Center for biomedical informatics Children's Hospital Informatics Program
Isaac Kohane

Isaac S. Kohane, MD, PhD

Professor of Pediatrics & Health Sciences Technology, Harvard Medical School

Director, Countway Library of Medicine, Harvard Medical School

Co-Director, HMS Center for Biomedical Informatics

PI, National Center for Biomedical Computing: Informatics for Integrating Biology to the Bedside

Director, Children's Hospital Informatics Program

Co-Director, Bioinformatics and Integrative Genomics training program

Director, Informatics Program, Harvard Catalyst (CTSA)


Institute of Medicine

American College of Medical Informatics

American Society for Clinical Investigation

Society for Pediatric Research



CHIP Miscellany

Atul Butte at CHIP

Longwood mon amour



103 years ago, the Flexner report provided damning evidence of the sad state of the medical education system, and its implications for the quality of care provided and the lack of quality in biomedical research. Today, there is a growing consensus that our healthcare system is a poor learner and poorer performer yet relentlessly growing in expense. A major contribution to this perfect storm of underperformance is lack of quantified instrumentation of our healthcare system as a whole and a fundamental bottleneck in translating our state-of-the-art biomedical science into efficient, expert and monitored clinical decision-making. The creation of an efficient biomedical information commons where all health data for all individuals, under personal control, can be analyzed is an essential precondition to bringing knowledge to the science and practice of medicine.

Two points in this regard:

1. Biomedical Discovery—Finding the True Names of Disease We may pride ourselves on our distance from the herbalists who discovered symptomatic relief for fever, yet the most common diseases (e.g. diabetes, depression) continue to this day to be characterized symptomatically. Unlike infectious diseases where the name of the disease is its cause, the true names of most common diseases, now at pandemic prevalence, remain unknown. This ignorance is reflected in our symptomatic, phenomenological treatments for these diseases. Now, computationally bringing together the entire spectrum of data types pertaining to human physiology, we have the opportunity to study diseases at multiple levels (e.g. genomic/molecular, healthcare system, behavioral, epidemiological and social web) to precisely triangulate the location of every individual with respect to diagnosis (who are you most alike) and prognosis (who will you be most alike). By virtue of its comprehensive and integrative perspective, biomedical informatics can help us find the true names of disease and thereby treat them effectively.

2. Clinical Care—Performed As if We Were Living in the Pre-Internet Era. Why is it easier to find out what experienced shoppers worldwide have had with the latest digital camera than it is to determine what adverse events patients have had with a particular drug? Why are blood tests and X-rays repeated needlessly? Why can one replace an application on an iPhone with a mouse click yet require a team of engineers to do it for an electronic health record system? Why are the latest algorithms for detecting domestic abuse or selecting the right genetic test not at the finger tips of all clinicians in all settings? Among the multitude of answers to these questions, two answers stand out: a) There are all too many parties who have a stake in data inertia and opacity. That is, they perceive that their business or science would be impeded or made less rewarding by allowing individual data to flow frictionlessly. b) Too many of the leaders in medical care have not made information processing, which is at the core of medicine, central to their design of their healthcare delivery systems. Instead, these are seen as details to be managed by the "IT" staff. If we are going to make our data work to improve our health and medical science, the change required is as much cultural as technical.

What I am working on now:

Selected References

Schmid PR, Palmer NP, Kohane IS, Berger B. Making sense out of massive data by going beyond differential expression. PNAS. 2012 Apr 10;109(15):5594-9.

Kohane IS, Drazen JM, Campion EW. A glimpse of the next 100 years in medicine. N Engl J Med. 2012 Dec 27;367(26):2538-9.

Kohane IS, Mandl KD, Taylor PL, Holm IA, Nigrin DJ, Kunkel LM. Medicine. Reestablishing the researcher-patient compact. Science. 2007 May 11;316(5826):836-7.

Butte AJ, Kohane IS. Creation and implications of a phenome-genome network. Nat Biotechnol. 2006 Jan;24(1):55-62.

Kho AT, Zhao Q, Cai Z, Butte AJ, Kim JY, Pomeroy SL, et al. Conserved mechanisms across development and tumorigenesis revealed by a mouse development perspective of human cancers. Genes Dev 2004;18(6):629-40.

Butte AJ, Tamayo P, Slonim D, Golub TR, Kohane IS. Discovering functional relationships between RNA expression and chemotherapeutic susceptibility using relevance networks. PNAS 2000;97(22):12182-6.

Full MEDLINE Bibliography (does not include computer science literature)