Case Study: Healthcare
Sentiment Analytics to Improve Patient Experience
A large healthcare provider that serves patients in Philadelphia and the surrounding communities in Pennsylvania and southern New Jersey wanted to improve patient experience by analysing patient sentiments extracted from free text comments in the patient feedback forms.
- Machine learning techniques are being used to automatically categorise opinions as either positive, neutral or negative and relate them to separate business aspects.
- The contexts are captured by a word embedding layer: word2vec.
- The output is then fed into the LSTM model, a type of Recurrent Neural Network, to classify texts into desired categories.
- SoftMax based attitude detection algorithm is then used to identify the user sentiments efficiently.
- The solution is providing an accurate assessment of patient opinions about different performance aspects of the hospitals.
- This is helping the client to identify key pain-points for patients through various stages of their healthcare journey.
- The client is witnessing a considerable increase in the positive sentiments along with an increase in the performance ratings.