Central nervous system infections (CNSI) are a major cause of morbidity and mortality in low-resource, high HIV-prevalence African settings. HIV has drastically altered the epidemiology of CNSI in high HIV-prevalence African settings and cryptococcal, tuberculous and pneumococcal meningitis comprise the majority of microbiologically-confirmed diagnoses. However, in high HIV-prevalence African settings up to half of patients with a CNSI do not receive a diagnosis when conventional diagnostics are used alone. The mortality in this group is high, 40% at 10 weeks, indicating the presence of significant undiagnosed pathology that needs improved diagnostic strategies to enable to the rapid initiation of appropriate and targeted treatment to achieve better outcomes. Enhanced diagnostics for CNSI, including monoplex PCRs for key CNS pathogens or rapid multiplex PCR platforms such as BioFIRE FilmArray-ME, are available and have transformed the investigation of CNSI in high-resource settings. However, access to these platforms is extremely limited in routine care in high HIV-prevalence Southern Africa where there is a continued reliance on conventional techniques, such as culture or microscopy with gram and Ziehl-Neelsen stains, which often lack sensitivity. Diagnostic delays also result from other components of the diagnostic pathway including delayed lumbar punctures due to a lack of consumables or concerns about potential contraindications to lumbar puncture prompting clinicians to perform often unnecessary neuroimaging. Data on the role of neuroimaging in resource-limited, high HIV-prevalence settings are extremely scarce although clinicians in the region face distinct challenges, including a broader differential diagnosis alongside severely restricted access to neuroimaging. In addition to the lack of a diagnosis meaning rapid diagnostic-driven treatment is not possible, the large burden of CNSI with no diagnosis results in a limited understanding of CNSI epidemiology in Southern Africa. This means empiric treatment choices are impossible, public health decisions regarding resource allocations are unguided, optimal drug procurement is challenging and both the impact and need for vaccination and prevention strategies is largely unknown. Robust data from comprehensive CSF analysis could determine optimal empiric treatment strategies in patients without a microbiologically confirmed CNSI diagnosis in low-resource settings, as well as supporting resource allocation and drug procurement choices, and evaluating the success of vaccination strategies. The overall aim of this thesis is to identify strategies to improve the diagnosis of CNSI in high HIV-prevalence settings. The specific aims of this thesis are to define the current epidemiology of CNSI in a high HIV-prevalence African setting, to determine the increase in diagnostic yield of CNSI following expansion of diagnostic capabilities, to describe the changes in epidemiology of common CNSIs in the region over time and to characterise the effect of computed tomography performed prior to lumbar puncture on patients with suspected CNSI. The epidemiology of CNSI in Botswana and Zimbabwe and the impact of enhanced diagnostics on diagnostic yield will be described using clinical and laboratory data from two prospective meningitis surveillance networks where enhanced diagnostics including BioFIRE FilmArray-ME and Xpert MTB/RIF Ultra were implemented into routine care alongside additional retrospective CSF analysis. Individual patient level data from other high HIV-prevalence African countries that have used BioFIRE FilmArray-ME will be combined in a systematic review and meta-analysis to further contextualise the impact of enhanced diagnostics in high HIV-prevalence African settings. These data will be complemented by national meningitis surveillance data from Botswana spanning 8 years which will be used to describe the changes in the detection of the most common CNSIs, cryptococcal and tuberculous meningitis, over time, following expansion of CNSI diagnostic capacity in routine care and increased ART coverage in Botswana. Finally, this thesis will use data from a tertiary hospital in Botswana to evaluate how computed tomography is used in routine clinical practice in high HIV-prevalence settings and characterise the effect of computed tomography performed prior to lumbar puncture on patients with suspected CNSI on delay to diagnostic lumbar puncture and treatment initiation. This thesis will be presented in 7 chapters including 4 published manuscripts and 2 chapters in the style of a research paper. Chapter I will serve as an overarching review of current challenges in the diagnosis of CNSI in high HIV-prevalence African settings and provides an overview of currently available diagnostic platforms for use in the investigation of suspected CNSI in the form a published narrative review. Chapter II describes the aims and objectives of this thesis and the methodology used to establish the two prospective meningitis surveillance networks in Botswana and Zimbabwe. Chapter III is a systematic review and meta-analysis of individual patient level data from all available studies from high HIV-prevalence African settings that have used the rapid multiplex PCR platform BioFIRE FilmArray-ME to evaluate the clinical utility of BioFIRE FilmArray-ME when used in settings with differing CNSI epidemiology from where BioFIRE FilmArray-ME was originally developed. Chapter IV consists of 2 manuscripts that used national meningitis surveillance data from Botswana to characterise trends in CNSI epidemiology following the rollout of improved CNSI diagnostics and scale-up of ART. Chapter V is a published manuscript evaluating the role of computed tomography in the investigation of patients with suspected CNSI in Botswana and assesses the impact of prior CT on time to diagnostic lumbar puncture and treatment initiation. Chapter VI uses prospective data captured following the introduction of enhanced diagnostic packages at two sites in Southern Africa to describe the epidemiology of CNSI in the region and evaluate the impact of these enhanced diagnostics on diagnostic yield. Chapter VII is a concluding chapter that summarises the key findings from the thesis and their implications.