Computerized system to provide medical diagnosis, prognosis, and treatment using more refined digital health records having improved context

1 . A method of improving medical diagnosis, prognosis, or treatment, comprising:

accessing a digital health record, wherein the digital health record comprises windows corresponding to different categories of information;

identifying, from the digital health record, an entry in which a replacement string or an additional string is to be suggested, based on an ambiguity of the entry or an inconsistency or a conflict with other data in the digital health record;

obtaining a first training set that comprises first suggestions that are accepted as being viable replacements or additions to the existing text;

obtaining a second training set that comprises second suggestions that are rejected as being nonviable replacements or additions to the existing text;

training one or more machine learning models using the first training set and the second training set;

predicting, by the one or more machine learning models, a particular string to suggest as a replacement or addition to existing text in the entry and one or more alternatives to the particular string based on respective probabilities of veracity of the particular string and the one or more alternatives;

in response to suggesting the particular string, presenting sources of data within the digital health record that support the particular string and the one or more alternatives to the particular string, wherein the sources of data correspond to one or more of the windows;

receiving an input, the input indicating whether the particular string or the one or more alternatives are accepted, or including a manually inputted string that is distinct from the particular string and the one or more alternatives;

updating the entry based on the received input;

determining or inferring any particular emphasis or any particular bias of the input towards a particular window of the digital health record over other windows of the digital health record;

training the one or more machine learning models according to the any particular emphasis or any particular bias;

determining a presence of a particular medical condition based on semantic segmentation and instance segmentation to determine first boundaries between different tissue types and second boundaries between common tissue types;

determining, based on the presence of the particular medical condition, a surgery to be implemented, the one or more machine learning models being trained according to particular medical conditions and particular measures to be implemented;

presenting, as an overlay over an existing window or a sidebar at a side of the existing window, a suggestion of the surgery;

receiving an acceptance of the suggestion of the surgery;

presenting a prompt to select a surgical machine associated with the surgery;

receiving an acceptance of the surgical machine;

following execution of a protocol by the surgical machine, receiving, from the surgical machine, a result associated with the execution; and

updating a portion of the digital health record according to the result.

2 . The method of claim 1 , wherein the determining a particular string to suggest comprises determining a particular string indicating a COVID-19 status of a patient, the COVID-19 status being determined based on whether a patient has breathing or speech difficulties, a respiratory rate, a PaO 2 and SpO 2 level, and an analysis of a chest X-ray of the patient.

3 . The method of claim 1 , further comprising:

receiving a textual input into the entry; and wherein the determining the particular string to suggest is simultaneous with the receiving of the textual input.

4 . The method of claim 1 , wherein the determining the particular string is based on a context of the entry.

5 . The method of claim 1 , wherein the digital health record comprises a problem list including one or more illnesses, one or more injuries, and one or more risk factors and a medication list including one or more drugs and one or more prescribed respective dosages; and

the determining the particular string is based on an item in the problem list and an item in the medication list.

6 . The method of claim 1 , further comprising:

identifying other entries in the digital health record to be updated based on the input.

7 . The method of claim 1 , wherein the presenting one or more alternatives to the particular string is in a drop-down list that enables switching between selection of the particular string or the one or more alternatives.

8 . The method of claim 1 , further comprising presenting a confidence level indicating a probability of relevance of the particular string.

9 . The method of claim 8 , further comprising presenting confidence levels indicating one or more probabilities of relevance of each of the one or more alternatives.

10 . The method of claim 1 , wherein the determining a particular string comprises importing a calculation of a parameter.

11 . The method of claim 10 , wherein:

the particular string comprises a dot-phrase or a native macro; and

the input indicates that the particular string is accepted; and the method further comprises:

populating a chart based on the input.

12 . The method of claim 1 , wherein, in response to receiving the input to override the particular string and the one or more alternatives, determining that the entry is to be maintained.

13 . The method of claim 1 , wherein:

the identifying an entry in which a replacement string or an additional string is to be suggested comprises identifying a numeral that is missing a subsequent unit of measurement.

14 . The method of claim 13 , wherein:

the determining a particular string to suggest comprises determining a particular unit of measurement to append to the numeral based on the context of the entry.

15 . The method of claim 1 , further comprising identifying a type of the received input; and

wherein the determining a particular string to suggest is based on the one or more machine learning models being trained by a feedback mechanism taking into account the type of the received input.

16 . The method of claim 12 , further comprising:

identifying a frequency at which the received input or the one or more alternatives is selected; and

wherein the determining a particular string to suggest is based on the one or more machine learning models being trained by a feedback mechanism taking into account the type of the received input.

17 . The method of claim 1 , further comprising:

identifying data in the entry having a reliability below a threshold; and

flagging the data.

18 . The method of claim 17 , wherein the identifying data in the entry having a reliability below a threshold comprises detecting data having a source other than a sensor reading or a calculation of a parameter.

19 . The method of claim 1 , further comprising presenting one or more respective sources of the one or more alternatives to the particular string.

20 . The method of claim 1 , wherein the identifying, from the digital health record, an entry comprises:

identifying a plurality of entries of which at least one of the entries conflicts or contradicts other entries; and

determining which of the entries is more likely to be inaccurate compared to another of the entries.