Clinical Speech Recognition and Natural Language Processing

In HealthTalk we are developing an automated reporting tool for any healthcare provider, saving at least 3 hours per day on your bookkeeping...

Healthcare provider or bookkeeper

Nowadays a healthcare provider is not much more than just a bookkeeper with a small side-job; taking care of people. And the amount of time spent performing administrative tasks has increased to about half of the physicians working day (1, 2). These administrative tasks reduce the physicians professional satisfaction (3) and adversely impact the physician-patient relationship. Other studies assessed the association between EHR usage and burnout, finding that increased EHR use, particularly outside the office hours, was associated with an increased risk of burnout. Recently, physicians have hired physician scribes to help alleviate administrative workloads, that is, people to handle administrative tasks, like summarising consultations.

Studies show positive results for using medical scribes, with physicians spending more time with patients in-person and fewer hours after hours at the Emergency Room. Although medical scribes may seem to be a perfect solution, they transfer workloads to other staff. As a result, out-of-pocket costs for care rise, and administrative costs remain significant. Recent opinions (4) described the need for what is called the digital scribe.

That is exactly why HealthTalk is using technologies like automated speech recognition (ASR) and natural language processing (NLP) to automate (parts of) clinical documentation. The proposed structure of a digital scribe includes a microphone recording the conversation, an ASR system that transcribes that conversation, and a collection of NLP models for mining or summarising the relevant information and providing it to a doctor. The extracted information can, for example, be used to generate clinical notes, to add billing codes, or use extracted information to support diagnoses.

See our example using MeSH to detect clinical structures:

Based on this we are now discussing with medical experts for the right business case. Also we will propose new projects in this area in the upcoming months. If you are interested please ping us.

1. Arndt, B. G. et al. Tethered to the EHR: primary care physician workload assessment using EHR event log data and time-motion observations. Ann. Fam. Med. 15, 419–426 (2017).
2. Tai-Seale, M. et al. Electronic health record logs indicate that physicians split time evenly between seeing patients and desktop medicine. Health Aff. 36, 655–662 (2017).
3. Rao, S. K. et al. The impact of administrative burden on academic physicians. Acad. Med.92, 237–243 (2017).
4. Quiroz, J. C. et al. Challenges of developing a digital scribe to reduce clinical documentation burden. Npj Digital Med.2, 1–6 (2019).