We have not yet reached the full potential of data science in healthcare. Hospitals, clinics, and private practices around the world are projected to produce 2,314 exabytes of data in 2020. Forrester estimates that up to 73% of any organization’s data are wasted because they aren’t analyzed.
Solutions for current data science in healthcare problems go beyond administrative efficiency desired by industry leaders. Improved data practices assist healthcare providers with insurance, legal compliance, and caseload issues. Most importantly, data science in healthcare can improve lives for patients around the world.
Dr. Atul Butte of the University of California, San Francisco put it plainly when he said:
“Hiding within those mounds of data is knowledge that could change the life of a patient, or change the world.”
This quote is a good mantra for anyone hoping to improve data science in healthcare. Understanding data science’s current limitations and almost limitless potential can help you reshape this crucial industry.
Key Problems for Data Science in Healthcare
The sheer volume of healthcare data flooding into hospital systems each day is overwhelming. Experts on data science in healthcare are currently addressing data storage and analysis issues to avoid cascading failures. Data management is not the only opportunity for software engineers and data scientists to help the healthcare industry.
The Brookings Institution published a report highlighting the industry’s data analytics obstacles. This report concluded that the biggest obstacles are:
- Sensitivity and customization of patient data and preferences
- Incompatible data conventions
- Traditional healthcare practices that aren’t aligned with data science
- Differing motivations for stakeholders in the healthcare industry
These obstacles are not insurmountable given data science’s applications in other fields. The report notes that a good starting point is the harmonization of Electrical Medical Records (EMR) systems. As a software engineer, you can use these obstacles as reminders of how your work can improve the future of healthcare.
Ensuring Privacy with Data Science in Healthcare
An important frontier for healthcare data science innovation is the balance between accessibility and privacy. Siloed or raw data doesn’t help nurses and doctors treat patients in time-sensitive situations. Data science can open this data while respecting legal and cultural restrictions on data availability.
Data science in healthcare must comply with local, state, and national laws guiding medical data. The most prominent of these laws is the Health Insurance Portability and Accessibility Act of 1996 (HIPAA). A 2003 rule applied to HIPAA made the following change summarized by the U.S. Department of Health and Human Services:
“A major goal of the Privacy Rule is to assure that individuals’ health information is properly protected while allowing the flow of health information needed to provide and promote high-quality health care and to protect the public's health and well being.”
Consumers are also concerned about how data science solutions will impact privacy. PricewaterhouseCoopers found the following data when surveying Americans about medical information:
- 71% of respondents placed privacy over access when considering medical tests.
- 68% of respondents are concerned about privacy when using app-based health tools.
- 66% of respondents are concerned generally about medical data privacy.
Policy changes and marketing to ease these concerns are only possible once the technology is ready for the market. Consider these data science tools as you ponder how to disrupt the healthcare technology field.
Anonymized Health Data
Artificial intelligence and machine learning provide avenues for turning individual records into anonymous data. An EMR system with anonymizing features prevents easy access to data in case of hacks. Anonymization is also invaluable to medical researchers who need real-world information to inform new treatments and equipment.
Drs. Khaled El Emam, Sam Rodgers, and Bradley Malin studied the likelihood of reidentification with anonymized records. Their study of Canadian birth records from 2005 to 2011 found reidentification probabilities ranged from 0.014% to 1%. This range spanned from a single data point to four data points related to the mother and child.
Data science in healthcare need not reach zero probability to reach its goals. El Emam, Rodgers, and Malin concluded:
“Anonymisation methods cannot ensure that the risk of re-identification is zero, but this is not the threshold that is expected by privacy laws and regulations in any jurisdiction. Strong precedents exist for choosing suitable probability thresholds for anonymising data. There is a need for anonymisation standards that can provide operational guidance to data custodians and promote consistency in the applications of anonymisation.”
Tools for Healthcare Data Privacy
Open data is an important tool for medical researchers and healthcare providers. Without realistic patient data, it is difficult to develop real-world solutions. Merging freely available data with anonymization concepts leads to innovations like Synthea.
The Standard Health Record Collaborative created Synthea to synthesize public health data into digital patient avatars. Synthea allows users to input their own data to create patient profiles. This AI-enabled tool allows researchers to interact with virtual patients without concerns about data privacy.
The future of data science in healthcare will not be a free-for-all for patient records. Public policy changes might address technological developments but are unlikely to change basic privacy principles.
Healthcare data science tools need to protect patient privacy as it spreads across multiple devices. Tools like Avalanche Mobility Center provide high levels of security to all devices within healthcare systems. These tools are also designed to protect against power surges, equipment failures, and other issues.
Analytics and Data Science in Healthcare
Producing raw data on patient vitals, medication impacts, and other care benchmarks are not enough to improve healthcare. Healthcare data science needs to follow with analytical methods and tools that extract value from medical records.
Deloitte conducted a roundtable of healthcare leaders to discuss the future of analytics. This discussion highlighted the following opportunities for data science in healthcare:
- Creating care scorecards for each patient based on inputted data
- Producing automated diagnostics that allow provider focus on patients
- Evaluating costs for patients and providers to balance needs with resources
- Implementing digital “nudges” to patients to follow physician recommendations
Dan Housman — the chief technology officer of ConvergeHEALTH — drilled down to the core reason for healthcare data science innovations:
“We’re looking for the right intervention at the right time – the biggest trend we’re seeing is the patient turning into a consumer and driver of their own health care. It’s creating interesting pressure on the system to drive their personalized analytics.”
Deloitte surveyed major healthcare providers about operational areas in need of innovation. Software engineers and data scientists can develop tools to meet the following provider needs:
- Population health (56% planned investments within one year of the survey)
- Clinical management (54%)
- Financial management (32%)
- Enterprise performance (30%)
- Research (30%)
Point-of-Care Applications for Data Science in Healthcare
Data science in healthcare needs to work on the ground level as well as the board room. Doctors, nurses, and other personnel can eliminate guesswork from their diagnoses with well-designed tools. Software engineers like you create practical uses for data science in healthcare used in clinical environments every day.
PinnacleHealth System’s development of a new data infrastructure provides a good example for point-of-care applications. The Pennsylvania-based provider created its Closed Loop Awareness System to predict patient outcomes based on documented behaviors. Steven Sepp from PinnacleHealth noted the following steps to create a successful system:
- Focus development on the provider’s areas in need of improvement;
- Work through data sets to find possible variations in system outcomes;
- Standardize the system to reduce the noise that obscures patient success
Financial pressures, cost-minded patients, and care expectations have driven advancement for data science in healthcare. Data experts and software engineers have produced innovations for the full cycle of care including:
- IQuity’s pilot program for early detection of multiple sclerosis
- Stanford ML’s neural network design for separating arrhythmia from standard heartbeats
- SeamlessMD’s post-operative program with patient self-evaluations
Data-driven solutions can be integrated into imaging systems, mobile devices, and diagnostic tools. A routine doctor’s appointment can be infused with data science from a check-in tablet for symptoms to a smart stethoscope. Data science in healthcare goes from isolated data points to constant data gathering with point-of-care technology.
Making Data Science in Healthcare Work for Users
Data science in healthcare can place power over health and well-being into the hands of individuals. A doctor prescribes medications, recommends fitness, and provides tips on living healthier lives. Data-driven technologies can help individuals follow physician mandates while moving toward healthier lives.
A major growth area in the past 10 years has been wearable devices. Devices like the Fitbit and the Apple Watch track fitness levels using wearer locations. We can look at the trend line in global wearable device usage to see potential avenues for healthcare technology:
- 325 million wearable devices in 2016
- 593 million wearable devices in 2018
- 835 million wearable devices in 2020
- 1.1 billion wearable devices in 2022
Substantial growth in device ownership creates openings for innovations in portable health tools. Data science experts can develop device features that provide ongoing analyses and recommendations. There are always improvements to be made in user interfaces and integration with other programs.
Mobile Phones and Apps
The iPhone, Samsung Galaxy, or Google Pixel in your pocket can stand in for a wearable device. Cracking the code on a mobile health app also opens a global market to innovators. Statista estimates that there will be 7.3 billion mobile phone users worldwide by 2023.
Widespread mobile phone ownership means that these devices will be essential to public health campaigns going forward. We can use the COVID-19 pandemic as an example of how mobile apps are promoting data science in healthcare. Developers and engineers have created the following apps used by public health officials:
- Apopka’s SafePass app to screen employees and customers during reopening
- Chicago’s Chi COVID Coach app for daily check-ins and localized health resources
- Harris County’s app for self-reporting symptoms and epidemiologist consultations
Mobile phone apps can also provide health resources to users in better times. Telehealth apps save trips to doctors’ offices and reduce costs. Future apps may match users to the right specialists and adjust diets based on health inputs.
Learning to Improve Data Science in Healthcare
Healthcare providers are trained to care for others, not develop data science solutions to their daily challenges. The healthcare industry needs help from software companies, tech firms, and individual innovators. You can step into the fray and help healthcare heroes save lives.
Your undergraduate work and early professional experiences are great initial steps toward data science work. A graduate degree in computer science sends you on the fast track to data science in healthcare careers. Baylor University’s Online Masters in Computer Science accelerates this progress further with its Software Engineering track.
Preparing for Healthcare Data Science Careers at Baylor University
Baylor University designed its Online MS in Computer Science to produce graduates who can solve tomorrow’s problems. The Software Engineering track assists your future work in healthcare data science with courses on:
- Software Verification and Validation
- Distributed Systems Development
- Advanced Object-Oriented Development
Rigorous courses from world-class faculty members bring these concepts to life. Program graduates are ready to design, maintain, and improve systems of all sizes.
Completing this innovative computer science degree boosts your career prospects thanks to Baylor University’s reputation. U.S. News & World Report placed the school high in rankings like:
- No. 42 in Most Innovative Schools
- No. 48 in Best Colleges for Veterans
- No. 79 in National Universities
Baylor University has taught future scientific leaders throughout its history. Gordon Teal helped produce the first silicon transistor used in modern computers after graduating in 1927. Allene Rosalinde Jeanes went from a 1928 graduate to the inventor of food additive Xantham gum.
You can learn how to join these innovators in scientific history by contacting an enrollment advisor today.