Gastrointestinal nematodes (GINs) are one of the major causes of productive losses in small ruminant farming in Brazil, especially the species Haemonchus contortus, which causes haemonchosis (Maia et al., 2013). Haemonchus contortus adheres to the mucosa of the abomasum and causes lesions that alter the secretion of hydrochloric acid and pepsinogen, promoting failure in the animal's protein digestion. However, the greatest damage is due to the fact this helminth is hematophagous. As a result, sheep with massive infection by this nematode develop anemia, causing a loss of productivity and a high mortality rate (Muchiut et al., 2019). Parasite resistance to anthelmintics is one of the biggest challenges in controlling GINs (Pereira et al., 2016). One of the main reasons for anthelmintic resistance development is the strategy of treating all animals at frequent fixed intervals (Kaplan et al., 2004). However, targeted selective treatments (TST) can be applied at the farm level to reduce the number of treatments administered, thereby increasing the proportion of the worm population in refugia (Burke et al., 2007, McBean et al., 2021), i.e., the proportion of the GIN population that is not exposed to anthelmintics (van Wyk, 2001).

Accordingly, TST is an efficient strategy to slow the establishment of parasite resistance (Costa et al., 2011). However, the main limitation to the adoption of this approach is the lack of efficient and economical means of identifying the animals requiring treatment (Kaplan et al., 2004). Therefore, tools for rapid and accurate diagnosis of anemia are essential for the efficiency of TST and the fight against haemonchosis (Muchiut et al., 2019).

Hematocrit is the main method for the clinical diagnosis of anemia in animals. This technique consists of centrifuging the sample and measuring the packed cell volume (PCV) of blood with the separation and determination of the fraction of red blood cells (erythrocytes) present in whole blood. PCV is considered an accurate method, but it requires sample preparation and laboratory tests (González and Silva, 2008). The time frame for applying the method is too long for TST since the report only reaches the producer after clinical analysis in the laboratory. Thus, it is not possible to diagnose the animals already contained in the pens and treat them immediately and selectively.

The FAffa MAlan CHArt (FAMACHA) method has been used to detect the degree of anemia in the field. Its results are associated with the color of the conjunctival mucosa of animals (van Wyk and Bath, 2002, van Wyk and Mayhew, 2013). However, technical training for the correct use of the card, and adequate time and implementation of the technique in situ are essential requirements for the success of the technique (Chagas, 2007). Thus, this method, although practical and useful, may be subject to failure due to subjectivity, reduced sensitivity or specificity (Burke et al., 2007).

Inaccurate diagnosis can lead to the spread of hemonchosis, loss of productivity, deaths, or lead to the excessive use of anthelmintics, favoring development of resistance (Pereira et al., 2016). Thus, the establishment of TST requires analytically accurate, fast and accessible diagnostic technology that can be applied in the field at the time of sample collection. Near-infrared spectroscopy (NIRS) employs electromagnetic radiation in the near-infrared region, with photons having wavelengths between 750 nm and 2,500 nm, corresponding to a wave number range between 4,000 cm-1 and 13,333 cm-1. The NIRS spectra are formed by the absorption of radiation by the samples, promoting the excitation of overtones and combinations of vibrational modes associated with the C-H, N-H, S-H, and O-H bonds. Mathematical methods such as multivariate analysis are necessary to ascertain the chemical information present in the spectra (Pasquini, 2003).

The NIRS technique has been widely used to determine the major constituents of various materials of agricultural interest since the 1980s, including the nutritional value of animal diets (Liu et al., 2017). It has also been evaluated for disease diagnosis in humans (Sakudo et al., 2005, Sakudo et al., 2009, Bahmani et al., 2009, Carmona et al., 2013, Paraskevaidi et al., 2018, Li and Wang, 2019, Ozawa et al., 2019, Raja et al., 2019) and in livestock (Andueza et al., 2014, Santos-Rivera et al., 2021). Regarding GIN infection, NIRS was studied to estimate the internal parasite burden of sheep (Dixon, R., Colditz, I., LeJambre, L., Lyndal-Murphy, M., Basier, B., 2013. Potential of near infrared reflectance (NIR) spectroscopy of faeces to estimate the internal parasite burden of sheep. 16th International Conference on Near Infrared Spectroscopy, IRSTEA - France Institut National de recherche en sciences et technologies pour l’environment et l’agriculture, Montpeiller, France, pp. 535–538), and recently it was also investigated for detection of H. contortus eggs or blood in sheep feces (Kho et al., 2020a, 2020b, 2020c). Therefore, this study aimed to predict PCV values from the collection of NIR spectra and create a classification and diagnostic model using the parameter PCV values, eggs per gram of feces (EPG) and mean daily weight gain (DWG).