Parasites are frequently used to test ecological and evolutionary hypotheses of coevolution, sexual selection, host immunocompetence, and pathogen virulence (Hamilton, 1982, Woolhouse et al., 1997, Otto and Nuismer, 2004, Mayer et al., 2016). Avian haemosporidian blood parasites are a common model system for these studies because they often infect multiple host species (Ricklefs and Fallon, 2002, Beadell et al., 2006, Fecchio et al., 2021), and hosts vary in their responses to infection (Atkinson and van Riper, 1991, Valkiūnas, 2005). Some hosts experience high parasite intensities and mortality when parasitized, while others suffer little morbidity and mortality, and many species maintain chronic low intensity infections for a number of years (van Riper et al., 1986, Valkiūnas, 2005). Variation in infections and host responses to infections influence transmission, evolution, and disease risk within host communities (Woolhouse et al., 1990, Woolhouse et al., 1997, Regoes et al., 2000). A crucial part of studying host-parasite interactions is accurately measuring a host’s infection status and the infection intensity.

Measuring infection prevalence and intensity is important in determining which host species are more likely to become infected (Beadell et al., 2006, Maria et al., 2009, Svensson-Coelho et al., 2013) and, for vector-borne diseases, estimating host infectiousness for vectors (Komar et al., 2003, Mackinnon and Read, 2004, Kilpatrick et al., 2007, Pigeault et al., 2015). The probability of detecting infection increases when infection intensity is higher, and infection intensity varies both among hosts and over the course of infection (van Riper et al., 1986, Moens et al., 2016). The parasitemia (the fraction of red blood cells infected) for most malaria parasites is highest during the acute phase of infection, which lasts for a few weeks starting approximately 1 week after infection (Atkinson et al., 2000, Atkinson et al., 2001, Dimitrov et al., 2015), and hosts may suffer morbidity or mortality during this time and thereafter. Subsequently, species and individuals will remain chronically infected, often at much lower infection intensities, for months or years (Jarvi et al., 2002, Valkiūnas, 2005). Accurately estimating infection prevalence in a species or population can be challenging because heavily infected hosts can have decreased activity (Yorinks and Atkinson, 2000), which may lead to undersampling, especially in studies that capture flying birds with mist nets (Valkiūnas, 2005, Mukhin et al., 2016), and low-level chronic infections can be difficult to document, depending on the detection technique used (Jarvi et al., 2002).

Parasite detection methods vary in their sensitivity (probability of detecting true positives) and in the information they provide. Malaria researchers have employed different detection methods over time, with earlier studies using microscopy of stained blood smears and more recent studies primarily using PCR-based molecular techniques (Garnham, 1966, Fallon et al., 2003, Zehtindjiev et al., 2008). Microscopy provides useful information about the morphology and life stages of haemosporidians (Jarvi et al., 2002, Fallon et al., 2003, Valkiūnas et al., 2008), but it generally has lower sensitivity than molecular techniques for detecting infection. This is especially true for low infection intensities, because the number of red blood cells that can be reasonably screened are far less than the numbers in an extracted DNA sample (Jarvi et al., 2002, Fallon et al., 2003, Fallon and Ricklefs, 2008). As a result, many studies now use either PCR (Beadell and Fleischer, 2005) or quantitative PCR (qPCR) to measure infection prevalence and the latter to estimate infection intensity (Neddermeyer et al., 2023, Paxton et al., 2023). However, integrating historical estimates of prevalence and intensity based on microscopy with more recent disease surveys that use qPCR requires a way to accurately integrate the measurements from these two methods.

Here, we develop a statistical relationship between Plasmodium relictum parasitemia quantified by microscopy and by qPCR to integrate past and current studies of avian malaria and examine patterns of infection among species. We analyzed blood samples both from experimentally infected birds and wild-caught birds in Hawai'i (USA), and examined the probability of detecting infection in eight common species found in Hawai'i that differ in parasitemia. We re-examined historical patterns of infection prevalence among these species using a large previous study that assessed infection using microscopy (van Riper et al., 1986). Finally, we combined parasitemia data from experimental infections and wild caught birds to assess whether parasitemias estimated from birds caught with mist nets are biased towards low results because birds with high parasitemias are less likely to be caught and sampled.