### Results

Time series of malaria admissions from 1980 to 1999, 1987 to 2000, and 1981 to 2000 were recorded for Kilgoris, Kisii, and Tabaka hospitals, respectively. These clinical data represent 171,312 admissions with a primary, coprimary, or coincidental diagnosis of complicated malaria over a total of 54 admission years. The Kilgoris, Kisii, and Tabaka hospitals managed an average of 2,243; 9,191; and 3,929 malaria admissions per year, respectively, for the duration over which records were available ( Table 1 ). Throughout the study period, the frequency of childhood admissions was on average twice that of adult admissions ( Table 1 ; Figure 1A,C,E). The average ACR calculated for all months was 0.46 for Kilgoris (14,079/30,793), 0.52 for Kisii (44,043/84,648), and 0.42 for Tabaka (23,692/55,871). The ACRs derived from monthly admissions are relatively constant throughout the duration of observation (Figure 1B,D,F). Several anomalies in these data were evident, however, particularly at Tabaka in 1985 and Kisii in 1999; we did not identify any obvious explanation for these exceptions in the time series, although they did not occur during periods of major epidemics at the sites^{[32]}. In further analysis, we focused on the primary pediatric clinical case data. We considered data from children to be more likely to give an accurate picture of local malaria transmission, as they are less likely to have developed functional immunity or to have traveled and acquired infections elsewhere.

Spider plots of adult, child, and total admissions and time series of adult:child ratio for three study hospitals in Kenya. Spider plots of malaria admissions in Kilgoris (A), Kisii (C), and Tabaka (E). The data are monthly averages for the 1980-1999, 1987-2000, and 1981-2000 time periods, respectively. Adult cases (≥15 years of age) are shown in blue, child cases (<15 years) are shown in red, and total cases in black. Time series plots of the monthly adult:child ratio data are also shown for Kilgoris (B), Kisii (D), and Tabaka (F) as the continuous black line. The dashed line represents the value of 1 where adult and child admissions are equal, as is to be expected in true epidemic conditions^{[39,40,41]}. The bold line is a 25-point (month) moving average of the adult:child ratio.

Spider plots of adult, child, and total admissions and time series of adult:child ratio for three study hospitals in Kenya. Spider plots of malaria admissions in Kilgoris (A), Kisii (C), and Tabaka (E). The data are monthly averages for the 1980-1999, 1987-2000, and 1981-2000 time periods, respectively. Adult cases (≥15 years of age) are shown in blue, child cases (<15 years) are shown in red, and total cases in black. Time series plots of the monthly adult:child ratio data are also shown for Kilgoris (B), Kisii (D), and Tabaka (F) as the continuous black line. The dashed line represents the value of 1 where adult and child admissions are equal, as is to be expected in true epidemic conditions^{[39,40,41]}. The bold line is a 25-point (month) moving average of the adult:child ratio.

Spider plots of adult, child, and total admissions and time series of adult:child ratio for three study hospitals in Kenya. Spider plots of malaria admissions in Kilgoris (A), Kisii (C), and Tabaka (E). The data are monthly averages for the 1980-1999, 1987-2000, and 1981-2000 time periods, respectively. Adult cases (≥15 years of age) are shown in blue, child cases (<15 years) are shown in red, and total cases in black. Time series plots of the monthly adult:child ratio data are also shown for Kilgoris (B), Kisii (D), and Tabaka (F) as the continuous black line. The dashed line represents the value of 1 where adult and child admissions are equal, as is to be expected in true epidemic conditions^{[39,40,41]}. The bold line is a 25-point (month) moving average of the adult:child ratio.

^{[39,40,41]}. The bold line is a 25-point (month) moving average of the adult:child ratio.

^{[39,40,41]}. The bold line is a 25-point (month) moving average of the adult:child ratio.

^{[39,40,41]}. The bold line is a 25-point (month) moving average of the adult:child ratio.

Time series of child admissions for the three study hospitals, Kenya. Time series of child admissions (<15 years of age) for Kilgoris (A), Kisii (B), and Tabaka (C) for 1980-1999, 1987-2000, and 1981-2000 time periods, respectively. The bold line is 25-point (month) moving average of the same data for child admissions.

Time series of child admissions for the three study hospitals, Kenya. Time series of child admissions (<15 years of age) for Kilgoris (A), Kisii (B), and Tabaka (C) for 1980-1999, 1987-2000, and 1981-2000 time periods, respectively. The bold line is 25-point (month) moving average of the same data for child admissions.

Time series of child admissions for the three study hospitals, Kenya. Time series of child admissions (<15 years of age) for Kilgoris (A), Kisii (B), and Tabaka (C) for 1980-1999, 1987-2000, and 1981-2000 time periods, respectively. The bold line is 25-point (month) moving average of the same data for child admissions.

The long-term data used in this analysis indicate that clinical cases of malaria occur every month at each hospital; acute seasonal peaks occur in June and July (Figure 1A,C,E). On average, one third of the total annual child malaria admissions were concentrated in these 2 months (35%, 32%, and 27% for Kilgoris, Kisii, and Tabaka, respectively).

The trends and interannual variation in pediatric malaria admissions at each facility are shown in Figure 2A-C. These graphs demonstrate clear, substantial between-year variation in child malaria admissions. The 2 years of highest case presentations were 1994 and 1998 for Kilgoris, 1996 and 1997 for Kisii, and 1997 and 1996 for Tabaka (the moving average line in Figure 2A-C clearly shows these years). Although Kisii and Tabaka showed similarities, little coherence occurred in peak years of child admissions between these sites and Kilgoris, despite their close geographic proximity. At each facility, pediatric malaria admissions rose substantially over the period of observation ( Table 2 ). In Kilgoris, deseasonalized child malaria admissions rose from 56 in January 1980 to 200 in December 1999, an increase of 256% over 20 years (p<0.001). Similar trends were observed at Kisii (32% increase from January 1987 through December 1999; p=0.019) and Tabaka (91% increase from January 1980 through December 1999; p<0.001).

In parallel with these significant rises in number of cases, estimates of the annual rate of natural population growth in the communities around the hospitals suggest that child populations have increased by 215%, 49%, and 77% during the same period at Kilgoris, Kisii, and Tabaka, respectively.

Emerging Infectious Diseases. 2002;8(6) © 2002 Centers for Disease Control and Prevention (CDC)

Cite this: Clinical Epidemiology of Malaria in the Highlands of Western Kenya - *Medscape* - Jun 01, 2002.

## Comments