Administration became an essential and time consuming part of healthcare these days. Doctors have to register everything they discover within the patient file. During an outpatient clinic visit the complaints, the medication used, and if possible the diagnosis and treatment plan have to be registered. This time is at the expense of the real care.
Together with LUMC and Google we are working on automating this process. By using speech-to-text technology we’re recording the conversation between doctor and patient and thereby saving all required information within the system automatically. All data will be saved within the patient file without doctor intervention – in a structured way – exactly the same for every department and patient. This doesn’t only save a lot of time, it also simplifies collaboration between medical departments.
Less administration & Data Science
Another advantage standardized data brings is: data analysis. Because normally doctors would enter the data themselves, manually, it isn’t only complicated to search within a file, it also makes data analysis almost impossible. Every doctor uses his own words and often in a different order. The focus of this initial project will be on reducing the administrative load, followed by more advanced data analysis. Machine Learning is better in recognizing patterns than humans will ever be. By comparing complaints, diagnoses and other patterns, it is likely that a diagnosis can be set in an earlier stage.
At LUMC we’re using the proven sprint methodology to come to results quickly. By devising, developing and testing a solution on a small scale in 20 days, we can quickly determine whether it can be deployed on a large scale. This way we can ensure within weeks if a solution really meets your expectations.
Wondering about the possibilities of speech recognition and machine learning for your organization? Fill out the form and we’ll get in touch. Or call Stefan, Chief machine learning: 020-3337500.
Get in touch with me!
Read more about this in Zorgvisie [Dutch-only]