I am a primary care clinician (GP, Urgent Care and previous CMO of Dept of Corrections) cautiously optimistic about the opportunity before Te Whatu Ora of digital health’s role (in particular interoperability) in addressing equity, improving access and outcomes. I attended the hui, and have reviewed and endorsed the ‘Simplify to Unify’ approach.
The HISO 10083:2020 Interoperability Roadmap had a July 2022 target to replace READ codes with SNOMED CT in primary care. (Unfortunately the link to the progress report on Te Whatu Ora’s webpage links to the old document).
SNOMED CT implementation had barriers in other countries too (e.g. Denmark, Ireland, etc.). I’m curious to hear your insights on the barriers primary care in New Zealand is experienced in replacing READ codes with SNOMED CT. Has ACC been the main hurdle? Has PHOs been approached and involved? Has the Royal New Zealand College of General Practitioners been involved? Has Te Whatu Ora prioritised resourcing change in primary care to adopt the tool to enhance interoperability?
At a practice level, I would hope that clinicians would never see READ or SNOMED-CT codes. I know it’s common, but I see it as a failure of IT and user interface design.
I know we have a huge installed base of systems that work with READ codes, ad Primary Care data that’s coded in READ, and that’s going to be a barrier for practices to migrate to SNOMED-CT.
In the informatics area, things are a little different. Autocoding between READ and SNOMED-CT may be a lot more acceptable as the higher error rate may be manageable.
Autocoding has improved a lot in the last few years (I’m told) and I expect it to continue to advance rapidly. Costs are also falling dramatically, although AWS charge almost 10X for SNOMED-CT coding as they do for READ coding.
Unfortunately, as Read Codes are not surrogate identifiers, many GPs have become accustomed to searching using the codes themselves. Furthermore, because the GP PMS implement both velocity coding (most frequently-used codes appear at the top of the pick lists) and allow user-entered synonyms much of the data coded in Read is useless beyond the description and cannot safely be mapped to SNOMED CT. Case in point, the largest survey performed on Read Coding in NZ Primary Care showed the most commonly-used Code is the one for ‘Miscellaneous’. Unfortunately, no end of dictates from various public sector organizations have shifted the dial on SNOMED CT adoption in NZ primary care and, IMO, nor will they until it’s linked to reimbursement.
Hi @WernerP, I can comment as a Patient Administration System administrator who has also implemented an ACC eClaiming app which uses codified diagnoses and a flexible dynamic search function.
Our PASes can sometimes accommodate multiple coding systems, such that a diagnosis can inherently be recorded as both a SNOMED and READ (or ICD10) code if necessary. Alternatively, standard mapping tables are used behind the scenes to transform what is recorded by a clinician into a different codeset for reporting, data collections, etc.
Good software would allow the user to search by descriptor keyword, partial keyword, or code, so if you’re having to type in actual entity identifiers because of poor search capabilities in the software, that’s pretty unfortunate.
As of next year, I believe Te Whatu Ora local areas will be introducing SNOMED coding to our emergency departments to record initial presenting complaints (and provisional diagnoses?). Given the vast number of SNOMED Disorders, I’ll be keen to see how this might be implemented in our ancient PASes without making the user experience terrible. I suspect we will need to arbitrarily limit the list of selecable Disorders to an agreed short list of a few hundred of the most commonly used.
Hi Peter, not quite true. Medtech uses 1:1 mapping of Read to Snomed, tho I agree the initial coding in Read not a great idea and relly need o go to native Snomed.
The search is actually quite good IMO, hsaving played wiht a lot of Snomed browsers.
In terms of the comment below the answer is refsets… Must be an ED refset, certainly there is a GP one.
Hi Richard, That survey (undertaken by HealthOne) was several years ago, so it’s good to hear that things have improved since then and that Medtech has implemented a good SNOMED Browser (I have also used a few of those over the years!). Reference Sets are certainly one answer to providing a good data entry UI, although there is a tension between the dictionary vs the phrase book approaches. There is certainly an excellent ED Refset in the NZ Edition, but I think that the GP one in the International Edition has been deprecated as it’s very challenging to produce a satisfactory subset for general practice. Regards, Peter
Sadly MT is still a Read browser… but 1:1 mapping in the tables.
Sad to gear the GP refset been deprecated, I was on the working group for that. Yes, not easy to do but I thought our process was pretty robust. I’d still advise using it in a GP setting in NZ. We collated multiple primary care data sets from the participating countries, concerted to Snomed if need (Read, ICD…) We did a frequency count and looked at synonyms 2 years work headed by University of Sydney with monthly meetings online.
I think we got down to about 4000 terms, with separate ‘reason for visit’ and diagnostic codes.
If Snomed has deprecated and no replacement then wth is going on with that organisation? Does primary care not count?
Not sure. Looking at the SNOMED International Release Site, the last version of the GP/FP Reference Set was published in July 2022. @alastairk will probably know the thinking behind this, including any alternatives (maybe the IPS/GPS RefSets?). Looking at that release, there are 4,507 members in the snapshot version.
The goal of getting all GP systems to SNOMED in place of Read codes remains, but we will need a new date
Since 2017 ACC has been able to receive SNOMED-coded claims (MSD from 2018) and this removed one barrier to SNOMED adoption by GP systems. ACC maintains a map from SNOMED to Read to make this possible with its claims systems. We continue to work with ACC on the move to SNOMED
We’re also currently working with PHOs and GPNZ on the Read to SNOMED shift. Clinicians and consumers are not likely to march in the streets in the cause of R2S, but to be a modern data-driven health system we need this change. We provide tools and materials such as the SNOMED NZ Edition, NZ Health Terminology Service and forthcoming NZ Patient Summary to make it as easy as possible for users and suppliers to upgrade to the SNOMED future, and put effort into the change management.
The GP/FP refset lives on with an annual release that I expect we’ll see in November. We’ll make it available via NZHTS. Fair to say opinions are divided on how much use to make of refsets (EDs and CanShare do extensively), but the important thing is you’re using SNOMED
Hi @hamish.luebbers. Thanks for explaining the backend. I’m interested to know if the ACC eClaiming app is used in primary care (or available to use).
As long as PASes supports multiple coding systems (and maintaining READ) the enablers to change will be weak.
Also, does ACC use a conversion app? My understanding is they are converting SNOMED CT codes back to READ? How long did it take to decommission fax machines in health services? It can be done. A primary care-specific approach is required and I don’t think the mahi should be lumped in with ED or hospital adoption of SNOMED CT.
IMO an agreed/shared targeted approach with the important stakeholders (GPNZ, RNZCGP, PHOs, ACC, Te Aka Whai Ora, Te Whatu Ora, Manatu Hauora, etc.) in primary care is required by doing a ‘barriers analysis’ (SWOT-like), a ‘work-done’ or progress analysis followed by a solutions plan. Key to the solutions plan should be incorporating change management principles. @alastairk is there a working group in place on this? What do you think is required to determine a new date?
ACC does not seem to be a “champion” of change from SNOMED CT to READ. Is there a missed opportunity here? I would be interested to know which barriers ACC have and how this might be overcome.
Most of our hospital EDs are now recording and reporting chief presenting complaint using SNOMED CT (a defined set of about 150 concepts) and many are doing the same for diagnosis (about 1500 concepts). So most ED visit data in NNPAC is SNOMED coded now.
I agree we need coordination and HISO is taking up this role, working with all the above parties (as well as NZ Defence Force and Corrections) as the R2S migration effort gets underway again. We’re creating a plan and will post regular updates on the TWO website. It would be great to tap your expertise Werner given your interest
The goal of getting all GP systems to SNOMED in place of Read codes remains, but we will need a new date
Thanks for those updates Alistair. Thse 1:1 Read to Snomed maps that NHS did all those years ago still useful! But agree native codig and reitre Read shd have happended years ago…
I’m afraid this specific ACC eClaiming app was built by a 3rd party vendor for the iPM PAS, and would need to be further developed/forked to accommodate other PAS/PMS integration.
However since we invested in this software, I believe ACC NZ have developed/are developing some APIs for querying and lodging claims online. This should make it easier for other software vendors to develop built-in ACC claiming within web based PMSes/PASes.
I’d like to think ‘Online ACC claiming’ would be an attractive future module on the product road map for the major primary health PMS vendors.
Our claim submission API accepts SNOMED trauma codes, but we translate them to READ codes internally for our automation and reporting. This translation loses granularity and we would like to transition to using SNOMED more generally, but the change is happening slowly. We will use FHIR and SNOMED for new APIs in the future.
Thanks for this reassurance of HISO’s role, Alastair. Happy to help as I’m getting my head around the different agencies’ positions, priorities and progress. Feel free to reach out.