Recommending drugs and medicines based on pre-existing conditions & patient profile

Summary

A study conducted in hospitals by the General Medical Council, UK and published by BBC, found out there are errors in the prescriptions done by doctors.* Among those wrong prescriptions, some of the errors are lethal. To reduce such mistakes, an AI-based matching algorithm can be used to suggest drugs to patients based on the patient profile(pre-existing condition). You might be hesitant to implement AI as a traditional development cycle of an AI-based matching algorithm takes almost 4 to 12 months to create such matching algorithms. However, Caboom can reduce such a traditional development cycle by one-fourth to one-twelfth. Caboom’s short development cycle includes several iterations to achieve high accuracy.

Scenario 

Emma is a patient of hypertension. During the cold, she has a nose blockage problem(decongestants). Pseudoephedrine is a popular drug both by doctors and over-the-counter. Pseudoephedrine is great for clearing up the congestion of bad colds, however, it shrinks the blood vessels, they raise the blood pressure. This is dangerous to patients with hypertension like Emma. Phenylephrine can be an alternative drug for Emma based on her pre-existing condition.

According to a report published by Pharma Publishing Media Europe, around 1 in 10 hospital prescriptions contain errors. A study done by General Medical Council(GMC) in 2009, UK found that out of 124,260 around 9% prescriptions have errors such as omitting drugs, incorrect doses, patient allergies not taken into account, and so on. Out of 9% of the error, 2% contain potentially lethal instruction.

Can AI assist doctors to recommend drugs?

There is a serious discrepancy between the doses given to the patients on doctors recommendation and the clinical pharmacists medical dose reaction study. In a conventional way, communication between the doctors and clinic pharmacists can minimize this kind of error. The sophisticated way to tackle this problem would be using Artificial Intelligence(AI) based solutions. AI can enable doctors to match the patient condition profile based on demographic information, allergies and disease diagnosis with the drug profile, based on drug reaction on drugs, food, allergies and disease and the adverse impact. AI solutions will remarkably reduce the errors. 

Caboom is a platform that enables the state of the art recommendation, where it can recommend the drug based on the personal profile to reduce the adverse impact on patients.

Figure 1: Drugs recommendation based on patient profile and drug profile

Figure 1: Drugs recommendation based on patient profile and drug profile


Caboom: Fast, Easy, and Efficient way of creating profile matching solution

Building an AI-based matching algorithm is about working on heavy math, finding patterns hidden in data, working with all the caveats, and power technology to meticulously combine everything to your use-cases. At Caboom, this complex and time-consuming task is automated with just a few clicks without hiring an in-house data science team in your organization. Caboom’s objectives are:

  1. To make AI recommendations accessible, simple, easy, and fun to even non-AI background people so that anyone can use this revolutionary tool to solve problems, grow businesses, and enrich their experiences.
  2. Replace or Augment in-house data science team and traditional AI recommendations development life cycle by the Caboom platform to reduce cost and time.
  3. Makes iterations cycle short to improve the performance of AI models.
Figure 2: Drugs recommendation based on patient profile and drug profile using Caboom

In a drugs-patient profile matching solution, the Caboom platform can replace the traditional development cycle to provide AI-based recommendation solutions at low cost and reduce time without compromising accuracy.

Figure 3: Traditional recommendation vs Caboom Recommendation


If you are a product manager, product owner or in a hospital management team without an AI/ML job title with a team of software developers looking to develop an AI-enabled recommendation system, then Caboom is the right platform to start. You can focus on your service, Caboom will take care of your AI model and its deployment for you.

* Article published by BBC

See instant AI recommendation results with caboom
Start with your data or a sample for instant results right away.
Request Access