Will healthcare play nice with data science?

Data science brings new understanding to healthcare and other fields. But does healthcare want to be told something different from such an understanding?

At its core data science is meant to challenge the common wisdom. Susan Lindquist, a geneticist and biochemist at MIT, states, “If you’ve been studying in a field all your life, having someone from a completely different field come and tell you something important could be rather irritating. It’s just human nature.” Part of the problem is a turf issue, she noted, but also a gap in understanding because experts in other fields “don’t get why my ideas work.”

Believe it or not there was a time in medicine when physicians were not physician-scientists. They thought of illnesses as an imbalance caused by bad air or evil spirits instead of looking at anatomy and empirical data.

The history of science is replete with such theories that only became accepted by the scientific community after a long, uphill battle.

Semmelweis (1846) asked healthcare providers to wash hands; over 130 years later CDC published the first national hand hygiene guidelines.
Jenner (1796), because he was a simple country doctor, experienced the prejudice of the medical world after discovering the vaccination for small pox; 50 years later the government of the day banned every treatment for smallpox except Jenner’s.
Snow (1849) refused to believe Cholera was contracted by bad air, but rather through the mouth; his “germ-theory” of disease was not widely accepted until the 1860s.

All the wonderful evidence-based science in the world cannot cover up some previous ideas that are part of our ugly history, including heroin cough syrup for children sold by Bayer & Co. or making Lysol the best-selling method of contraception during the Great Depression. Even with these unfortunate cases in our recent history, there are still many scientists whose ideas struggle to find acceptance.

Are you ready for a revolution?

A quick review of history will demonstrate that every 50 years there is a revolution in healthcare based on the trends of that period.

In the 1870s we had germ theory of disease and promotion of public health efforts. In the 1920s we discovered penicillin and propelled forward the use of medication as treatment for disease. In the 1970s we began randomized, controlled trials which ushered in a period of evidence-based medicine. Now approaching the 2020s, we are set for another revolution: using data science to empower physicians’ work — and most significantly — improve patient outcomes.

I get it: undergraduate studies, medical school, residency and fellowships certainly add up to expertise cultivated over many years; and for that reason, some healthcare providers may have a hard time with receiving outside help. Many of these providers’ view excursions into the medical field by non-medical individuals as intruding. However, Data science does not discriminate by field of study and finds data patterns in the most unexpected places.

The sooner and more broadly professionals within healthcare accept data science as beneficial to their cause, the sooner healthcare could reduce rising mortality rates and out-of-control medical costs. Data science, machine learning and Big Data are not a panacea, but significant approaches still being underutilized — if at all — in modern healthcare.

Data science should be leveraged to make progress in areas that concern many patients and hospital CFOs:

Table 1. Top 20 most expensive conditions treated in U.S. hospitals, all healthcare insurance, 2011

Rank	Category	Hospital Cost, U.S. $, in millions	National Costs, %	Number of Hospital Discharges
1	Sepsis	20.3	5.2	1,094,000
2	Degenerative joint disease	14.8	3.8	964,000
3	Complications of device, implant, or graft	12.9	3.3	699,000
4	Births	12.4	3.2	3,818,000
5	Acute myocardial infarction	11.5	3.0	612,000
6	Back problems	11.2	2.9	667,000
7	Pneumonia	10.6	2.7	1,114,000
8	Heart Failure	10.5	2.7	970,000
9	Heart Disease	10.4	2.7	605,000
10	Respiratory failure	8.7	2.3	404,000
11	Circulation of blood to the brain	8.3	2.2	597,000
12	Irregular heartbeat	7.6	2.0	795,000
13	Complications of medical care	6.9	1.8	529,000
14	COPD	5.7	1.5	729,000
15	Rehab services	5.5	1.4	420,000
16	Diabetes with complications	5.4	1.4	561,000
17	Gallstones	5.1	1.3	469,0000
18	Hip fracture	4.9	1.3	316,000
19	Mood disorders	4.9	1.2	896,000
20	Unspecified kidney failure	4.7	1.2	498,000
Total for top 20 conditions		182.3	47.1	16,755,000
Total for all hospitalizations		387.3	100	38,591,000

Source: Agency for Healthcare Research and Quality (AHRQ), Center for Delivery, Organization, and Markets, Healthcare Cost and Utilization Project (HCUP), Nationwide Inpatient Sample (NIS), 2011

What does healthcare in the 2020s look like?

Currently, modern medicine treats the 84-year-old diabetic patient with hypertension similarly to the 43-year-old athlete with hypertension, based on both being grouped together in the same clinical trial. Let data science help personalize care by learning what worked best previously for millions of similar patients. This level of customized care offers the promise of better and more applicable treatment and outcomes.

Forward-thinking healthcare providers can take advantage of data science, robust tools, and clear processes for intervention today. There is no need to wait for 2020.

Data science represents an opportunity for many types of innovators — MD, RN, MBA, etc. Data science and data-driven healthcare represent the potential for better outcomes and lower mortality rates for patients. Brace for a revolution in healthcare where we all have the opportunity to help and everyone has a stake.

What are you saying?

The future of healthcare is now. We have what we need for the next revolution in healthcare. Embrace new technology, new methods and large amounts of data. Realize data science is built on science and should be a large part of patient care.

Next read this:

Damian Mingle is the Chief Data Scientist at WPC Healthcare and an expert in analytics. He leads a team of professionals transforming data into meaningful strategic insights.

Security vs. innovation: IT's trickiest balancing act