Interface for patient-provider conversation and auto-generation of note or summary
A computer-implemented method includes receiving, by a computing device, a particular textual description of a scene. The method also includes applying a neural network for text-to-image generation to generate an output image rendition of the scene, the neural network having been trained to cause two image renditions associated with a same textual description to attract each other and two image renditions associated with different textual descriptions to repel each other based on mutual information between a plurality of corresponding pairs, wherein the plurality of corresponding pairs comprise an image-to-image pair and a text-to-image pair. The method further includes predicting the output image rendition of the scene.
1. A machine learning based method for automatically processing a note based on a conversation between a patient and a healthcare provider, comprising:
automatically recognizing, in a transcript of an audio recording of the conversation, words or phrases spoken by the patient, wherein the words or phrases relate to one or more medical topics relating to the patient;
generating a billing code associated with a given medical topic of the one or more medical topics, wherein the billing code is based on the automatically recognized words or phrases;
generating the note in substantial real time with a rendering of the audio recording, and populating the note with the automatically recognized words or phrases, wherein the generating of the note comprises applying a machine learning model to generate at least a portion of the note in prose; and
displaying, by a computing device, the billing code alongside a portion of the note that includes the automatically recognized words or phrases.
2. The method of claim 1 , further comprising:
displaying, by the computing device, a textual description to support the billing code, wherein the textual description is based on the transcript and the billing code.
3. The method of claim 1 , wherein the generating of the billing code comprises:
providing information related to a medical symptom by providing one or more labels for phrases associated with the billing code.
4. The method of claim 1 , further comprising:
generating the transcript using a speech-to-text engine in substantial real time with the rendering of the audio recording.
5. The method of claim 4 , further comprising:
training the speech-to-text engine using supervised learning, wherein the training is based on labeled training speech data to recognize medical-related terminology in speech, wherein the medical-related terminology comprise one or more of symptoms, medications, or human anatomical terms.
6. The method of claim 4 , wherein the automatically recognizing of the words or phrases is performed by a named entity recognition model, the method further comprising:
receiving the text generated by the speech-to-text engine; and
recognizing, in the text, medically relevant words or phrases spoken by the patient.
7. The method of claim 6 , further comprising:
training, using deep learning word embedding, the named entity recognition model based on training data comprising one or more of a corpus of medical text books, a lexicon of known medical ontologies, or a collection of annotated medical encounter transcripts.
8. The method of claim 1 , wherein the note comprises one or more classifiers associated with the billing code, and wherein the one or more classifiers is indicative of a medical attribute associated with the billing code.
9. The method of claim 1 , wherein the displaying of the billing code and the textual description further comprises:
receiving, via the display of the workstation, an indication to toggle to a billing tab displayed by the display of the workstation; and
displaying, in response to the indication to toggle to the billing tab, the billing code and the textual description in a display associated with the billing tab.
10. The method of claim 1 , wherein the words or phrases are placed into appropriate categories or classifications in the note.
11. The method of claim 9 , wherein the note is editable to perform one or more of an addition, a removal, or an edit of, the categories or classifications in the note.
12. The method of claim 1 , further comprising:
displaying, by the computing device, a minimized icon on a display of an electronic health record screen which, when activated, toggles to the audio recording of the conversation and the transcript thereof.
13. The method of claim 1 , further comprising:
automatically generating suggestions of alternative words or phrases for words or phrases in the transcript; and
providing one or more tools to approve, reject or provide feedback on the generated suggestions to thereby edit the transcript.
14. The method of claim 1 , further comprising:
generating at least one of a suggestion of a topic to follow-up with the patient, and a suggestion of a clinical problem; and
displaying, by the computing device, the at least one of the suggestion of the topic, and the suggestion of the clinical problems.
15. The method of claim 1 , further comprising:
identifying an inaudible word or phrase in the audio recording of the conversation;
automatically generating a suggestion of one or more alternative words or phrases to replace the inaudible word or phrase; and
displaying, by the computing device, the one or more alternative words or phrases to replace the inaudible word or phrase.
16. The method of claim 15 , further comprising:
generating a confidence score associated with the suggestion of the one or more alternative words or phrases, and
wherein the displaying of the one or more alternative words or phrases comprises displaying the associated confidence score.
17. The method of claim 1 , further comprising:
determining, from the audio recording, an incorrect or incomplete mention of a medication by the patient;
automatically generating a suggestion of one or more alternative medications to replace the incorrect or incomplete mention of the medication; and
displaying, by the computing device, the one or more alternative medications.
18. The method of claim 1 , further comprising:
linking words or phrases in the note to relevant parts of the transcript from which the words or phrases in the note originated.
19. A server for automatically processing a note based on a conversation between a patient and a healthcare provider, comprising:
one or more processors; and
memory storing computer-executable instructions that, when executed by the one or more processors, cause the server to perform operations comprising:
automatically recognizing, in a transcript of an audio recording of the conversation, words or phrases spoken by the patient, wherein the words or phrases relate to one or more medical topics relating to the patient;
generating a billing code associated with a given medical topic of the one or more medical topics, wherein the billing code is based on the automatically recognized words or phrases;
generating the note in substantial real time with a rendering of the audio recording, and populating the note with the automatically recognized words or phrases, wherein the generating of the note comprises applying a machine learning model to generate at least a portion of the note in prose; and
displaying, by a computing device, the billing code alongside a portion of the note that includes the automatically recognized words or phrases.
20. An article of manufacture for automatically processing a note based on a conversation between a patient and a healthcare provider comprising one or more computer readable media having computer-readable instructions stored thereon that, when executed by one or more processors, cause the one or more processors to carry out operations comprising:
automatically recognizing, in a transcript of an audio recording of the conversation, words or phrases spoken by the patient, wherein the words or phrases relate to one or more medical topics relating to the patient;
generating a billing code associated with a given medical topic of the one or more medical topics, wherein the billing code is based on the automatically recognized words or phrases;
generating the note in substantial real time with a rendering of the audio recording, and populating the note with the automatically recognized words or phrases, wherein the generating of the note comprises applying a machine learning model to generate at least a portion of the note in prose; and
displaying, by a computing device, the billing code alongside a portion of the note that includes the automatically recognized words or phrases.