Course Recap

Course Recap

Congratulations on finishing the content portion of this course!

Let's put everything you did into perspective!

• In the EHR Data security and Analysis part of this course, you learned about Data Security and Privacy. Including the key terms and regulations, you need to be aware of when dealing with healthcare-related data. You also analyzed healthcare data through exploratory data analysis and evaluated how demographically representative your input data is before starting to build models.
• You then dove into the deep end of the pool with EHR Code sets where you worked with Diagnosis, Procedure, and Medication codesets. You can now interpret and work with these types of codes in your healthcare-related projects. Further, you can categorize and group EHR dataset codes for further analysis and use in machine learning.
• After code sets, you learned to address the different EHR dataset levels, at the Line, Encounter, and Longitudinal levels. You also used your Dataset splitting lumberjack skills to ensure that you don't leak data into your model training and give it features that might compromise the use of your model in production. You also use the TensorFlow Feature column API for Feature engineering.
• We also eventually got to the fun stuff, Building, Evaluating, and Interpreting Models with a focus on Bias and Uncertainty. You used your Feature Engineering skills to build simple Tensorflow models using DenseFeatures. Then you used evaluation metrics like Brier Scores to evaluate your model. You can also now conduct a Demographic Bias Analysis using Aequitas to ensure your models are less negatively biased. Then, you used Uncertainty Estimation to estimate model predictions more accurately. Finally, you interpreted your models using model agnostic methods like Shapley Values.

Working with EHR data has never been more critical than it is today. So it's time to put the new skills you have learned into practice to help provide better insights, predictions, and healthcare for patients.

Best of Luck on the Project and Your Future!!