"HES Specialty" vocabulary - Two codes are merged (& and Null)

[adelmestri]

Describe the problem in content or desired feature In the "HES Specialty" vocabulary the following row should be split into 2, "Class" should be "Physicial Specialty" and "Domain" should be "Provider" (not "Observation"): 44777731 | & | null = Other maternity event | Observation | Non-standard | Valid | Observation | HES Specialty

As the link below shows: https://digital.nhs.uk/data-and-information/data-tools-and-services/data-services/hospital-episode-statistics/hospital-episode-statistics-data-dictionary in the "HES Specialty" lookup used for Hospital Episode Statistics specialties, the following codes/descriptions are reported, which have been merged together in the Athena vocabulary:

"Null = Other Maternity Event" "& = Not Known"

Moreover, from the link above, these codes represent "The specialty in which the consultant was working during the period of care. It can be compared with MAINSPEF, the specialty under which the consultant is contracted. Prior to 1 April 1996, this data item contained the code for the sub-specialty. From April 2004 a new list of treatment specialities was introduced, which describes the specialised service within which the patient was treated." This means that they are attributes of the providers associated to an observation (not the observation itself) and define their specialty, so the domain should be "Provider" and the class "Physician Specialty" for both

How to find it In Athena, select the "HES Specialty" vocabulary, click on "Code": the first row is the one that needs to be split into 2 and corrected

Expected adjustments The 1 row/concept reported above should be split into two as follows 44777731 | & | Not known |Physician Specialty| <Standard?> | Valid | Provider | HES Specialty

| null | Other maternity event | Physician Specialty| | Valid | Provider | HES Specialty **Screenshots**  **Additional context** Add any other context about the problem here. 44777731 & [null = Other maternity event](https://athena.ohdsi.org/search-terms/terms/44777731) Observation Non-standard Valid Observation HES Specialty

[cgreich]

@adelmestri:

Not sure I understand what you are trying to do. So, there is a funny code "&" which apparently stands for "other maternity event", which really is not a specialty and is not listed in the official NHS website. Why should this be split? What are the two codes? What are the two concept_names?

[adelmestri]

@cgreich: The two codes merged into 1 are the following:

"Null = Other Maternity Event" "& = Not Known"

as reported in the NHS lookups provided by NHS digital, which I shared above: HES APC TOS tab, MAINSPEF in row 166 ("The specialty under which the consultant is contracted.") and TRETSPEF in row 299 ("The specialty in which the consultant was working during the period of care.") fields, column H, look at the bottom. Unfortunately, there are discrepancy in the NHS documentations, but these values are present in the data and the spreadsheet is kept updated.

If you look at the screenshot I shared is looks that "&" means "Null = Other maternity event" which is wrong and makes no sense ("&" means "Not Known" and "Null" means "Other maternity event")

All the codes in the "HES Specialty" vocabulary should have "Class" = "Physicial Specialty" and "Domain" = "Provider" (not "Observation") because they are used as attributes for the consultants who provided the service associated to an event: the consultant is already an attribute of the event. I agree that the wording chosen by NHS sometimes is confusing, but this is how these "HES Specialty" vocabulary codes are used in the HES APC data.

[cgreich]

I am trying, @adelmestri. The website you linked to does not contain the word "specialty" anywhere. The only thing is a download of a spreadsheet HES+TOS+V2.01.xlsx, which also does not contain any specialty information. Where is this lookup? I only found the lookup I linked.

It's fine to split up that code. But we need to map them accordingly. Two questions:

1. What specialty does "Other Maternity Event" stand for? Obstetrics?
2. HES specialties are coded. What code does "Other Maternity Event" use? You left that out.

Happy to create a non-standard concept 44777731 | & | Not known |Physician Specialty| null | Valid | Provider | HES Specialty with no "Maps to" relationship, as flavors of Null have to be represented by the concept_id=0. In fact, all we have to do is to remove the "Other Maternity Event" from the name and fix the Concept Class.

[adelmestri]

@cgreich: If you open the Excel spreadsheet you mentioned, go to the "HES APC TOS" tab (there are many tabs), and search for "MAINSPEF" which is in row 166 ("The specialty under which the consultant is contracted.") and for "TRETSPEF", which is in row 299 ("The specialty in which the consultant was working during the period of care."): the look ups are in column H for both. Look at the bottom of the lists to find the 2 lookups I mentioned. To answer your questions:

1. What specialty does "Other Maternity Event" stand for? Obstetrics? AD: Yes, perfect
2. HES specialties are coded. What code does "Other Maternity Event" use? You left that out. AD: the lookup report "Null" as code. If you are unfappy to use that, just remove the "Other Maternity Event" from the current name and fix the Concept Class and the Concept Domain please. Thank you for your help

[adelmestri]

@cgreich: there are other concepts in "HES Specialty" vocabulary which are not in the correct Class/Domain: would you be OK if I send you a list of those?

[cgreich]

Ok. Looks like this is a bigger construction site.

First, there are missing codes. I assume you want to add them?

108 | Spinal Surgery 109 | Bariatric Surgery 111 | Orthopaedic 113 | Endocrine Surgery 115 | Trauma Surgery 223 | Paediatric Epilepsy 230 | Paediatric Clinical Pharmacology 240 | Paediatric Palliative Medicine 250 | Paediatric Hepatology 263 | Paediatric Diabetes 264 | Paediatric Cystic Fibrosis 270 | Paediatric Emergency Medicine 324 | Anticoagulant 327 | Cardiac Rehabilitation 328 | Stroke Medicine 329 | Transient Ischaemic Attack 331 | Congenital Heart Disease 333 | Rare Disease 335 | Inherited Metabolic Medicine 342 | Pulmonary Rehabilitation 343 | Adult Cystic Fibrosis 344 | Complex Specialised Rehabilitation 345 | Specialist Rehabilitation 346 | Local Specialist Rehabilitation 347 | Sleep Medicine 348 | Post-COVID-19 Syndrome 431 | Orthogeriatric Medicine 461 | Ophthalmic and Vision Science 505 | Fetal Medicine 657 | Prosthetics 658 | Orthotics 659 | Dramatherapy 660 | Art Therapy 661 | Music Therapy 662 | Optometry 663 | Podiatric Surgery 670 | Urological Physiology 673 | Vascular Physiology 675 | Cardiac Physiology 677 | Gastrointestinal Physiology 725 | Mental Health Recovery and Rehabilitation 726 | Mental Health Dual Diagnosis 727 | Dementia Assessment 730 | Neuropsychiatry 920 | Diabetic Education Null | Other Maternity Event

Second, you want to review the domains. Right now, we have the following non-specialty domain records:

190 | Not a treatment function | Observation 319 | Respite care | Visit 424 | Well babies (care given by the mother / substitute, with nursing advice if needed) | Visit 500 | Not a treatment function | Observation 510 | Antenatal clinic | Visit 520 | Postnatal clinic | Visit 600 | Not treatment functions | Observation 620 | Other than maternity | Observation 902 | Community health services dental | Visit 950 | Nursing episode | Visit 990 | Joint consultant clinic | Visit & | null = Other maternity event | Observation

You are saying they in fact are all specialty of a provider, rather than a visit/observation?

[adelmestri]

Exactly @cgreich: I did not want to overload you with requests, but you are spot on. Thanks!

[cgreich]

And what exactly is:

324 | Anticoagulant 329 | Transient Ischaemic Attack 331 | Congenital Heart Disease 333 | Rare Disease 343 | Adult Cystic Fibrosis 348 | Post-COVID-19 Syndrome

These are specialties? Or conditions of the patients?

[cgreich]

And what exactly is:

324 | Anticoagulant 328 | Stroke Medicine 329 | Transient Ischaemic Attack 331 | Congenital Heart Disease 333 | Rare Disease 343 | Adult Cystic Fibrosis 348 | Post-COVID-19 Syndrome

These are specialties? Or conditions of the patients?

[adelmestri]

@cgreich: they are all Specialties of the doctors who treated the patients:

In "TRETSPEF" (The specialty in which the consultant was working during the period of care.) your list is reported as:

324 = Anticoagulant Service (From 1 April 2004) 328 = Stroke Medicine Service (From 1 April 2004) 329 = Transient Ischaemic Attack Service (From 1 April 2004) 331 = Congenital Heart Disease Service (From April 2013) 333 = Rare Disease Service (From 1 April 2021) 343 = Adult Cystic Fibrosis Service (From 1 April 2004) 348 = Post-COVID-19 Syndrome Service (From 1 April 2021)

meaning that the consultant has the specialty provided by the service

[cgreich]

I see. We would have to link them up to the hierarchy, so they become part of more general specialties, since in the rest of the world there is no such thing as "anticoagulant service". I guess that would be under Internal Medicine.

But again, there are more missing from the two tables combined. And you are saying that these things are specialties. Need to sort that out.

[cgreich]

Exactly @cgreich: I did not want to overload you with requests, but you are spot on.

You are too kind. :-)

But seriously. These visits and observations are not visits and observations, but actually badly described specialties of providers? So, "Postnatal clinic" is not a visit to a postnatal clinic, but a provider who cares for newborn babies?

[adelmestri]

@cgreich: exactly! These fields are attribute of episodes, for which lots of other information are provided, including a id for a consultant. These lookups describe the consultants specialties. I guess it is because visit costs are based on the consultant specialty, and these are administrative data.

[TinyRickC137]

Great conversation going on here :)

Let's summarise what needs to be done 1) Add missing codes 2) Review and change domains (list required) 3) Split codes: '&' and 'NULL' (written in varchar)

Did I miss something?

@adelmestri would you mind helping us by submitting community contributions with these changes?

[adelmestri]

Thanks @TinyRickC137. I discuss this quite a bit with @cgreich and sent him my contributions as there were so many to put them here. I have no problems in cut&copy them. @cgreich: would you be happy with that?

[cgreich]

Like this:

New Speciality	relationship_id	Existing Specialty
Anticoagulation	Is a	Cardiovascular Disease
Stroke Medicine	Is a	Cardiovascular Disease
Transient Ischemic Attack	Is a	Cardiovascular Disease
Congenital Heart Disease	Maps to	Adult Congenital Heart Disease
Rare Disease	Is a	Internal Medicine
Adult Cystic Fibrosis	Is a	Pulmonary Disease
Post-COVID-19 Syndrome	Is a	Internal Medicine

?

[adelmestri]

@cgreich : Not sure if your comment was for me or for @TinyRickC137 Anyway, if it was for me, I am not in a position of working on this directly, sorry

[aostropolets]

I discuss this quite a bit with @cgreich and sent him my contributions as there were so many to put them here. I have no problems in cut&copy them. @cgreich: would you be happy with that?

@adelmestri, if you would like to contribute, there is a formal process that you can follow. Here is the link to the guidelines: https://github.com/OHDSI/Vocabulary-v5.0/wiki/Community-contribution:-non%E2%80%90drug-vocabularies-p.I.

In general, issues placed in the backlog and prioritized against other maintenance needs. HES Specialty vocabulary is not on the roadmap and we do not have an estimate on when this request will be fulfilled. Submitting a formal contribution would enable faster incorporation of the proposed changes.