explanations for this apparent decrease in efficacy are the Hawthorne effect,19 and selective survival; high-risk children in control villages may die in the first year, and the surviving cohort contributing to the second year consists of a healthier selection of older children. Time to re-treatment with insecticide is unlikely to be a contributing factor because most delayed re-treatment occurred in Gem, not Asembo.18 It is unlikely that delayed acquisition of immunity resulting in increased malaria morbidity in older children played a role since no shift in the age distribution of SCVs over time was observed. There was also no evidence that the reduction in malaria resulted in increased morbidity due to other non-malaria related illnesses (replacement morbidity).
It is unlikely that our results were influenced by the type of health facility or the type of antimalarial treatment issued by these facilities. The health facilities were equally dispersed between intervention and control villages and none exclusively served children randomized into either intervention of control. Data from this, and the baseline study of diagnosis and prescribing practices,16 illustrated that the 86% of the sick children visiting health facilities received chloroquine, which was the first-line treatment of uncomplicated malaria at the time of the study. Sulfadoxine-pyrimethamine was made available as first line treatment of uncomplicated, microscopically confirmed Plasmodium falciparum malaria to children participating in longitudinal monitoring in the 19 cohort villages.13 However, the villages were randomized equally between intervention and control arms in both the cohort and remaining villages.14
Although none of the clinics had facilities for microscopy or assessment of hemoglobin concentrations, we found a significant reduction in the proportion of SCVs diagnosed to have malaria, but no reduction in the proportion of diagnosis for other common childhood illnesses including respiratory tract infection, diarrhea, measles, or helminth infection. This overall result is consistent with findings of detailed cross-sectional and longitudinal surveys in the same population,20 suggesting that statistics gleaned from peripheral health facilities can provide useful estimates of efficacy of public health interventions in this rural setting. An ITN trial in coastal Kenya also showed that records from Kilifi district hospital could be used to detect a reduction in crude hospital admission rates, malaria admissions, and severe malaria admissions. An important advantage of passive surveillance through routine facilities is that an estimate of the effect of an intervention on actual numbers of children seeking health care may be obtained.
This study also illustrates the limitations of statistics from rural health facilities for monitoring of illnesses requiring laboratory diagnosis. Malaria is one of the main causes of severe anemia in this area, particularly in young children.2123 In concurrent cross-sectional surveys, the prevalence of any anemia (hemoglobin level <11 g/dL) and severe anemia (hemoglobin level < 7 g/dL) was 70% and 6.9%, respectively, with 34% of children exhibiting detectable pallor.23 In the same study, a reduction of 39% in severe anemia in children less than three years old was associated with ITN use,20 and in a birth cohort of 900 infants a reduction of 60% was observed.23 However, no reduction in anemia was reflected in health statistics. This was most likely due to the low prevalence of anemia detected (2%) in the absence of laboratory capabilities. With this diagnostic limitation, it is unlikely than anemia can effectively be used as a key indicator for measuring the impact of a malaria control intervention through rural health facilities.
Two important conclusions can be drawn from this study. First, and most important, is the observation that ITN use was associated with a reduction of 27% in SCVs at peripheral health facilities, reflecting substantial amelioration of the malaria burden on both the children of Asembo and the health care system of Asembo. Second, we confirmed that passive surveillance in peripheral health facilities provided information on disease trends that was sufficiently sensitive to detect a substantial ITN-related reduction in malaria-associated SCVs (though not anemia) among pre-school children.
Acknowledgments: We thank the health facility recorders and data entry staff for their contribution to this study, and Caleb Ouma and Gideon Ogude for their supervisory assistance. We are grateful to the residents of Asembo for their cooperation. Dr. Kevin DeCock is thanked for his comments on the manuscript. Kathleen A. Wannemuehler acknowledges administrative support from the Oak Ridge Institute for Science and Education (Oak Ridge, TN). This paper is published with the permission of the Director of the Kenya Medical Research Institute.
Financial support: The bed net project was funded by the United States Agency for International Development.
Disclaimer: The opinions or assertions contained in this manuscript are the private ones of the authors and are not to be construed as official or reflecting the views of the U.S. Public Health Service or Department of Health and Human Services. Use of trade names is for identification only and does not imply endorsement by the U.S. Public Health Service or Department of Health and Human Services.
Authors addresses: Penelope A. Phillips-Howard, Kathleen A. Wannemuehler, Margarette S. Kolczak, Feiko O. ter Kuile, John E. Gimnig, Kris Olson, and William A. Hawley, Division of Parasitic Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Mailstop F-22, 4770 Buford Highway, Atlanta, GA 30341. Bernard L. Nahlen, Roll Back Malaria, World Health Organization, Avenue Appia 20, 1211 Geneva 27, Switzerland. Jane A. Alaii, Amos Odhacha, and John M. Vulule, Centre for Vector Biology and Control Research, Kenya Medical Research Institute, PO Box 1578, Kisumu, Kenya.
| REFERENCES |
|
|
|---|
This article has been cited by other articles:
![]() |
P. E. OKELLO, W. VAN BORTEL, A. M. BYARUHANGA, A. CORREWYN, P. ROELANTS, A. TALISUNA, U. D'ALESSANDRO, and M. COOSEMANS Variation in malaria transmission intensity in seven sites throughout Uganda. Am J Trop Med Hyg, August 1, 2006; 75(2): 219 - 225. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. ADAZU, K. A. LINDBLADE, D. H. ROSEN, F. ODHIAMBO, P. OFWARE, J. KWACH, A. M. VAN EIJK, K. M. DECOCK, P. AMORNKUL, D. KARANJA, et al. HEALTH AND DEMOGRAPHIC SURVEILLANCE IN RURAL WESTERN KENYA: A PLATFORM FOR EVALUATING INTERVENTIONS TO REDUCE MORBIDITY AND MORTALITY FROM INFECTIOUS DISEASES Am J Trop Med Hyg, December 1, 2005; 73(6): 1151 - 1158. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. A. PHILLIPS-HOWARD, F. O. TER KUILE, B. L. NAHLEN, J. A. ALAII, J. E. GIMNIG, M. S. KOLCZAK, D. J. TERLOUW, S. K. KARIUKI, Y. P. SHI, S. P. KACHUR, et al. THE EFFICACY OF PERMETHRIN-TREATED BED NETS ON CHILD MORTALITY AND MORBIDITY IN WESTERN KENYA II. STUDY DESIGN AND METHODS Am J Trop Med Hyg, April 1, 2003; 68(90040): 10 - 15. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. A. PHILLIPS-HOWARD, B. L. NAHLEN, M. S. KOLCZAK, A. W. HIGHTOWER, F. O. TER KUILE, J. A. ALAII, J. E. GIMNIG, J. ARUDO, J. M. VULULE, A. ODHACHA, et al. EFFICACY OF PERMETHRIN-TREATED BED NETS IN THE PREVENTION OF MORTALITY IN YOUNG CHILDREN IN AN AREA OF HIGH PERENNIAL MALARIA TRANSMISSION IN WESTERN KENYA Am J Trop Med Hyg, April 1, 2003; 68(90040): 23 - 29. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. ARUDO, J. E. GIMNIG, F. O. TER KUILE, S. P. KACHUR, L. SLUTSKER, M. S. KOLCZAK, W. A. HAWLEY, A. S. S. ORAGO, B. L. NAHLEN, and P. A. PHILLIPS-HOWARD COMPARISON OF GOVERNMENT STATISTICS AND DEMOGRAPHIC SURVEILLANCE TO MONITOR MORTALITY IN CHILDREN LESS THAN FIVE YEARS OLD IN RURAL WESTERN KENYA Am J Trop Med Hyg, April 1, 2003; 68(90040): 30 - 37. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. A. PHILLIPS-HOWARD, K. A. WANNEMUEHLER, F. O. TER KUILE, W. A. HAWLEY, M. S. KOLCZAK, A. ODHACHA, J. M. VULULE, and B. L. NAHLEN DIAGNOSTIC AND PRESCRIBING PRACTICES IN PERIPHERAL HEALTH FACILITIES IN RURAL WESTERN KENYA Am J Trop Med Hyg, April 1, 2003; 68(90040): 44 - 49. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. O. TER KUILE, D. J. TERLOUW, S. K. KARIUKI, P. A. PHILLIPS-HOWARD, L. B. MIREL, W. A. HAWLEY, J. F. FRIEDMAN, Y. P. SHI, M. S. KOLCZAK, A. A. LAL, et al. IMPACT OF PERMETHRIN-TREATED BED NETS ON MALARIA, ANEMIA, AND GROWTH IN INFANTS IN AN AREA OF INTENSE PERENNIAL MALARIA TRANSMISSION IN WESTERN KENYA Am J Trop Med Hyg, April 1, 2003; 68(90040): 68 - 77. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. A. ALAII, H. W. VAN DEN BORNE, S. P. KACHUR, H. MWENESI, J. M. VULULE, W. A. HAWLEY, M. I. MELTZER, B. L. NAHLEN, and P. A. PHILLIPS-HOWARD PERCEPTIONS OF BED NETS AND MALARIA PREVENTION BEFORE AND AFTER A RANDOMIZED CONTROLLED TRIAL OF PERMETHRIN-TREATED BED NETS IN WESTERN KENYA Am J Trop Med Hyg, April 1, 2003; 68(90040): 142 - 148. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. WISEMAN, W. A. HAWLEY, F. O. TER KUILE, P. A. PHILLIPS-HOWARD, J. M. VULULE, B. L. NAHLEN, and J. MILLS THE COST-EFFECTIVENESS OF PERMETHRIN-TREATED BED NETS IN AN AREA OF INTENSE MALARIA TRANSMISSION IN WESTERN KENYA Am J Trop Med Hyg, April 1, 2003; 68(90040): 161 - 167. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. A. HAWLEY, F. O. TER KUILE, R. S. STEKETEE, B. L. NAHLEN, D. J. TERLOUW, J. E. GIMNIG, Y. P. SHI, J. M. VULULE, J. A. ALAII, A. W. HIGHTOWER, et al. IMPLICATIONS OF THE WESTERN KENYA PERMETHRIN-TREATED BED NET STUDY FOR POLICY, PROGRAM IMPLEMENTATION, AND FUTURE RESEARCH Am J Trop Med Hyg, April 1, 2003; 68(90040): 168 - 173. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |