Predictive Medicine

It would be a fantastic world if, instead of waiting to see if people got sick, we could proactively stop illnesses from becoming a problem in the first place. While such innovation requires massive investment, the long-term benefits will be groundbreaking and change the healthcare world as we know it.

As artificial intelligence (AI) continues to proliferate industries and markets, predictive medicine is entering the realm of possibility. Health systems can predict the risk of an individual in developing a condition, then offering preventative measures to help bring them to good health. The aim will be to reduce disease and mitigate costly treatments significantly.


What is predictive Medicine?

Predictive Medicine assesses the probability that an individual will develop a disease in the future. Laboratory tests and genetic examinations will analyze the current health of an individual, along with social data, to review it against research and likely outcomes. The field is relatively new as computing power and algorithms have only recently become powerful enough to process the vast amount of data that the field demands.


How does Predictive Medicine work?

Machine learning algorithms can take a pool of data, detecting and analyzing differences in genes, proteins, or other cell components to decide the likelihood of developing a disease or whether responses to treatment are plausible.  Following the completion of the Human Genome Project in 2003, scientists can now map the entire human genome. Even though DNA has been around for the best part of half a century, the project created hype around the ability to unlock cures for diseases that were never thought possible.

Genomic data can be obtained much quicker today than before the project, with huge insight to be understood from the information.


Predictive algorithms will flag risk factors to physicians and patients, allowing them to work together on a solution.  For example, if the data insights suggest an increased risk of heart attacks, the patient may benefit from regular appointments with a cardiologist.


How Predictive Medicine revolutionizes healthcare

Early diagnosis is critical to recovery. Over 90% of bowel cancer patients will survive for more than five years when it is discovered during the initial stages. Similarly, 80% of lung cancer patients will survive for at least one year when diagnosed at an early stage, against 15% at an advanced stage.

Predictive medicine can take early diagnosis to the next level. For example, patients could start treatment prior to any symptoms developing. Such a practice might help them to change their lifestyle and ideally prevent the disease from prevailing. The world population is aging, and healthcare budgets are tight, meaning any reduction of treatments is well-supported by governments.

The key benefits of Predictive Medicine are:

  1. Clinical trials can be completed using machines and can negate animal testing. The cost is reduced while practices are more ethical.
  2. Healthcare professionals can have machine-based training for diagnosis and avoid making costly mistakes on humans.
  3. The unnecessary use of drugs are reduced by reviewing the individual data of each patient. They are given the treatment that is right for them, rather than making assumptions on a condition.
  4. The failure rate of medicinal products are reduced by predicting the compounds that are likely to succeed and those likely to fail. Atomwise provides an AI tool that makes the task of identifying potential drugs more efficient. The algorithms will predict the binding affinity of molecules so Atomwise can recommend those with the best binding affinity for a disease protein. The prediction engine can handle millions of possibilities.

Predictive Medicine can transform the healthcare industry if the challenges associated to it can be resolved. We talk about those challenges later in this post.


Predictive Data Mining

Data mining is a term that is increasingly being used in medical fields over the last few years. In short, it is the use of methods to analyze vast amounts of data. In predictive medicine, the goal is to derive models that can support decision-making by discovering new patterns or relationships that provide a clear result to data analysts.

A study by Bellazzi and Zupan shows predictive data mining in action. They use two types of algorithm to form a prediction regarding a sample set of patients who went through hip arthroplasty surgery.  In the dataset, 20 patient’s records show attributes of overall health, timing, and complications that relate to the surgery. The data passes through two frameworks, a naïve Bayesian classifier, and a decision tree, inducing predictive models from the information.

Using the training data, the algorithms will then predict the likely outcome of future cases. You can read more about the study here.

Applications of Predictive Medicine

With vast volumes of medical data available, there is the potential for countless applications of predictive medicine. Below are some of those that are already in place or that companies have been experimenting with.

Testing Risk

A risk test is the most common form of predictive medicine. Using data, it looks at whether a patient exhibits the risk factors that may lead to disease. For example, there is a far greater propensity for a 60-year old heavy smoker to develop lung cancer or emphysema than a 20-year old non-smoker.

Diagnostic Tests

A diagnostic test helps a doctor to determine whether they should confirm or refute a diagnosis. When a diagnosis is tentative, additional predictive tests will help to understand the root cause of an issue rather than doctors relying on assumptions.

Screening Newborns

Shortly after birth or sometimes pre-birth, a newborn screening test can predict potential genetic disorders. US state law makes taking blood samples of every newborn baby mandatory, meaning the technique is widespread across the country. During pregnancy, parents can go through testing to ascertain whether they carry any gene mutations that could cause genetic disorders. Even before conception, parents can decide whether they should try for a child with predictive tests.

Consumer Testing

Sites like 23AndMe allow people to test their genes without going to a physician. Although they cannot be as granular as other types of predictive medicine, it gives consumers better access and more data privacy. For example, during Covid-19, 23AndMe can offer related tests to help put consumer’s minds at ease.


Case Study – BaseHealth

In 2015, the health data platform, BaseHealth, raised $3.6 million to develop the Genophen health assessment service. Using clinical data, lifestyle data from wearable devices, and family data, the service can calculate the user risk of developing specific diseases. The system will also explain the top factors contributing to the results across forty of the most common complex conditions.

The BaseHealth platform also offers users ways that they can modify behaviors to reduce their risk over the course of a lifetime.



Challenges of Predictive Medicine

It will take time for everyone to embrace predictive medicine with many existing challenges.

False Positives

Imagine that predictive medicine achieves an accuracy rate of 99.9% in diagnosing diseases. Even with that, one in a thousand people will still undergo the stress of believing they are at risk of an illness that never develops. The problem here is that medication and expense could go into prevent something that would never happen in the first place. If those medications have bad side effects for the patients, predictive medicine could end up harming the quality of their life.


All forms of artificial intelligence (AI) and predictive algorithms raise the challenge of responsibility. For example, if the algorithm does falsely predict cancer, who can the patient hold accountable for the wrong decision? Healthcare providers will need to think about the policies and procedures they put in place to mitigate such questions.

Personal Data

Medical data is highly personal, and the general public still doesn’t have enough trust in organizations that use it. An issue with machine learning algorithms is that they are kept in a “black box,” making them difficult to interpret. The healthcare industry needs to focus on explainable AI to gain credibility and ensure that patients know how they arrive at decisions.


There is a possibility that predictive medicine could start generating biased algorithms. For example, if the data says that people of black origin are always at risk of a specific disease, it will continue to use that in decision-making. In communities that are densely populated with people of similar races, this can create bias and algorithmic stereotyping. Data quality should be a priority in predictive medicine, and collection methods be carefully evaluated.



The future of medicine lies with prediction and precision. The field provides several opportunities for pioneering providers to take healthcare to a new level. Computing power is still increasing, and algorithms are becoming more complex, which means there is still much potential improvement for predictive medicine. New technology such as edge computing, 5G, quantum computing, and cognitive AI will see the market expand exponentially in the next decade.

Disclaimer:  The author of this text, Jean Chalopin, is a global business leader with a background encompassing banking, biotech and entertainment.  Mr. Chalopin is Chairman of Deltec International Group,

The co-author of this text, Robin Trehan, has a bachelor’s degree in economics, a master’s in international business and finance, and an MBA in electronic business.  Mr. Trehan is a Senior VP at Deltec International Group,

The views, thoughts, and opinions expressed in this text are solely the views of the authors, and do not necessarily reflect those of Deltec International Group, its subsidiaries, and/or its employees.