What does big data have to do with hematology?

More than one thinks. Hematology is moving into a digital age where data is a goldmine of innovations and insights. While moving into a digital age is exciting and comes with some fantastic opportunities, there are some real challenges. Today, healthcare organizations are challenged with actualizing tangible outcomes of IT investments, organizational adoption and change management, and securing patient data. These organizations also rely on their educated and experienced staff to be the supercomputers of consistent clinical decisions. But sometimes people do not make the best supercomputers. We have limitations, especially when overworked or rushed and we find ourselves frustrated.

This is why laboratories and healthcare organizations need to advance and shape healthcare with technology. The opportunities in hematology are limitless. The world is moving fast into this digital age with real-time CBC monitoring, predictive disease alerts, and true data integration. However, is healthcare ready to take on these challenges? Absolutely.

How are healthcare organizations advancing healthcare?

Hematology data is one of the richest information sources in the laboratory for clinical decision-making. It can tell a deep patient history with dozens of parameters and even help predict the likelihood of disease states. The next step of true advancement and innovation is a deep integration of laboratory data.

Consider iron deficiency anemia’s (IDA) predictive value. Today, IDA may occur when there are low amounts of iron in the body to make hemoglobin. With low amounts of iron, fewer, smaller RBCs are created. The typical CBC will indicate low amounts of hemoglobin (Hgb), Hematocrit (Hct), and Red Blood Cells (RBCs). The Mean Corpuscular Volume (MCV) and Mean Corpuscular Hemoglobin (MCH) are present in the normal range but often drift lower due to the smaller RBCs (microcytic). Most laboratories have rules or algorithms that pull in these five parameters of the CBC. However, true clinical decision power is increased when it includes other clinical chemistry parameters like ferritin, total iron-binding capacity (Transferrin), and transferrin saturation (TSAT). Combining these clinical parameters provides a sharper focus on the interpretation and improves clinical efficacy.

Another example is acute Leukemia. White blood cell (WBC), platelet count, and hemoglobin may be used to identify abnormal samples that would require visual investigation of the cells. Advanced laboratories would have algorithms based on agreed-upon thresholds that would trigger manual differential to a digital imaging system.

This digital imaging system uses complex artificial intelligence (AI) to count, identify, and pre-classify cells, reducing ambiguity and time for clinical decisions. What can maximize these innovations is fine-tuning these algorithms. These parameters can be maximized for full clinical impact using machine learning and AI.

So, how can this be accomplished?

Refinement of these algorithms needs lots of data, not just text data or excel spreadsheets. It needs a systematic, quality-centered process to clean, standardize, and organize this data, ensuring robustness. However, this can be difficult. Standards for managing and reporting quality data for health research simply do not exist. Also, this type of research does not fare well with shortcuts. These shortcuts may ultimately lead to compromises that lead to poor patient care. The center of this research always needs to be patient-centric. Also, the skills required for this type of research are often beyond the laboratory’s skill set.

Data science is a growing field where data scientists are trained to build databases and use standard frameworks and a programing language called “R.” Many healthcare systems are putting together teams of data scientists and healthcare professionals to approach this challenge. Once the team is together, a clinical challenge is identified, the data is organized in the proper format, and the algorithm creation process can start.

New big data research indicates that stress can show up in the CBC, particularly in WBCs. Nevertheless, WBC alone is not an accurate indicator, one needs to also consider other factors. Cholesterol levels, interleukin-6 (IL-6), and C-Reactive Protein (CRP) all factor into stress sensitivity. When other layers like sleep, BMI, working hours, mobility, gender, and zip code, are incorporated, it can paint an even better picture. Stress.

Having this algorithm in alignment with annual health checkups could translate to disease prevention and promoting mental health.

What is the ultimate dream?

The ultimate dream is to pull data beyond the laboratory, like connecting pharmacy data, radiology, anatomic pathology, even personal online purchase history. We are starting down this path with healthcare system standardization of hospital information systems (HIS) and breaking down the barriers between interoperable databases.

Advancing healthcare is not easy because all this data comes in complex and unique formats. Cracking this code can tip the scales with more accurate clinical decision-making instead of relying on arbitrary interpretation.

So what hematology innovation is trending?

Daily wearables are now common and some people are even wearing two. These wearables currently can monitor heart rate, temperature, glucose, and sleep, but some are now advancing to monitor RBCs and WBCs levels. What does that even mean to an average person? Does everyone understand what high or low RBC or WBC levels mean?

Over time, it will be a standard, just like understanding heart rate. People today impacted by anemia can take advantage of this innovation. Monitoring RBCs does not stop there. Understanding how real-time RBC fluctuations can indicate menstrual cycle timing, early indications of cancer, the health of a pregnancy, or even uncover something as dire as bleeding in the digestive tract. Weight loss surgery that sometimes causes iron deficiency is a real problem, and real-time monitoring helps maintain a healthy diet. One atypical use is to monitor RBCs when on vacation in exotic areas. Parasites found in some parts of the world cause blood loss and having a wearable may prevent a dream vacation from turning into a nightmare.

Monitoring WBCs has its advantages, too. In general, one can trend an uptick in WBCs that may represent the start of a cold or flu. Immune-compromised people can monitor their WBC levels and take preventive action before seeing an enormous spike or dip in their WBC levels. Cancers like leukemia have abnormally high or low indications of WBCs, too. Temporarily, these wearables may be used for pre-surgery to evaluate readiness. They are even pulling this data in real-time to a physician’s tablet.

But will this technology flood patients to doctors and hospitals? Managing large amounts of data, especially data transparent to the end-user, will need a solution. Perhaps one consideration is for call centers to be established to triage escalations, thus preventing increased admissions. Another perspective is this leads people to live healthier, catching problems sooner, stopping them before they are critical, thus preventing increases in hospital admission.

The technology is still new and in development. Universities and private companies worldwide are racing to build this technology since its application can drastically improve the quality of life for so many.

It is not just about the hardware that we can wear on our fingers, wrists, or even ears. It is also about the data collected. Both laboratory-generated data and real-time data collected from wearables help the patient. One advantage wearables already have is that it’s an already adopted technology. It’s more common to see someone with a smart wearable than a traditional watch. These devices to truly make an impact just need some new features and maybe it’s a full CBC on your wrist.

Big data is the future of hematology

Ready or not, hematology is moving into this digital age. This article addresses many unique and exciting future concepts, but they all come with real challenges. Early adopters need to be fearless and take a big risk. They will do the heavy lifting, and the impact is limitless. They will also need to avoid pitfalls by collaborating with teams of data scientists and healthcare professionals. They will need to continue approaching problems systematically and carefully, creating new clinical decision rules and technology that provide consistent, robust care while reducing the burden on our healthcare professionals. Advancing and shaping the future of healthcare with data is still in its infancy. It is an exciting time to be innovating and learning. 

Daniel Johnson, MBA, is the Integrated Solutions Group Marketing Manager for Sysmex America.