Embracing Data Science is the Future of Healthcare

machine learning
Dmitri Adler and John Nader    
Icon clock
   July 2022         
machine learning        Blog
financial services data analytics

Impressive progress is happening at the crossroads of healthcare and data science which we explored in part one of this three-part series. An abundance of healthcare data and advanced technologies are opening doors to nearly limitless possibilities for streamlining operations, combatting the challenges of the day, and improving patient outcomes.

In this article, the second in our series probing data science in the healthcare and life sciences industries, we will investigate data science’s expanding role in healthcare and the tools that help providers optimize their data assets and avoid common pitfalls. In addition to my examination of data science training and technologies in the field, Data Society’s Co-Founder, Chief Operating Officer, and General Counsel John Nader will contribute his knowledge of the legal and ethical issues that accompany the expansion of healthcare data applications.

Data Science Applications for Improved Healthcare

financial data science

Once a relative laggard in adopting data science technologies, today’s healthcare industry is making up for the lost time. As professionals in the sector increasingly embrace the tools and techniques to leverage their abundant data assets, their data science skills must likewise grow. Healthcare practitioners can only reap the full benefits of data literacy and navigate the perils that data science applications present if they are prepared with the necessary training. Let’s start with a review of some of the data science tools that improve patient care and increase the need for data science training in the healthcare workforce.

Widespread Digitization and the Healthcare Experience

As today’s consumers become more adapted to the convenience and accessibility of digital transactions, their expectations for healthcare encounters have changed. Digitization is helping healthcare organizations create patient touchpoints that more closely resemble their experiences in other areas. Here are some examples:

  • Patients—increasingly empowered by sensors and mobile devices—can now play more active roles in managing their wellness journeys by accessing, reporting, and tracking information about their health through wearable monitors and other technologies.  
  • Broader sharing of records helps healthcare providers connect disparate data points relevant to a patient’s history, reducing the need for patients to provide the same information repeatedly or take the same tests multiple times and limiting the likelihood of data input errors. 
  • Digitization can provide patients with pharmacy experiences that more closely match the convenience they’ve grown accustomed to, giving text reminders of prescription refills and electronic access to medication information. 
  • Offering online scheduling options allows patients to set up and change appointments without placing calls between certain hours and waiting on hold for assistance. Streamlining the scheduling process saves time and reduces frustration for patients and healthcare scheduling staff.   
  • Digital technologies that offer self-service capabilities for billing and payment functions ease the strains of medical billing for both healthcare staff and patients. Giving patients easy access to billing details and options for online payments allows them to anticipate and understand their medical payment responsibilities, prepare for expenses, and create manageable remittance plans.  

AI and ML in Inpatient Settings

Machine learning (ML) and artificial intelligence (AI) capabilities have the power to increase operational efficiency and worker responsiveness in inpatient care settings. By automating the repetitive tasks hospital employees have historically managed, these technologies free staff members to focus on patient care and other functions that demand human attention. This workforce support is especially timely given the healthcare industry’s staffing shortage.

In addition, AI tools can help administrators anticipate staffing and equipment needs, driving their resource management decisions. AI-driven predictive analytics and real-time insights into new admissions, likely discharges, and patient treatment requirements help administrators prepare for probable scenarios. Thus informed, hospital administrators can anticipate demands and optimize staff scheduling, reduce patient wait time, limit waste, and plan for the appropriate availability of hospital beds, rooms, and supplies.        

financial data science

Identifying Appropriate Courses of Treatment

With the advent of wearable devices and other new data sources, physicians have access to a broader pool of insights to inform their decisions about patient care. These new healthcare data sources can illuminate additional dimensions of patient health, including social determinants, that provide physicians with a more detailed and personalized view of a patient’s overall wellness and health-related vulnerabilities. Before a patient even presents symptoms, practitioners with the tools to mine the breadth of patient data and apply predictive analytics can foresee conditions, take preventative measures, and even provide precision care. Data science technologies can also aid doctors in making more accurate diagnoses, a critical and highly consequential step in patient care. A 2014 report by BMJ Journals stated that diagnostic errors impact an estimated 12 million US adults yearly. The  Society to Improve Diagnosis in Medicine further estimates that misdiagnoses contribute to 40,000 to 80,000 deaths in US hospitals annually. 

Contributing to the increased personalization and accuracy of patient care is the rise of data sharing. Facilitated in part by the implementation of interoperability standards that promote uniform terminology and normalization of healthcare data for industry-wide use, the freer exchange of data within and across organizations enables practitioners to connect the disparate dots of a patient’s healthcare background to develop a more thorough understanding of individual histories and needs. However, data sharing presents healthcare organizations with additional responsibilities and advantages.

financial data science

Avoiding Misinterpretation and Misapplication of Data
The expanding pool of medical data drives dramatic improvements in patient care. Still, the information healthcare data yields can be meaningless at best or harmful in the absence of skilled data cleaning, stewardship, analytics, modeling, and reporting. Here are some examples of data pitfalls that practitioners can avoid given appropriate data science training:

  • Skillful data cleaning and stewardship are essential to avoiding inaccurate analysis. Practitioners must develop the data science knowledge to discern between reliable data sets and data sets that are too fragmented to yield reliable results. It stands to reason that the more healthcare professionals rely on data to inform their decisions, the more critical it is for them to have the skills to recognize whether their data is reliable.
  • To use data accurately and effectively, practitioners must be aware of the limitations of a data set and understand what data they can use to answer a given question. In addition, they must know the data’s source and context of the collection. Such practices help prevent the introduction of sampling bias when training AI algorithms, which can lead to skewed results and misleading information. 
  • An essential lesson in statistical interpretation, understanding the difference between causation and correlation, can safeguard analysts from drawing erroneous conclusions about the relationship between two variables. 
  • After conducting an accurate analysis of a data set, an analyst is still vulnerable to producing misinformation based on the results. For this reason, it is important for employees who generate the reports to receive training in data visualizations and storytelling to avoid the pitfalls of misleading visualizations and misleading statistics.

Sharing Patient Data: Critical Legal and Ethical Considerations

The increasingly smooth flow of data within and across healthcare organizations is without a doubt a mark of progress. However, with this positive trend comes a heightened need for practitioners in the field to navigate the legal and ethical challenges that arise alongside the rise in sharing and usage of patient records.

Healthcare data is proliferating at an explosive rate and is changing hands more readily, thanks partly to the large-scale adoption of Electronic Health Records (EHRs) and Electronic Medical Records (EMRs). In 2019, 89.9% of office-based physicians reported using EHRs or EMRs. This proliferation of healthcare data, the sensitive nature of patient records, the high stakes of medical errors, and the damaging impact of ethical transgressions demands scrupulous attention in the healthcare industry. The legal and ethical considerations associated with the exchange and use of patient data boil down to matters of compliance, privacy, data integrity, and a profound understanding of issues surrounding data science applications.   

Data Governance

I would offer a single phrase to encapsulate the critical mechanisms practitioners should implement as a line of defense against legal and ethical lapses—data governance. Many of the same potential pitfalls of data use in healthcare that Dmitri mentioned above apply to legal and ethical risks and the prospect of faulty analysis. And similar solutions can mitigate threats in all of these areas. Skillful data governance demands a thorough knowledge of current local, state, and federal guidelines governing the sharing and usage of patient data and strict adherence to policies and practices that guarantee compliance with these guidelines. Resources dedicated to data quality and security provide safeguards against breaches, biased analyses, and regulatory violations. By investing in these resources, organizations can help insulate themselves from legal and ethical challenges and ensure sound practices in the following vital areas of data usage:

data analytics courses
  • Data Quality - Garbage in, garbage out is an oft-repeated phrase in data analytics. There is a good reason for that. As a fundamental and immutable rule, which I believe cannot be emphasized enough, the quality and integrity of a data set will impact any output it produces. To avoid analytical errors resulting from inaccurate data and drive better decision-making, practitioners must first follow data governance best practices to ensure that their data is complete, accurate, and properly contextualized.  
  • Data Security - Safeguarding the confidentiality of patient records is among the most critical responsibilities for healthcare providers and is protected in the US under the Healthcare Insurance Portability and Accountability Act of 1996 (HIPAA). In addition to protecting organizations against the legal implications of compliance violations, data governance that includes vigorous data security systems provides fortification against the costly prospect of cyberattacks and other rising forms of medical data breaches.
  • Data Storage and Retention - Germane to data security, healthcare data storage, retention, and destruction are governed by laws and regulations at the federal, state, and local levels. A sound data governance program will include a detailed retention schedule for medical records that covers the full data life cycle from creation to destruction and protocols that comply with the various requirements for applicable jurisdictions.
data analytics courses

In addition to the established practices that all healthcare organizations should follow, I would also emphasize the need for organizations to adapt readily to current issues surrounding healthcare data usage and any changes in guidelines or legislation. Emerging technologies create dynamic environments as guidelines, and ethical standards are developed in response to new concerns. There are a few legal and ethical issues related to healthcare that I believe will remain fluid as they face further scrutiny in the coming years: 

  • Bias introduced into predictive analytics through unrepresentative sampling and the question of accountability: At what point should human judgment override AI? 
  • The ongoing debate surrounding the use of standardized Unique Personal Identifiers for healthcare records.
  • The unintended sharing of healthcare data with third parties.

With great progress come great challenges. Data usage and sharing in healthcare are rapidly evolving, and requirements for safeguarding these processes will develop in tandem. Therefore, I stress the importance of ongoing vigilance as healthcare organizations traverse this uncharted terrain. However, given the remarkable benefits of data science applications and the opportunities they present for future advances in the field, it is clear that these resources are well worth the investment in data governance expertise they require.

Toward a Data-Driven Healthcare Horizon

The incredible advancements that the healthcare industry is making, and will continue to make, through data science technologies offer tremendous hope for the future of patient care. Although the potential challenges, some of which we outlined briefly above, may seem overwhelming, it is encouraging to know that they are surmountable given awareness of broader social implications, thoughtful data governance, and effective data science training across organizations. The technologies propelling healthcare organizations toward new frontiers at accelerating rates, in the hands of a workforce equipped with the skills and knowledge to leverage them responsibly, promise to improve patient outcomes and the experience of all healthcare stakeholders for generations to come. 


Toward a Healthcare Data Revolution


With more data science applications in healthcare around the globe, it is clear the industry's transformation has already begun. However, achieving the data maturity required to leverage these capabilities effectively does not come without significant infrastructure, culture, and education challenges.

We'll explore how to start a data science movement that impacts:

  • Hospital Operations
  • Personalized Medicine
  • Patient Self-Care
  • Public Health

Subscribe to our newsletter

Data Society provides customized, industry-tailored data science training solutions—partnering with organizations to educate, equip, and empower their workforce with the skills to achieve their goals and expand their impact.

cross linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram