Leveraging Microsoft Premonition to Predict Epidemics

Introduction

Envisaging and averting epidemics is paramount, given the intensifying health threats worldwide. Outbreaks of various illnesses remain one of the primary concerns regarding public health. Even with the existing modern technologies and innovative approaches, the world still has to deal with high rates of epidemics. For instance, during the COVID-19 outbreak, every nation felt the adverse impact. The coronavirus outbreak struck when the world was unprepared to stop its widespread. In this regard, it had an unforgettable impact as many people lost their lives, and the global economy remained crippled. The existing mechanisms aimed at identifying such threats proved ineffective as there was a delay in sharing accurate information to curb the virus before it spread to almost every part of the world. The COVID-19 pandemic and its impacts serve as a lesson, necessitating more robust approaches to deal with such scenarios effectively. When global health organizations cannot tackle challenges like information collection and analysis delays, widespread transmission and deaths become inevitable. Such cases prompted Microsoft Corporation to develop a platform to revolutionize the healthcare sector. Microsoft Premonition empowers relevant stakeholders to predict epidemics in real-time and implement effective interventions. The platform’s capabilities make it the most effective tool for providing a guaranteed solution, as it facilitates early detection of any threats to public health. As a result, it allows timely response to mitigate potential outbreaks. Unlike conventional reactive approaches, Microsoft Premonition is an effective tool to transform public health by foreseeing and mitigating outcomes of disease outbreaks.

Background

In the contemporary world, disease outbreaks and the rapid spread of viruses such as coronavirus bother many stakeholders due to the outcomes. In addition, the less efficient system for disease surveillance remains a challenge because it cannot detect possible outbreaks in due time, which could give room for suitable interventions. As stated, the existing surveillance approaches fail because of the unstandardized data, making it difficult to disseminate the needed information urgently. In other words, it ultimately obstructs the addressing of the rising health challenges. Besides, time and lack of real-time data will be unavoidable unless there are standardized data (Burkhardt et al., 2021). Resolving these issues by providing a way in which the data interoperability and exchange becomes straightforward determines data standardization. While global health bodies such as the Centers for Disease Control and Prevention (CDC) prioritize public health, they focus on overcoming this problem by ensuring data consumption normalization. The CDC has implemented new enabling strategies and tools (FHIR). The FHIR is a standard model known for its unique feature of effectively using data arrangement and distribution. It represents a complex variety of “resolvable and expandable data structure specifications” called resources (Alterovitz et al., 2020, p.1). Analogously, internet data sharing can be attributed to FHIR (2023). In this context, introducing FHIR is essential for smooth operator data exchange. Azure Fabric can easily integrate data from various sources, such as other devices or countries, and help different actors review and examine such data. Besides, Azure Fabric creates ways to allow people to observe disease outbreaks and accommodate them whenever they happen.

The protocols of emerging standards require positive action against risks to public health. In particular, the compatible union of the platforms would promote predictive analysis instead of outdated reactive analysis. However, for a public health group, engaging in a holistic platform and well-structured data organization is necessary to get more detailed and timely information. For example, the concept of one-lake regularizes information storage in the “Open Delta Parquet Format.” As a result, this ensures data openness and distribution, which reduces redundancy and increases efficiency (Hernández, 2024). With this, public health establishments can foresee and prevent outbursts of diseases before becoming worse. Therefore, such a pre-emptive tactic for envisaging and dealing with epidemics is paramount. It shows a substantial shift in the model to allow practitioners to be aware of emerging threats and enforce specific interventions.

Integrating FHIR, Microsoft Fabric, and Premonition is a unified approach. Microsoft Premonition is useful in foretelling the spread of diseases and viruses. Choney (2022) holds that Microsoft Premonition involves changing the model from responding to medical conditions to dealing with pathogens and assessing their evolution. These emerging tools can help detect imminent threats earlier, address them swiftly, and create suitable interventions before outbreaks hit hard and leave adverse outcomes. Premonition comprises “artificial intelligence, robotic sensing platforms, predictive analytics, and cloud-scale metagenomics” (Choney, 2022). With this, public health organizations can “monitor disease-carrying animals autonomously, gather samples robotically, and scan them from any biological threats genomically” (Choney, 2022). Premonition ensures that healthcare entities and other relevant stakeholders comprehend different aspects of epidemics. Besides, adopting FHIR is a significant move to ensure that these organizations share information seamlessly among themselves and that such data is interoperable (Garza et al., 2020). Access to this information allows public health organizations to restructure the data from multiple sources to make informed decisions. Likewise, Azure Fabric contributes to the efficacy of predictive analysis, supporting other platforms in providing significant solutions and addressing public health concerns effectively. These platforms provide the CDC with critical insights to adopt a proactive approach to managing epidemics.

Fast Healthcare Interoperability Resources (FHIR)

FHIR is a significant standard in data analytics as it allows seamless data sharing and interconnectivity that cannot be compromised. The concept proposed by Odisho et al. (2020) is that the FHIR Standard makes data exchange feasible within applications from different health facilities. It is an excellent resource that prevents hits such as data losses from happening. FHIR cannot cover the following: “faxing of referrals, looking up the patient in EHR, documenting demographic information, updating insurance and protocols, filling patient’s information, phoning a patient back, and logging a referral” (Odisho et al., 2020, p.408). Among the most crucial factors under consideration is ensuring that the information is accurate and that the organizations receive data they can trust. Implementing these standards can pave the way to speedy data gathering and deploying robust information. FHIR allows for effective communication among medical professionals from different specialties. With FHIR, there is a whole lot to gain, from smooth data exchange to better interoperability.

Streamlining Data Exchange

Implementing the Fast Healthcare Interoperability Resources (FHIR) standards enables facilitated through the FHIR standards. Efficient surveillance of epidemics deals with the need for access to and selfless sharing of crucial information. HL7 FHIR is indispensable in keeping up with trends and preventing disease outbreaks because it sends data seamlessly between the various constituents of healthcare networks. Utilizing the FHIR standards allows the CDC to collect data from different resources most efficiently, so they can now make data-driven decisions easily. FHIR can present information in its standard form, thus eliminating complexity in data integration. It eventually led to its acceptance as one of the most efficient tools that give considerable insights to quickly detect the unintentional spread of diseases. Furthermore, FHIR greatly benefits healthcare organizations by promoting data fusion from multiple sources to create an integrated dataset that enhances data mining.

Moreover, the conventional method contributes to the gathered data’s quality and information-richness. Additionally, they can design corresponding risk evaluation mechanisms that could help reduce the risks of public healthcare in their earliest stage. Conclusively, the role FHIR plays in the exchange of information gets captured in the statement, which brings positivity to the process of epidemic surveillance. CDC will use FHIR for data remodeling and choose the correct data to input into the processes. This approach will allow the CDC and other healthcare providers to reveal the threats and defend the public at an earlier time. Using standardized data will eliminate the need to waste time, especially when sharing the data with other organizations. The CDC may adopt the FHIR to respond to the emergency public health concern.

Enhancing Interoperability

Furthermore, FHIR promotes interoperability. Epidemic surveillance systems become efficient if people exploit information sharing as readily as possible using interoperable data. It contributes to the view that the information flows uninterrupted and that the communication between the key parties is efficient. It necessitates a standardized approach, which is the vital element of interoperability. Vorisek et al. (2022) state that FHIR is the main framework that companies should develop for interoperable data. An example is FHIR, which concentrates on setting up common principles and ways everyone should follow. On this basis, the model will produce a standardized language that healthcare organizations can use to record data accurately. Hence, adhering to FHIR standards is of utmost importance because it allows different stakeholders to harmonize their systems or strategies so that information exchange can go smoothly with no difficulties. Communicating obstacles can only be there without the full integration of all parties, making them join hand in hand in achieving the intended objectives. Consequently, they establish a secure communication channel, enabling them to react speedily to health crises.

The FHIR, the interoperability standard of healthcare information (Rosenau et al., 2022), can construct a conducive environment. For instance, it provides a location for different pieces to monitor and prevent disease transmission. Healthcare providers can collaborate with other organizations to capitalize on FHIR advantages so that data can be shared and used meaningfully for their benefit. This cooperation allows them to tap into the expertise and resources needed to make an impact on the given surveillance activities. Hence, FHIR fosters interoperability between entities and organizations, allowing them a coordinated approach to epidemiological scrutiny. When these stakeholders share interoperable information and cooperate when required, they can accomplish the set goals, and public safety remains a priority. Their success occurs due to the interconnectedness, which enhances collective efforts to foresee epidemics and develop effective countermeasures. Working together presents these stakeholders with an opportunity to share resources and insights. They can effectively leverage such benefits to deal with public health challenges. Therefore, relevant entities can adhere to the standardized format and established communication protocols to overcome hurdles. They can embrace a coordinated approach to handling public health issues.

Integration of FHIR Genomics Standards

Incorporating FHIR genomics can also boost epidemic surveillance interventions. They significantly contribute to the monitoring efforts by creating a new venue where individuals can capture genomic data and share it with the key stakeholders to prevent the spread of communicable diseases. Alterovitz et al. (2020, p. 405) define FHIR genomics as  “a subset of the emerging Health Level 7 FHIR standard and targets data from increasingly available technologies such as next-generation sequencing”. There are valuable genomic data, such as pathogens’ evolution, transmission, and host vulnerability. Exploring such data allows individuals to gain critical insights into various facets of illnesses. Thus, the CDC can rely on the information generated from FHIR genomics to succeed in its surveillance efforts. Individuals with access to genomic data tend to have crucial insights concerning the pathogens, making it easier to develop effective countermeasures. It can act as the aftermath they look for as they can look at their genetic data. The CDC can study and mine data on the genetic makeup of disease and obtain valuable characteristics of the pathogen, which include resistance to drugs and the strength of the host immune system. This information helps in planning the control program to facilitate the prevention of the spread of the disease.

Genetics can help by giving the correct information in the public sector, such as new medicines that successfully address instances of diseases. By exercising this information, the officials can handle outbreaks caused by infectious diseases. An example can be provided for clarification. These drugs do not respond to drugs detecting and characterizing genetic changes. Through this data, they will gain a view of the achievements in fighting COVID-19. Additionally, they may apply this identification to assess public health issues through the conduct of both in reaction to actual events. In contrast, their primary role includes reducing the burden in the healthcare system. Genetic information can serve as a weapon for CDC in determining what caused an epidemic outbreak and using that knowledge to form a superior strategy that brings about the rapid solution of a problem to the public.

Leveraging Microsoft’s FHIR Azure Server

The deployment solution must include the integration of Microsoft FHIR Azure Server into the system. Also, Azure Server FHIR has a paramount position in terms of storing and managing healthcare data. Also, through this platform, organizations gain access to the interoperable data used to supervise epidemics. Interoperability, according to Godlovitch and Kroon’s (2022) definition, refers to the ability of systems to distribute and use information. The Azure platform can process, assess, and store vast volumes of data as a source of data for public health agencies’ analytics. This server has not only the capability to store structured data but also the ability to store tons of unstructured data due to its scalability. Hence, this approach will enable a stable information transfer between the edge and cloud (Hassan et al., 2022). It can create an opportunity for health facilities to use this channel to reach them with the relevant information that will make them capable of responding to the threats of disease outbreaks. The analytical potential of these platforms will enable the CDC to get a personification of the information about infectious diseases. In turn, it can utilize the system to make up-to-date decisions and take necessary measures to handle disease outbreaks and spread. FHIR Azure is well known for its outstanding scalable and flexible functionality. The result is that healthcare organizations can design how they manage their data and surpass any barrier. The scalability provides the ability to cope with data volume fluctuations and processing in terms of their dimensions. Therefore, the healthcare establishment will be ready and prepared to contribute towards following the patterns of the epidemics. Azure ensures a unified experience by incorporating other valuable tools such as Fabric and Premonition; this leads to an integrated epidemic monitoring approach. It is due to their ability to use FHIR capabilities and cloud technologies to enhance their ability to track patients. Consequently, these platforms and tools help automate the emergency response process using predictive analytics and proactive response tactics.

Azure Fabric

Azure Fabric’s architecture is one of a kind since it can collect data from suppliers and then perform predictive analysis at the cloud scale, rendering it unbeatable. It is a systemic approach that simplifies the complex steps of the process, which are comprehensive and step-by-step (Hernandez, 2024). The fabric of systems, of which Microsoft Azure is the distributed platform, helps with package, deployment, and management of microservices that are scalable and reliable at a steady pace (Cassidy et al., 2022). It is an enterprise-ready analytical platform that has the capabilities to solve the challenges of launching and maintaining a cloud platform application. A fabric on Azure comes in handy when the idea is to deal with numerous infrastructure complications and instead direct all the energy into handling simple, manageable, and scalable processes.

In response to an epidemic situation, Azure Fabric is the at-hand tool that allows data sharing among different sources. Azure Fabric’s powers would help them combine data sets from different cloud systems. Public health agencies can graph them for a thorough epidemic forecasting analysis. It is mainly the case when epidemiological data undergoes integration, enabling us to research the behavior of the disease at the finest level and use it to predict future outbreaks and address them promptly. By integrating with other cloud platforms, such as GCP and AWS, Azure Fabric becomes its platform. With the business capacity for integrating data from multiple Lands of cloud formats, it will bring significant value for deeper analysis. They can also consider importation data coupled with different data sources as tools for public health agencies, which enhances the capability of epidemic forecasting. The Azure platform will leverage numerous features that will assist public health agencies with their career smart and sophisticatedly with the function of advanced analysis and model projection. Settled by this one, the members of such communities will soon have the ability to extricate analytics on the combined data records and detect many epidemic details, patterns, and correlations, thus bringing insight and timely warnings.

Seamless Data Integration across Cloud Platforms

Azure Fabric would keep track of the spread of the pandemic. It is an invaluable tool because it can quickly blend cloud-based information from different platforms for predicting epidemics (Hernández, 2024). The existence of data silos is among the most prominent problems currently faced in the exchange and effective utilization of information. It has, in turn, led to delayed responses to emerging threats. In the view of Kumar (2023), data silos mean the isolation of data sources, such as the storage and processing of information in separate and incompatible systems. It generates unevenness and waste. In the context of the Azure Fabric, these challenges are mission-critical because they involve unifying information from many sources, such as clinical records, population demographics, and lab data. Unlike this, data source consolidation allows public health agencies to access valuable data inputs for comprehensive epidemiological surveillance. The ubiquitousness of information is one of the biggest pros of Microsoft weave (Hernández, 2024). Hence, clinical records are essential data sources that provide detailed information about patient’s demographics, symptoms, and health history. However, data from the laboratory identifies the species of the pathogen and its features. The demographics of community populations provide valuable information on the most vulnerable and health care disparities. By infusing these different data sources into Azubri Fabric, one can have a detailed analysis touching the root of infectious outbreaks’ success.

Furthermore, combining data from different cloud systems is an essential function for Fabric Azure to build a shared data environment for epidemic surveillance. It allows the establishment of smooth connections between data, breaking down data silos and making collaboration possible from various sources of information. Therefore, this approach helps stakeholders bolster their early detection, monitoring, and response mechanism to epidemics and other infectious diseases as they become more coordinated and punctual.

Interoperability with Other Cloud Providers

One of the most essential features of Azure Fabric is that it can interoperate with other cloud environments like AWS and GCP. It is significantly serving the society affected by epidemics. Today, characterized by technological advancement, people have been exposed to data. Many cloud platforms use various types of formats and protocols. Azure Fabric can work with the most popular cloud providers; thus, one’s data has no obstacle traveling from one cloud to another (Joyce, 2023). Take, for instance, interoperability that enhances public health agencies in predictive analysis processes through a higher number of data inputs. This skill is the capacity to attain data from several cloud services and then use the information to read the outbreak. AWS has access to genomic sequencing data, but GCP has the advantage of having real-time feedback. Another impressive capability of the Azure Fabric in integrating these datasets is the depth of temporal knowledge of the disease occurrence and the making of those actions required for the benefit of public health.

On the other hand, this platform acts as a communication channel and implements integration with other cloud providers, simplifying public health stakeholders’ work. Therefore, the interoperability of the data systems and integration of the platforms in Azure Fabric help it become a collaborative system where policymakers, researchers, and public health officers can share valuable data and unite to counteract disease outbreaks. The initiative strengthens epidemiological surveillance and the health systems against any emerging threats.

Empowering Predictive Analysis

With Azure Fabric in the hands of public health organizations, they can use it as a weapon to perform elaborate and targeted precision analytics. Machine learning, advanced analytics, and AI-enabled virtual sensing are the essential components of predictive medicine that enable early detection and management of health problems with the help of machine learning, AI, and machine learning (Basulo-Ribeiro & Teixeira, 2024). Azure Fabric and the data system help the specialists better understand disease in real time and apply tailored interventions if necessary. A comprehensive approach helps public health organizations notice the first symptoms of disease outbreaks early before they grow into large-scale epidemics. Through aggregating and analyzing data from broad-based sources, Azure Fabric aids in identifying patterns that indicate a rise in communicable diseases. Predictive health architecture is a vital tool in developing the forecasting of illnesses and significantly controlling and fighting pandemics and epidemics (Basulo-Ribeiro & Teixeira, 2024). When movements and confessions through such an active approach to surveillance are suspected, public health organizations can act quickly, implement preventive measures, and channel resources where necessary to prevent the spread of diseases.

Also, Azure Fabric can provide real-time tracking of the spreading patterns of diseases. This way of tracking disease progression and assessing the intervention measures’ efficacy gets completed in real-time. Incoming data, which may come from several sources, is routinely analyzed by the Azure Fabric so that public health authorities can receive updated information about infectious disease transmission patterns. Technological innovations have changed healthcare models from reactive to preventive, increasing healthcare accessibility and minimizing chronic illness load (Basulo-Ribeiro & Teixeira, 2024). However, this constant monitoring allows for immediate response to the fluctuations of the trends and thus enables the agencies to refine their approach to limit the spread of the diseases.

In addition, the predictive analysis of Azure Fabric permits public health organizations to see the future path of epidemics and to be ready for new challenges. The system can use advanced modeling techniques and machine learning algorithms to identify high-risk places, vulnerable individuals, and possible hotspots for disease spread. From such insight, public health organizations can rapidly deploy the necessary preventive measures, such as targeted vaccination campaigns and quarantine recently used to curb the spread of COVID-19, thereby preventing the outbreaks from getting out of control. Therefore, leveraging various datasets within an integrated framework can enable public health organizations like the CDC to stay ahead of epidemics, saving many lives and safeguarding public health.

Predictive Analytics with Microsoft Premonition

History of Project Premonition

Since its inception in 2015, Microsoft has tested Project Premonition’s capabilities in diverse habitats, from the Florida Keys to the remote Tanzania forests in Africa. The company develops the Premonition systems in the “Premonition Proving Ground,” an advanced “Arthropod Containment Level 2” (ACL-2) facility that supports the raising, observing, and digitizing of wild mosquitoes to create identification algorithms and assess device designs (Choney, 2020). Another hub is the Microsoft Redmond campus, where the researchers computationally scan collected samples for pathogens. Since the initiative of alleviating environmental risks before they result in epidemics is a cross-industry and cross-disciplinary effort, Microsoft collaborated with academic institutions to extensively comprehend the underlying science. The National Science Foundation awarded the company “Convergence Accelerator” grant (Baru et al., 2022). This award comprises academic alliances from the University of Pittsburg, the University of Washington’s Institute of Health Metrics and Evaluation, John Hopkins University, and Vanderbilt University. In its award speech, the NSF affirmed that “This project will provide long-lasting contributions to human health and pandemic preparedness” and highlighted that “as deep biome data exponentially scales, the life sciences will become overwhelmed with genomic information. Convergence must lead to new methods to efficiently harness these data and autonomously derive insights” (Choney, 2020, p.1). Microsoft also allied with industrial partners such as Bayer. The Premonition Project aimed to partner with other corporations seeking to prevent or eradicate infectious diseases. For instance, leading firms in vector control are working toward doing away with malaria by 2040 (World Health Organization, 2020). In 2018, the World Health Organization highlighted that almost half of the global population was susceptible to malaria (Talapko et al., 2019), prompting the need for diverse solutions, including proactive ones. The Global Vegetable Seeds and Environmental Science at Bayer president, Jacqueline Applegate, asserted that “Microsoft Premonition gives us the opportunity to be able to get a much more realistic perspective” (Choney, 2020, p.1). The project will enable companies like Bayer to leverage resources, information tools, and data in novel approaches to enhance prescription and advance vector control mechanisms for the most considerable impact. Microsoft Premonition shifts the paradigm from reactive to proactive because instead of responding to known pathogens’ effects after they occur, they will be captured and studied as they evolve. Besides, if the COVID-19 pandemic that disrupted vector-control operations worldwide was anything to go by, having a biothreat predictive system is more than essential.

Working Evidence

Pioneered in 2015 by Microsoft Research, the Premonition Project was aimed to leverage technological innovation in addressing global challenges. The project builds on recent computing and robotics milestones alongside rapid, portable sequencing to transform mosquito gathering and evaluation to determine possible pathogens (Peckham & Sinha, 2019). The research center employed drones to covey internet-configured smart traps with sequencing abilities to the fields and extract them back. Ostensibly, the company sought to incorporate cloud computing, mobile laboratories, and drones in identifying the existence, occurrence, and movement of vector-borne diseases to predict outbreaks before entering the human population. Project Premonition’s significance is evident from its focus on establishing computer vision and machine learning to afford drones intricate autonomy and the ability to make efficient decisions independently (Smith & Shum, 2018). Besides, the project has situated drones as a vital factor in a broader, all-encompassing digital world founded on data collection, novel algorithms, and computational power that drive rapid processing. The Premonition Project has created and tested a system of drone-deployed artificial intelligence, genome sequencing, and ingenious traps to “detect the presence and movement of vector-borne diseases and predict outbreaks before they spill over into human populations” (Peckham & Sinha, 2019, p.1207). As part of an “automating field biology” system, alive mosquitoes are effectuated in a socio-technological-ecological formation that expands the possibility of disease governance, forecasting, and monitoring (Tirado & Cano, 2020, p.126). Thus, Microsoft’s Premonition encompasses operational ecologies that integrate and advance disease-carrying species in automated technologies constructed to understand and intervene in disease ecologies. Various aspects of the Premonition Project highlight its exceptional predictive analytics ability, posing significant possibilities in averting disease outbreaks and epidemics like COVID-19.

First, mosquitoes present an explicit sentinel, characterized by their nature to draw blood from animal species, existing without any restrictions, achieving the feat of getting blood at comparatively high speed and over a broad range that many health workers cannot access (Guillot et al., 2021). A mosquito’s life span averages approximately 20 days, can consume 2.5 ll per blood-based meal, and fly over multiple miles in a widely geographically distributed region (Ohm, 2018). Since they naturally reside in urban and rural settings, they can sample diverse animal species and their pathogens using their advanced olfactory systems to locate even the well-hidden ones (Yan et al., 2021). To this effect, their action can take advantage of classical entomological techniques. Second, the Premonition Project leverages wide-ranging smart traps to program field biology. The existing mosquito traps are inefficient, given their low throughput and requirement of daunting manual human processing, because they have to be placed in the fields and monitored in the environment for 12 to 18 hours (Liu et al., 2016). This aspect makes the project miles more efficient than the existing insect control systems. The autonomous drone-based deployment characterizing the project ensures high throughput. The system comprises traps laid by drones in areas that human beings cannot access, collect the data, and transmit it over massive distances much quicker than conventional epidemiological systems. The scientists then sift through loads of samples from the population collected data to conduct metagenome analytics and determine risks to the population. The unified outcome from the merging of all the factors is a circuit comprised of various stages, including “(a) animals maintaining pathogens in their environment; (b) mosquitos collecting blood and sampling pathogens; (c) drones deploying autonomous traps and collecting and transporting mosquitos; the drone covering long distances; (d) laboratories collecting mosquitoes and sequencing genes and detecting pathogens; (e) finally, global pathogen maps of genes in space and time showing the movement of potential pathogens and hosts” (Tirado & Cano, 2019, p.13). The most vital and primary objective pervades the entire infrastructure is identifying animal-housed pathogens before they affect humanity. As a result, this knowledge allows the development and deployment of intervention measures to predict and alleviate possible infectious outbreaks before they happen, lowering the projected economic and human health threat they create.

Predictive Analytics

Premonition’s robotic sensing configuration will capture, accumulate, amass, and analyze data regarding the trivial, frequently ostensibly invisible threats. The hundreds of millions of existing sensor networks that gather data in predicting the weather, power grid-related information to help unload balancing, and traffic-related data to aid in prediction cannot capture vital things (Rahman et al., 2013). The director of Microsoft Premotion, Ethan Jackson, highlighted, “These life forms we’re talking about are invisible to basically all those sensors we’ve deployed across the globe. And that’s pretty incredible when you think about it, that we have such a huge blind spot about what’s in the environment” (Čirjak et al., 2022). In Harris County, ten robotic smart traps were trained and deployed to collect relevant mosquitoes, and they recorded a 90% accuracy. The metagenomic analyses detected the viruses and microorganisms in the samples of mosquitoes and then determined the animal species on which they fed. The Premonition will afford Harris and similar counties a sensor network at scale, offering “continuous biological, situational awareness” that can perceive the environment and see what is happening in real-time, a provision that does not exist as asserted by weather predictors. The advanced Premonition will provide a future where researchers can detect pathogens like Zika and suppress them swiftly and equitably across a wide geographical distribution. It will also assess novel genomic abilities and highlight known and emerging pathogens from species samples, which now the world recognizes as particularly essential for infections like COVID-19.

The novel epidemiology the Premonition Project affords scientists merges the establishment of multi-drone networks with artificial intelligence (AI) systems’ actionable insights. Since the devices are programmable, they pose numerous benefits, from developing autonomous action systems to creating airplane assistance. In this case, we will only focus on leveraging their predictive analytics aspect to alleviate epidemics. The novel relationships arising from the drone’s employment also bear significant insights. As evidenced by bio-epidemiology, animal species function as sentinels but now go through a form of tracking and incorporation in a broader system than that to which they are accustomed and to which their makeup proves significant in data collection (Tambo & Xiao-Nong, 2014). Additionally, researchers can now readjust and reprogram the animals’ biological actions, a feat that is unachievable with the standard sentinel (Beckoff & Pierce, 2009). Alterations to their biological systems that occur naturally and instinctively no longer bear much weight compared to the initiative of a sentinel functioning in a comprehensive artificial device. The connection between the deployed drones and the animal species allows researchers to conjure the diagonal link between the solely vertical and visual tele-epidemiology action and the purely horizontal operations of creating bio-epidemiology-specific information (Tirado & Cano, 2020). Thus, as the Premonition name indicated, this novel bio-tele-epidemiology merge can proactively react to the outset of epidemics among human populations. All the initiatives exemplify the importance of movement in knowledge production. The chemical and biological information provided by visual knowledge or sentinels made easily accessible by satellites is integrated into a logic that houses mobilization vectors and trajectories. The prospect of predicting epidemics like COVID-19 rests explicitly on this movement (Pettorelli, 2019). Microsoft’s Premonition asserts this point because this surveillance system’s outcome is a map displaying animal hosts and pathogenic genes in movement generated from the mass movement of data, drones, and mosquitoes (Tirado & Cano, 2020). A logic of action results from these factors, whose basis surpasses space and time and extends to motility. This movement of mosquitoes, drones, and several other elements is proportioned with its primary goal – the infectious vectors –even though outbreaks do not care for borders and necessitate joint global efforts, they haze the nation-state’s political borders. Therefore, the novel variations of epidemiological surveillance driven by drones are a vector that requires renegotiating concepts like sovereignty.

Surveillance

The Premonition drones are introducing novel anatomy to surveillance, which depends on movement, extends beyond biological and physical hindrances, and does not necessitate prior visibility. According to Iqbal et al. (2021), drone system surveillance transcends time, evidenced by its ability to store and retrieve data from databases that enable drone anatomy modification and course alteration. As the Premotion professionals outline, drones do not just carry movement; their operations extend to immediate intervention and transformation of infectious vectors’ courses and flow through their independent movement (Carrasco-Escobar et al., 2022). In other words, they completely change the emergency. By installing a pre-signal or pre-symptomatic signal, the drones attempt to forestall the occurrence of an infection, and if it fails in this regard, they attempt to get ahead of the transmission. To achieve this feat, the drone’s course joins, combines, or merges with that of the harmful organism. This form of capture ultimately involves crossing the gorge between the surveillance device and the harmful organism. Although bio and tele-epidemiology logic seeks to identify the threat’s movement completely, the drone takes up a form of capture or merge. In this case, the threat’s course mixes with the drone’s flow. Microsoft’s Premonition highlights this feature; the resulting maps from the surveillance systems activities entirely rely on trajectories and drone phases mentioned before, including the location of the smart traps, the space from which the samples were obtained, and when they were transported, among other factors (Tirado & Cano, 2020). As a result, surveillance in this model generates merging and assembly surfaces with the courses of the drone and health emergency.

This merging is pivotal to averting global epidemics. When the drone combines its course with that of the health emergency, the threat is diverted, detached from its typical movement context and trajectory, and translated into discreet flows that the drone determines. This intervention and course reassembly primarily create a new object using the threat. Researchers can then analyze this new entity from another perspective and develop intervention measures through statistics, comparison, and analytics. A vital factor is that the threat is established, captured, and studied before it poses any problems for the human population. A reassembly from this process is coupled with diverse entities and varying contexts and eventually grows into a section of other logic (Tirado & Cano, 2020). The result is the development of a novel course that disembodies the threat, transforms it into actionable data, and ultimately into an interface constituting linked surfaces between the order of the drone and the threat. The established assembly depends on technological advancements, which create and record discreet observations and develop a risk-order range. The movement shifts from a simple trackable unit to a drone-determined flow. The assembly surface exceeds the notions of risk and the threat’s space and time; the researchers place it in an event determined by the drone’s operations. At this point, they no longer aim to monitor an infectious vector’s surreptitious outbreak, enact immediate intervention measures, collect data on the impacted cases, and categorize them to attempt risk formalization and representation. Now, with the drones’ incorporation, the objective is to identify “with which of the drone’s actions the element of contagion works, with which maps of pathogens movement, with which of the drone’s action-movements it connects with the element of contagion and how they assemble. Surveillance takes on a new guise of being-assembled-together or being-assembled with on the plane or image of movement” (Tirado & Cano, 2020, p.16). This enables the prediction of infectious outbreaks.

Leveraging FHIR Genomics

Recent technological advancements in bioinformatics, deep sequencing, and high throughput have resulted in metagenomic strategies for quick, exact resolution of intricate ecological samples, as evidenced by Microsoft Premonition. It leverages a “Bayesian mixture model-based” (BMM) metagenomics pipeline able to determine known species taxa levels and approximate new species exiting in a single specimen (Pastusiak et al., 2023). The pipeline employs 10 tera-base genomic reference database and cloud-scale statistical machine learning to swiftly “(1) build probabilistic assignments from reads to species based on sequence similarity, (2) refine species probabilities for ambiguous reads by computing a global statistical model across all reads, and (3) identify novel, unexpected, and contaminant genetic material by aligning against all taxa with available (partial) genomic references, i.e. without a priori assumptions on which taxa might be present in a sample and without limiting the analysis to a small subset of genomic references (e.g. to only pathogens for computational reasons)” (Pastusiak et al., 2023, p.3). Consequently, Premotion’s metagenomics pipeline can be utilized to analyze publicly available datasets and those collected by drones to identify constituent species in each sample.

However, the CDC’s lack of data standards that delay information accrual might hinder the Premonition’s reach and effectiveness.

The acceleration into the study of inheritable diseases by next-generation sequencing has significantly increased the data volume, which is computationally challenging to sort through and analyze. They require specialized bioinformatics pipelines to assist in cleaning and processing before gene mapping, conditions replicating to the Premonition Project’s objectives (Oguzie et al., 2023). This secondary utilization of healthcare data in machine learning and analytics presents a vital value proposition for employing wide-ranging data sources such as genomic data and the FHIR format clinical data. Since the Premotion Project requires massive loads of data to aid in predictive analytics for determining and helping prevent epidemics, the FHIR will be a valuable integration to streamlining the process. The scientists can employ FHIR data to train machine learning models and examine how they can predict the human race’s susceptibility to being infected by a specific vector-borne disease.

The Premotion-FHIR integration should not be an uphill task because the CDC already leverages the Microsoft Azure FHIR server. The Azure architectural design merges “synthetic FHIR clinical data from Azure Health Data Services (using FHIR to Synapse Sync Agent OSS) and real publicly available 1000 Genomes Project data on Azure Synapse Analytics” (Cosgun, 2022, p.1). This design incorporates data, AI tools, a secure cloud, and interoperability, allowing researchers to analyze clinical data and genomics (Cosgun, 2022). An example application of this combination entails using data in FHIR bundles and then employing the “FHIR to Synapse Sync Agent” open-source solution to generate external views and tables to access the FHIR data from Azure Health Data Services to Azure Data Lake (Cosgun, 2022). This enables researchers to conduct Machine Learning and Analytics on FHIR data in Synapse Analytics Workspace.

Microsoft’s Premonition aims to perceive pathogens among insect species early before infecting humans. It achieves this objective by classifying insects such as mosquitoes as a device that can locate animal species and sample their blood, given that they feed on blood. The project employs robotic mosquito smart traps and drones to obtain mosquitoes from the environment and analyze their components for existing pathogens (Ravi et al., 2016). Researchers integrate gene sequencing techniques to deter pathogens from mosquitoes’ bolls sampled from the environment and then employ computational tools to search for known and unknown pathogens in the sequenced material (Maljkovic et al., 2020). This is where the FHIR standard comes in because identifying pathogens is critical to determining the type of prospective infectious outbreak and from which animal species it will most likely emanate. A CDC collaboration and Microsoft Azure tools pose a revolutionary strategy for predicting infectious outbreaks. The CDC can harness the predictive analytics power of Microsoft Premonition, incorporate the stunning data collection FHIR capabilities, and unify the data with Azure Fabric to enable it to shift from a mere reactive response to a proactive approach in analyzing infectious outbreaks. This methodology aids in locating pathogens and studying the impact of their behavior before they attack human populations. If the researchers successfully avert the outbreak, they may never get the chance to attack.

Since the CDC already has experience with FHIR data bundles, integrating them into the Project Premonition for more precise results should be achievable. The Azure API for FHIR enables interoperation and data sharing over the cloud (Nan & Xu, 2023). The Premonition Project has cloud-scale metagenomics that will leverage this data for Predictive Analytics. Its analytics pipelines utilize cloud-scale computing, taking advantage of the state-of-the-art Microsoft Azure’s enhancements in Azure Data Lake and Azure IoT (Stewart, 2024). Since the Premonition’s pipelines scan trillions of genomic materials from the collected environmental samples, integration of FHIR genomics will assist in interpreting the biological samples from the animal species to help evade biological threats. Moreover, the CDC experiences data collection complications because of a lack of standardization, resulting in delayed acquiring vital information (Collier & Molina, 2019). Leveraging FHIR standards can smoothen data-gathering procedures from various sources, minimize collection sequences, and fast-track data velocity. Since the CDC already harnesses Microsoft’s FHIR Azure server, incorporating FHIR Genomics standards extends the scope of data collection, especially in genomics, which will be extremely valuable in gene sequencing. The extra data source also augments epidemic prediction models, advancing their precision and analytics authority.

A Premonition Integration with Predict

Project PREDICT was a “United States Agency for International Development” (USAID) funded epidemiological research that aimed to provide early warnings of looming pandemics. The project was launched in 2009 as a response strategy to the influenza A virus subtype H5N1 (M). The USAID director at the time, Dennis Carroll, designed and oversaw the development, assisted by the University of California’s Jonna Mazet (McNeil, 2019; Morrison, 2018). Project PREDICT gathered over 140000 biological samples from diverse animal populations, encompassing 10000 bats and other 2000 assorted mammals (Baumgaertner & Rainey, 2020). The data collection was centralized to “hot interfaces” that characterized areas with high biodiversity, human-animal interactions, those that could support the spread of infections, and those densely populated by human beings (Baumgaertner & Rainey, 2020). Its virus hunting spanned several regions from the Amazon basin to the Congo basin. The project considerably reinforced laboratory diagnostic and global surveillance abilities for known and unknown viruses within numerous vital virus clusters such as coronaviruses, paramyxoviruses, influenza viruses, and filoviruses.

However, while PREDICT offers valuable insights into the discovery of known and new viruses, its capabilities fall short in predicting and successfully aversion epidemics such as The recent COVID-19. Extensive research affirms the limitations of the project and suggests better positioning for broader impact. According to Carlson (2020, p.1), much as PREDICT almost certainly uncovered hundreds of potential zoonoses, “their true zoonotic potential is almost impossible to assess, leading to the surprising statistic that the program only led to one conclusive discovery of a zoonosis, the Bas-Congo virus.” To this effect, human infection observation is the only valid means of differentiating the 10000 possibly zoonotic mammalian viruses from their 40000 low-risk equivalents. This limitation presents an excellent opportunity for integration with Microsoft Premonition. A collaboration of these two groundbreaking projects can be phenomenal and groundbreaking in not only studying human infections but also studying known and new pathogens to alleviate these infections before they occur. Microsoft’s statistical and machine-learning techniques can leverage the data generated by PREDICT to predict zoonotic potential. This merged model can help determine gaps in wildlife sampling, identify viral traits with the highest human transferability, pinpoint viral categories with the most significant zoonotic potential, and locate wildlife reservoirs.

Besides, the merger could address the concerns surrounding PREDICT’s inefficiencies. One virologist highlighted that it is insufficient to merely catalog the virus due to the difficulty of predicting spillover infections (Morrison, 2018). Additionally, considering the frequent virus mutation, some might mutate to novel hosts or die out before any studies have been conducted. These inadequacies led to the closure and funding cessation for PREDICT after only five years of operation. This shutdown garnered significant criticism from various stakeholders and even policymakers. An excellent case in point is the letter from Senators Elizabeth Warren and Angus King in 2020, stating that “The rise of 2019-nCoV heightens the need for a robust, coordinated, and proactive response to emerging pandemics – one of the roles that PREDICT played” (United States Senate, 2020, p.3). The senators were right, and PREDICT's revival with a Premonition integration can adequately address their concerns. Microsoft Premonition’s cloud-based, real-time data-collecting smart traps and analytics capabilities can seamlessly address these concerns. The smart traps can process loads of PREDICT-collected data and make real-time decisions, transport the data over the cloud, and allow pathogen analytics to drive epidemic prediction and prevention.  

Conclusion

Regarding previous surveillance systems, Microsoft’s Premonition provides significant changes and valuable insights. For instance, researchers are no longer simply tracking infectious vectors independently afforded by bio and tele-epidemiology; the sensor drones allow them to take direct action in real-time. Additionally, the project has contributed considerably to the redefinition of surveillance devices’ social-technical networks, as evidenced by the progressive loss of importance in the human aspect. The sensor drones are driving the world from monitoring global flows such as efforts to define the course of the COVID-19 virus – to creating these flows through their direct action. Microsoft’s Premonition presents a future where drones and animals can work collaboratively to impact other animal-pathogen-carrying species to modify their contact with other living organisms, migratory movements, or population sizes.

References

Alterovitz, G., Heale, B., Jones, J., Kreda, D., Lin, F., Liu, L., Liu, X., Mandl, K. D., Poloway, D. W., Ramoni, R., Wagner, A., & Warner, J. L. (2020). Fhir Genomics: Enabling Standardization for precision medicine use cases. Npj Genomic Medicine, 5(1). https://doi.org/10.1038/s41525-020-0115-6

Ayaz, M., Pasha, M. F., Alahmadi, T. J., Abdullah, N. N. B., & Alkahtani, H. K. (2023). Transforming Healthcare Analytics with FHIR: A Framework for Standardizing and Analyzing Clinical Data. Healthcare (Basel, Switzerland), 11(12), 1729. https://doi.org/10.3390/healthcare11121729

Baru, C., Pozmantier, M., Altintas, I., Baek, S., Cohen, J., Condon, L., Fanti, G., Fernandez, R. Jackso, E., Landman, B., & Zhang, P. (2022). Enabling AI innovation via data and model sharing: An overview of the NSF Convergence Accelerator Track D. AI Magazine, 43(1), 93-104.

Basulo-Ribeiro, J., & Teixeira, L. (2024). The Future of Healthcare with Industry 5.0: Preliminary Interview-Based Qualitative Analysis. Future Internet, 16(3), 68.

Baumgaertner, E. & Rainey, J. (2020). Trump administration ended pandemic early-warning program to detect coronaviruses. Los Angeles Times. https://www.latimes.com/science/story/2020-04-02/coronavirus-trump-pandemic-program-viruses-detection

Bekoff, M., & Pierce, J. (2009). Wild justice: The moral lives of animals. University of Chicago Press.

Burkhardt, H. A., Brandt, P. S., Lee, J. R., Karras, S. W., Bugni, P. F., Cvitkovic, I., Chen, A. Y., & Lober, W. B. (2021). StayHome: A FHIR-Native Mobile COVID-19 Symptom Tracker and Public Health Reporting Tool. Online journal of public health informatics, 13(1), e2. https://doi.org/10.5210/ojphi.v13i1.11462

Carlson, C. J. (2020). From PREDICT to prevention, one pandemic later. The Lancet Microbe, 1(1), e6-e7. https://doi.org/10.1016/S2666-5247(20)30002-1

Carrasco-Escobar, G., Moreno, M., Fornace, K., Herrera-Varela, M., Manrique, E., & Conn, J. E. (2022). The use of drones for mosquito surveillance and control. Parasites & vectors, 15(1), 473. https://doi.org/10.1186/s13071-022-05580-5

Cassidy, T. (2022, July 14). Overview of Azure Service Fabric - Azure Service Fabric. Azure Service Fabric | Microsoft Learn. https://learn.microsoft.com/en-us/azure/service-fabric/service-fabric-overview

Choney, S. (2020). Predicting epidemics like the weather: How Microsoft Premonition can help in the global fight against disease outbreaks. Microsoft. https://news.microsoft.com/source/features/innovation/microsoft-premonition/

Čirjak, D., Miklečić, I., Lemić, D., Kos, T., & Pajač Živković, I. (2022). Automatic pest monitoring systems in apple production under changing climatic conditions. Horticulturae, 8(6), 520. https://doi.org/10.3390/horticulturae8060520

Collier, A. R. Y., & Molina, R. L. (2019). Maternal mortality in the United States: updates on trends, causes, and solutions. Neoreviews, 20(10), e561-e574. https://doi.org/10.1542/neo.20-10-e561

Cosgun, E. (2022). Combine and Explore FHIR Server and Genomics data in Azure Synapse Analytics. Microsoft. https://techcommunity.microsoft.com/t5/healthcare-and-life-sciences/combine-and-explore-fhir-server-and-genomics-data-in-azure/ba-p/3298335

Garza, M. Y., Rutherford, M. W., Adagarla, B., Eisenstein, E., Kumar, K. R., Zimmerman, K. O., Topaloglu, U., & Zozus, M. (2021). Evaluating site-level implementations of the HL7 FHIR standard to support eSource data exchange in Clinical Research. Studies in Health Technology and Informatics. https://doi.org/10.3233/shti210188

Godlovitch, I., & Kroon, P. (2022). Interoperability, switchability and portability: Implications for the cloud. WIK-Consult Report.

Guillot, C., Bouchard, C., Berthiaume, P., Mascarenhas, M., Sauvé, C., Villeneuve, C. A., & Leighton, P. (2021). A portrait of sentinel surveillance networks for vector-borne diseases: a scoping review supporting sentinel network design. Vector-Borne and Zoonotic Diseases, 21(11), 827-838. https://doi.org/10.1089/vbz.2021.0008

Hassan, M., Zmij, K., Azhygulov, K., & Sitaula, S. (2022). Cloud Computing Services and Microsoft Azure. Why Microsoft Azure?

Hernández, M. (2024, January 7). Microsoft Fabric: The future of data integration in the cloud. Medium. https://medium.com/@marinohdezb/microsoft-fabric-the-future-of-data-integration-in-the-cloud-bb1be51960f2

Iqbal, M. J., Iqbal, M. M., Ahmad, I., Alassafi, M. O., Alfakeeh, A. S., & Alhomoud, A. (2021). Real-time surveillance using deep learning. Security and Communication Networks, 2021, 1-17. https://doi.org/10.1155/2021/6184756

Joyce, P. (2023). Interoperability in Intercloud. NUI Galway.

Kumar, S. (2023). Data Silos A Roadblock for AIOps. arXiv preprint arXiv:2312.10039.

Liu, Z., Gao, J., Yang, G., Zhang, H., & He, Y. (2016). Localization and classification of paddy field pests using a saliency map and deep convolutional neural network. Scientific reports, 6(1), 20410. https://doi.org/10.1038/srep20410

Maljkovic Berry, I., Melendrez, M. C., Bishop-Lilly, K. A., Rutvisuttinunt, W., Pollett, S., Talundzic, E., Morton, L., & Jarman, R. G. (2020). Next generation sequencing and bioinformatics methodologies for infectious disease research and public health: approaches, applications, and considerations for development of laboratory capacity. The Journal of Infectious Diseases, 221(Supplement_3), S292-S307. https://doi.org/10.1093/infdis/jiz286

McNeil Jr. D. G. (2019). Scientists Were Hunting for the Next Ebola. Now the U.S. Has Cut Off Their Funding. The New York Times. https://www.nytimes.com/2019/10/25/health/predict-usaid-viruses.html

Morrison, J. (2018). Can Virus Hunters Stop the Next Pandemic Before It Happens? Smithsonian Magazine. https://www.smithsonianmag.com/science-nature/how-to-stop-next-animal-borne-pandemic-180967908/

Nan, J., & Xu, L. Q. (2023). Designing interoperable health care services based on fast healthcare interoperability resources: Literature review. JMIR Medical Informatics, 11(1), e44842. https://doi.org/10.2196/44842

Odisho, A. Y., Lui, H., Yerramsetty, R., Bautista, F., Gleason, N., Martin, E., Young, J. J., Blum, M., & Neinstein, A. B. (2020). Design and development of referrals automation, a smart on FHIR solution to improve patient access to specialty care. JAMIA Open, 3(3), 405–412. https://doi.org/10.1093/jamiaopen/ooaa036

Oguzie, J. U., Petros, B. A., Oluniyi, P. E., Mehta, S. B., Eromon, P. E., Nair, P., Adewale-Fasoro, O., Ifoga, P. D., & Happi, C. T. (2023). Metagenomic surveillance uncovers diverse and novel viral taxa in febrile patients from Nigeria. Nature Communications, 14(1), 4693.

Ohm, J. R. (2018). Mosquito feeding and fitness: An exploration of mosquito diets and behaviors that impact mosquito fitness, with applications for mosquito mass rearing and control. The Pennsylvania State University.

Pastusiak, A., Reddy, M., Chen, X., Hoyer, I., Dorman, J., Gebhardt, M., Carpi, G., Norris, D., Pipas, J., & Jackson, E. (2023). A metagenomic analysis of the phase 2 Anopheles gambiae 1000 genomes dataset reveals a wide diversity of cobionts associated with field collected mosquitoes. https://orcid.org/0000-0002-4631-2777

Peckham, R., & Sinha, R. (2019). Anarchitectures of health: Futures for the biomedical drone. Global Public Health, 14(8), 1204-1219. https://doi.org/10.1080/17441692.2018.1546335

Pettorelli, N. (2019). Satellite remote sensing and the management of natural resources. Oxford University Press.

Rahman, A., Liu, X., & Kong, F. (2013). A survey on geographic load balancing based data center power management in the smart grid environment. IEEE Communications Surveys & Tutorials, 16(1), 214-233. https://doi.org/10.1109/SURV.2013.070813.00183.

Ravi, P., Syam, U., & Kapre, N. (2016, November). Preventive detection of mosquito populations using embedded machine learning on low-power IoT platforms. In Proceedings of the 7th Annual Symposium on Computing for Development (pp. 1-10). https://doi.org/10.1145/3001913.3001917

Rosenau, L., Majeed, R. W., Ingenerf, J., Kiel, A., Kroll, B., Köhler, T., Prokosch, H. U., & Gruendner, J. (2022). Generation of a Fast Healthcare Interoperability Resources (FHIR)-based Ontology for Federated Feasibility Queries in the Context of COVID-19: Feasibility Study. JMIR medical informatics, 10(4), e35789. https://doi.org/10.2196/35789

Smith, B., & Shum, H. (2018). The future computed. Artificial Intelligence intelligence and its role in society.

Stewart, B. (2024). Architecting IoT Solutions on Azure. ““O’Reilly O’Reilly Media, Inc.”

Talapko, J., Škrlec, I., Alebić, T., Jukić, M., & Včev, A. (2019). Malaria: the past and the present. Microorganisms, 7(6), 179. https://doi.org/10.3390/microorganisms7060179

Tambo, E., & Xiao-Nong, Z. (2014). Acquired immunity and asymptomatic reservoir impact on frontline and airport Ebola outbreak syndromic surveillance and response. Infectious diseases of poverty, 3, 1-11. https://doi.org/10.1186/2049-9957-3-41

Tirado, F., & Cano, P. T. (2020). Drones and epidemiology: A new anatomy for surveillance. BioSocieties, 15(1), 115-133. https://doi.org/10.1057/s41292-019-00144-w

United States Senate. “Letter from Senator Elizabeth Warren and Senator Angus S. King, Jr., to USAID Administrator Mark Green.”

Vorisek, C. N., Lehne, M., Klopfenstein, S. A., Mayer, P. J., Bartschke, A., Haese, T., & Thun, S. (2022). Fast Healthcare Interoperability Resources (FHIR) for Interoperability in Health Research: Systematic Review. JMIR Medical Informatics, 10(7). https://doi.org/10.2196/35724

World Health Organization. (2020). Malaria eradication: benefits, future scenarios, and feasibility: a report of the Strategic Advisory Group on Malaria Eradication. World Health Organization.

Yan, J., Gangoso, L., Ruiz, S., Soriguer, R., Figuerola, J., & Martínez‐de la Puente, J. (2021). Understanding host utilization by mosquitoes: Determinants, challenges, and future directions. Biological Reviews, 96(4), 1367-1385. https://doi.org/10.1111/brv.12706