tap to open menu
Ten Algorithms that are changing healthcare

10 Algorithms that are Changing Health Care

The basic definition of an algorithm is simple: a step-by-step set of instructions for solving a problem or performing a task. A recipe for chocolate chip cookies is an algorithm. The “chunking” or “decimal bus stop” methods of doing long division are algorithms. Far more complex algorithms are behind some of the greatest advances in world history.

Today, algorithms are everywhere. They are the brains behind smart phones, email and WiFi. Without them there would be no online shopping or downloading MP3s—–never mind, Facebook or Google. Invisibly working in the background, they can predict crime and stock prices, serve up movies we might like and connect us with people we know. They’ve even led to the discovery of new planets. In fact, Christopher Steiner, author of the book Automate This: How Algorithms Came to Rule Our World, calls those who design algorithms, “the preeminent entrepreneurs of this generation.”

With this in mind, we asked ourselves: What are the game-changing algorithms in health care? No ranking is perfect. Ours is bound to stir dissent. But don’t let us be the judge. Tell us, what’s in your top 10?

1. Fourier Transform: Enhancing our Senses

The fourier transform has been dubbed one of most important algorithms of our time. It’s a mathematical technique for breaking complex signals into basic components. It allows technicians, for example, to see voltage fluctuations in a wire connecting a microphone to a loud speaker. Because it reduces a signal to a short list of numbers, it’s also used to squeeze audio and image files into portable packages (MP3’s and JPEG’s). Without it, medical imaging wouldn’t exist. Magnetic resonance and ultrasound machines couldn’t turn raw data into pictures that enable doctors to see inside our bodies to diagnose and treat bleeds and broken bones, tears, tumors and more.

2. TCP/IP: The Power behind the Internet

It’s almost impossible to imagine a world without the Internet, or to describe the myriad ways in which the Internet has changed health care. But there would be no Internet without TCP/IP, or Control Protocol/Internet Protocol. It’s the lingua franca for computers. Back in the late ‘60s there were no desktop computers, only massive mainframes––most of them owned by universities and the military. Because they weren’t built to a common standard, data could only be transferred manually and in an agreed-upon format. TCP/IP was created by the U.S. government to get systems talking.

3. RSA: The Encryption Algorithm

If it weren’t for RSA, an acronym based on the names of its three inventors, health care records might still be locked in filing cabinets, mailed and faxed. Developed in the 1970s by the military to defend against hackers, this algorithm allows for the secure transmission of digital data. It was one of the first practicable encryption algorithms––and encryption is key to the secure sharing of electronic health records.

4. MUMPS: Health Care’s Operating System

The Massachusetts General Hospital Utility Multi-Programming System–or M– was developed in an animal lab in the late 1960s. It’s a computer programming language made for the health care industry, and still used today by many hospitals and banks. It was one of the first languages to enable computers to run multiple programs simultaneously. Today it powers the entire Veterans Health Administration’s clinical records management system and Epic, America’ largest electronic health record software company.

5. Probabilistic Data Matching: The Clinician-Scientist’s Best Friend

If you’re a doctor treating John Silver, of what use to you are the electronic medical records of Jon Silver? Many computer searches are deterministic, a byte-to-byte comparison with zero tolerance for typographical or data entry errors. Probabilistic algorithms look for various bits of information in medical records, and then rank them according to their likelihood of belonging to John. They’re used to retrieve clinical data and aid in research. The probabilistic algorithm, Niave Bayes Classifier, for example, is used to update the probability estimate––or provide additional evidence––for a research hypothesis. Paired with genetic sequencing it allows biologists to better understand the evolutionary relationships among species or populations––to trace the phylogenetic relationships within major branches of Darwin’s tree of life

6. BLAST: Analyzing Gene and Protein Sequences

High-throughout sequencing has ushered in a new age of genetic discovery, making it possible to cheaply and quickly find mutations among the 3 billion base pairs of the human genome. Identifying mutations is just the first step, though. It falls to biologists, aided by computer algorithms, to make sense of the growing body of data, to work out which genes and proteins confer disease and how. Chief among those algorithms is the search tool, BLAST (Basic Local Alignment Search Tool). BLAST, a search algorithm, accomplishes this by comparing a sequence to a library or database of sequences and relevant scientific papers. Publications about BLAST hold the 12th and 14th spots in a list of the 100 top-cited science papers of all time, according to the journal, Nature. BLAST is being surpassed, though, by Clustal, a similar program for aligning multiple sequences at once, according to Nature.

7. Neighbor-Joining: Phylogenetics

No. 20 on Nature’s list of top-cited science papers is a study explaining the “neighbor-joining” algorithm, which, when paired with genetic sequencing, allows biologists to better understand the evolutionary relationships among species or populations––to trace the phylogenetic relationships within major branches of the tree of life. Phylogenetic trees are used in drug development to, for example, identify closely related, naturally occurring chemical compounds suspected to have medicinal value. Phylogenetic trees of pathogens help scientists understand the adaptive evolution of bacteria, viruses and parasites––how they infect hosts, subvert immune systems and resist treatment.

8. Google Search: Page Ranking

We do dozens of Internet searches every day using Google or Bing. Whatever the platform, it’s a complex algorithm known as “page rank” that does the work by scouring the Internet for pages containing the key words you enter and then ranking them based on factors, such as their location or their frequency of use. “Googling” answers our burning questions, or at least gives us a start. But has it helped or hurt health care? It’s certainly democratized it and put more information within easy reach of patients.

9. Medical Algorithms: Decision Aids for Better, Safer Care

Airplane pilots use checklists to safeguard against mistakes and rely on formulas to plot the right speed and trajectory to get the plane safely to its destination. With increasing awareness of medical errors––believed to be responsible for more than 100,000 deaths a year––health professionals are now using these same strategies to guide delivery of care. A medical algorithm can be as low-tech as a look-up table or decision tree (if symptoms A, B and C are evident, then use treatment X). Or it can be as complex as the programming behind mechanical ventilators. Medical algorithms remove some of the uncertainty from medical decision-making and improve the efficiency and accuracy of provider teams. They’re developed by providers for providers, and they’re evidence-based and data-driven.

10. Health Scores: Quantifying Illness

Scoring systems, like Apgar for evaluating a newborn’s condition at birth or APACHE for determining the severity of patients in intensive care, help physicians monitor and predict a patient’s prognosis based on a multitude of factors, from heart rate and oxygen levels to neurological reflexes. It’s a system for seeing a patient holistically, a prognosis or wellness meter.