JALT Meeting - Rolling Agenda and Minutes 2023

From CCMDB Wiki
Jump to navigation Jump to search

List of items to bring to JALT meeting

Add to this by adding the following to the article where the problem is documented:

{{DiscussTask | JALT
* <question details>}}

(this will bring it to Task if not addressed at JALT)

or

{{Discuss | JALT
* <question details>}}

(this will not bring it to Task) Toggle columns: Last modified

wiki page question Last modified
wiki page question Last modified
Change of remaining location names from "our" names to EPR/Cognos names JALT - Is there anything here we want to do before SF? Or that still needs to be done at all? Ttenbergen 09:42, 2023 July 6 (CDT)
  • What happens to the ICU Previous Location, Pre-admit Inpatient Institution, Dispo or even Service Location - should they be changed too by the new COGNOS ICU locations? Example current STB_ACCU is SBGH-CCUO in COGNOS, STB_CICU is SBGH_ICCS, STB_MICU is SBGH_ICMS. Should the old labels remain? We need to think hard for its implications to queries of linking and/or matching tables before implementing any change. --JMojica 16:33, 2022 February 2 (CST)
    • It would be nice to have this consistent, and yet you are correct that this would tie into a lot of things. I think the benefits of making it consistent win out, though especially when it comes to also thinking about this in terms of that metadata we discussed the other day. Even if we keep the (possibly identical) data in both s_tmp and s_dispo for now, we would then be able to use that metadata table for both. This would require thinking through the details. Julie, I think it only involves you and me, so maybe we should discuss at our wiki meetings? Ttenbergen 13:44, 2022 February 8 (CST)
      • Julie and Tina discussed:
  • We use the 4 fields Previous Location, Pre-admit Inpatient Institution, Dispo and Service/Location also to map patient flow between laptops, and we very much don't use Cognos values for this (e.g. HSC_Med). We need to retain this ability to use the entries for linking but would also make them the same as Cognos where possible. So we need to keep our "own" values for this for locations where we collect.
  • We decided to use manually split CC entries e.g. HSC_MICU vs HSC_SICU since Julie reports in those increments, ie it is hard to pull apart a stay in two ICU types if it is collected as one record. We don't want to lose that.
  • We would still like to change these own values to the "modern" values where we use legacy terms, eg. STB ICMS vs STB MICU. As long as we make a clean transition between old and new, or change all old, that should not be a problem, but we need to account for it.
  • We could use the Cognos values for all places where we don't collect, e.g. if a pt comes from Ward HSC_A1 and Cognos lists that as HSC-GA1, we could just enter that. However, for locations we don't collect we currently aggregate this to HSC_ward. Do we want the extra detail? It would be easier to enter but might be harder to interpret and possibly even harder to work with for collectors.
  • If we want to keep our proprietary value for locations where we collect, and keep aggregate ones for locations where we don't collect, I am not sure which locations that then leaves where we would use the Cognos values?
      • Julie, do you agree to that summary? If so, there may be nothing to discuss with Lisa, since we will need to leave this as is. If I am missing something pls update and then pass on to Lisa for her take. Ttenbergen 16:56, 2022 March 23 (CDT)
        • agree. pass to lisa. --JMojica 15:27, 2022 June 8 (CDT)
  • I think this is no longer an issue, unless we are looking to change how we collect this, which I am not in favor of Lisa Kaita 12:23, 2022 August 24 (CDT)
    • Even though this is no longer an issue, we should keep the above 5 summary issues here for future reference. --JMojica 13:38, 2024 March 12 (CDT)
2024-03-12 6:38:18 PM
Query cardiac arrest throughout admission JALT Review after 2023-09-15
  • Lisa flagged that, if we do this for Cardiac Arrest, we should really do it for other dxs as well. And if we did that, it could result in a lot of work since it would need to be mediated by Pagasa for now. So we decided to see where the SF implementation goes and review the definition of this check once we have a centralized tool where the data collector would not need to mediate this. Ttenbergen 15:43, 2023 July 13 (CDT)
  • 2023-07-25 3:49:53 PM

    _

    _


    JALT Meeting – June 28, 2023

    • Present: J,A,L,T
    • Emails: LKaita@hsc.mb.ca; allan.garland@umanitoba.ca; JMojica@hsc.mb.ca; TTENBERGEN@hsc.mb.ca
    • Minutes prepared by: AG
    • Action items in BOLD

    1. With regards to the replatforming plan to PwC/Salesforce, we discussed how to store comorbid diagnoses in a separate patient-level object.

    • Issues involved include: (a) while most comorbids are permanent, some can come and go, even more than once, and (b) there ARE errors made in recording comorbid diagnoses and these tend to be propogated forwards indefinitely ("chart lore").
    • We agreed on a plan for a Salesforce Comorid "object" which for every comorbid ICD-10 code, there would be 1 or more entries/rows.
      • Each row will have a Diagnosis field, and 2 date fields. The first date is the admission date when that comorbid is first recorded. The second date would remain empty as long as that comorbid continues to be recorded, but if there comes an admission date at which it is no longer present this second date would contain that admission date. Neither of these dates should be considered to reliably identify the onset or ending of the presence of that comorbidity.
    • Complex example: Patient has CKD Stage 5 (N18.5) with 1st recording of it as a comorbid is at first ever ward admission on 2/3/2005.
      • On that date this comorbidity object would gain an entry as follows:
        • Code-----DateA-----DateB
        • N18.5----2/3/2005
      • At the ward admission on 4/5/2007 there is no recording of CKD as a comorbid. The collector searches and finds out that the patient had a kidney transplant in Ontario in 2006, and that's why it's gone, i.e. it really went away. Now the object should be:
        • Code-----DateA-----DateB
        • N18.5----2/3/2005--4/5/2007
        • Z94.0----4/5/2007 <---- this is: Past history, transplanted kidney
      • But, sadly, the transplant fails and at the next admission (to ICU now) on 6/7/2010 the CKD is back so now it'll read:
        • Code-----DateA-----DateB
        • N18.5----2/3/2005--4/5/2007
        • Z94.0----4/5/2007
        • N18.5----6/7/2010

    2. We need to go over/remember why we previously decided that in identifying comorbid diagnoses we would use Admit Diagnosis and Acquired Diagnosis.

    • Emailed Allan: You wrote up the decision to use all buckets for Charlson in a series of edits on 2019-02-13 in Charlson_Comorbidities_in_ICD10_codes. I don't see an entry corresponding to that date in a Task meeting and we didn’t have JALT meetings then. Do you have a record of what might have been discussed then that caused you to update it? Ttenbergen 12:34, 2023 July 5 (CDT)

    3. We began a discussion about records with no labs at all.

    • This could be real (e.g. very short admissions, with labs done in ED before coming to ward or ICU).
    • But a problematic reason is when the person has 2 PHINs -- as occurs with a person initially a John/Jane Doe who is then identified.
      • In this case, we have concrete examples of 1 PHIN/MRN having no labs with the actual labs from this encounter being in the 2nd PHIN/MRN.
    • After discussion we decided to handle such "lab empty" situations as follows: Pagasa to identify them -->
      • if LOS<6 hours ignore the problem as it's probably real
      • if LOS>6 hours Pagasa to check in the EHR to see if this person/encounter actually has 2 PHINs/MRNs. If so she will collapse the 2 records into one, and ensure that the info for the actual labs be sent to Alun Carter with the next request.

    JALT Meeting – April 19, 2023

    • Present: J,A,L,T
    • Emails: LKaita@hsc.mb.ca; allan.garland@umanitoba.ca; JMojica@hsc.mb.ca; TTENBERGEN@hsc.mb.ca
    • Minutes prepared by: AG
    • Action items in BOLD

    1. After further work and careful consideration our long-standing concept of limiting some ICD10 codes to specific bins (i.e. admit, acquired, comorbid), we recognized that doing so would have some positives, but that it also has some negatives. Most importantly though, was the recognition that even though some codes would not make sense in a given bin as a standalone item, it could make sense as a linked item. THUS, we agreed to cancel this planned work.

    2. We need to update out Privacy Impact Assessment. Allan & Tina are working on this and have a Zoom meeting planned with Rosanne and Dan to discuss some items on 2023-04-26.

    4. Update on Charlson comorbidity coding. Allan sent the final version to Tina, who will work with Julie to implement it.

    5. New item: There are locations (e.g. HSC-GA7S) whose level of care changes over time. For example it has variously been a ward level and an IICU level. Thus we recognized that we need level of care to be a variable that can change over time. Tina to work on creating machinery to support that.

    6. We discussed the request of Nelson Prudencio for ICU data.

    • Nelson is a data person in the Shared Health Business Performance Group. He is apparently interested in comparing the indicators on our quarterly reports to the electronic data sources available to Shared Health.
    • We agreed that Nelson should be invited to the next Steering Committee meeting. Allan emailed Rosanne today about this.

    JALT Meeting – March 22, 2023

    • Present: J,A,L,T
    • Emails: LKaita@hsc.mb.ca; allan.garland@umanitoba.ca; JMojica@hsc.mb.ca; TTENBERGEN@hsc.mb.ca
    • Minutes prepared by: AG
    • Action items in BOLD

    1. Final decision about how to deal with delays (either in admission from ED, or for other transfers) -- given that the administrators (Bojan, Roseanne, Ebi) all agree that short delays are inconsequential, we agreed that from this point onwards, for all types of delays, our reporting will simply provide actual delays. No allowances will be made such that delays less than some value will be ignored.

    2. Further work on whether/how to decide on limiting some ICD10 codes to specific bins (i.e. admit, acquired, comorbid). We agreed on a multiple stage process:

    • First Allan will peruse the Excel file Julie/Tina made that indicates for each ICD10 code the # (%) in each of the 3 bins. This will be used to prescreen for items that might be prolematic -- DONE.
    • Second, Lisa will assign 1 or more data collectors to go through and make preliminary "DISALLOW" decisions on them.
    • Third, those disallowed in stage#2 will be reviewed for a final decision.

    3. We need to update out Privacy Impact Assessment. Tina/Allan to meet April 5 at 9am for this purpose. Follow-up meeting with Roseanne and Dan booked for 2023-04-26.

    4. Update on Charlson comorbidity coding. Tina resent this email to Allan and he responded on March 26, 2023.

    5. Issues were raised about coding of possible repeat episodes of AKI. Allan has modified the AKI template regarding this.

    JALT Meeting – Feb 2, 2023

    • Present: J,A,L,T
    • Emails: LKaita@hsc.mb.ca; allan.garland@umanitoba.ca; JMojica@hsc.mb.ca; TTENBERGEN@hsc.mb.ca
    • Minutes prepared by: AG
    • Action items in BOLD

    1. Update regarding those records where ER is the first boarding loc but the patient is there for only a few minutes, or even for Zero minutes.

    • Per last time, we recognized that these are real when the patient is actually in an ER bed.
    • What we don't yet know is about the cases (of which Julie found 147) where patient is admitted from ED triage without being put in an ED cot.
    • The crux of this issue is exactly how our users utilize this information. If, as we suspect, it is to identify opportunities for substantially reducing such ER delays, then short delays (e.g. <10 mins for sure, and prob even <20 mins) are irrelevant as they're not likely remediable. If this is the case, then for collection and reporting purposes we should only report on those with ER delays >some threshold (e.g. 30 mins).
    • Allan emailed Bojan, Carmen and Roseanne about this today ---> and they replied that indeed they only care about delays that are substantial.
    JALT
    
    • For CC Reports, ER Delay for LE 30 minutes, ignore (treat as zero) and GT 30 minutes, use actual ER Delay. N is number of cases where previous location is from own ER, CC reports Mean ER Delay as Total ER delays/N and Median ER based on N. Prior to Oct 2020, we calculate ER delay as Arrive Dttm - Accept Dttm. But there are cases where Accept Dttm is missing, therefore ER Delay is missing and these missing are not included in N. Are all these correct? --JMojica 17:36, 2023 February 13 (CST)
    • For Medicine Reports, I am reporting LOS (mean, std, total, min, max) per Boarding LOC which includes ER . Should the GT 30 mins threshold be applied for ER boarding loc or use actual ER LOS? do we need to follow the same rule for both CC and Med? --JMojica 17:36, 2023 February 13 (CST)
    • SMW


    • Cargo


    • Categories

    2. We need to update out yearly PIA report. Tina will email the last one to Allan.

    3. Discussion about the intention throughout the province to start collecting (voluntarily) on indiginous status and LGBTQ status. We recognized that this would be very interesting data to have in our databases.

    • Allan related that "they" are working on a governance plan, so Tina will put a reminder on this item to revisit it December 2023.

    4. Information suggests that the Dept of Medicine is (again) contemplating ceasing to fund the Medicine database.

    • Allan spoke today with Bojan about this, particularly about the importance of ensuring that IF they do so, that the ICU Database will be protected. The plan is that Allan and Bojan will meet with Dr. Renner and Hana to discuss these issues.

    5. Question about whether to include medicine patients on a soon-to-open ward at HSC, that might be a mixed (medicine and non-medicine patients) ward.

    • Answer: Yes, but only include those on Medicine service.

    6. We were informed that Hi Obs at HSC has recently had some patients on non-Medicine services (Respirology, Neurology). We agreed that we will only include the Medicine patients in our database.

    7. Update on Charlson comorbidity coding. Allan to respond to the email sent by Tina on Jan 26, 2023.

    JALT Meeting – January 11/12, 2023

    • Present: J,A,L,T
    • Emails: LKaita@hsc.mb.ca; allan.garland@umanitoba.ca; JMojica@hsc.mb.ca; TTENBERGEN@hsc.mb.ca
    • Minutes prepared by: AG
    • Action items in BOLD

    1. RESOLVED ITEM: Followup about APACHE 2 Scoring. We now believe it is correct.

    2. Visit_Admit_DtTm differences within same admission

    • Tina reports that she has not received any such issues from Pagasa, and it's not clear if (a) there haven't been any vs. (b) there have been some but Pagasa has been dealing with them herself.
    • Tina to ask Pagasa about this. If 'b' then we DO want them sent to Tina so she can send them to Chastity to see if they can be fixed withing ADT/Cognos.
      • Tina emailed Pagasa about this December 8, 2022 but hasn't heard back, and again 2023-01-25.

    3. RESOLVED ITEM: Regarding the new infrastructure to separately record both Dispo DtTm field (for when the patient is no longer under our care) and BedHeldEnd DtTm (for when the assigned bed is released).

    • This only applies to situations when patient vacates the bed but it is "held" for expected or possible return of the patient.
      • The main such situations are the patient does NOT return after: (a) going elsewhere for a procedure, or (b) leaving AMA and the bed is held for a day or 2 hoping she/he will return. Which of these occurred is recorded in TEMP.
    • Today we recognized another such situation: Patient living at St.Amant had a PEG, went back to St. Amant hoping he'd do well, but the hospital bed was held for a day just in case. As this situation isn't covered by recording AMA or Procedure in TEMP, we agreed to add another possible TEMP entry for such occurrence, to be named something like "Other" or "Weird Reason". Tina has added this option to CCMDB as part of CCMDB.accdb Change Log 2023#2023-01-25.
    • Julie also inquired about how to now report about bed-days and occupancy. After discussion, we agreed that she will alter reporting as follows:
      • Primarily report on days patients were in beds being cared for in those beds (i.e. from Admit DtTm until Dispo DtTm.
      • But also report the ADDITIONAL cumulative days when beds were held with the expectation patients would return to them, but did not, as directly above.

    4. RESOLVED ITEM: Repeat item, regarding how to record time spent waiting for transfer (after Transfer Ready DtTm):

    • We had previously decided to: (a) ignore such transfer delays <2 hrs, but (b) for all delays >=2 hrs to use the actual delay time.
    • But this seems to conflict with the prior national Vital Signs Monitoring program (which is now defunct) which substracted 2 hrs from all actual delay intervals.
    • Allan's rationale for not subtracting 2 hrs is as follows: While there are seemingly obligate delays in transferring a patient out (sending site getting the patient ready to travel and ensuring personnel are available, getting bed cleaned an ready in the accepting site and ensuring personnel are available) the concept of any given interval for these to occur is counter to the concept that we can and should always be striving to make out systems more efficient and reduce those delays.
    • Allan asked Bojan if this plan is OK, and he said yes.

    5. RESOLVED ITEM: How to deal with transfer delay in a recent case where patient in ICU was made ACP/C and then palliative ---> didn't die and indeed improved --> reversed the ACP/C and patient left ICU alive 7 days later.

    • After discussion, we agreed that this is an extremely rare occurrence, and that rather than make a new rule for it, we will stick with the existing rule that transfer delays be calculated from when first transfer ready until the patient leaves. And also, this example is not fundamentally different from when a patient is ready to leave but gets sicker and the transfer is cancelled.

    6. RESOLVED ITEM: Cases (esp at St. B ICUs) where ER is the first boarding loc but within a few minutes (or sometimes simultaneous with) that boarding loc, a 2nd boarding loc of ICU is recorded.

    • Though we first thought this is a phenomenon of incomplete charts, Julie checked and it is NOT. After discussion we determined that this is a real phenomenon that occurs when ICU/ward team only agrees to accept the patient once a bed is available, and that once that happens the patient IS quickly transferred from ED to ICU or ward.

    7. In looking into #6, Julie noticed that there are 2 sorts of admissions about which she is concerned:

    • (1) Direct admit without passing through ED at all. These are real, and almost certainly are direct admissions from clinics to wards. So not a problem.
    • (2) Admit from ED triage without being put in an ED cot. Again, these may well be real, and if so likely are admissions to wards (rather than ICUs). Julie will send a few of these to Lisa, who will check on them.

    Previous

    For earlier minutes see JALT Meeting - Rolling Agenda and Minutes 2022