Leveraging the Power of Data Science in Healthcare
By Arunima Rajan
We explore the transformative power of data in revolutionising population health, showcasing its pivotal role in enhancing patient treatment strategies, advancing predictive analytics, and refining public health initiatives.
The Dawn of Data-Driven Healthcare
In March 2020, the world was brought to a standstill by the COVID-19 pandemic, marking the beginning of a crisis unlike any other in modern history. Yet, amidst this unprecedented challenge, a silver lining emerged: a deluge of data flowing from every corner of the globe, offering insights into the virus's spread, patient outcomes, and the efficacy of various treatments. This immense dataset, powered by the collective efforts of healthcare professionals and supported by significant funding from leading institutions like the US National Institutes of Health (NIH) and the UK's Health Data Research (HDR UK), has not only reshaped our approach to managing the current pandemic but has set the stage for a transformative leap in healthcare.
The Role of Predictive Models in Modern Medicine
Kristin Kostka, a leading figure focused on health data science and informatics at Northeastern University discusses the power of predictive models that use real-world data—essentially, information collected from everyday healthcare activities. This data enables healthcare teams to leverage existing health information from a variety of patient populations to improve care for their patients. An example is how patients in different populations experiencing chronic diseases may respond better to one treatment versus another. Kostka's involvement in the SEEK COVER project, aimed at assessing individual risks for COVID-19, is a prime example of the benefits of this data-driven approach focused on diverse patient populations.
The Emergence of Health Information Exchanges (HIEs)
“Moving forward, we will see Health Information Exchanges (HIE) play a vital role in making health information more accessible by linking patient care across different healthcare providers and sharing diagnostic and treatment information for improved patient care overall. In fact, the Department of Health and Human Services recently set national standards for health data exchanges in order to move this type of data share forward in a safe, protected way,” she adds.
Challenges in Chronic Disease Management
In this context Luigi Vacca, head of AI and Machine Learning at tallywell™, underscores the significance of leveraging data insights and predictive analytics in healthcare. He highlights the success of predictive models in the US and UK in identifying individuals at risk for chronic diseases, leading to improved healthcare outcomes and cost reductions. Despite the progress, challenges like data inconsistency, privacy concerns, and the digital divide need addressing to fully harness data-driven healthcare's potential.
The COVID-19 pandemic has accelerated the adoption of data analytics in healthcare, facilitating better management of the virus spread and improving patient care through real-time data analysis. Vacca is optimistic about the future of healthcare, noting the transformative impact of emerging data technologies, AI-driven models, and HIEs. These advancements enable more effective disease prediction, prevention, and patient care coordination, paving the way for improved healthcare outcomes globally. However, overcoming existing challenges is crucial for the full realisation of data-driven healthcare's benefits.
Advancing Public Health Strategy with Data Analytics
Vasudev Bailey is the Managing Partner of ARTIS Ventures, a techbio focused venture capital firm. He notes that population studies are complex. “Clinical trials need to capture not just genetic, but also societal diversity. Information on what a study involves and how participant data will be collected and used must be presented in a way that is not only comprehensive but widely accessible. Public health officials have a responsibility to adequately simplify the necessary guidance,”he adds.
Bailey highlights the critical role of data in shaping policy decisions and guiding public health practices. He points out that data is essential for determining the target demographic for services, identifying the locations where these services are most needed, and deciding the extent of their coverage. Public health management fundamentally depends on detailed, multi-dimensional data that reflects the diverse characteristics of target populations. In recent years, a range of public health efforts have been underpinned by systematically collected data at the population level. Such efforts include the creation of vaccination sites tailored to the needs of rural communities, guided by vaccine uptake statistics, the use of epidemiological data to track the spread of COVID-19, and the informed decision-making regarding the adoption or coding of healthcare innovations based on health technology assessments (HTA) or economic analysis, such as for Ozempic. Additionally, the implementation of mandatory calorie labelling in restaurants, driven by data on Body Mass Index (BMI) and eating habits, and the subsequent evaluation of its impact, serve as further examples of data's pivotal role in public health initiatives, Bailey explains.
Addressing Social Determinants of Health
Bailey points out that social elements play a pivotal role in determining health outcomes, noting that the area where a person lives can often provide a more accurate forecast of their health than their genetic makeup. He explains that among the various factors influencing health—such as genetics, access to healthcare, individual behaviours, and socio-environmental conditions—it is the social and environmental aspects that contribute as much as 20% to an individual's risk of health issues, whereas healthcare access accounts for about 10%. Emphasising the importance of incorporating these social determinants, especially income, into health policy is essential for addressing the needs of the entire population. Bailey further mentions that studies, such as those conducted as part of the National Family Health Survey in India, are vital. He highlights India's rich genetic and social tapestry, illustrating this with the potential of genetic testing in tribal communities to identify susceptibility to genetic conditions like sickle cell anaemia. Such proactive measures could lead to early intervention, significantly improving lives and potentially saving billions through enhanced productivity and healthcare efficiencies.
“And then there’s the issue of voluntary participation- how do we encourage clinical trial participation in a manner that’s not coercive and is fully voluntary? In the US, for example, the Tuskegee study conducted between 1932 and 1972, where 600 African American men were involuntarily enrolled in a Syphilis study with the lure of receiving “free medical care,” resulted in deep mistrust of public health officials that lingers within African American communities to this day,” he adds.
Importance of Data Privacy and Security
Bailey emphasises that safeguarding data security, ensuring confidentiality and privacy for patients, and upholding data integrity remain persistent challenges in the public health sector. He advocates for the thorough anonymization of patient data and the strengthening of cybersecurity measures to boost confidence in population studies, not only among participants but also among professionals who depend on the reliability of the data collected.
Bailey points out that the safeguarding of data, preserving patient confidentiality, and upholding data integrity continuously present significant challenges within the public health domain. He notes that by prioritising the anonymization of patient information and strengthening cybersecurity protocols, population studies can secure increased trust from both the participants and the specialists who depend on the reliability of the research findings.
The Impact of Data on Population Health
Bailey explains that through the analysis of wastewater—specifically, by identifying virus particles in sewage discharged from residences—health officials at the state and national levels were able to track the spread of COVID-19 in real-time during periods when the testing infrastructure was overstretched. Additionally, he mentions that in the initial stages of the COVID-19 pandemic, prior to the availability of vaccines, a significant number of people received timely alerts on their smartphones about possible exposure to the virus, which effectively facilitated testing and helped in controlling the spread.
Expanding Beyond COVID-19
“Scaling digital health infrastructure and extending data generation to other disease areas will not only improve health outcomes for populations, but also will provide a massive healthcare savings opportunity for governments – according to a McKinsey report, this opportunity could equate to $1.5 trillion to $3 trillion in savings a year by 2030,”he concludes.
The Indian Perspective on Data-Driven Healthcare
Surveys such as the National Family Health Survey (NFHS), Annual Health Survey, and those conducted by the National Nutrition Monitoring Bureau are invaluable tools for assessing the health status, demographic trends, and nutritional profiles of populations across India. “While these surveys provide valuable insights into various health indicators, they do have inherent limitations such as lag time in data availability, potential for sampling biases, and limited granularity. In the era of rapidly evolving healthcare dynamics, there is undoubtedly a growing need for real-time data to enable timely interventions, monitor emerging health trends, and inform evidence-based policymaking,” adds Rajeev Srivastava, Head West Zone- Healthcare Government Business of Samsung Electronics.
Srivastava points out that the COVID-19 pandemic has indeed underscored the critical importance of data-driven approaches in managing population health. “With the emergence of the pandemic, there has been a noticeable shift towards leveraging data for better healthcare delivery, resource allocation, and decision-making processes. The National Digital Health Mission (NDHM) initiated by the Government of India is a landmark initiative aimed at transforming the healthcare ecosystem through the seamless exchange of health information. It holds immense potential to revolutionise healthcare delivery by promoting interoperability, ensuring data privacy, and empowering individuals with greater control over their health data. The NDHM has the potential to catalyse innovation, improve healthcare access, and enhance the overall efficiency of the healthcare system in India,” he explains.
He also notes that India’s private sector plays a pivotal role in leveraging data for population health management. Private healthcare entities possess valuable clinical data, technological capabilities, and innovative solutions that can complement government efforts in improving healthcare outcomes. Successful examples of private sector contributions include the development of
advanced analytics platforms, telemedicine solutions, and health informatics systems that enable efficient data aggregation, analysis, and predictive modelling. Collaborative partnerships between public and private stakeholders can foster synergies, promote knowledge exchange, and drive transformative changes in healthcare delivery models.
Collaboration between Clinicians and Technologists
Srivastava also notes that collaboration between clinicians and technology teams is indispensable for the successful implementation of healthcare technology solutions. Clinicians bring valuable domain expertise, clinical insights, and end-user perspectives that are essential for designing user-friendly, clinically relevant solutions. By actively involving clinicians in the development and deployment phases, healthcare organisations can mitigate the risk of administrative failures, ensure alignment with clinical workflows, and enhance user acceptance. Effective communication, interdisciplinary teamwork, and mutual respect for each other's expertise are key enablers of successful collaboration between clinicians and technology teams.
Empowering Health Equity and Innovation
Camille Cook, Sr. Director at LexisNexis Risk Solutions, emphasises the transformative role of data analytics in health management, highlighting the shift towards innovative data handling through technology like AI and machine learning. These advancements enable early detection and targeted healthcare interventions, improving response times and decision-making. Cook notes that leveraging these technologies enhances predictive modelling, patient engagement, and the development of patient-centric strategies, marking a significant evolution in population health outcomes.
She emphasises the significance of factoring in Social Determinants of Health (SDOH) beyond conventional medical records to enhance health policies. It's essential to acknowledge the role of economic stability and the availability of fundamental necessities in mitigating health inequities. Utilising SDOH information enables precise measures to assist at-risk populations and advance health fairness. Current strategies are geared towards using this knowledge to better access and results, including the growth of telehealth services and community-based health programs. She also highlights the paramount importance of ensuring data privacy and security within health analytics, pointing out tokenization as a technique to anonymize information while adhering to HIPAA and GDPR regulations, thereby safeguarding patient confidentiality and sustaining trust.
She underlines how predictive analytics enable healthcare organisations to foresee and address potential health crises by analysing past data, leading to targeted preventive actions and efficient resource use. This approach not only aims to limit disease spread but also improves health outcomes. Highlighting the crucial collaboration between healthcare experts and data scientists,
she notes this partnership enhances decision-making and cost efficiency. Additionally, technological advancements have transformed public health management by facilitating the integration of various data sources and advanced analytics, breaking down data silos and allowing for informed policy decisions, thus providing essential insights for addressing global health challenges in a more comprehensive manner.
Future Directions and Challenges
“Interoperability remains a formidable challenge for the successful implementation of healthcare solutions in India. To overcome this challenge, stakeholders must prioritise the adoption of standardised data exchange protocols, interoperability frameworks, and open application programming interfaces (APIs). Embracing internationally recognized standards such as HL7 FHIR (Fast Healthcare Interoperability Resources) can facilitate seamless data exchange and interoperability across disparate healthcare systems. Additionally, fostering a culture of collaboration, promoting data sharing agreements, and incentivizing interoperable solutions can accelerate progress towards achieving interoperability goals in India's healthcare ecosystem. In conclusion, the convergence of data, technology, and healthcare holds immense potential to transform the landscape of population health management in India. By harnessing the power of data-driven insights, fostering collaborative partnerships, and embracing interoperable solutions, we can collectively advance towards the shared goal of ensuring equitable access to high-quality healthcare for all,” adds Srivastava.
Nilesh Chandra is a partner in Healthcare Strategy and Data Analytics, Lead for US Healthcare Analytics at PA Consulting Group, an innovation and transformation consulting firm. Nilesh Chandra says, “A very effective mechanism for integrating disparate data is through the development of clinical data registries that span multiple health systems, and often span entire medical specialties across the country.” What are the main barriers to achieving this integration, and how can they be overcome? Nilesh notes, "A few barriers exist. First, many health systems are resistant to sharing their data and the value proposition to answer – ‘why should we submit our data to a registry,’ is not well articulated. Second, healthcare data is messy, it takes a good amount of effort to collate, structure and normalise data from multiple systems and sources.”
Nilesh Chandra says that “AI and big data analytics can quite significantly narrow health disparities across the globe. Even relatively simple pattern matching tools and meta data based data transformation approaches can significantly reduce the time and effort in normalising data from multiple sources, freeing up resources and time to pursue analysis rather than data collection. For example, in our work developing clinical data platforms for clients, we have implemented automated routines and are able to ingest data in a fully automated manner.”
He adds, “The first step really has to be the collection of representative data, in near real-time. And then we can build AI models and other signal detection mechanisms to identify potential health issues. Since COVID-19, many cities have implemented sewage testing mechanisms to identify the degree of community spread. It is a highly credible mechanism. However, the data collection isn’t automated, isn’t easily scalable and requires consistent funding and effort.”
“We need to reduce the time it takes to generate insight and get the results published in peer reviewed journals that are widely read. Furthermore, we need new and different mechanisms to disseminate knowledge within healthcare. During COVID-19, it took a very short period of time to transfer cutting edge knowledge around the world (giving patients an OTC steroid and laying them on their side improved lung function and reduced mortality). That only happened because for a unique moment in time, the entire attention of the health community was focused on this one threat. Our goal should be to get similar advances and discoveries every day, even in normal times, in the hands of clinicians as quickly as we were able to do during the COVID-19 pandemic,” he concludes.