Age-incidence and age-prevalence patterns

This report aims to:

Explore the relationship between age and incidence of malaria
Gain insights to help with future model development.

Age-incidence relationships

Griffin et al. (2014)

Data: prev_inc_griffin2014

The data come from an analysis by Griffin et al. (2014) describing age-specific patterns of disease at different transmission intensities. This analysis collates data from 14 studies containing prevalence and incidence estimates, as well as entomological inoculation rates (EIRs) from 23 distinct sites. The majority of the studies repoeted incidence data over a whole number of years, but where they did not, the authors adjusted the time at risk based on the length of the malaria season as reported in each publication. The studies that recorded cases using active case detection (ACD) were categorized into daily or weekly ACD, and the rest of the studies cases were recorded via passive case detection (PCD), i.e. on presentation at the health facility.

Below we can take a look at the variables included in the dataset:

head(prev_inc_griffin2014)

##   country        reference                  site_name site_index age0 age1
## 1 Senegal Diallo_etal_1998 Senegal_Diallo_Dakar_south         36    0    2
## 2 Senegal Diallo_etal_1998 Senegal_Diallo_Dakar_south         36    2   10
## 3 Senegal Diallo_etal_1998 Senegal_Diallo_Dakar_south         36   10   15
## 4 Senegal Diallo_etal_1998 Senegal_Diallo_Dakar_south         36   15   20
## 5 Senegal Diallo_etal_1998 Senegal_Diallo_Dakar_south         36   20  100
## 6 Senegal Diallo_etal_1998 Senegal_Diallo_Dakar_south         18    0    1
##         type numer denom case_detection    meanEIR     sd_hi     sd_low
## 1 prevalence     1   908     weekly ACD 0.05000001 0.1112771 0.02246645
## 2 prevalence    22  6245     weekly ACD 0.05000001 0.1112771 0.02246645
## 3 prevalence     4  3256     weekly ACD 0.05000001 0.1112771 0.02246645
## 4 prevalence    19  2869     weekly ACD 0.05000001 0.1112771 0.02246645
## 5 prevalence    14  6097     weekly ACD 0.05000001 0.1112771 0.02246645
## 6  incidence     0    21     weekly ACD 0.05000001 0.1112771 0.02246645
##   alpha(prop treated) beta(prop treated) plot_index           plot_title
## 1              4.6875            20.3125          1 Senegal, South Dakar
## 2              4.6875            20.3125          1 Senegal, South Dakar
## 3              4.6875            20.3125          1 Senegal, South Dakar
## 4              4.6875            20.3125          1 Senegal, South Dakar
## 5              4.6875            20.3125          1 Senegal, South Dakar
## 6              4.6875            20.3125          1 Senegal, South Dakar

The authors explored incidence by age in each site. They found that there is a pronounced peak in young children in some of the high transmission sites. In lower transmission sites the distribution of cases widens to include older ages. The magnitude and location of this peak does however vary across sites. The figure below plots incidence in <20 year olds across the different sites, with the panels ordered by increasing transmission intensity/EIR. Confidence intervals are calculated here using the CI_inc() function (from SIMPLEGEN package).

prev_inc_griffin2014 %>% 
  filter(type == "incidence") %>%
  mutate(age_mid = (age1 + age0) / 2,
         incidence = numer / denom) %>% 
  do(cbind(., CI_inc(.$numer, .$denom))) %>%
  filter(age_mid < 20) %>% 
  ggplot(aes(x = age_mid, y = incidence, group = case_detection, color = case_detection)) +
    geom_point() +
    geom_errorbar(aes(ymin = lower, ymax = upper)) +
    labs(x = "Age (in years)",
         y = "Incidence (clinical episodes per person per year") +
    facet_wrap(~reorder(plot_title, meanEIR), scales = "free_y") +
    expand_limits(x = 0, y = 0) +
    theme_minimal()

Shazia Ruybal-Pesántez

Last updated: 30 Mar 2023

Age-incidence relationships

Griffin et al. (2014)