Emerging tools are harnessing the power of big data to improve healthcare
By Ketan Paranjape and Paul Lambert
In the United States, elements of the healthcare system are continuing to push forward with the notions of value-based healthcare, focus on the patient, and improved access to healthcare for all. Nevertheless, there remains ample room for improvement.
For example, the United States has a long way to go to reduce its healthcare costs to parity with those of other developed nations.1 And despite the utility of diagnostics for clinical decisionmaking, clinical labs continue to be challenged to prove their value within the healthcare system and to gain recognition as resources that can significantly improve clinical, operational, and financial outcomes.
All this is about to change. The convergence of information technology with the abundance of data collected in every corner of the healthcare system on every individual patient presents an opportunity to shed new light on where clinical and operational improvements can and should be made. Such insights can guide providers toward developing new strategies at every level, from the physician-patient interaction on up.
Such a potential to transform healthcare finds precedents in the ways that ‘big data’ has transformed the consumer experience in a variety of fields. For examples, one need only consider how Amazon has personalized shopping, or how Uber has reimagined personal transportation. Such learnings are now also ready to be applied to healthcare, enabling leaps toward improved outcomes.
This article reviews how new and emerging digital tools are empowering health networks to tap into the wealth of diverse data available in the lab, to collaborate more efficiently within the system and beyond, and to fine-tune processes in order to achieve significant operational and financial efficiencies. New tools that provide clinical decision support and help networks optimize the use of operational and clinical data are moving the day-to-day practice of precision medicine closer to reality and placing the goals of greater quality, cost efficiency, and access within reach.
The Opportunity
In healthcare systems nationwide, data collection is a round-the-clock activity. Healthcare professionals have become accustomed to using electronic medical records (EMRs), laboratory information systems (LISs), picture archiving and communication systems (PACS), and a wide variety of databases within and specific to such independent settings as the clinical laboratory, point-of-care environments, and genetic sequencing facilities. The amount of data generated by such systems—as well as their diversity of sources, formats, and quality issues—can be overwhelming. And many such databases are created, maintained, and updated in different areas of the healthcare system, involving diverse stakeholders.
Efforts to aggregate and create value from data originating from varied sources have so far not been very successful. The underlying reasons for not translating such rich data resources into useful information are many, including the lack of enabling digital tools, concerns about patient privacy and consent to allow use of their data to address broader institutional objectives, regulatory considerations, and simply the human and financial resources required to implement solutions. The ‘silos’ culture and tendency of departments to operate independently, setting their own rules and guidelines, as well as the diverse configurations of EMRs from various vendors, are further reasons for the lack of coordination and data sharing.
Traditional and narrowly defined roles—such as the notion that physicians order tests and labs merely report results—can limit the degree to which labs are motivated to leverage their databases, identify trends, and glean insights that can guide physicians in mapping out patient care strategies. As in any area of innovation, building the confidence and trust of the end-user—in the case of clinical labs, typically the ordering physician—is critical.
In this context, the evolving opportunity for digital diagnostics can be defined as using digital technology to enhance the value of diagnostic data by presenting it in a more comprehensive manner—for instance, presenting a more holistic or longitudinal picture of the patient, or offering both a macro and micro view of lab operations to better understand workflow and resource utilization—and augmenting this rich dataset with tools that make it easy and efficient to use the data.
Figure 1. Using a tablet interface to access the Viewics diagnostics solution from Roche Diagnostics, Indianapolis.
Key drivers and enablers of digital diagnostics include technological advances such as Cloud computing, which reduces the burden on an institution’s infrastructure, and new business models such as software as a service (SaaS), which make access to cost-effective and scalable tools a reality. Instrument-agnostic software accessible via multiple devices—including mobile devices—further enhances the value of digital diagnostics by promoting use and accelerating the growth of the knowledge base (Figure 1).
Powering Clinical Decision Support
Digital diagnostics can significantly influence clinical decisionmaking by providing a range of tools for the aggregation, presentation, and interpretation of patient data, including tapping into deidentified patient pools or curated tumor mutation datasets, matching patients to clinical trials, and facilitating multidisciplinary collaboration. When patient data from multiple modalities—such as clinical chemistry, biomarkers, genomics, and imaging—are aggregated and presented in context, the physician is provided with an information-rich overview of the patient (Figure 2).
Figure 2. A screenshot from the Navify tumor board solution by Roche Diagnostics, Indianapolis. Having an at-a-glance view of patient data and images from different diagnostic modalities facilitates multidisciplinary collaboration in clinical decisionmaking, as in the case of tumor boards. Some software tools can also pull information from published studies, clinical guidelines, and clinical trial registries.
And this is only the beginning. Add to these elements immediate access to information about targeted therapies, current clinical guidelines, and the clinical significance of genetic variants, and the physician will have a powerful tool at hand to evaluate treatment options. Learnings from pooled, deidentified patient data can also be used to guide clinical decisionmaking. One such example is a clinical genomics database linking deidentified EMRs from 30,000 patients at more than 200 US cancer centers with anonymized results from genomic tests.2
Digital diagnostics systems can also help physicians streamline the process of finding clinical trials for which their patients may be eligible. The systems can automatically search clinical trial registries, providing the computational power to match patients’ profiles—including their genomic data—so that an initial assessment of eligibility can be made. Making it easier for physicians to match their patients to clinical trials not only benefits patients with access, it can also facilitate patient recruitment, often a rate-limiting step for clinical trial sponsors. Such connections are particularly beneficial in oncology, where precision medicine strategies are being used more widely than in other medical specialties.
Decision support tools can also improve the efficiency and quality of multidisciplinary collaboration in the treatment of complex diseases such as cancer. Gathering the diversity of information about a patient, often from different databases within the health system, is time-consuming and prone to such problems as overlooked or duplicate information or out-of-date data.
A recent study showed that digital tools can improve tumor board preparation by generating substantial time savings and higher user satisfaction. Based on observations over 41 sessions, preparation time was reduced by 53% for oncologists, 11% for radiologists, 11% for pathologists, and 7% for surgeons. Such improved efficiency translates directly into more time spent on evaluating options and making decisions to benefit the patient. Study participants also reported statistically significant increases in their ratings regarding ease of use and satisfaction.3 Digital tools to support multidisciplinary collaboration also hold great potential in areas such as sepsis in acute care, or diabetes in chronic disease management.
Driving Systemwide Performance from the Lab
Clinical labs have long used analytics to improve operations within the lab—for example, optimizing workflows and improving turnaround times. Digital diagnostics takes such efforts one step further.
Clinical labs are rich with data that can guide fact-based decisions to improve operational, financial, and clinical performance throughout their institutions. Using digital tools, lab and hospital management can access such data to optimize the lab infrastructure and its processes, reducing waste and improving operational and financial performance. Managers can use lab data to establish benchmarks and monitor performance over time within an individual lab, across labs within the health system, and even outside the lab.
Within the lab, digital diagnostics systems can connect preanalytic, analytic, and postanalytic modules to achieve the dual goals of streamlining sample flow from ordering to archiving, and monitoring how the lab is performing in real time based on key metrics. A good example of the use of lab data in direct process improvement is autoverification, which uses preset criteria to determine whether a result can be released without human review, improving workflow and reducing turnaround time and errors. Another example is the use of moving averages as a quality control tool to alert the lab about any issues that arise between scheduled quality control checks, and to prompt timely corrective action.
Real-time monitoring of such key metrics as test volume, turnaround times, pending tests, and quality control represents an important resource for driving process improvement. The utility of real-time monitoring data for problem-solving can be enhanced by a combination of dashboards to visualize trends, reports to provide snapshots (eg, for verification of compliance), or alerts to trigger prompt action, as well as ready access to such updates within and outside the lab via various devices. One example of how such analytics can help improve lab operations is the ability to anticipate instrument and reagent needs based not only on volume but also on the timing or seasonality of specific tests, enabling the lab to optimize capacity as well as to fine-tune other needs, such as sample transportation, so that all of a lab’s operations are working to maximize efficiency.
Clinical labs are also beginning to recognize that they control a wealth of data that can be used for analytics with an impact beyond the laboratory. For example, labs can use digital tools to elucidate individual physician test-ordering practices, identify disparities, and look for ways to standardize testing for specific international classification of diseases (ICD) codes to help optimize the use of testing across a network. Such analytics can help healthcare networks to track whether physicians are complying with the appropriate frequency of a patient’s HbA1c testing, for instance, or to identify discrepancies in the ordering of cardiac biomarker tests for patients who present to the emergency department with chest pain.
Such analytics can help to streamline lab processes and the use of lab resources, and can also facilitate the sharing of patient experiences and insights among physicians in order to identify best practices in patient management. Likewise, lab management can use such analytics to identify and address trends in under- or overutilization of testing, contributing further to improved patient care and financial outcomes (Figure 3).
Figure 3. Digital tools for managing lab and healthcare system performance can access data on frequency of testing, organized by physician and by assay, to highlight which physicians are mostlikely to be outside the norm as defined by guidelines or best practices, and where education and adjustment of utilization are needed.
Last, but certainly not least, labs can use digital tools to monitor and analyze individual patient test data to help physicians identify opportunities for patient education. Flagging favorable or unfavorable trends in a patient’s lipid levels or blood glucose status, for instance, can help to encourage healthy behavior.
Toward the Future
With their power to provide depth and breadth of information at every level in the healthcare system, starting with the patient, digital diagnostics represent an enabling tool for precision medicine and its increasing focus on fine-tuning patient care to the molecular and genomic level. Advances in computing, machine learning, and artificial intelligence will continue to drive and accelerate digital diagnostics. Support will also come from the ongoing cultural shift away from ‘silos’ and toward collaboration and the development of solutions that are agnostic to specific instruments or software platforms.
The speed at which such new tools are being developed will undoubtedly demand a fresh approach on the regulatory front. FDA’s digital health software precertification pilot program, a part of its digital health innovation action plan, is an important step toward collaborating with industry to create a framework and process for precertification, and ultimately to streamline the path to market for digital health innovations.4
Just as important is the human factor—gaining the trust of healthcare stakeholders, especially physicians, and training them on the new tools. Today’s lab consultancies will need to add digital diagnostics to their competencies so that they can guide clinical labs toward implementing the new tools, and thereby strengthen the role of clinical laboratories as key resources within the health system.
Ketan Paranjape is vice president for diagnostics information solutions, and Paul Lambert is vice president for commercial IT operations, at Roche Diagnostics, Indianapolis. For further information, contact CLP chief editor Steve Halasey via [email protected].
References
- Papanicolas I, Woskie LR, Jha AK. Healthcare spending in the United States and other high-income countries. JAMA. 2018;319(10):1024–1039; doi: 10.1001/jama.2018.1150.
- Agarwala V, Khozin S, Singal G, et al. Real-world evidence in support of precision medicine: clinico-genomic cancer data as a case study. Health Aff. (Millwood). 2018;37(5):765–772; doi: 10.1377/hlthaff.2017.1579.
- Krupinski EA, Comas M, Gallego LG. A new software platform to improve multidisciplinary tumor board workflows and user satisfaction: A pilot study. J Pathol Inform. 2018;9:26; doi:10.4103/jpi.jpi_16_18.
- Digital Health Software Precertification (Pre-Cert) Program [online]. Silver Spring, Md: FDA, 2018. Available from: www.fda.gov/medicaldevices/digitalhealth/digitalhealthprecertprogram/default.htm. Accessed December 17, 2018.
