Metagenomic Detection of Bacterial Zoonotic Pathogens among Febrile Patients, Tanzania, 2007-20091.

Bacterial zoonoses are established causes of severe febrile illness in East Africa. Within a fever etiology study, we applied a high-throughput 16S rRNA metagenomic assay validated for detecting bacterial zoonotic pathogens. We enrolled febrile patients admitted to 2 referral hospitals in Moshi, Tanzania, during September 2007-April 2009. Among 788 participants, median age was 20 (interquartile range 2-38) years. We performed PCR amplification of V1-V2 variable region 16S rRNA on cell pellet DNA, then metagenomic deep-sequencing and pathogenic taxonomic identification. We detected bacterial zoonotic pathogens in 10 (1.3%) samples: 3 with Rickettsia typhi, 1 R. conorii, 2 Bartonella quintana, 2 pathogenic Leptospira spp., and 1 Coxiella burnetii. One other sample had reads matching a Neoerhlichia spp. previously identified in a patient from South Africa. Our findings indicate that targeted 16S metagenomics can identify bacterial zoonotic pathogens causing severe febrile illness in humans, including potential novel agents.

Prevalence and risk factors for Q fever, spotted fever group rickettsioses, and typhus group rickettsioses in a pastoralist community of northern Tanzania, 2016-2017.

BACKGROUND: In northern Tanzania, Q fever, spotted fever group (SFG) rickettsioses, and typhus group (TG) rickettsioses are common causes of febrile illness. We sought to describe the prevalence and risk factors for these zoonoses in a pastoralist community. METHODS: Febrile patients ≥2 years old presenting to Endulen Hospital in the Ngorongoro Conservation Area were enrolled from August 2016 through October 2017. Acute and convalescent blood samples were collected, and a questionnaire was administered. Sera were tested by immunofluorescent antibody (IFA) IgG assays using Coxiella burnetii (Phase II), Rickettsia africae, and Rickettsia typhi antigens. Serologic evidence of exposure was defined by an IFA titre ≥1:64; probable cases by an acute IFA titre ≥1:128; and confirmed cases by a ≥4-fold rise in titre between samples. Risk factors for exposure and acute case status were evaluated. RESULTS: Of 228 participants, 99 (43.4%) were male and the median (interquartile range) age was 27 (16-41) years. Among these, 117 (51.3%) had C. burnetii exposure, 74 (32.5%) had probable Q fever, 176 (77.2%) had SFG Rickettsia exposure, 134 (58.8%) had probable SFG rickettsioses, 11 (4.8%) had TG Rickettsia exposure, and 4 (1.8%) had probable TG rickettsioses. Of 146 participants with paired sera, 1 (0.5%) had confirmed Q fever, 8 (5.5%) had confirmed SFG rickettsioses, and none had confirmed TG rickettsioses. Livestock slaughter was associated with acute Q fever (adjusted odds ratio [OR] 2.54, 95% confidence interval [CI] 1.38-4.76) and sheep slaughter with SFG rickettsioses case (OR 4.63, 95% CI 1.08-23.50). DISCUSSION: Acute Q fever and SFG rickettsioses were detected in participants with febrile illness. Exposures to C. burnetii and to SFG Rickettsia were highly prevalent, and interactions with livestock were associated with increased odds of illness with both pathogens. Further characterisation of the burden and risks for these diseases is warranted.

Serological and molecular evidence of Brucella species in the rapidly growing pig sector in Kenya.

BACKGROUND: Brucellosis is an emerging yet neglected zoonosis that has been reported in Kenya. Epidemiological data on brucellosis in ruminants is readily accessible; however, reports on brucellosis in pigs remain limited. This study sought to detect Brucella infection in pig serum by both serological and molecular techniques. Serum from 700 pigs randomly collected at a centralized abattoir in Nairobi region, Kenya were screened in parallel, using both Rose Bengal Test (RBT) and competitive Enzyme-Linked Immuno-sorbent Assay (cELISA) for antibodies against Brucella spp. All sera positive by RBT and 16 randomly selected negative samples were further tested using conventional PCR targeting bcsp31 gene and real-time PCR (RT-PCR) assays targeting IS711 and bcsp31 genes. RESULTS: A prevalence of 0.57% (n = 4/700) was estimated using RBT; none of these samples was positive on cELISA. All RBT positive sera were also positive by both PCRs, while two sero-negative samples also tested positive on RT-PCR (n = 6/20). Brucella abortus was detected in four out of the six PCR positive samples through a real-time multiplex PCR. CONCLUSION: The detection of antibodies against Brucella spp. and DNA in serum from slaughterhouse pigs confirm the presence of Brucella in pigs. Therefore, investigation of the epidemiology and role of pigs in the transmission of brucellosis in Kenya is needed. Further targeted studies would be useful to systematically quantify and identify the spp. of Brucella in pigs.

Transmission ecology of canine parvovirus in a multi-host, multi-pathogen system.

Understanding multi-host pathogen maintenance and transmission dynamics is critical for disease control. However, transmission dynamics remain enigmatic largely because they are difficult to observe directly, particularly in wildlife. Here, we investigate the transmission dynamics of canine parvovirus (CPV) using state-space modelling of 20 years of CPV serology data from domestic dogs and African lions in the Serengeti ecosystem. We show that, although vaccination reduces the probability of infection in dogs, and despite indirect enhancement of population seropositivity as a result of vaccine shedding, the vaccination coverage achieved has been insufficient to prevent CPV from becoming widespread. CPV is maintained by the dog population and has become endemic with approximately 3.5-year cycles and prevalence reaching approximately 80%. While the estimated prevalence in lions is lower, peaks of infection consistently follow those in dogs. Dogs exposed to CPV are also more likely to become infected with a second multi-host pathogen, canine distemper virus. However, vaccination can weaken this coupling, raising questions about the value of monovalent versus polyvalent vaccines against these two pathogens. Our findings highlight the need to consider both pathogen- and host-level community interactions when seeking to understand the dynamics of multi-host pathogens and their implications for conservation, disease surveillance and control programmes.

Mobile Phones As Surveillance Tools: Implementing and Evaluating a Large-Scale Intersectoral Surveillance System for Rabies in Tanzania.

Katie Hampson and colleagues describe their experience of developing and deploying a large-scale rabies surveillance system based on mobile phones in southern Tanzania.

Integrating serological and genetic data to quantify cross-species transmission: brucellosis as a case study.

Epidemiological data are often fragmented, partial, and/or ambiguous and unable to yield the desired level of understanding of infectious disease dynamics to adequately inform control measures. Here, we show how the information contained in widely available serology data can be enhanced by integration with less common type-specific data, to improve the understanding of the transmission dynamics of complex multi-species pathogens and host communities. Using brucellosis in northern Tanzania as a case study, we developed a latent process model based on serology data obtained from the field, to reconstruct Brucella transmission dynamics. We were able to identify sheep and goats as a more likely source of human and animal infection than cattle; however, the highly cross-reactive nature of Brucella spp. meant that it was not possible to determine which Brucella species (B. abortus or B. melitensis) is responsible for human infection. We extended our model to integrate simulated serology and typing data, and show that although serology alone can identify the host source of human infection under certain restrictive conditions, the integration of even small amounts (5%) of typing data can improve understanding of complex epidemiological dynamics. We show that data integration will often be essential when more than one pathogen is present and when the distinction between exposed and infectious individuals is not clear from serology data. With increasing epidemiological complexity, serology data become less informative. However, we show how this weakness can be mitigated by integrating such data with typing data, thereby enhancing the inference from these data and improving understanding of the underlying dynamics.

Complete genome sequence of Brucella abortus isolated from a human blood culture sample in Tanzania.

Brucella abortus causes infections in humans and livestock. Bacterial isolates are challenging to obtain, and very little is known about the genomic epidemiology of this species in Africa. Here, we report the complete genome sequence of a Brucella abortus isolate cultured from a febrile human in northern Tanzania.

Brucellosis in low-income and middle-income countries.

PURPOSE OF REVIEW: Human brucellosis is a neglected, underrecognized infection of widespread geographic distribution. It causes acute febrile illness and a potentially debilitating chronic infection in humans, and livestock infection has substantial socioeconomic impact. This review describes new information regarding the epidemiology of brucellosis in the developing world and advances in diagnosis and treatment. RECENT FINDINGS: The highest recorded incidence of human brucellosis occurs in the Middle East and Central Asia. Fever etiology studies demonstrate brucellosis as a cause of undifferentiated febrile illness in the developing world. Brucellosis is a rare cause of fever among returning travelers, but is more common among travelers returning from the Middle East and North Africa. Sensitive and specific rapid diagnostic tests appropriate for resource-limited settings have been validated. Randomized controlled trials demonstrate that optimal treatment for human brucellosis consists of doxycycline and an aminoglycoside. Decreasing the burden of human brucellosis requires control of animal brucellosis, but evidence to inform the design of control programs in the developing world is needed. SUMMARY: Brucellosis causes substantial morbidity in human and animal populations. While improvements in diagnostic options for resource-limited settings and stronger evidence for optimal therapy should enhance identification and treatment of human brucellosis, prevention of human disease through control in animals remains paramount.

Trends of frequency, mortality and risk factors among patients admitted with stroke from 2017 to 2019 to the medical ward at Kilimanjaro Christian Medical Centre hospital: a retrospective observational study.

OBJECTIVE: The burden of stroke has increased in recent years worldwide, particularly in low-income and middle-income countries. In this study we aim to determine the number of stroke admissions, and associated comorbidities, at a referral hospital in Northern Tanzania. DESIGN: This was a retrospective observational study. SETTING: The study was conducted at a tertiary referral hospital, Kilimanjaro Christian Medical Centre (KCMC), in the orthern zone of Tanzania. PARTICIPANTS: The study included adults aged 18 years and above, who were admitted to the medical wards from 2017 to 2019. OUTCOME: The primary outcome was the proportion of patients who had a stroke admitted in the medical ward at KCMC and the secondary outcome was clinical outcome such as mortality. METHODS: We conducted a retrospective audit of medical records from 2017 to 2019 for adult patients admitted to the medical ward at KCMC. Data extracted included demographic characteristics, previous history of stroke and outcome of the admission. Factors associated with stroke were investigated using logistic regression. RESULTS: Among 7976 patients admitted between 2017 and 2019, 972 (12.2%) were patients who had a stroke. Trends show an increase in patients admitted with stroke over the 3 years with 222, 292 and 458 in 2017, 2018 and 2019, respectively. Of the patients who had a stroke, 568 (58.4%) had hypertension while 167 (17.2%) had diabetes mellitus. The proportion of admitted stroke patients aged 18-45 years, increased from 2017 (n=28, 3.4%) to 2019 (n=40, 4.3%). The in-hospital mortality related to stroke was 229 (23.6%) among 972 patients who had a stroke and female patients had 50% higher odds of death as compared with male patients (OR:1.5; CI 1.30 to 1.80). CONCLUSION: The burden of stroke on individuals and health services is increasing over time, which reflects a lack of awareness on the cause of stroke and effective preventive measures. Prioritising interventions directed towards the reduction of non-communicable diseases and associated complications, such as stroke, is urgently needed.

Prevalence and risk factors for human leptospirosis at a hospital serving a pastoralist community, Endulen, Tanzania.

BACKGROUND: Leptospirosis is suspected to be a major cause of illness in rural Tanzania associated with close contact with livestock. We sought to determine leptospirosis prevalence, identify infecting Leptospira serogroups, and investigate risk factors for leptospirosis in a rural area of Tanzania where pastoralist animal husbandry practices and sustained livestock contact are common. METHODS: We enrolled participants at Endulen Hospital, Tanzania. Patients with a history of fever within 72 hours, or a tympanic temperature of ≥38.0°C were eligible. Serum samples were collected at presentation and 4-6 weeks later. Sera were tested using microscopic agglutination testing with 20 Leptospira serovars from 17 serogroups. Acute leptospirosis cases were defined by a ≥four-fold rise in antibody titre between acute and convalescent serum samples or a reciprocal titre ≥400 in either sample. Leptospira seropositivity was defined by a single reciprocal antibody titre ≥100 in either sample. We defined the predominant reactive serogroup as that with the highest titre. We explored risk factors for acute leptospirosis and Leptospira seropositivity using logistic regression modelling. RESULTS: Of 229 participants, 99 (43.2%) were male and the median (range) age was 27 (0, 78) years. Participation in at least one animal husbandry practice was reported by 160 (69.9%). We identified 18 (7.9%) cases of acute leptospirosis, with Djasiman 8 (44.4%) and Australis 7 (38.9%) the most common predominant reactive serogroups. Overall, 69 (30.1%) participants were Leptospira seropositive and the most common predominant reactive serogroups were Icterohaemorrhagiae (n = 20, 29.0%), Djasiman (n = 19, 27.5%), and Australis (n = 17, 24.6%). Milking cattle (OR 6.27, 95% CI 2.24-7.52) was a risk factor for acute leptospirosis, and milking goats (OR 2.35, 95% CI 1.07-5.16) was a risk factor for Leptospira seropositivity. CONCLUSIONS: We identified leptospirosis in approximately one in twelve patients attending hospital with fever from this rural community. Interventions that reduce risks associated with milking livestock may reduce human infections.

Q fever and early pregnancy failure: a Scottish case-control study.

Q fever is a bacterial disease that passes between animals and humans and causes disease in both. The disease has been associated with pregnancy complications including miscarriage. This study was undertaken to identify if Q fever exposure was correlated with miscarriage in 369 women attending a pregnancy support unit in Edinburgh. The women in the study were in two groups, the miscarriage group with 251 women who had experienced a miscarriage and a control group of 118 women who had not experienced miscarriage. Three women were found to be positive for Q fever antibodies, suggesting that they had previously been exposed to the infection and all of them were from the group who had experienced miscarriage. The study indicates that Q fever is relatively rare in women attending an urban Scottish hospital suggesting that the infection is not a major cause of miscarriage in this population. However, as Q fever antibodies could only be found in women within the miscarriage group, it suggests that the infection cannot be ruled out as a potential cause of miscarriage in individual cases.

Target-enrichment sequencing yields valuable genomic data for challenging-to-culture bacteria of public health importance.

Genomic data contribute invaluable information to the epidemiological investigation of pathogens of public health importance. However, whole-genome sequencing (WGS) of bacteria typically relies on culture, which represents a major hurdle for generating such data for a wide range of species for which culture is challenging. In this study, we assessed the use of culture-free target-enrichment sequencing as a method for generating genomic data for two bacterial species: (1) Bacillus anthracis, which causes anthrax in both people and animals and whose culture requires high-level containment facilities; and (2) Mycoplasma amphoriforme, a fastidious emerging human respiratory pathogen. We obtained high-quality genomic data for both species directly from clinical samples, with sufficient coverage (>15×) for confident variant calling over at least 80% of the baited genomes for over two thirds of the samples tested. Higher qPCR cycle threshold (Ct) values (indicative of lower pathogen concentrations in the samples), pooling libraries prior to capture, and lower captured library concentration were all statistically associated with lower capture efficiency. The Ct value had the highest predictive value, explaining 52 % of the variation in capture efficiency. Samples with Ct values ≤30 were over six times more likely to achieve the threshold coverage than those with a Ct > 30. We conclude that target-enrichment sequencing provides a valuable alternative to standard WGS following bacterial culture and creates opportunities for an improved understanding of the epidemiology and evolution of many clinically important pathogens for which culture is challenging.

Brucellosis testing patterns at health facilities in Arusha region, northern Tanzania.

BACKGROUND: Brucellosis is listed as one of six priority zoonoses in Tanzania's One Health strategic plan which highlights gaps in data needed for the surveillance and estimation of human brucellosis burdens. This study collected data on current testing practices and test results for human brucellosis in Arusha region, northern Tanzania. METHODS: Retrospective data were extracted from records at 24 health facilities in Arusha region for the period January 2012 to May 2018. Data were captured on: the test reagents used for brucellosis, procurement and testing protocols, the monthly number of patients tested for brucellosis and the monthly number testing positive. Generalised linear mixed models were used to evaluate relationships between health facility characteristics and the probability that brucellosis testing was conducted in a given month, and the proportion of individuals testing positive. RESULTS: Four febrile Brucella agglutination tests were used widely. The probability of testing for brucellosis in a given month was significantly associated with an interaction between year of testing and facility ownership. Test probability increased over time with more pronounced increases in privately owned as compared to government facilities. The proportion of individuals testing positive for brucellosis was significantly associated with facility type and district, with individuals tested in hospitals in Meru, Monduli and Ngorongoro districts more likely to test positive. CONCLUSIONS: Febrile Brucella agglutination tests, known for their poor performance, were the mainstay of brucellosis testing at health facilities in northern Tanzania. The study indicates that historical data on human brucellosis in Arusha and other regions are likely to provide an inaccurate measure of true disease burden due to poor performance of the tests used and variation in testing practices. Measures to address these identified shortcomings could greatly improve quality of testing and surveillance data on brucellosis and ultimately inform prevention and control of this priority disease.

An outbreak of Rift Valley fever among peri-urban dairy cattle in northern Tanzania.

BACKGROUND: Human and animal cases of Rift Valley fever (RVF) are typically only reported during large outbreaks. The occurrence of RVF cases that go undetected by national surveillance systems in the period between these outbreaks is considered likely. The last reported cases of RVF in Tanzania occurred during a large outbreak in 2007-2008. METHODS: Samples collected between 2017 and 2019 from livestock suffering abortion across northern Tanzania were retrospectively tested for evidence of RVF virus infection using serology and reverse transcription quantitative polymerase chain reaction (RT-qPCR). RESULTS: A total of 14 RVF-associated cattle abortions were identified among dairy cattle in a peri-urban area surrounding the town of Moshi. RVF cases occurred from May to August 2018 and were considered to represent an undetected, small-scale RVF outbreak. Milk samples from 3 of 14 cases (21%) were found to be RT-qPCR positive. Genotyping revealed circulation of RVF viruses from two distinct lineages. CONCLUSIONS: RVF outbreaks can occur more often in endemic settings than would be expected on the basis of detection by national surveillance. The occurrence of RVF cases among peri-urban dairy cattle and evidence for viral shedding in milk, also highlights potentially emerging risks for RVF associated with increasing urban and peri-urban livestock populations.

Inter-epidemic Rift Valley fever virus infection incidence and risks for zoonotic spillover in northern Tanzania.

Rift Valley fever virus (RVFV) is a mosquito-borne pathogen that has caused epidemics involving people and animals across Africa and the Arabian Peninsula. A number of studies have found evidence for the circulation of RVFV among livestock between these epidemics but the population-level incidence of infection during this inter-epidemic period (IEP) is rarely reported. General force of infection (FOI) models were applied to age-adjusted cross-sectional serological data to reconstruct the annual FOI and population-level incidence of RVFV infection among cattle, goats, and sheep in northern Tanzania from 2009 through 2015, a period without reported Rift Valley fever (RVF) cases in people or animals. To evaluate the potential for zoonotic RVFV spillover during this period, the relationship between village-level livestock RVFV FOI and human RVFV seropositivity was quantified using multi-level logistic regression. The predicted average annual incidence was 72 (95% Credible Interval [CrI] 63, 81) RVFV infections per 10,000 animals and 96 (95% CrI 81, 113), 79 (95% CrI 62, 98), and 39 (95% CrI 28, 52) per 10,000 cattle, sheep, and goats, respectively. There was variation in transmission intensity between study villages, with the highest estimated village-level FOI 2.49% (95% CrI 1.89, 3.23) and the lowest 0.12% (95% CrI 0.02, 0.43). The human RVFV seroprevalence was 8.2% (95% Confidence Interval 6.2, 10.9). Human seropositivity was strongly associated with the village-level FOI in livestock, with the odds of seropositivity in an individual person increasing by around 1.2 times (95% CrI 1.1, 1.3) for each additional annual RVFV seroconversion per 1,000 animals. A history of raw milk consumption was also positively associated with human seropositivity. RVFV has circulated at apparently low levels among livestock in northern Tanzania in the period since the last reported epidemic. Although our data do not allow us to confirm human RVFV infections during the IEP, a strong association between human seropositivity and the FOI in cattle, goats, and sheep supports the hypothesis that RVFV circulation among livestock during the IEP poses a risk for undetected zoonotic spillover in northern Tanzania. We provide further evidence for the likely role of raw milk consumption in RVFV transmission from animals to people.

Prospective cohort study reveals unexpected aetiologies of livestock abortion in northern Tanzania.

Livestock abortion is an important cause of productivity losses worldwide and many infectious causes of abortion are zoonotic pathogens that impact on human health. Little is known about the relative importance of infectious causes of livestock abortion in Africa, including in subsistence farming communities that are critically dependent on livestock for food, income, and wellbeing. We conducted a prospective cohort study of livestock abortion, supported by cross-sectional serosurveillance, to determine aetiologies of livestock abortions in livestock in Tanzania. This approach generated several important findings including detection of a Rift Valley fever virus outbreak in cattle; high prevalence of C. burnetii infection in livestock; and the first report of Neospora caninum, Toxoplasma gondii, and pestiviruses associated with livestock abortion in Tanzania. Our approach provides a model for abortion surveillance in resource-limited settings. Our findings add substantially to current knowledge in sub-Saharan Africa, providing important evidence from which to prioritise disease interventions.

Multisectoral cost analysis of a human and livestock anthrax outbreak in Songwe Region, Tanzania (December 2018-January 2019), using a novel Outbreak Costing Tool.

OBJECTIVES: We applied a novel Outbreak Costing Tool (OCT), developed by the US Centers for Disease Control and Prevention (CDC), to estimate the costs of investigating and responding to an anthrax outbreak in Tanzania. We also evaluated the OCT's overall utility in its application to a multisectoral outbreak response. METHODS: We collected data on direct costs associated with a human and animal anthrax outbreak in Songwe Region (December 2018 to January 2019) using structured questionnaires from key-informants. We performed a cost analysis by entering direct costs data into the OCT, grouped into seven cost categories: labor, office, travel and transport, communication, laboratory support, medical countermeasures, and consultancies. RESULTS: The total cost for investigating and responding to this outbreak was estimated at 102,232 United States dollars (USD), with travel and transport identified as the highest cost category (62,536 USD) and communication and consultancies as the lowest, with no expenditure, for the combined human and animal health sectors. CONCLUSIONS: Multisectoral investigation and response may become complex due to coordination challenges, thus allowing escalation of public health impacts. A standardized framework for collecting and analysing cost data is vital to understanding the nature of outbreaks, in anticipatory planning, in outbreak investigation and in reducing time to intervention. Pre-emptive use of the OCT will also reduce overall and specific (response period) intervention costs for the disease. Additional aggregation of the costs by government ministries, departments and tiers will improve the use of the tool to enhance sectoral budget planning for disease outbreaks in a multisectoral response.

Molecular epidemiology of Brucella species in mixed livestock-human ecosystems in Kenya.

Brucellosis, caused by several species of the genus Brucella, is a zoonotic disease that affects humans and animal species worldwide. Information on the Brucella species circulating in different hosts in Kenya is largely unknown, thus limiting the adoption of targeted control strategies. This study was conducted in multi-host livestock populations in Kenya to detect the circulating Brucella species and assess evidence of host-pathogen associations. Serum samples were collected from 228 cattle, 162 goats, 158 sheep, 49 camels, and 257 humans from Narok and Marsabit counties in Kenya. Information on age, location and history of abortion or retained placenta were obtained for sampled livestock. Data on age, gender and location of residence were also collected for human participants. All samples were tested using genus level real-time PCR assays with primers specific for IS711 and bcsp31 targets for the detection of Brucella. All genus positive samples (positive for both targets) were further tested with a speciation assay for AlkB and BMEI1162 targets, specific for B. abortus and B. melitensis, respectively. Samples with adequate quantities aggregating to 577 were also tested with the Rose Bengal Test (RBT). A total of 199 (33.3%) livestock and 99 (38.5%) human samples tested positive for genus Brucella. Animal Brucella PCR positive status was positively predicted by RBT positive results (OR = 8.3, 95% CI 4.0-17.1). Humans aged 21-40 years had higher odds (OR = 2.8, 95% CI 1.2-6.6) of being Brucella PCR positive compared to the other age categories. The data on detection of different Brucella species indicates that B. abortus was detected more often in cattle (OR = 2.3, 95% CI 1.1-4.6) and camels (OR = 2.9, 95% CI 1.3-6.3), while B. melitensis was detected more in sheep (OR = 3.6, 95% CI 2.0-6.7) and goats (OR = 1.7, 95% CI 1.0-3.1). Both B. abortus and B. melitensis DNA were detected in humans and in multiple livestock host species, suggesting cross-transmission of these species among the different hosts. The detection of these two zoonotic Brucella species in humans further underpins the importance of One Health prevention strategies that target multiple host species, especially in the multi-host livestock populations.

"He Who Relies on His Brother's Property Dies Poor": The Complex Narratives of Livestock Care in Northern Tanzania.

Background: Endemic zoonoses have important impacts for livestock-dependent households in East Africa. In these communities, people's health and livelihoods are severely affected by livestock disease losses. Understanding how livestock keepers undertake remedial actions for livestock illness has the potential for widespread benefits such as improving health interventions. Yet, studies about livestock and human health behaviours in the global south tend to focus on individual health choices. In reality, health behaviours are complex, and not solely about individualised health experiences. Rather, they are mediated by a range of "upstream" factors (such as unequal provision of services), which are beyond the control of the individual. Methods: This paper presents qualitative research conducted from 2014 to 2019 for a study focused on the Social, Economic, and Environmental Drivers of Zoonoses in Tanzania (SEEDZ). Qualitative data were collected via focus group discussions, community meetings, informal interviews, formal in-depth interviews, observations and surveys that addressed issues of health, disease, zoonotic disease risks, and routes for treatment across 21 villages. Thematic analysis was carried out on in-depth interviews and focus group discussions. Conceptual analyses and observations were made through application of social science theories of health. Findings: Livestock keepers undertake a range of health seeking strategies loosely categorised around self and formal treatment. Two key themes emerged that are central to why people make the decisions they do: access to resources and trust in health care providers. These two issues affect individual sense of agency which impacts their ability to act to improve livestock health outcomes. We suggest that individual choice and agency in veterinary health seeking decisions are only beneficial if health systems can offer adequate care and health equity is addressed. Significance: This study demonstrates the value of in-depth qualitative research which reveals the nuance and complexity of people's decisions around livestock health. Most importantly, it explains why "better" knowledge does not always translate into "better" practise. The paper suggests that acknowledging and addressing these aspects of veterinary health seeking will lead to more effective provision.

Molecular detection of Coxiella burnetii infection in small mammals from Moshi Rural and Urban Districts, northern Tanzania.

Coxiella burnetii is an obligate intracellular bacterium that causes Q fever, a zoonotic disease of public health importance. In northern Tanzania, Q fever is a known cause of human febrile illness, but little is known about its distribution in animal hosts. We used a quantitative real-time PCR (qPCR) targeting the insertion element IS1111 to determine the presence and prevalence of C. burnetii infections in small mammals trapped in 12 villages around Moshi Rural and Moshi Urban Districts, northern Tanzania. A total of 382 trapped small mammals of seven species were included in the study; Rattus rattus (n = 317), Mus musculus (n = 44), Mastomys natalensis (n = 8), Acomys wilson (n = 6), Mus minutoides (n = 3), Paraxerus flavovottis (n = 3) and Atelerix albiventris (n = 1). Overall, 12 (3.1%) of 382 (95% CI: 1.6-5.4) small mammal spleens were positive for C. burnetii DNA. Coxiella burnetii DNA was detected in five of seven of the small mammal species trapped; R. rattus (n = 7), M. musculus (n = 1), A. wilson (n = 2), P. flavovottis (n = 1) and A. albiventris (n = 1). Eleven (91.7%) of twelve (95% CI: 61.5-99.8) C. burnetii DNA positive small mammals were trapped within Moshi Urban District. These findings demonstrate that small mammals in Moshi, northern Tanzania are hosts of C. burnetii and may act as a source of C. burnetii infection to humans and other animals. This detection of C. burnetii infections in small mammals should motivate further studies into the contribution of small mammals to the transmission of C. burnetii to humans and animals in this region.