Chapter 7: Introduction to Section II: Integration of and Bringing Big Data to Practical Usability

This Section II on use cases in technology brings you to the operability aspects of data, focusing on user (i.e., patient) interaction with interfaces that bring their data into clinical usability. Data sharing and the digital applications on smart phones and wearable technology have the potential for supporting actual care needs. But, while all of us use apps in one form or another, how much do we know about who is on the receiving end and what they are doing with our data? And inversely, how good are we at inputting data and judging our health and quality of life?

The BigMedilytics (BML) project focused on bringing technology into use. Section II offers a series of four approaches, with divergent points of focus and models. Each chapter brings its own insights and lessons.

In a herculean effort and with access to circa four million integral health records over a 5-year period, Navarro-Cerdán and coworkers (Chapter 8: Effects of comorbidities (chronic illness) on hospitalization and mortality risks) developed models for 19 chronic illnesses and how care decisions and all-cause 5-year mortality could be assessed in cohorts and potentially for a single patient. An example might be a middle-aged male with COPD gold 3, poor cardiac function, and now peripheral vascular disease with acute ischemia in the toes of a foot: Would hospital admission, peripheral bypass surgery, or limited amputation support him and his residual, small ‘action radius’, as well as his 5-year life expectancy? How would you approach such a discussion with a patient? You will find the tools they describe interesting and relevant in terms of user interfaces.

Kidney failure and dialysis have a major impact on quality of life, the ability to travel, and life expectancy. Receiving a donor kidney offers a new lease on life. You would expect that the recipient would be totally compliant with everything they could do to support and maintain the donor kidney. However, this does not seem to be true, with poor therapy compliance being a relevant factor in transplantation failure. Duettmann and colleagues (Chapter 9: eHealth and telemedicine for risk prediction and monitoring in kidney transplantation recipients) took a telemedicine and eHealth approach to empowering the patient and reducing failure rates by improving recipient support (and compliance) by self-registration monitoring, leading to input in risk prediction models that could be used as decision support.

The authors, working in Germany, describe the difficulties of data sharing in a federal state and in setting up and using a system in a multidisciplinary setting. They developed a robust dashboard that could then be integrated into the Electronic Medical Record (EMR). The mobile phone app created interesting logistics and responsibility queries: response time, accuracy, access, and the ability for ready and early decision-making.

Since non-adherence to the use of the app could be seen as a risk in and of itself, the study spent time searching for and overcoming resistance to app use and interface issues. You will find a focused narrative offering a framework suitable for many organ-specific approaches.

In this chapter, Grossi (Chapter 10: Remote monitoring to improve gestational diabetes care) describes the user case for eHealth in the form of a mobile phone-based interactive real-time app for use by pregnant women who have developed diabetes during their pregnancy. Gestational Diabetes Mellitus (GDM) has an extensive range of complications, both in the immediate future and in the more extended future, as well as for the unborn child. Careful management and tight control of blood sugar levels and ready supportive access to healthcare professionals support good outcomes and were implemented using a self-reporting system.

Grossi offers extensive insights into the developmental aspects of the app and into the choices made to gain valid information. Key needs for success are described, as are the data offered to show the reader that the system works.

In this study, a prognostic model was also developed and implemented. The study was able to show a reduction in the need to visit the outpatient department, reduce overhead costs, and improve blood glucose level stability.

A great example of a user case is where data management, different cohorts, and data sharing using self-reporting can show benefit.

Chronic Obstructive Pulmonary Disease (COPD) is an international chronic epidemic, reaching all layers of society and not only causing a major disruption in the quality of life, but also creating a serious load on healthcare. Particularly in countries with larger rural areas, the burden of this disease is increased by the large investments needed to reach and be seen by a healthcare professional.

In this chapter, Pickering (Chapter 11: Monitoring wellness in chronic obstructive pulmonary disease using the myCOPD app) approaches a new aspect of the use of self-reporting and big data: the ability to use subjective measures. Different from reporting objective numbers, reporting perception introduces new complexities. Where in Chapter 9, Duettmann and colleagues want to use telemedicine and eHealth to mitigate the decline in self-care, in this chapter aspects such as external influence, the individual’s – and potentially changing – self-assessment of wellness, even as a function of seasons (i.e., it is fall and chilly, my COPD should be/will give me more trouble) is taken into account.

The chapter describes and analyzes two substudies to offer the reader insights into the clinically relevant and useful model of bringing data to clinicians and its potential to alter an illness’s natural course and thus improve quality of life.

Privacy regulations, of which GDPR is only one, have been developed to protect individual. One can imagine that having access to medical data could allow the industry to tune, revise, or innovate more accurately. Being able to gain such insights is, however, complex, as (corporate) interests may not be parallel with the (individual) data holder.

In this chapter, Spini and coworkers (Chapter 12: Privacy-preserving techniques for analysis of medical data: secure multi-party computation), working from TNO (the Dutch Organization for applied scientific research, and thus as a neutral external party) describe a project involving a University Hospital and a large insurance company. They used the real-life scenario of heart failure patients, a chronic illness potentially requiring both admission and technological intensive care. One can imagine that an insurance company well versed in an epidemiologic approach to reimbursement strategies would be more than happy to have detailed insights into such a cohort. A dataset of Achmea and Erasmus MC, once intersected and combined, could, for example, be used to train a prediction model that would identify high-impact lifestyle factors for heart failure and thus, in turn, recognize high-risk heart failure patients.

The authors offer insight into their modeling, a number of algorithmic approaches, extensive references, and explain at some length how such a solution can be reached. For example, a third party can hold the database(s) and perform the calculations – such that the party supplying the data is not actually giving it to the interested other party.

Where privacy regulations have been seen as major hindrances to big data analytics, Spini and coworkers offer a safe and robust strategy to resolve this in a pragmatic fashion.

Section II describes a range of situations in which (big) data, self-reported or userdriven data generation, can be used to improve, strengthen, and intensify healthcare. While three of the chapters use a telemedicine/eHealth app-based approach, the concepts are generalizable. The different foci, from monitoring and prediction to early intervention to quality-of-life support, potentially offer other interested parties handholds in further development within their niche. Each chapter has its own learnings and is self-supporting.

Healthcare professionals, managers, the industry, and primarily the individual (patient) will recognize them as only partially tapped resources and methodologies described in Section II. It should stimulate and challenge all the stakeholders to continue and intensify their efforts to bring these technologies into practical, safe use.