Clinical Epidemiology is the basic science that underlies the development of tools for clinical diagnoses, valid prognoses, and effective interventions to prevent, treat and manage illness. These activities are key to improving population health. Within the Center for Clinical Epidemiology and Biostatistics, we tackle many questions in a unique way through collaborative thinking and team science. For instance, we generate answers to questions about big health concerns that confront identifiable subpopulations by classifying those people precisely—integrating information about their genetics and biology, their environment, and how they live day to day—then leveraging all these types of high-dimensional data to promote scientific discovery. Because we bring the benefits of “deep phenotyping” into our research, the work we engage in can fundamentally influence people’s lives.
The plan for the Chronic Renal Insufficiency Cohort (CRIC) Study 2018 is a case in point about the value of this collaborative team science approach. CRIC has been called the “Framingham Study of kidney disease”; to investigate the “hidden” epidemic of chronic kidney disease (CKD), this multi-decade study, which I chair, has enrolled 5500 people to date. In its next stage, it will expand recruitment among underrepresented populations bringing into the study 500 American Indians and an additional 125 participants of Hispanic ethnicity. We will gather dense multidimensional data in home and community settings from 3000+ people with CKD—expanding our use of the tools of biomedical informatics and statistical analysis of high dimensional data to optimize scientific discover from those data. One in seven Americans is burdened by CKD, which promotes many types of severe morbidity. Right now we have insufficient information about the disease’s causes and consequences. Yet with biostatisticians, epidemiologists and informaticians working together, we will move toward a deeper understanding that can reduce suffering while using fewer resources. We will work to identify high-risk phenotypes, determine which existing interventions are likely to work best for whom, and design novel interventions where they are needed.
CRIC 2018 will center on two lines of investigation. For one, we will build on something we know from prior observation: People with CKD have an extraordinarily high risk of cardiovascular illness. As study participants go about their everyday activities, we will use various mobile-health devices to measure their physical activity and physiological parameters, and to generate EKG data. Some of these measures yield complex information even taken alone—the data on location and body position are a great example. But we will take aim at a further challenge. Currently we see a high rate of illness events—often devastating ones—among people with CKD. Informatics expertise will help us to distinguish the most relevant aspects of each type of data and combine them into a multidimensional matrix. We can then identify subgroups of people who are edging toward such a health “cliff.” We can set the stage to implement and test interventions to help keep those people on safe ground.
We start with sound epidemiological design: Who wears these monitors, for how long? How do we make sure to represent all the sub-communities of CKD—by age, sex, race; by co-morbidities? Epidemiologists, in conjunction with biostatisticians, have planned this up front. Utilizing modern-day tools of biostatistics, we will understand what we can and cannot infer from these large arrays of data.
CRIC 2018’s second line of inquiry will detect progressive kidney dysfunction. In their homes, participants will repeatedly obtain finger-stick blood samples to measure creatinine (an elevated level signals impaired kidney function) and will collect urine to measure protein levels. We’ll look at both in conjunction with what else is happening: Did that person run a race, or change medications? Our hypothesis: Declines in kidney function that we haven’t observed before—short episodes that may result from, say, a drug—injure the kidney. We think that in CKD, kidney function declines over time not just because that is its inevitable course, but because of the cumulative effect of these incidents.
What can this mean for that one in seven among us? If we can pick out the people at highest cardiovascular risk, we can get them into preventive therapy. When we find people whose kidney dysfunction is continuing to progress, we can learn what causes this, and try to intervene—and we can reduce the harm when episodes of sudden decline do occur.
CRIC is about to enter its fourth phase, years 18 to 23 of continuous funding from the National Institute of Diabetes and Digestive and Kidney Diseases—a duration rarely seen in the NIH. The study has contributed more than 150 papers to the medical literature, with a like number underway. Active studies in Germany, Japan, China and India, among other countries, model themselves on CRIC.
The CRIC Study was built on the foundation of sound principles of epidemiology and biostatistics. Over the past 20 years, CRIC, along with the broader scientific community, has progressively incorporated the Informatics tools of machine and deep learning. To meet the challenges ahead during this next phase of the CRIC Study, our Penn-based CRIC team has expanded to include additional expertise in Informatics (John Holmes, PhD), Biostatistics (Douglas Schaubel, PhD) and Cardiology (Rajat Deo, MD, MTR). With our colleagues across the nation, we stand poised to continue making important and clinically relevant discoveries that improve the health of all people with kidney disease within all US including populations that are disproportionately afflicted.
Read more about the history of CRIC in Penn Medicine Medicine Magazine.