{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/ca120823bd8e41fb90ed23c223eae27e\" frameborder=\"0\" width=\"2730\" height=\"2047\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":2047,"width":2730,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":2047,"thumbnail_width":2730,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/ca120823bd8e41fb90ed23c223eae27e-e6f5792204591ce8.gif","duration":537.896,"title":"Using AI to Fill Burdensome Medical Forms","description":"In this walkthrough, I demonstrate how to use Doximity's generative AI to bootstrap the form-filling process — no fancy EHR integration required.\n\nThe example form is New York State's CDPAP, the form used to get patients home health aide services. I upload the form directly into Doximity and write a simple prompt: fill out the attached form using the patient information provided below, and flag anything you can't complete due to missing data. I paste in de-identified patient notes from a prior visit and let the AI run.\n\nThe output is solid. It pulls the primary diagnosis, secondary diagnoses with ICD codes, a description of the current condition, the treatment plan, functional limitations, dietary needs, home therapies, and contributing factors — essentially the clinical meat of the form. I then ask it to update the form with the proper ICD codes based on the primary diagnoses, and it returns accurate codes that I spot-check against a separate AI tool.\n\nSubscribe to Healthcare Huddle: https://www.healthcarehuddle.com/upgrade"}