Big data analytics
Big data analytics typically involves high-performance analytics technologies, data mining, predictive models, machine learning, natural language processing, and statistical algorithms to process big data – data sets characterized by high levels of volume, variety, veracity, and velocity. Leveraging big data helps extract meaningful information essential to making faster and better decisions (Dash, Shakyawar, Sharma, & Kaushik, 2019)). Healthcare big data analytics plays a major role in the delivery of personalized and value-based care, clinical risk management, care variability mitigation, and standardized internal and external reporting (Huang, Mulyasasmita, & Rajagopal, 2016; Peddoju , Kavitha, & Sharma, 2019).
With growing adoption of eHealth, mHealth, telehealth, health monitoring wearables, patient portals, and other healthcare information systems, the volume, variety, veracity, and velocity of consumer, patient, and clinical data will constantly increase. Appropriate healthcare big data analytics is needed to extract actionable insights and make strategic decisions based on such data (Peddoju et al., 2019).
In the developing world, pediatric pneumonia is identified as one of the major causes of deaths among children (Peddoju et al., 2019). To avoid life-threatening pneumonia-induced complications, it is important to identify the children at risk of treatment failure in a timely manner. In case pneumonia symptoms are detected, prompt medical attention should be sought to ensure accurate and timely diagnosis is made and proper treatment given. Peddoju et al. (2019) highlight the need to analyze the history of patients’ records, doctors’ expert opinions and previous treatment for similar symptoms, and cloud-based diagnostic reports (such as X-ray and blood tests) to support quick and accurate decisions regarding pediatric pneumonia interventions.
Role of Network and/or Telecommunications Technologies
Peddoju et al. (2019) recommend the use Mobile Cloud and Big Data analytics to supporting the collection, storage, and analysis of large and complex healthcare data towards generating actionable insights. In the context of treating pediatric pneumonia, cloud computing is identified as one of the major supporting technologies. Generally, cloud computing refers to an interconnected network of diverse types of devices that deliver three main categories of services, namely software-as-a-service (SaaS), platform-as-a-service (PaaS), or infrastructure-as-a-service (IaaS). Users acquire the cloud-based services on-demand (Peddoju et al., 2019). The approach proposed by Peddoju et al. (2019) assumes that all stakeholders (especially the doctors, staff, and patients) are able to share information via the cloud. Furthermore, the cloud is used to interconnect diverse doctors’ information, especially diagnosis and treatment reports.
In the studied case, the cloud is primarily used as the basis for storing the childhood pneumonia patients’ data. The patient-related information captured in each hospital is stored in the cloud. Then, the entire collection of pneumonia data is integrated via the cloud. The data stored in the cloud can be readily shared and accessed by all stakeholders across the world at any time. Importantly, the pneumonia treatment data is stored in the cloud so that any doctor can use it to make quick and precise decisions regarding how to treat pediatric pneumonia cases at hand. Doctors are empowered to choose the best possible course of actions from the readily available data. Furthermore, the cloud-based data storage approach allows doctors to readily share their views opinions regarding pediatric pneumonia symptoms and treatment interventions. To accomplish this, doctors are allowed to update the cloud-based repository with any relevant information. Therefore, use of a cloud-based information repository could contribute to knowledge creation and sharing.
The primary motive for adopting cloud-based treatment is its quick service provisioning, on-demand usage patterns, and significant cost savings. Cloud computing contributes to the implementation of an integrated IT infrastructure that enhances the overall data sharing potential between hospitals, laboratories, and government agencies among other key healthcare stakeholders. Moreover, the cloud implies an opportunity to amass data generated from eHealth and mHealth technologies. Integrating healthcare applications, services, and data into the cloud could support remote patient monitoring (Peddoju et al., 2019). Consequently, the overall need for hospital admission is significantly reduced.
Critical Success and/or Failure Factors
Peddoju et al. (2019) opine that learning from the past could create opportunities for better results in the future. Using Mobile Cloud and Big Data analytics, it is possible to proactively identify the pneumonia symptoms among children so that timely and precise treatment can be given. Consequently, the proactive intervention against pneumonia could reduce the incidence of life-threatening pneumonia cases. In addition, unnecessary hospital admissions and re-admissions can be avoided. In turn, significant healthcare cost savings could be realized.
Maintaining different forms of health-related data, including structured, semi-structured, and unstructured, implies an opportunity to leverage big data and gain greater insights into healthcare issues of interest. Notably, close to 80% of health data is inherently unstructured, yet it could contain important cues related to healthcare needs and potential interventions. A study conducted by McKinsey Global Institute showed that creative and effective use of big data in the U.S. healthcare could generate value in excess of $300 billion every year. Cost reduction is the primary form of value created by healthcare big data analytics (Belle et al., 2015). To have a comprehensive perspective of disease states, an aggregated strategy should be used whereby structured and unstructured clinical and non-clinical data is analyzed. Without strong aggregation of all available data sets, it would be difficult to consider every potential marker for medical assessment (Belle et al., 2015). Therefore, all forms of health data should be leveraged to increase the quality of generated insights and associated healthcare delivery decisions.
Digitization of patient data by healthcare organizations plays a major role in enhancing the ease and accuracy of data analysis. In addition, digitization makes it easy to collect, store, and distribute data across diverse digital platforms (Peddoju et al., 2019). However, to derive optimal value from the vast digital data, a long-term storage solution should be implemented. In addition, fast and accurate statistical algorithms are needed to facilitate decision-assisting automation. For example, medical images constitute a major source of data often used to support diagnosis, assessment, and planning missions. Magnetic Resonance Imaging (MRI), X-Ray, computed tomography (CT), ultrasound, fluoroscopy, and photo-acoustic imaging are some of the common clinical imaging techniques. Nevertheless, uncompressed images often require large storage and processing capacities (Belle et al., 2015; Peddoju et al., 2019).
Cloud computing is essential to ensuring “anytime, anywhere” data availability. Consequently, cloud solutions support enhanced accessibility and sharing of health data. The cloud ensures information confidentiality and availability through mechanisms such as multi-factor authentication, data encryption, data tokenization, and strict SLAs (Bamiah, Brohi, Chuprat, & Brohi, 2012; Peddoju et al., 2019). The importance of cloud computing in healthcare is underscored by the need to centrally store patients’ data in the cloud from where it can be readily accessed. Centralized storage implies that healthcare providers have a single “source of truth” as the basis for becoming more data-driven and patient-centric. Centralized data storage and management minimizes the incidence of errors. Furthermore, cloud computing provides on-demand, elastic, and cost-efficient data services.
Cloud computing and service-oriented architecture (SOA) may be used to efficiently provide scalable, interoperable, and integrated healthcare IT infrastructures. With cloud computing and SOA, it is possible to deliver integrated telemedicine, medical imaging, and electronic medical records (EMR) applications and services. In addition, an integrated IT infrastructure supports the activities of all the key stakeholders involved in healthcare services. For example, an integrated cloud-based application creates an opportunity for enhanced collaboration between patients, healthcare practitioners, researchers, insurers, and public health officials (Belle et al., 2015; Peddoju et al., 2019). Consequently, the incidence of data inconsistences and medical errors is considerably reduced.
The distributed nature of cloud environments necessitates robust information security and data protection measures. With cloud-based services, there are significant data breach threats and risks. The issue is especially complicated when personal data need to be stored or shared. Even worse, working with health-related big data could make it difficult to detect incidents of stealth data theft (Bamiah et al., 2012). In addition, the successfulness of cloud-based pediatric pneumonia treatment is significantly dependent on how well sensitive data is protected against potential breaches against data confidentiality and integrity. As an example of a data protection technique, Belle et al. (2015) cite iDASH as a viable approach to integrated data analysis and sharing while anonymizing biomedical data for privacy preservation purposes. The Health-e-Child consortium is another effort geared aimed at supporting secure integrated data analysis and sharing (Belle et al., 2015).
The proposed cloud-based data storage approach to pediatric pneumonia monitoring and treatment requires doctors to share their views and update the centralized repository accordingly. While this could contribute to the cause of knowledge creation in the healthcare domain, it brings about the challenge of introduction of unreliable information (Peddoju et al., 2019). While vast cloud-based medical repositories could help identify potential treatment procedures, it could be difficult to identify false information contained in those repositories. Therefore, trust is a fundamental cloud computing requirement. This makes it important to devise strategies for evaluating the trustworthiness of doctors’ opinions. A rating functionality should be implemented to allow doctors to appraise the pneumonia intervention suggestions made by colleagues. This could help overcome the challenges caused by erroneous data. Furthermore, the credibility of doctors making recommendations may be difficult to verify. To overcome credibility concerns, doctors’ professional qualifications and contact details ought to be provided to facilitate clarification. Another challenge takes the form of inadequate technical and human resources to support healthcare big data initiatives (Peddoju et al., 2019).
Network and/or Application Strategy Utilized in the Case
Figure 1 illustrates the healthcare system proposed by Peddoju et al. (2019). The proposed healthcare model comprises of the following layered components (Peddoju et al., 2019):
- The cloud computing platform: To store and integrate childhood pneumonia patients’ data obtained from different sources. In addition, the cloud provides a scalable and on-demand infrastructure to support the ever-increasing big data digitization and processing needs.
- Childhood pneumonia patient information: Typically include digitized genomic, clinical, patient behavior, healthcare publication, and administrative and business data that make up healthcare big data. Predicting childhood pneumonia cases requires an aggregated strategy whereby structured, semi-structured, and unstructured data that stem from diverse clinical and non-clinical modalities is used to indicate the need for clinical intervention.
- Big data integration and analytics tools and algorithms: Required for ingestion, integration and clustering, and distributed processing of structured and unstructured childhood pneumonia patient’s data across multiple cloud-based nodes. Hadoop that uses MapReduce is an example of a tool that supports rapid analysis and transformation of large data sets. Other analytics that are employed in the proposed system include Cassandra, MongoDB, Informix, Ingres, LucidDB, Teradat, and Interbase.
- Data discovery and visualization: Visual design tools for processing patient data. Importantly, doctors need to readily update the cloud-based database with relevant information. The system may implement Predictive Modeling Mark-up Language (PMML) for visualizing big data analytics results and reports.
- Predictive analysis: Powerful algorithms required for fast generation of decision-supporting information. Applicable algorithms may include classification, regression, association, and clustering.
- Enterprise and ad-hoc monitoring and reporting tools: Stakeholders (doctors, staff, patients, and parents and/or guardians) will publish information via the cloud, and access reports generated by big data analytics via the system’s presentation layer.
Figure 1: Application strategy used in the case (Peddoju et al., 2019)
The proposed approach to childhood pneumonia monitoring entails relying on the history of patients’ records, doctors’ experience and previous treatment for similar symptoms, and diagnostic reports (such as imaging data and blood tests) from the cloud in addition to big data analytics tools and techniques to enable doctors to make quick and precise pediatric pneumonia treatment decisions. Primarily, cloud computing is used to allow for “anytime, anywhere” data availability and accessibility. In addition, the cloud provides a centralized health-related data repository that can be readily updated by authorized doctors. In Peddoju et al. (2019)’s model, big data analytics tools and techniques are implemented in the cloud. These include aspects of data preparation, modeling, and distributed processing as well as predictive analysis, data discovery, and visualization.
A number of improvements can be made to the pediatric pneumonia monitoring and treatment approach proposed by Peddoju et al. (2019). To start with, IoT technology ought to be integrated with the proposed healthcare model to optimize the use of real-time data streams. Here, IoT-based devices will automatically monitor the patients’ health conditions and send the information to the cloud-based data integration and analytics platform. Then, a pneumonia monitoring system will integrate the patient-related data via the cloud. An analysis of the continuously generated patients’ data could help gain crucial insights into patients’ health statuses and provide relevant diagnostic and treatment information to concerned patients and parents or guardians in a timely manner. With automated IoT-based pneumonia monitoring capability, it would be possible to provide real-time alerts and more personalized medical care. Figure 2 shows the modified diagram that illustrates the improved solution.
Figure 2: Application strategy with the proposed improvement
Bamiah, M., Brohi, S., Chuprat, S., & Brohi, M. N. (2012, December). Cloud implementation security challenges. In 2012 International Conference on Cloud Computing Technologies, Applications and Management (ICCCTAM) (pp. 174-178). IEEE.
Belle, A., Thiagarajan, R., Soroushmehr, S. M., Navidi, F., Beard, D. A., & Najarian, K. (2015). Big data analytics in healthcare. BioMed research international, DOI: https://doi.org/10.1155/2015/370194
Dash, S., Shakyawar, S. K., Sharma, M., & Kaushik, S. (2019). Big Data in Healthcare: Management, Analysis and Future Prospects. Journal of Big Data, 6(1), 54-73.
Huang, B. E., Mulyasasmita, W., & Rajagopal, G. (2016). The path from big data to precision medicine. Expert Review of Precision Medicine and Drug Development, 1(2), 129-143.
Peddoju, S. K., Kavitha, K., & Sharma, S. C. (2019). Big Data Analytics for Childhood Pneumonia Monitoring. In Web Services: Concepts, Methodologies, Tools, and Applications (pp. 1129-1145). IGI Global.