Feature Requests

Psychiatrist using dictation; word-for-word setting: I find that heidi is generating text far from my original dictation/meaning and not flagging its creative generation of text. EGs: 'seen with danielle smith, community nurse' became 'seen with Dr Daniel Smith'. 'N-H-I DABxxxx' became 'date of birth 200x/0x/0x', 'thought content; predominant themes of...' became 'thought content and thought form' (I didn't say 'thought form' in that sentence, never use the phrase 'thought content and thought form' as they are different concepts, and never use that phrasing in dictation as I use 'thought process' not 'thought form'). This isn't a template issue, it is the pretrained data it is referencing, and its level of confidence at generating a solution that is congruent to that, even in the face of ambiguity. The problem with this is that the generated text has correct syntax/spelling, and has the appearance of semantically coherence - but it not semantically correct. This leads to more effort in proofreading - as it doesn't stand out as an obvious error. I have 3 questions: Is it possible to change which pretrained data it is using - eg a generic english dataset, or 'professional document' dataset rather than a medical/psychiatric dataset. Is it possible to have a warning for ambiguous text generation - eg if the dictation seems ambiguous/atypical then some flag for that is used (eg a placeholder term, question mark, highlighting of the best guess, or several possibilities etc), rather than an unflagged relatively 'creative' generated solution. Is it possible to have stricter ambiguity thresholds beyond the 2 options of 'smart' or 'word-for-word'. eg really-really-strict word-for-word. Could the clinician change their language use? A comment about psychiatry: this might be a solution in simplistic checklist style psychiatric documents, which are common in managed care, production line type settings. In other settings, expertise and style of language use is fundamental part of the specialty. Expert psychiatrists do not want to adopt generic conventional language use. Changing my language use to average/conventional phrasing is not an option. I would resign before that happens.

I would like to request to reduce the size of the floating dictation mic that is now part of the features. It is very large and takes up a lot of room. Especially on laptops where the EMR itself is very bulky and taking up the whole screen, I often have to move the dictation mic over to the side and back depending on what I'm trying to click on. Options would be to make it a horizontal option which would be easy to tuck away in a corner or to have a minimize feature so you can see that it is still running,.

When I dictate I add the client name in the client details at the top but Heidi doesn't use this and frequently spells the client's name incorrectly (esp if an unusual name) despite it being provided. It seems to use the name provided in transcription but not dictation - I don't understand why but it would be great if this could be fixed.

Please remove the constant request to allow access to microphone on MacOS Desktop App (standalone dedicated Heidi app) - I use Heidi on my Mac to review letters and not dictate them so I do not want the microphone to automatically be on whenever the Heidi App is running. I have reverted to using the webpage on Chrome as I can't stand the pop up request to allow microphone every time I click on a new patient.

Here’s a clearer and more professional version: --- The user is asking if there is an indicator or notification when the microphone is not detecting any voices.

Quite often our consultants will use the translation function to translate the letter into the patients native tongue language. However we also need to keep the original copy in English for our records. it has a copy function but will not let them paste. It would be helpful if translating into another language we can keep the original letter without it translating everything

Can I make presets for every dictation? I want a letter back to referring clinic set (on best) for every session. I don't want to have to select that every time, and then change it from fast to best.

can we have option to dictate rather than having to type when we create new document from patient consultation