Release Notes | IMO Developer Portal

Date	release summary
10/2/2023	Updated MSD to create mappings between CVX, RxNorm and NDC: Provides more robust results by returning CVX codes alongside or in addition to RxNorm and NDC codes. These “crosswalked” mappings will appear when customers search by description within the Medication domain, and also when searching by any of the supported code sets (RxNorm, NDC, or CVX) Updated Synonyms/labels and CIRS within the lab matching logic: Provides more robust content to improve matching quality when customers send lab inputs that contain synonyms of LOINC descriptions. Additionally, this update provides customers with our most up to date clinical inductive rules (CIR) created by our clinical team, and these rules help produce more accurate matches based on the LOINC parts identified within the input term. Updated Lab matching logic to utilize the latest version of LOINC (2.76): Allows for customers to receive match results based on the latest version of LOINC content available.
9/4/2023	Precision Normalize now supports ICD-O-3 within the Problem domain. With this new feature, customers can normalize input terms along with IMO, SNOMED or ICD codes within the Problem domain and receive ICD-O-3 attributes/codes for applicable IMO results. By incorporating ICD-O-3 code sets, IMO can enhance its capabilities for oncology-related clinical documentation, coding, and data analysis, leading to improved patient care, research, and interoperability. 1. These codes represent a dual classification system designed to capture a topological code and a histological code for each neoplasm. We have added new attributes in the search and detail payloads for Topography, Morphology, Histological Grading and Immunophenotype Designation. Precision Normalize has added a new flag within the Problem domain to clearly distinguish IMO lexicals that are eligible for further refinement using IMO’s modifier workflow within Core. 1. This new flag is a boolean variable indicating whether the lexical can be further refined using IMO's ICD-10-CM calculator/modifier workflow data. When true, it indicates that a term with more ICD-10-CM relevant specificity can be retrieved by submitting the lexical to the IM.
8/21/2023	Normalize has added CVX Notes, MVX codes, and MVX status to the metadata when CVX codes are returned. Customers will be able to receive additional metadata in their response for CVX Notes, MVX codes, and MVX status. This will allow for easier reporting and transmission of CVX related information.
7/24/2023	Continued Improvements to Medication Parser and fuzzy matching Improves scoring when frequencies (Ex. QID, Twice a day, 2xday, etc.) are present within the input term. Improves handling of parentheses within the medication parser Improves handling and scoring of the abbreviation of OR when it applies to route. Resolved issue with missing Map_Context in the ICD-10-CM code mappings Provides better clarity on why ICD-10-CM codes were not returned in the response. Customers will see either "NON-CODABLE" or "INSUFFICIENT INFORMATION" in the response if no ICD-10-CM codes were returned in the Problem domain. Resolved issue with Discrepancy Check not functioning properly for input_code_first preference Provides correct detailed information to the customer regarding discrepancy checks performed when using the input preference of input_code_first.
6/26/2023	Added exact matching logic for LOINC long and short common names. This provides better matching logic within the Lab domain when customers send us LOINC common names as input Medication Parser matching improvements to include: Better handling of the word 'and' when it separates ingredients Better handling of the special character () Overall value: Customer facing value to help produce more quality matches when input terms contain asterisks or when ingredients are separated by the word 'and.' Removed duplicate precise ingredients.* Improves scores for matches and no longer indexes precise ingredients as 2 ingredients instead of 1. Content has been updated/refreshed for the following areas: LOINC and other Lab compendiums (ARUP, Labcorp, Mayo) Enhanced lab ontology synonyms Allows for customer input terms to match against the latest content.
5/29/2023	Normalize engine now supports SNOMED-CT-US terminology. Customers can now normalize both input terms and input codes. Input codes should be normalized using the code set SNOMED-CT as the SNOMED-CT-US extension is tied to the international version. Medication matching has been improved to provide more accurate scoring when the input string doesn't contain ingredients and normalize returns the correct match that now includes ingredient information. In the example below, previous scored low, but now produces a higher score: Example input term: NORVASC 5 MG 1 TAB PO QD (refill) Matched term: amlodipine 5 mg oral tablet (norvasc)
5/15/2023	Normalize engine now includes full support for ICD-10-DSM terminology. This includes: Licensed customers can search by input term and normalize will return the proper lexical with mappings to ICD-10-CM and DSM when both are available. DSM lookup by Code will not return DSM content in the metadata, and will only return the ICD-10-CM code. Additional code mappings to ICD-9 and SNOMED will also be provided The normalized output for ICD-10-PCS codes has been updated to accurately reflect code group relationships so that customers can easily navigate which and how many codes need to be selected for proper coding.
5/1/2023	Normalize will now limit the total number of NDC codes displayed within the output. If the matched lexical contains up to 25 NDC>RxNorm code mappings, all 25 NDC codes will display in the output. This covers approximately 82% of all RxNorm codes. If there are greater than 25 NDC code mappings, the output will display additional map context of "Maps to more than 25 NDC product codes."
4/4/2023	Tie breaking logic and rules have been added to provide a more consistent matching experience when a given term previously had multiple potential matches.
3/20/2023	Enabled Fuzzy title matching for all RxNorm synonym titles. This increases opportunities for improved string matching, and customers will likely experience an increase in True Positive match results, along with a decrease in False Positive results.
3/6/2023	Medication improvements to include: In an effort to create more consistent match results, we have created logic to query titles when they are tied to more than one RxCUI, and then prioritize which is returned. Indexes have been updated to expose more titles and allow for them to be searchable. This provides expanded matching capabilities and allows for customers to receive better match results.
2/6/2023	Normalize has implemented two separate features to help address False Positive results caused by some RxNorm synonyms. 1. Normalize no longer returns non-drug ingredients as a match Example: If the input term was hot soapy water, we previously were returned water as the match scored at 1.0. Water is not a valid ingredient to return, as it will never be administered as a drug. We do still keep those ingredients for multi-ingredient drugs and try to match based on that scenario. 2. Normalize no longer uses wrong RxNorm synonym data. RxNorm has some bad synonyms on their end, or synonyms that are mapped to multiple RxCUIs. Because of this, bad matches result when we use the synonyms as RxNorm has them listed. Therefore, Normalize will essentially ignore that synonym translation when attempting to match.
2/1/2023	The lab normalize project has created a list clinical inductive rules (CIR) for most of the tests found in a Comprehensive Metabilic Panel (CMP), Complete Blood Count with Differential (CBC DIFF), a subset of COVID tests, and a few additional chemistry tests. The CIR were inserted into our lab normalize pipeline, situated primarily between the LOINC parts identifier and the LOINC code resolver with a few overriding features where needed. When the rules were applied to a known data set, it was noted that for select tests, the process was able to resolve to a single LOINC code.
1/23/2023	Global transformations are now in production for the Medication domain. These transformations will allow for Normalize to better interpret abbreviated drug and chemical names that are present within input terms and improve matching results. Remediated the following issues: Missing ICD10 mappings within Problem domain Initial icident resolved on 1/5/23, but additional error handling was added after issue was resolved. Link to executive summary File API was returning duplicate records in Medication Domain This was caused due to 6 CVX records that have an identical RxNorm code mapped to the same lexical and 2 CVX codes mapped ot the same lexical. Coding has been updated to handle these scenarios appropriately.