US-20260128137-A1 - Automated Intelligent Actionable Edit System Using Machine Learning

US20260128137A1US 20260128137 A1US20260128137 A1US 20260128137A1US-20260128137-A1

Abstract

Techniques for recommending actions for resolution of errors associated with medical coding records are described. In an example, a plurality of edits corresponding to a medical coding record is identified, where the plurality of edits is indicative of resolutions to at least one error associated with the medical coding record. The plurality of edits is then analysed using a code editing model to rank the plurality of edits, where the code editing model is trained based on the statistical information indicative of relevance of the plurality of edits for resolution of errors associated with the medical coding record. The ranked plurality of edits is then provided along with a set of recommendation actions for resolution of the error associated with the medical coding record.

Inventors

Jeramie P. Naef

Assignees

SOLVENTUM INTELLECTUAL PROPERTIES COMPANY

Dates

Publication Date: 20260507
Application Date: 20251105

Claims (20)

1 . A system comprising: at least one processor; a computer-readable medium comprising instructions that, when executed by the at least one processor, causes the system to: identify a plurality of edits corresponding to a medical coding record, the plurality of edits being indicative of resolutions to at least one error associated with the medical coding record; analyse the plurality of edits using a code editing model to rank the plurality of edits, the code editing model being trained based on the statistical information indicative of relevance of the plurality of edits for resolution of errors associated with the medical coding record; and provide the ranked plurality of edits along with a set of recommended actions for resolution of the error associated with the medical coding record.
2 . The system of claim 1 , wherein to rank the plurality of edits, the at least one processor causes the system to: compute a confidence score for each of the plurality of edits; and rank each of the plurality of edits based on the confidence score.
3 . The system of claim 2 , wherein the statistical information comprises contextual information associated with the plurality of edits, the contextual information including at least one of: a category of the plurality of edits, a severity level of the plurality of edits, an experience level of a coder who created the plurality of edits, a state of resolution indicating whether the plurality of edits led to resolution of the at least one error associated with medical coding record, and a count of utilization of the plurality of edits for resolution of the at least one error associated with medical coding record.
4 . The system of claim 3 , wherein the at least one processor further causes the system to: determine whether the statistical information includes the contextual information corresponding to at least one edit from the plurality of edits; and generate the contextual information for the at least one edit based on determining that the statistical information does not include the contextual information corresponding to the at least one edit.
5 . The system of claim 4 , wherein the at least one processor further causes the system to analyse text of the at least one edit to identify a correlation between the at least one edit and the error associated with the medical coding record.
6 . The system of claim 5 , wherein to analyse the text of the at least edit, the processor causes the system to: parse the text into one or more segments that correspond to at least one other edit indicative of a resolution to the at least one error associated with the medical coding record; and generate contextual information corresponding to at least one edit based on past instances of utilization of the at least one other edit for the resolution to the at least one error.
7 . The system as claimed in claim 4 , wherein to generate the contextual information for the at least one edit, the processor causes the system to: generate an edit identifier for each edit, wherein the edit identifier uniquely identifies the edit, and the edit identifier is based at least in part on one or more parameters included in text of the edit, wherein the one or more parameters comprise at least a numerical parameter and a classification title, and wherein the classification title specifies an edit type, and the numerical value is unique for edits of each edit type.
8 . The system of claim 4 , wherein the at least one processor further causes the system to determine whether the statistical information includes the contextual information corresponding to at least one edit based on presence of an edit identifier associated with the at least one edit.
9 . The system of claim 8 , wherein the at least one processor further causes the system to: determine whether the edit identifier associated with the at least one edit matches an edit identifier associated with a first edit in the statistical information; and compare text of the at least one edit with text of the first edit on determining that the edit identifier associated with the at least one edit matches the edit identifier associated with the first edit.
10 . The system of claim 1 , wherein the at least one processor further causes the system to: receive a user selection of at least one recommended action from the set of recommended actions; update the statistical information to include the user selection of the at least one recommended action for resolution of the error associated with the medical coding record; and train the code editing model based on the updated statistical information.
11 . A computer implemented method comprising: identifying, by a medical code editing system, a plurality of edits corresponding to a medical coding record, the plurality of edits being indicative of resolutions to at least one error associated with the medical coding record; analysing, by the medical code editing system, the plurality of edits using a code editing model to rank the plurality of edits, the code editing model being trained based on the statistical information indicative of relevance of the plurality of edits for resolution of errors associated with the medical coding record; providing, by the medical code editing system, the ranked plurality of edits along with a set of recommended actions for resolution of the error associated with the medical coding record.
12 . The method of claim 11 , wherein ranking the plurality of edits comprises: computing a confidence score for each of the plurality of edits; and ranking each of the plurality of edits based on the confidence score.
13 . The method of claim 11 , wherein the statistical information comprises contextual information associated with the plurality of edits, the contextual information including at least one of: a category of the plurality of edits, a severity level of the plurality of edits, an experience level of a coder who created the plurality of edits, a state of resolution indicating whether the plurality of edits led to resolution of the at least one error associated with medical coding record, and a count of utilization of the plurality of edits for resolution of the at least one error associated with medical coding record.
14 . The method of claim 13 , further comprising: determining whether the statistical information includes the contextual information corresponding to at least one edit from the plurality of edits; and generating the contextual information for the at least one edit based on determining that the statistical information does not include the contextual information corresponding to the at least one edit.
15 . The method of claim 13 , wherein generating the contextual information comprises analysing text of the at least one edit to identify a correlation between the at least one edit and the error associated with the medical coding record, wherein the text of the at least one edit is analysed based on natural language processing (NLP).
16 . The method of claim 15 , wherein analysing the text of the at least edit comprises: parsing the text into one or more segments that correspond to at least one other edit indicative of a resolution to the at least one error associated with the medical coding record; and generating contextual information corresponding to at least one edit based on past instances of utilization of the at least one other edit for the resolution to the at least one error.
17 . The method of claim 14 , wherein generating the contextual information for the at least one edit comprises: generating an edit identifier for each edit, wherein the edit identifier uniquely identifies the edit, and the edit identifier is based at least in part on one or more parameters included in text of the edit, wherein the one or more parameters comprise at least a numerical parameter and a classification title, and wherein the classification title specifies an edit type, and the numerical value is unique for edits of each edit type.
18 . The method of claim 14 , further comprising: determining whether the edit identifier associated with the at least one edit matches an edit identifier associated with a first edit from the plurality of edits in the statistical information; and comparing text of the at least one edit with text of the first edit on determining that the edit identifier associated with the at least one edit matches the edit identifier associated with the first edit.
19 . The method as claimed in claim 11 , further comprising: receiving a user selection of at least one recommended action from the set of recommended actions; updating the statistical information to include the user selection of the at least one recommended action for resolution of the error associated with the medical coding record; and training the code editing model based on the updated statistical information.
20 . A non-transitory computer readable medium comprising computer-readable instructions that when executed cause a processing resource of a computing device to: identify a plurality of edits corresponding to a medical coding record, the plurality of edits being indicative of resolutions to at least one error associated with the medical coding record; analyse the plurality of edits using a code editing model to evaluate the plurality of edits, the code editing model being trained based on the statistical information indicative of relevance of the plurality of edits for resolution of errors associated with the medical coding record, wherein to evaluate the plurality of edits, the instructions cause the processing resource to: compute a confidence score for each of the plurality of edits; and rank each of the plurality of edits based on the confidence score; and provide the ranked plurality of edits along with a set of recommended actions for resolution of the error associated with the medical coding record.

Description

BACKGROUND Healthcare practitioners provide numerous services depending on the patient's requirements. To keep track of the services provided to a patient, a medical chart is prepared for the patient, where the medical chart tracks information of every medical diagnosis and treatment performed on the patient. Edits to the medical chart can be made by specialists to address issues in the medical chart. BRIEF DESCRIPTION OF DRAWINGS The following detailed description references the drawings, wherein: FIG. 1 illustrates a computing system for recommending actions for resolution of errors associated with medical coding records, in accordance with an example of the present subject matter, FIG. 2 illustrates the computing system for recommending actions for resolution of errors associated with medical coding records, in accordance with another example of the present subject matter, FIG. 3 illustrates a method for recommending actions for resolution of errors associated with medical coding records, in accordance with an example of the present subject matter, and FIG. 4 illustrates a non-transitory computer-readable medium for recommending actions for resolution of errors associated with medical coding records, in accordance with an example of the present subject matter. Throughout the drawings, identical reference numbers designate similar, but not necessarily identical, elements. The drawings provide examples and/or implementations consistent with the description; however, the description is not limited to the examples and/or implementations provided in the drawings. DETAILED DESCRIPTION Medical coding records may need to be modified for various reasons. In operation, the medical coding records are usually identified manually by referring to a list that includes various medical coding records along with their respective descriptions. However, manual identification and inclusion of the medical coding records is prone to errors. If ignored, such errors can prove costly for the providers and my ultimately impact the quality of medical care. For instance, the inclusion of an incorrect medical coding record in an invoice may lead to rejection of insurance claims for reimbursement by an insurance company. As another example, the misidentification of a medical code may prevent systems specifically configured to address preventable complications from presenting relevant information to a medical professional, leading to increased risk of complications or other adverse medical outcomes. To alleviate the above-mentioned problems, a practice of editing the medical coding records is adopted, where potential errors associated with a medical coding record are identified and various edits that could be made to the medical coding records are presented to the users. However, in known claim editing processes, a user is usually presented with numerous edits, such that, the user may have to sift through numerous edits for identification of a relevant edit. Sifting through such numerous edits causes edit fatigue as it becomes difficult to know which edits are important to the user. Accordingly, the user tends to get confused as to which edit should be addressed first. Further, the user may also at times choose to ignore the presented edits as finding the solution to the edit can be tedious, challenging and time consuming. All of these factors contribute to errors in conventional editing techniques. According to examples of the present subject matter, techniques for recommending actions for resolution of errors associated with medical coding records are described. In operation, a plurality of edits corresponding to a medical coding record are identified, where the plurality of edits is indicative of resolutions to at least one error associated with the medical coding record. The plurality of edits is then analysed using a code editing model to evaluate the plurality of edits, where the code editing model is trained based on the statistical information indicative of relevance of the plurality of edits for resolution of errors associated with the medical coding record. In an example, the evaluation of the plurality of edits involves computing a confidence score for each of the plurality of edits and ranking each of the plurality of edits based on the confidence score. The ranked plurality of edits is then provided along with a set of recommendation actions for resolution of the error associated with the medical coding record. Ranking the plurality of edits based on the statistical information and providing the ranked plurality of edits along with the set of recommended actions facilitates a user in initiating an appropriate action for resolution of errors associated with medical coding records, thereby improving the efficiency involved in resolution of the errors associated with the medical coding records. The above aspects are further described in conjunction with the figures, and in the associated description below. It should be noted