19/11/2018


Data leads:

Discussion/Action

#
Discussion/Action
Responsible
Outcome
Status
1Resolve Duplicates - TISDEV-4841Resolve the duplicate invalid GMC numbers, such as 'Retired,' and make them 'UNKNOWN'Data leads

Discussed on DQ on 04/06:

  • Agreement re changing '0', 'Retired', 'TBC' or 'u' to UNKNOWN in the GMC/GDC/PH number
  • Changing blank to UNKNOWN not recommended as it may legitimately be blank because they have a GDC or PH number
  • What we should do to those with blank/null in all 3?
  • JH: Among those none have programmes and very few have placements. There are Staff, Both, Contact, Admin.
  • JW: 
    • Admin and Contact = N/A ?
    • Staff and Both = UNKNOWN ?
      is that the agreement?

04/06: The following has been agreed:

  • GMC for Contacts, Staff and Both = UNKNOWN
  • GMC for Admin = N/A

DS: List of persons with same GMC's, check the DW to see if some of those have been resolved?

TISDEV-4841 to be paused from releasing to Prod until David Short (Unlicensed)has done his investigation of the ones that could potentially be merged. Merge Rules to be shared. To create tickets following that.

11/06: Ongoing and will probably be ready mid-week.

18/06: On hold until the document uploads are available on People records? The idea being if de-duplicated before the documents have been added, we won't be able to attach the documents to the deduped person records. 

Get an update from Ben/Anita regarding document management in a couple of weeks. 

02/07: Benjamin Wittonare you able to give an update on this?

09/07: Ben: We are expecting the document delta from Hicom this week. The next step will be to get them onto Azure and then apply the agreed meta tags.  Do let me know if you need any other details/updates. 

16/07: Delta received. Document work to continue in this sprint. Following which David can work his magic on the duplicates. David Short (Unlicensed)  does your  initial work need to be revisited? and will there be any dev requirement related to your dup work? We will need to include this in Planning so will need to know (smile) Complexity and volume - David may need to work with Dev in next sprint in order to prepare this. Prioritisation required. 

23/07: Work in progress, script modified. (David Short (Unlicensed),  bullet points of what's required to bring into the sprint as User Stories)

13/08: Document Management migration complete - when can work start on the duplicate records David Short (Unlicensed) ?

20/08: DS: Scripts being worked on. Next step to create the stories in Jira. 

03/09: David: Looking at having the scenarios by end of the week. 

10/09: DS confirmed this is a trainee on his list too -  TISNEW-1894 - Getting issue details... STATUS

10/09: DS demo'ed some 'strong match' scenarios that can be safely de-duped. Sue/ Joanne Watson (Unlicensed): impact on leave manager when de-duped.

Output of Round 1 of de-duplication:

https://healtheducationengland.sharepoint.com/:x:/r/sites/WPI-North/Shared Documents/TIS/North (all)/TIS Person De-Dup Project/Duplicates Output v1.xlsx?d=w70c77e6caa5b4cc1a2408663a1e77fd6&e=4%3a3feb8d869af74e61aeae63ea2a9baff4&at=9

17/09: Tickets for dev to start looking at. (AR and DS) 

Considering other priorities in the next sprint, we are looking at mid to late October.

24/09:  TISNEW-2057 - Getting issue details... STATUS  - DS to send the rules to add to the ticket. (This has been added to Sprint 59 which is due to start on Thursday - do we have all the info we need?)

01/10: Methodology attached to the ticket. David Short (Unlicensed) to separate out only the records that need to be merged for Round 1 to attach to the ticket.  

15/10: TISNEW-2057 has been pulled out of sprint due to impact on leave manager records. Merging on the TIS side would leave orphan records on the Intrepid/TWS side.

22/10: Awaiting confirmation on the above regarding the correct records out of the duplicates to get rid of.

29/10: List received from David - will add to the ticket and work with devs to get it into a forthcoming sprint.

05/11: Awaiting confirmation on the merge rules from Data leads, as the work involves merging data to the survivor person record then removing the duplicate.

DS: There are 2 main merge rules that need to be reviewed by other data leads and may require some tweaking.

Person Merge Rules (DRAFT)  (David Short (Unlicensed)James HarrisAlice ThompsonSue Barrand (Unlicensed))

1.All associated records (placements, programme/curriculum memberships, ARCPs, etc.) belonging to a duplicate person record need to be transferred to the surviving person record.

2.Where the surviving person record has a field containing a NULL or Zero-Length-Strings value but the same field in the duplicate person record contains a non-NULL/ZLS value, this value should be copied into the surviving person record. This should exclude address fields due to possible town/county/postcode mismatches.

  • Use the address that are against the survivor records. The script to be tested on UAT first with the data leads. 
  • All associated records (placements, programme/curriculum memberships, ARCPs, etc.) regardless of whether they are different on the duplicate Person record, bring across to the Survivor records. 
  • Data leads to specify the fields for which we do (2) and those to exclude e.g. Address. DS will provide this.
  • Documents: Pull all the docs from the Duplicate across to the Survivor records regardless of whether these are duplicates.  
  • GMC/GDC/PH that should be kept on the survivor records to be confirmed. DS: Basically, after a person record merge, if a GMC/GDC/PH number is specified, that is the value that should be assigned to the GMC/GDC/PH field of the surviving record. In the case of blanks, the current value (whatever that they be) of the surviving record is retained.

12/11: The merge rules have been updated on the ticket and David has provided attachments for:
- what GMC/GDC/PH to use on the surviving record
- list of front end fields where values can be taken from duplicates to surviving where there is a blank in the surviving record
Could you all confirm please?
TISNEW-2057 - Getting issue details... STATUS

  • Added rule for Sensitive data - All rules confirmed by data leads. 




2Identification and removal of Consultant data on TISAll

13/08: 

  • Need to ID those Consultants who aren't on TIS as trainers and remove them
  • Those who are on a trainers we need to understand what information we should surface on their records e.g. Assessment information perhaps isn't required.
  • All open for discussion and agreement - Ben to set the scene
  • Other regions to review Alice's scripts. 
  • How were consultants recorded on intrepid?  - Simply, not in a consistent way.  Some Trusts might have used the consultant role to define these records, others may have applied a training/non-training grade to a record.  
  • Benjamin Witton to update with some bullet points of what's been discussed. (not about the mulberry bushes) 
  1. We have been asked to remove consultant/non-trainee data from the system.  The focus is those records where there is noinvolvement in the management of a trainee.
  2. There are four parts involved in this work:
    1. Define the rules/cases to identify the records which should be removed from the system.  We want to achieve this incrementally, so we should start with easy win's and then work our way to the more difficult cases.
    2. Create the SQL scripts, run the SQL scripts and verify the results and that those records should be removed from the system.
    3. Introduce a process where the TIS Development Team are able to take the results and simply and efficiently remove the records from the system (without too much back and forth).
    4. Confirm with the owner the data to be removed if they want a copy of the data.  
  3. Senior colleagues want to see progress by 22nd August.  We can satisfy this if we can show our approach and even any preliminary results.  

Additional note - during my catch up with Hicom, a proposed solution from Hicom included looking at the role which a record is associated with  (i.e. does the record have a role of 'DR in training' or 'GP Tutor' etc.).  I am seeking assistance from the NDW Team to produce an output which shows all records and with what role they have  (e.g. clinical supervisor, medical director, leave manager etc.).  There is an option to work with Regional Data Quality Leads/Local Teams to confirm which roles are they actively using in TIS.  

Actions from the meeting:

  • Alice Brindle to share SQL script with Chris Norman.
  • Alice Brindle to effectively provide Chris Norman a handover of the SQL. 
  • Benjamin Witton to provide an update to both James Harris and David Short.  
  • Benjamin Witton to catch up with Chris Norman and Adrian Ashley to identify potential case/rules.


Additional actions:

  • Define with Joanne the roles which should be in TIS.  

20/08: CN: ~20000 records potentially identified. Approx. half from KSS and NW.  

Removal of non-training grades to follow after removal of consultants - Create Jira ticket to be prioritised. 

03/09: Validate 'survivor Ids' on TIS against Intrepid Ids for Leave manager and CPPS

10/09: BW:  Waiting on duplication work to be done, then look at.

Dental role approval and accreditation process - piece of work required to align the process. e.g. YH prefixed roles.

Benjamin Witton: Publish on DQ channel for data leads to review or pass to colleagues/quality team to review.e.g. definitions against the roles. 

17/09 - (Regarding Roles) - BW update (by BW) - Starter for ten published to Data Quality Leads to review in their respective Regions and Local Teams.  Return asked for by Monday 24th September.  Link to access the starter is below:

 https://healtheducationengland.sharepoint.com/:x:/g/TIS/EY1dIsn6lNFPtpRL9Jz8ncgBLTbJhw1Py3cg2DcYPHREVg?e=Lc1zgK  

24/09: Jira Tickets to be created. Benjamin Witton to share work re Roles on DQ channel and follow up.

01/10: TISNEW-2083

22/10: Dependency on the De-dupe work. Follow up with Chris after the de-dupe work has taken place. This is to identify which Consultant data belongs to which Trust so that they can have this if the trust wants.

29/10: Linked to the removal of the consultant information from TIS we also need to address the issue of the non-trainee posts. So a further piece of work needs to take place to identify and hard delete these posts. Will need to take a copy of this so that (like with the non-trainee records) we can pass this to the trusts if needed.JW to create ticket related to the deletion of posts.

19/11 - Once de-dupe is done, check the full list of People from Chris. Please review against the surviving records (IDs) on TISNEW-2057.


3Trust access - role logicAll / Andy Horton (Unlicensed)James Harris

Objective:

Gain understanding of what the local office requirements for trust access should be based on - NPN seems to be the incorrect basis as in TWS it is based on Training and/or Employing body

Email conversation excerpts: 

07/09:
James:
- Having the logic based on NPN would require us to update thousands of posts numbers and the ESR positions in the trust.
- I was discussing this with Gwilym earlier and we think the logic should be an OR statement with EmployingBody and TrainingBody.
07/09:
- Joanne/Paul: Currently on TIS: Placement > Site > Trust determines the access

- JH: This is what was on Intrepid and works for me
10/09:
- Andy: concerns over placement > site > trust logic
Q - Site reference table - Is site attributed to only one parent trust

Q - So if we are using the Trust of the placement site – then we can still only grant access to a post & its occupant to that one trust

Q - e.g. a Severn GP trainee goes to Royal United Hospital (RD130) for a placement. The parent trust in the reference table for that site would be Royal United Hospitals Bath NHS Foundation Trust (RD1). We would also want however for Gloucestershire Hospitals NHS Foundation Trust (RTE) to have access as the lead employer. The only way I’d picture that happening would be through the ‘OR’ statement on Employing Body and Training Body that James mentioned below.

Q - small pockets of users that are trust based but have their data profiles restricted by Programme. We have done this for the new GP lead employer in Thames Valley as Pennine Acute Trust won the tender as well as some local trust based administrators that only look after GP. This allowed us to prevent them from seeing trainees that were not relevant to them at the trust they are based but also an easy way of seeing the trainees out in GP practices as they don’t really have a parent trust, we have the CCG in that field for GP practices. I suspect we may need to find a workaround to this in the short term perhaps using reporting.

Joanne Watson (Unlicensed) / Ashley Ransoo - Requirement for TIS: (Ticket to be created and get an estimate from Dev)

TISNEW-1987 - Getting issue details... STATUS

People - If either the trainnee's EmployingBody OR TrainingBody for their Current Placement Post match the Trust, then make available to the Trust.

Posts -  If either the User is at a trust which is the EmployingBody OR TrainingBody of the Post, then make available to the Trust.

17/09:

JW: 

...very rough guideline of the complexity of amending the current logic on the trust role was given an 8 (out of a possible 10). This is rather large and I have concerns that we might not make the deadline of the end of next week for the role to be amended, tested and users issued with their login details.

So, with the above in mind I considered what else we could do if we approach this from another angle.  Please could I ask you to consider the following - instead of changing the trust role we leave it as it currently is - so placement>post>site>trust but for those users who are lead employers or who just need to see the trainees and posts from a number of trusts we give them the access to that range of trusts so they can see the entire trainee landscape they should. We would obviously need this info from yourselves but it would save us a lot of dev time.

This isn't elegant but should get us across the line for the 1st October. I then suggest that I ask Ashley to speak to you all asap and get full understanding of the different roles in trust-land so that we can correctly create roles that are needed e.g. the one to allow users to only see trainees based on programme etc. We can then get these roles created and rolled out.


GW/AH: There might be an IG Issue as they would be able to see trainees that they should not be able to see. 

JW: Will be discussed at pre-planning. This will be taken into the next sprint. Review progress on next call. 

24/09: Being worked on in Current Sprint. Joanne Watson (Unlicensed)/ Alistair Pringle (Unlicensed) - To get an update from dev

01/10: Trust role on Prod. Need to create the trust access role based on Programme next.

22/10: Andy will confirm the details ahead of next sprint planning. 

29/10: Andy confirmed that the logic was acceptable for the South. Forwarded to LaSE and North for info. Ticket will go into Sprint 61. JW to speak to Simon/Paul about a bulk addition of these users/

05/11: In progress in current sprint

12/11: How many different user roles are we eventually going to have for Trusts access? - To review with the devs/ UM

19/11: Yes, if there is a requirement to, would need to be ticketed up as and when.



4Rob - New approach with data leads/data group:Rob/Data Leads

To discuss with the existing data leads...

Data Group
• X4 Regional Data Leads; NDW link; Service Manager
• Self-ordering team
• Transparency
     o One action list
     o MS Teams
     o Slack

• Ownership of the data model in TIS inc. multi-professional
• Manage business-facing reference tables and have a change control process
• Simplify reference data and standardise to HEE / DH / ESR standards
• Management of data issues as and when
• Develop Dashboards
     o Data Quality
     o Management information
• Participate as data experts in any process alignment, research or development work
• Help identify, rationalise and simplify data flows and data handling e.g. bulk uploads, ESR etc
• Prioritise data activity
• Work with NDW to ensure data relevance
• Develop and support national reports and extracts
• Undertake testing where there is a data

05/11: BW when is this going to get formalised and therefore move this call entirely to the new 'TIS Data Group'.

12/11: Data quality dashboard? Alistair Pringle (Unlicensed) to speak to Rob. 


5

New GP programme 



Hi Joanne & Alistair,

Not sure which one of you it is best to approach to ask this question hence addressing you both J

We have a new programme of targeted GP trainees starting in February and need to look at a TIS solution for recording these trainees.  Not sure if you are aware of them but they are all trainees who have failed GP training in the past and are now coming back in to training for 18 months in a GP practice post they are spread across the local offices and there are only 58 in total but HEE national have asked for one Lead Employer arrangement for them.  This could mean reporting problems as the Lead Employer (it will be St Helens and Knowsley) will need to have them as a separate list.  I can think of 2 ways of doing this, the first would be for the GMC to approve a national programme but then we would need to add this as a national programme to TIS so each local office could choose the programme when adding in their trainees.  The original agreement with the GMC was that we would add the trainees to the existing training programmes but if we do this we would need something in TIS that we could flag so we can pull one report of all of the trainees, maybe something like Military trainees with it being an option in programme membership to say Targeted GP Training.  Whatever the solution we need something sorting pretty quickly as we need to get local offices to record them correctly.  Could you let me know your views on what would be the easiest solution for TIS.

12/11: Alistair Pringle (Unlicensed) - Enabling a new programme membership type for the GP Programme - Not feasible as they could be substantive as well.

Add a new programme would be the preferred approach by the data leads. 

19/11: With Amanda.


6Chris Norman's Apex tool analysis. Issues highlighted relate to data discrepancies between TIS and TIS_interim - Data discrepancies

There are lots of data discrepancies, do they need to be further investigated by the data leads?

/wiki/spaces/TISDEV/pages/519831557

  1. Investigate why ~64000 placements in TIS-Interim have no NPN where they did in Intrepid
  2. Investigate why ~64000 placements in TIS-Interim have no PostStatus where they did in Intrepid - likely to be same reason as missing NPNs?
  3. Whole Time Equivalent discrepancies - We have 102819 WholeTimeEquivalent with 2 decimals in TCS and 102159 in Intrepid with > 2 decimal places. To have a more accurate comparison on this I suggest to round the values before comparing.
  4. There are other discrepancies where data counts suggest they exist in Intrepid_Interim only but not on TIS_Interim but not highlighted in the spreadsheet. These are of smaller numbers,e.g. GDCNumber. Should there be investigation tasks for these?....

12/11: Get an update from Chris Norman (Unlicensed) when he is next on the call

19/11: work superseded, was relevant when data was still flowing from Intrepid to TIS. CLOSED.

CLOSED
7Programme, Post occupancy report from Tableau.
19/11: Chris to check the report on Tableau.