Best solution for telling optimal timing and sequence RRS feed

  • Question

  • Hello,

    I am a little new to ML, but am wondering what will be the best algorithm to help with our problem. 

    We have terrabytes of healthcare data and are looking to use it to create a new ML algorithm.

    Below is a description of the situation.

    - Each healthcare claim is 30 days.

    - During that 30 days, there are a series of nursing visits by various types (Registered Nurse, speech language pathologist, physical therapy, etc.)

    - We also have a field for outcomes (died, admitted to hospital, still patient, discharged home).

    - There is also one more classification field which determines the type of patient by clinical group (Behavioral Health, Complex Nursing, wound, etc.)

    What we are looking to do, is determine the optimal series of nursing visits during that 30 day period for patients that are still a patient or discharged to home. Example output below.

    Clinical Group         Sequencing

    Complex Nursing     Day 1: Registered Nurse    Day 8: Physical Therapist      Day 12: Physical Therapist   

    Wound                   Day 1: Speech Language Pathologist        Day 20: Physical Therapist

    Behavorial Health    Day 1: Registered Nurse    Day 4: Physical Therapist    Day 8: Physical Therapist   Day 30: Registered Nurse

    Wednesday, November 4, 2020 7:14 PM