Malaria misdiagnosis substantially contributes to the underestimation of global human brucellosis incidence

Abstract

Background

Brucellosis is a neglected and re-emerging zoonotic disease of nearly worldwide distribution. Recently, a new model utilizing publicly available data, voluntarily provided by countries to the World Organization for Animal Health (WOAH), estimated a global annual incidence of at least 1.62–2.10 million cases. However, global and regional knowledge gaps, partially attributed to lack of diagnostic capability and disease unawareness, result in misdiagnosis.

Methods

This study incorporated misdiagnosis to assess the potential impact on incidence estimates. Specifically, we examined the potential impact of misdiagnosing brucellosis as malaria. Incidence rates modeled from WOAH data comprised a baseline estimate for brucellosis rates. A range of potential misdiagnosis rates based on World Bank malaria data and scientific literature were utilized to estimate the number of misdiagnosed cases. We assumed a conservative misdiagnosis range of 0.25-4.00% based on clinical studies that reported a 4–11% misdiagnosis rate.

Results

We show that malaria misdiagnosis can significantly impact global annual incidence estimation of brucellosis. Rates of 0.25% increase brucellosis incidence by 0.34-5.4 M cases, while rates of 4% increase these estimates by 2.43-7.45 M cases.

Conclusions

Human brucellosis misdiagnosed as malaria can significantly impact global annual incidence estimates. Enhanced efforts are needed to identify misdiagnosed cases in countries where both diseases are endemic.

Background

Brucellosis, a zoonotic bacterial disease caused by Brucella spp., is known to affect multiple domestic animals and mammalian wildlife species and remains endemic in many parts of the world [1,2,3,4,5]. Of the currently recognized species, three are highly virulent to livestock and humans: Brucella abortus primarily infects cattle, B. melitensis primarily infects goats and sheep, B. suis primarily infects swine [4]. The disease in livestock is of significant economic importance and poses a public health risk [2, 3, 6]. Infection in ruminants can cause abortion, weak offspring, and infertility, as well as reduction of milk production [2, 4]. Conventional transmission of the bacteria amongst animals occurs during contact with infected reproductive secretions, placentas, or aborted fetuses [4, 7]. Disease in humans is typically characterized by non-specific influenza-like illness that can manifest as undulant fever, sweats, fatigue, and malaise, and can resemble some of the most commonly acquired infectious diseases in resource-limited regions, such as malaria and typhoid [1, 2, 4]. Furthermore, undulant fevers, arthritis, myocarditis, and neuropathies can occur amongst chronic cases of human brucellosis [1, 4]. Humans are generally exposed to Brucella spp. during consumption of unpasteurized milk products and the handling of contaminated tissues, such as aborted livestock placentas and other reproductive excretions [4]. These exposure pathways make raw milk product consumers, livestock owners, abattoir workers, and veterinarians at high risk for disease acquisition within endemic areas [4]. Recently, it has been estimated that at least 1.62–2.10 million new human cases occur each year worldwide [5]. However, this number is believed to be considerably lower than the actual incidence due to significant drawbacks associated with routine diagnostics and disease surveillance. These include lack of disease awareness by clinicians and public health workers, financial constraints that impede appropriate surveillance systems, and the unavailability of easy-to-implement diagnostic tests [3, 6].

To better assess the potential impact that human brucellosis misdiagnosis has on the annual brucellosis case incidence, this study uses a combination of publicly available human brucellosis incidence data reported for each country globally by the World Organization for Animal Health (WOAH) and incorporates a range of potential malaria misdiagnosis rates. Malaria was selected for this study because it is considered to be one of the most commonly misdiagnosed diseases in areas where both diseases are endemic. It also has the most complete annual incidence information available for symptomatically similar diseases. Our findings suggest that misdiagnosed cases significantly influence the underestimation of overall global annual human brucellosis incidence and should be considered when evaluating its disease impact.

Methods

The most recent estimates of global and regional annual human brucellosis incidence and reported malaria incidence were used for this study. Due to the sharp decline in information available during the COVID-19 pandemic [8], data from the five-year pre-pandemic timeframe (i.e., 2014–2018) were employed. Briefly, human brucellosis presence (i.e., endemicity) was based on a recent study that utilized three data sources: (1) reported livestock information that indicates the presence of B. abortus, B. melitensis, and B. suis among the 182 WOAH member states, (2) the reported human case count (RCC) data compiled by WOAH, which demonstrate the presence of human brucellosis within each country (without differentiation of the Brucella species causing disease), and (3) rural human population counts within these countries (i.e., those with the highest likelihood of contact with livestock) provided by the World Bank [5]. Malaria endemicity status was estimated on open-source information provided by the World Bank. Specifically, the “incidence of malaria (per 1,000 population at risk)” dataset [9] was used to determine the endemicity of the disease (i.e., official reports of the presence of malaria in humans, for at least 3 of the 5 years, by the World Bank). Subsequently, “high-risk” populations for both diseases were defined and identified. High-risk population identification for human brucellosis was based on scientific evidence that demonstrates an elevated risk of acquiring brucellosis in rural areas due to more frequent consumption of raw milk and direct handling of infected animals [4, 6, 10]. Therefore, in this study, the “high-risk” population was defined as the rural population within the areas where both brucellosis and malaria are known to be present. Rural populations within endemic areas were extracted by utilizing the World Bank datasets “Population Estimates And Projections” [11] and “Rural population (% of total population)” [12] to calculate rural population size from each of the countries (see Additional file 1). From these populations, the “incidence of malaria (per 1,000 population at risk)” dataset [9] was used to calculate incidence of malaria (see Additional file 1). To understand the global distribution of these cases, all countries were stratified into 4 regions: (1) Africa, (2) the Americas, (3) Asia, and (4) Europe (see Additional file 1). Oceania was excluded due to the inability to model this region’s human brucellosis incidence caused by: (1) incomplete reporting, (2) the small number of countries (7 total) and at-risk population numbers (7.6 million) involved, and (3) the small number of RCCs (132 RCCs over five years) compared to the rest of the world [5].

Calculation of misdiagnosis was performed by taking the 5-year average of reported malaria cases from each country and subsequently assigning different scenarios (rates) in which misdiagnosis could occur (see Additional file 1). Specifically, rates were set incrementally at 0.25% (1 out of 400 reported cases of malaria misdiagnosed), 0.50% (1/200), 1.00% (1/100), 2.00% (1/50), and 4.00% (1/25). A conservative range of 0.25–4.00% was selected based on previous clinical studies where patients initially diagnosed as having malaria were subsequently correctly diagnosed as having brucellosis; importantly, in these same studies, a 4–11% misdiagnosis rate was calculated [13, 14]. Cases resulting from these misdiagnosis rates were then allocated to their respective country and added to the regional Bayesian hierarchical estimates of human brucellosis incidence that were previously modeled [5] (see Additional file 1).

Model validation was conducted using Kenya. This country was chosen due to the presence of clinical studies that address the misdiagnosis of human brucellosis as malaria, a documented history of brucellosis surveillance neglect prior to 2012 [15], and creation in 2012 of a national One Health (OH) initiative and office called the Zoonotic Disease Unit (ZDU). The ZDU prioritizes brucellosis surveillance and control [13, 15,16,17] and has generated new data regarding current human brucellosis status in the country. To determine whether the new number of cases from this study provided an accurate representation of the number of actual misdiagnosed cases, a comparison was made between (1) modeled estimates of 0.25–4.00% of malaria cases reported pre-prioritization (2008–2012) added to the number of reported human brucellosis cases pre-prioritization, and (2) reported annual brucellosis incidence between 2014 and 2018. The expectation was that the newly calculated numbers generated by the model (2008–2012) would be approximately equal to the number of reported brucellosis cases post-prioritization (2014–2018). Since it is well established that brucellosis is endemic to Kenya, and in the absence of factors like vaccination or test-and-slaughter programs, the endemic equilibrium prevalence/incidence is unlikely to change significantly over this timeframe. Therefore, the data used should remain applicable over time.

To ensure best reporting practices, the study was conducted utilizing the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) [18]. All datasets are available upon request.

Results

The most recent global incidence of human brucellosis estimate is between 1.62 and 2.10 million new cases per year [5]. This number is significantly higher than the previously presumed 500,000 new cases [5, 8], which highlights the value of utilizing improved epidemiological and statistical approaches to better predict incidence. However, the more recent approximation is based solely on voluntarily reported, publicly available human and livestock data compiled by WOAH [5]. Detailed analysis of the data also highlighted incomplete reporting at global and regional levels, which can be partially attributed to a lack of diagnostic capability and disease awareness by local clinicians. These factors led to misdiagnosis and underreporting [5, 8]. It is well established that human brucellosis is commonly misdiagnosed as malaria, typhoid, or other acute febrile illnesses in locations where these diseases are endemic [13, 14]. Misdiagnosis usually occurs due to the nonspecific signs and symptoms shared between these diseases, including fever, chills, sweats, aches, and lack of energy [4]. To include misdiagnosis as a parameter, a variable was incorporated in the incidence model that represented a series of different malaria misdiagnosis rates. Utilizing this new information, the revised model provides updated global and regional estimates of human brucellosis.

Global distribution of countries at risk for both malaria and human brucellosis (2014–2018)

Figure 1 represents the location of (1) countries in which livestock brucellosis is endemic [5], (2) countries in which malaria is endemic, and (3) countries in which both livestock brucellosis and malaria are endemic. Between 2014 and 2018, 73.7% (129/175) of the countries in the world were endemic for at least 1 of the 3 zoonotic Brucella species in livestock (Fig. 1A). Among African countries, 88.7% (47/53) reported the presence of the disease with the exception of Central African Republic and the island countries reporting disease freedom; see Fig. 1A. Within Asia, 83.7% (41/49) of countries reported disease presence with the exception of Bahrain, Brunei Darussalam, Cyprus, Japan, Maldives, Singapore, Taiwan, and Uzbekistan reporting disease freedom; see Fig. 1A. Within the Americas, 77.4% (24/31) of nations were endemic, mainly within central and south America, with only Canada and most of the West Indies remaining disease free (Fig. 1A). In Europe, 40.5% of countries (17/42) maintain infections in livestock, primarily in the Mediterranean/Eastern countries, as well as Belgium, Germany, and France (Fig. 1A). Countries that were identified to be free of brucellosis in livestock were excluded from the remainder of the analysis. Figure 1B represents the locations where malaria is endemic and in which the disease is mainly spread across tropical and sub-tropical parts of the world (Fig. 1B). Worldwide, 48.0% (84/175) of the countries are endemic with 83.0% (44/53) of Africa, 58.1% (18/31) of the Americas, 44.9% (22/49) of Asia, and 0.0% (0/42) of Europe (Fig. 1B). Malaria disease-free zones are primarily located in Europe, North America, Northern Africa and Asia, and the southern most countries of South America (Fig. 1B). Countries identified to be free of malaria were excluded from the remainder of the analysis. Following the exclusion of all the countries free of either brucellosis and/or malaria, Fig. 1C demonstrates that 95.2% (80/84) of countries endemic with malaria are also endemic with brucellosis in livestock. Specifically, 100% (22/22) of the countries in Asia, 95.5% (42/44) of the countries in Africa, and 88.9% (16/18) of the countries in the Americas were endemic with both diseases (Fig. 1C). These are the countries where human brucellosis is most likely to be misdiagnosed as malaria.

Although our methodology does not modify the incidence rates of brucellosis in Europe and other countries where malaria is absent, their previously estimated cases of human brucellosis are included in the baseline global annual incidence rates.

Fig. 1

Global distribution of countries reporting (A) zoonotic Brucella spp. at least 3 times within the 5-year period (Yellow), (B) malaria (Blue), and (C) brucellosis and malaria (Red) (2014–2018). White = Do not report the presence of the respective disease or excluded (i.e., Oceania) due to the inability to model human brucellosis incidence in this region caused by (1) incomplete reporting, (2) the small number of countries and at-risk population numbers involved, and (3) the small number of reported case counts compared to the rest of the world

Potential misdiagnosis of human brucellosis as malaria (2014–2018)

We identified the number of malaria cases within rural populations in countries that were endemic for both brucellosis and malaria (Fig. 2A). Between the years 2014 and 2018, there were an estimated 133.78 million reported cases of malaria in rural areas worldwide (Fig. 2A). In Africa, 123.21 million new cases were reported, accounting for 92.17% (123.21/133.78 million) of the total number of people diagnosed worldwide (Fig. 2A). These cases originated only from the sub-Saharan sub-region because countries in North Africa are free of malaria (Fig. 1B). Asia accounted for 7.55% of malaria cases (10.10/133.78 million), and the Americas for 0.27% (0.37/133.78 million). These data highlight the significant influence that misdiagnosis of malaria can have within the African region as well as the likely influence that Africa has on the overall number of human brucellosis cases globally (Fig. 2A).

Fig. 2

(A) Estimated number of malaria cases per year, within rural populations in countries that reported the presence of brucellosis in livestock by region (2014–2018). (B) Calculated 5-year average number of human brucellosis cases potentially misdiagnosed as malaria, at different rates of misdiagnosis (0.25–4.00%) by region. (C) Cumulative global and regional 5-year average incidence of human brucellosis (cases potentially misdiagnosed as malaria added to modeled baseline estimates) (2014–2018). Solid lines represent misdiagnosis estimates at different rates and dashed lines represent Bayesian modeled baseline estimates based on available WOAH data [5]

The estimation of human brucellosis cases potentially misdiagnosed as malaria was subsequently calculated by utilizing the reported malaria cases from each country and assigning different rates in which misdiagnosis might potentially occur (0.25–4.00%). Figure 2B represents the estimated number of human brucellosis cases misdiagnosed as malaria each year, at the global and regional level. Interestingly, a 0.25% misdiagnosis rate increases the global incidence estimate by 0.34 million new cases per year. This increase is primarily due to the contribution of Africa (0.31 million cases; see Fig. 2B). At the upper end of the spectrum, a 4.00% misdiagnosis rate increases the global incidence estimate by 5.40 million new cases per year. A majority (4.93 million) of the misdiagnosed cases are estimated as coming from Africa (Fig. 2B). Interestingly, Asia (0.03–0.40 million) and the Americas (0.00–0.01 million) have a much lower influence in the overall count when misdiagnosis is considered, compared to Africa (Fig. 2B).

To calculate the new global and regional human brucellosis incidence estimates, the newly “recognized” cases described above (Fig. 2B solid lines) were added to the unadjusted baseline estimates (Fig. 2B dashed lines and Fig. 2C). Utilizing the Bayesian modeled baseline estimates, the revised estimate of global incidence increased to approximately 2.43–7.45 million new cases per year (Fig. 2C). The influence on the total number of potential new human cases globally, in order by region, is Africa (0.47–5.40 million), followed by Asia (1.62–2.03 million), and the Americas (0.02–0.04 million). The minor difference between the global and Africa results, coupled along with the substantial difference of the global estimates from Asia and the Americas, further highlights that Africa is the primary region of interest when considering the influence of malaria misdiagnosis on the overall global annual incidence of brucellosis. Although misdiagnosis as malaria likely occurs in other regions, the only region that substantially impacts the overall outcome of the global annual incidence is Africa. It is important to note that without misdiagnosis, Asia accounts for the majority of estimated human brucellosis cases worldwide (Fig. 2C). However, as the rate of misdiagnosis increases above 1%, Africa is likely responsible for a larger number of new cases each year (Fig. 2C).

Model validation

To validate the model, a sub-Saharan African country (Kenya) was selected. Selection was based on the endemicity of both human brucellosis and malaria, the availability of scientific studies confirming misdiagnosis of brucellosis in local healthcare facilities, and a recent national brucellosis surveillance, prevention, and control initiative through the ZDU. The ZDU has trained physicians to better identify and report new cases. In addition, the ZDU has attempted to establish coordination structures and partnerships that promote OH in the country, to strengthen surveillance, detection, prevention, and control of specific zoonoses in both humans and animals, and to encourage research, field training, and mentorship of veterinary, medical, and public health trainees [17]. Two studies within this region have demonstrated a 4–11% misdiagnosis rate based on physicians initially diagnosing patients with malaria and subsequently correcting their diagnoses to brucellosis [13, 14]. Figure 3 represents (1) modeled estimates of human brucellosis cases misdiagnosed as malaria between 2008 and 2012, when human brucellosis reporting was not conducted within Kenya (solid black line), (2) actual reported human brucellosis cases provided by the country to WOAH (2014–2018) (red dashed line represents the average), along with (3) an area showing where the model would fit if misdiagnosis was between 4 and 11%, as shown in scientific literature [13, 14] (green shaded area).

Fig. 3

Model validation. Between 2008–2012, 2.1 million cases of malaria and no cases of human brucellosis were reported in rural Kenya. (a) The solid black line represents modeled 5-year average estimates of human brucellosis cases misdiagnosed as malaria between 2008 and 2012, prior to brucellosis prioritization (0.25–16.00% of 2.1 million malaria cases = 5,286–338,286). (b) The green zone represents the area where the model would fit if prior to brucellosis prioritization misdiagnosis was between 4–11% of 2.1 million malaria cases (84,571–232,572), as shown in scientific literature [13, 14]. (c) The red dashed line displays the average reported human brucellosis cases between 2014–2018 provided by the country to the World Organization for Animal Health (WOAH), after brucellosis prioritization (106,525 = 5.03% of 2.1 million malaria cases) fitting within the green zone. (d) The area under the green zone represents the 5-year average estimates of human brucellosis cases misdiagnosed as malaria between 2008 and 2012, prior to brucellosis prioritization (0.25–4.00% of 2.1 million malaria cases = 5,286–84,571), indicating a conservative modeled estimate

In Kenya, between 2008 and 2012, there were no cases of human brucellosis reported to WOAH, while 2.1 million cases of malaria were reported by public health authorities. Within the years immediately following the OH initiative and the creation of the ZDU (2014–2018), a significant number of new brucellosis cases were diagnosed (annual incidence ranging between 96,571 and 126,836 per year; average = 106,525). During the same period 2.7 million cases of malaria were reported [9]. When the model was applied using the 0.25–4.00% misdiagnosis rate during this timeframe, only between 5,286 and 84,571 new cases were calculated, suggesting that these rates should be considered an underestimate (Fig. 3). However, when a 5.03% rate is applied, it coincides with the average cases (106,525) between 2014 and 2018 (Fig. 3). This number is consistent with studies that estimate a 4–11% misdiagnosis rate [13, 14]. Strikingly, if the misdiagnosis rate was 11%, as calculated at the upper end in those studies, the number of newly identified cases would be as high as 232,572 per year (Fig. 3). These data suggest the model is likely not overestimating the actual number of misdiagnosed cases. Unfortunately to this date, there are no other countries in Africa that have generated data similar to Kenya’s to further validate the use of this approach.

Discussion

This study provides insight into the potential effect that misdiagnosis of human brucellosis as malaria can have, highlighting the current underestimation of new human cases worldwide. Fortunately, modeling has become an essential tool for enhancing the analysis of imperfect data, enabling the estimation of key epidemiological parameters even when direct measurements are unavailable, or data are limited and imperfect [19]. An incorporation of potentially misdiagnosed cases to the previously modeled estimates was prompted by our field practice of veterinary and public health within resource-limited settings. This experience presented an eye-opening reality in which human brucellosis in the tropics and subtropics is commonly misdiagnosed. These observations are supported by a growing body of scientific literature that demonstrates an inability of physicians to correctly diagnose this endemic bacterial disease [13, 14].

Strikingly, this new estimate reveals that at least 2.43–7.45 million new cases likely occur each year globally, compared to the previous estimates of 1.62–2.10 million based on self-reported data provided by WOAH [5]. It is important to mention that the 0.25–4.00% misdiagnosis rates employed in this study are intended to represent plausible rates of misdiagnosis, and the actual rate can be modified amongst countries and regions based upon local conditions and knowledge. Probabilistic modeling could strengthen the robustness of the findings; however, incorporating the 5–11% range into such a model would likely inflate the results, causing the lower bound estimate to surpass the upper bound of the 0.25–4.0% range. Instead, we used a range of potential misdiagnosis rates to demonstrate the increasing impact of misdiagnosing brucellosis as malaria. The 0.25–4.0% range reflects our uncertainty about what we consider the lower end of the misdiagnosis scale.

A limitation of this work is that the model does not take into account other causes of diagnostic errors which routinely occur, including misdiagnosis of other infectious diseases (e.g., typhoid, meningitis, and pneumonia), the lack of disease awareness that prevents accurate diagnosis (e.g., physicians do not consider the possibility of brucellosis), sub-standard assays (e.g., low sensitivity and specificity, or incorrect usage), and inadequate reporting structures (e.g., adequately diagnosed individuals not accounted for) within resource-limited settings [1, 20]. Additionally, while malaria and brucellosis have distinct case definitions, their signs and symptoms are often similar. The US CDC provides internationally accepted case definitions, requiring the detection and specific identification of the malaria parasite or Brucella bacterium and/or confirmation through a nucleic acid test [21,22,23]. However, both diseases exhibit non-specific symptoms that overlap with other acute febrile illnesses, such as fever, chills, sweats, headaches, and muscle pain [21,22,23]. Moreover, many resource-limited countries in sub-Saharan Africa lack brucellosis surveillance and diagnostic capabilities, including culture testing, which is expensive, technically challenging, has low sensitivity, and poses safety risks without adequate resources, equipment, and trained personnel. Consequently, many countries have not adopted the internationally accepted case definition for brucellosis. To our knowledge, no reliable data exist on the extent to which physicians apply proper case definitions for either brucellosis or malaria. Moreover, given the extremely high prevalence of malaria, many co-infections likely go unidentified and undocumented.

It is worth highlighting that human brucellosis incidence is not routinely reported across Africa [8] and when tests are conducted, low sensitivity and specificity can impact the results of studies by inflating or deflating the reported incidence. Although disease misdiagnosis was incorporated into our estimates, the assumed misclassification rates of 0.25–4.00% might be lower than the actual rates, which could also vary across regions. These estimates of incidence should be interpreted with caution, because every 1.00% increase in misdiagnosis in Africa adds an additional 1.23 million new cases. For example, in a study on febrile patients that seek medical treatment in Kenya, 11% of the total cases initially diagnosed as malaria were later identified as brucellosis [13]. Similar findings have been reported from hospitals in Tanzania [14], which provides evidence that the misdiagnosis rate for brucellosis could be much higher than 4%. If these higher misdiagnosis rates are correct and validly applied more broadly, global brucellosis incidence rates could be in the tens of millions. The rates used in this model are intended to represent an average misdiagnosis rate across all countries where both diseases are endemic and are not intended to imply that each country has the same rate of misclassification.

Model validation was used to demonstrate that a comprehensive surveillance strategy can help identify cases that were previously misclassified when testing was not performed. If data from another country endemic to both malaria and brucellosis were available, particularly including information from before and after the establishment of a brucellosis surveillance program, we would have included it as well. Unfortunately, to the best of our knowledge, this information is unavailable. Given that testing is not conducted in most of Africa, this validation may be applicable to other sub-Saharan African countries endemic to both malaria and brucellosis. Accordingly, to account for the possibility that other endemic countries without control programs and brucellosis testing may have an endemic equilibrium significantly lower than Kenya’s on average, a range of 0.25–4.00% was used.

Interestingly, in some situations, human brucellosis outweighs malaria among acute febrile patients. For example, a cross-sectional study in Ethiopia indicated that 15.8% of febrile patients had brucellosis, while 4.3% had malaria [24]. Furthermore, other countries in Africa have a human seroprevalence similar to or higher than Kenya (i.e., as high as 10.8% in the at-risk population), which might suggest similar or higher annual incidence. For example, at-risk populations in South Africa have recently shown seroprevalence as high as 10.7% [25], and a meta-analysis indicates the seroprevalence in Nigeria is as high as 17.6% [26]. Yet, these countries do not actively prioritize and routinely report human brucellosis incidence [15]. To elucidate this, between 2014 and 2018, Nigeria acknowledged the presence of human brucellosis but did not report any cases [15], while it had the highest malaria rate globally, with approximately 28 million cases reported annually in rural areas [9, 11, 12]. A misdiagnosis rate of 0.25–4.00% would translate to roughly 70,000 to 1.12 million brucellosis cases per year. In endemic countries, a balance must be maintained with enough new cases to sustain equilibrium, as births add new non-infected individuals and deaths remove the infected. In Nigeria, a rough estimate suggests that a misdiagnosis rate of 2.34% would be required to sustain a prevalence of 17.6%. At a 0.25% misdiagnosis rate, 70,000 new cases would only maintain a prevalence of 0.02%. Furthermore, in South Africa, which typically reports fewer than one human brucellosis case annually [15], the number of new brucellosis cases needed to sustain a 10.7% prevalence would surpass the total reported malaria cases. A misdiagnosis rate of 0.25–4.00% in South Africa would equate to approximately 85 to 1,358 cases per year. These examples further demonstrate that the model likely does not overestimate misdiagnosed cases, with Nigeria’s 2.34% fitting well within the modeled range, and more brucellosis cases needed to maintain endemicity in South Africa than the total reported malaria cases.

Although global funds for malaria prevention, control, and elimination are well documented (totaling $3.3 billion USD in 2020), we are unaware of an estimate of annual human brucellosis expenditures. When compared to the overall number of new malaria cases (225 million urban and rural), 7 million new human brucellosis cases (3.1% of malaria) may seem slight, but this incidence is in the range of the estimates of typhoid fever (9.2 million) [27], meningitis (2.8 million) [28], and other acute febrile illnesses of importance. If 3.1% of the funds allocated to malaria were redirected towards brucellosis surveillance, prevention, and control, there would be an annual investment of $102.3 million. Unfortunately, the annual investment to combat the human brucellosis burden is likely to be significantly less. It is our hope that this analysis will result in an awareness that misdiagnosis can have a major impact on the outcome of regional and global estimates of the annual human brucellosis incidence. Additionally, commitment of resources is needed to train physicians and public health personnel on identification and differentiation of these diseases, as well as to cultivate better diagnostic capabilities worldwide. If ignored, the continued neglect of this disease will perpetuate the underestimation of the population affected by brucellosis and will result in a continued waste of resources on the treatment of misdiagnosed illnesses, thus prolonging human suffering.

Conclusions

Human brucellosis misdiagnosis has a significant impact on estimated global annual incidence. Our results suggest that within Africa, if 1 patient out of every 100 diagnosed with malaria actually has brucellosis, then the global baseline estimate of human brucellosis incidence should be amplified by 1.2 M cases. Increased emphasis must be placed on the correct disease diagnosis in countries that are endemic for both malaria and brucellosis.