Statistician

From CCMDB Wiki
Jump to navigationJump to search

p:Julie Mojica is the full time Statistician for the Critical Care and Medicine Database.

She provides Statistical Analysis and Reporting using our data.

The Statistician is a member of the Steering Committee & the Task Team.

Questions on wiki

Here is a list of questions driven by the Template:Discuss that have been flagged for Julie.

58

 QuestionModification date"Modification date" is a predefined property that corresponds to the date of the last modification of a subject and is provided by Semantic MediaWiki.Last editor is"Last editor is" is a predefined property that contains the page name of the user who created the last revision and is provided by Semantic MediaWiki.
Panelling or Discharge PlanningThat link no longer goes anywhere, the heading is not on that page. Can the reference be deleted from here, or do we need to review? And, how will this affect the use of the Category:Awaiting/delayed transfer codes? Ttenbergen 14:51, 2018 September 6 (CDT)
28 November 2018 04:53:25Ttenbergen
Previous Location field
  • "We are aware that this may affect categorization under APACHE II but will collect like this for now. This will need to be dealt with when we move to ICD10." - Julie, I don't know what this comment is about, I just came across it when cleaning out things we need to take care of in going to ICD10. I can't think of what we would need to take care of here, if you can't either please take the comment out.
21 December 2018 16:10:01Ttenbergen
Check pre acute consistentwhat exactly do we want to check for? Please also have a look at the stuff below that doesn't specifically have your name. This requested check ties into a bunch of things and if we want the check we need to be sure that instructions stay consistent and lose ends are tied up.
There was a previous attempt to address some of this in Care levels in the community; this page and it need to be consistent, and consistently linked from the relevant field definitions. If we can get it short enough we might make a template to apply the instructions to each of the field pages.
How does Chronic Health Facility fit into this?
There was talk about comparing Postal Codes to known PCH Postal Codes. Since these might Include other buildings at the same site that are not PCHs, this check can at best be a soft check. Please add the list of these postal codes here.
  • from a data perspective, what do you mean by "admitted directly"? If I were to build a check, where would I find that? OR maybe I don't need to know, but then I need to have a definition of what combination of data would be an error.

Integrity check


  • ... unless they are discharged somewhere else entirely, like another ward. So what do we really mean with this? That they can't come from one PCH and go to another or maybe "home" after all?
    • I realize this maybe hard to do. what I mean here is that if one is already a PCH resident, when leaving the hospital, the dispo location must be a PCH location too. or is a patient is already in CHF, the destination when leaving the hospital must either be a CHF or another PCH.


  • Need to look at the PCH Postal code data.

It may be relevant to this check that we have ICD10 Imprisonment/incarceration and other info in Prison / Jail / Correctional Institution.
21 December 2018 16:09:26Ttenbergen
CAM positive (TISS Item)
  • Did these ever get better? Ttenbergen 12:44, 2017 May 11 (CDT) Does anything need to be done about this? Ttenbergen 20:15, 2018 November 27 (CST)
6 February 2019 21:10:31Dr. Allan Garland
Reporting from ICD10/CCI
  • Different procedures would be listed with the same CCI code; will Julie easily interpret and utilize CCI codes for reporting?
  • Do we care that we will not be able to differentiate between a Blakemore tube from an Upper GI scope with banding or hemostasis, when in CCI they both look the same: (T) Stomach, pylorus... and Control of Bleeding. --LKolesar 14:11, 2018 May 1 (CDT)
    • discussed at task 14:08, 2018 June 20 (CDT), Julie to review what she needs and we will discuss again Ttenbergen 14:08, 2018 June 20 (CDT)
21 March 2019 16:11:32Ttenbergen
Transfer-for Organ Transplantation
  • Do we use this in a specific report? If not we should probably consider dropping it, it's an odd thing to collect.Ttenbergen 23:00, 2012 December 12 (EST)
    • we have 12 in ICU database to date. 2 coded in 2011 and 1 in 2010......rest random back to 1994.
      • Julie, do you use this? Do we need to do anything about this DX? A counterpart does not exist in ICD10. Ttenbergen 19:49, 2018 November 27 (CST)
28 November 2018 01:50:05Ttenbergen
Notes field
Yes, I use the primary diagnosis for the reason of readmission even if the record status is incomplete. --JMojica 09:07, 2018 December 6 (CST)
20 December 2018 19:46:33Ttenbergen
Drug or biological substance/agent NOS, adverse effect
  • Had code T88.7 when we first started but was later changed to this code. I have deleted the original code from the tables, but it might still linger in other references. Once you are sure you don't need this info, please delete the comment. Since this would only exist in test data we don't need to keep it long-term.
1 December 2018 05:45:16Ttenbergen
HD (Hemodialysis)
  • I need clarification on this one-if a patient is either dialysed in Emerg, or on a different ward, (ie. anytime during the current hospitalization prior to coming to your area), are we to enter this only as an admit CCI?Mlagadi 10:26, 2019 February 11 (CST)
    • Please have a look at Admit Procedure, it should explain whether to code a procedure or not. There is no explicit exception for Dialysis. If the info there does not give you the answer you need, please elaborate. If the info is there, then, would you have looked there? How can we make clearer to look there? Ttenbergen 11:53, 2019 February 13 (CST)


  • We used to have a special procedure for indicating if a renal transplant patient received dialysis post transplant under the tasks. If we are now only coding HD once, it will often be missed if the patient required HD post renal transplant. Do we care about collecting this information? There is the post procedural renal failure code that will be coded in the acquireds, but this does not automatically mean that they received dialysis.
    • AG REPLY -- I don't see any special reason to need this info. Unless Julie tells us it's being requested, I think we can do without it.
      • Julie, what are your thoughts on this? Ttenbergen 11:53, 2019 February 13 (CST)
13 February 2019 17:53:56Ttenbergen
Deceased patients
  • Is that really what you want? It will give fact that patient died, but miss new location.


  • Is that really what you want? It will give occupancy but miss actual time of death.


to be done likely after DSM: add destinations to organ donor deaths.
1 Julie asks:

Once that is implemented, I can set up Check dx implying death across encounters.
Correcting suspect links also needs to be dealt with when this is done. And needs to be documented.
7 February 2019 17:31:06Pagasa Torres
Readmission Rate to ICU
  • Is this the same as Re-admission? Ttenbergen 12:03, 2017 July 5 (CDT)
    • similar but specific with ICU. I will do a separate one for Med and delete the Re-admission after.
10 March 2019 04:33:09Ttenbergen
Comfort Care
  • We may be able to stop this when ICD10 comes; but continued collection wont break anything.


  • Julie, can you add here why this can stop when ICD10 comes? Is it because we will start collecting Palliative care? Because that is not really the same definition...

We will need to update a the reference to this in Palliative_care#This_code_vs_Comfort_Care once decided.

1 February 2019 17:50:15Ttenbergen
Resistance to antimicrobials, methicillin (anti-staph penicillins)
  • Julie, the above question specifically affects some projects you work with as well - do you think unifying this rule will be a problem for any of them?


Also affected are :

Does anyone think making this one rule for all will be a problem?

31 October 2018 20:50:53Trish Ostryzniuk
APACHE Comorbidities in ICD10 codesDx grouping
  • Need to update from Allan's email 2018-11-26, but he said he would need to review this in light of the changes that had been made to ICD10 and CCI since he and Julie discussed. Ttenbergen 00:36, 2018 November 27 (CST)
21 March 2019 18:07:29Ttenbergen
S ICD10 Blocks table
Primary Admit Diagnosis
  • They are used in the periodic quarter and fiscal year reports of both the Critical Care and Medicine Programs. (Julie)
    • are the two I linked to above those reports? Ttenbergen 14:03, 2015 April 20 (CDT)

    • How will the primary admit dx involving ICD10 be handled - another query? or be combined to Primary_admit of old dx? --JMojica 09:17, 2019 January 31 (CST)
      • I have built query Query L_ICD10_primary and documented it. Julie, once you have found this and read it, please delete.


    • The query Primary_admit of CFE contains multi records per D_ID. These are the records with admit dates before or on Dec 31, 2018 and are still in the unit by Jan 1, 2019. The L_Dxs of these cases have all the same priority number.
    1 February 2019 22:09:33Ttenbergen
    Instructions for importing a batch of DSM DataSomething is still not right with the code for reconnecting, Tina needs to look into. Ttenbergen 17:04, 2018 May 17 (CDT)
    • This could also be true where no labs were sent for, eg a patient who dies shortly after arrival. In the past we would have entered a "no labs" for these. Do we want to do something similar? It would have to be Pagasa that does it. Might be a lot of extra work. Need to review. (ex. wrong D_ID when exported but found it error and so fixed it before the data for import comes back). For now we do not have an entry like that. And it might not be worth it - what would Pagasa do to check that the no-labs are legit?
    14 February 2019 22:09:22Ttenbergen
    Check Inf Potential Infection must have pathogen or alt combined code
    • This would cause extra collection work, since there are quite a few potential infections (most of the NOS codes are potential infections), and a lot of them would not usually be infections. Do we really want to do this or can we decline it?
    1 March 2019 19:10:53Ttenbergen
    Person ID field20 March 2019 22:11:44Ttenbergen
    Medical Assistance In Dying
    • When we started out this dx used code U23, but then as of 2018-07-17 ICD10 actually added a code for this so we changed ours to that code. I don't really think we are interested in keeping that very early test data, so this comment can probably just go, and we can delete them. I am removing the code from our s_ICD10 table.
    31 December 2018 15:58:59Dr. Allan Garland
    L ICD10 APACHE Dx query
    • You and Allan discussed what should be on the list. At some point we will need to integrate the result into this query. Did you end up including Acquireds? Since the first 24hrs might include them, but they might happen later, and the difference is not clear from Dx_Date? Ttenbergen 20:20, 2018 November 24 (CST)
    25 November 2018 02:20:14Ttenbergen
    Data dictionarysomething went wrong with this query and it has no data
    Attempt at a easier to follow data dictionary. Is this what you had in mind?
     Yes, this is what I have in  mind. Thanks.  Some suggestions: 
    
    • I just notice some start dates are not the actual start dates - It is important for the users to know how far back the data are available so they can decide the covered period of their study. is it possible to change the date to actual earliest collection start date (not 1 Jan 1900). --JMojica 10:05, 2019 January 3 (CST)
      • The start and end dates are stored on the individual pages. If the list shows 1900 then they were not filled in. If you know what they are and fill them in then this page will list them. I can also change the default if-not-filled value in templates to something other than 1900-01-01. I just needed a value for ranges and filters to work. Ttenbergen 14:49, 2019 January 3 (CST)
    • add a column for program (Critical care only or Medicine only or both).
      • I can do that, but the table is already getting quite wide for a web page. Can we lose any of the columns we currently have? The info you mean is always visible in the pages themselves as well, in case that's sufficient. But, yes, we can tweak what the tables should show. It is done in Template:DataDictionaryQuery (details visible once you edit it...) and removing fields would be easy enough. I can show you how to add fields. Or I can add them once we confirm what we want.

    Tina has changed the ICD10 and CCI templates to use a startdate of 2019-01-01 and will change other default dates as I receive dates Julie wants me to use.
    14 January 2019 11:34:54Ttenbergen
    STB E6
    • are you still running these? Ttenbergen 21:06, 2018 November 24 (CST)
    12 March 2019 16:54:59Pamela Piche
    Bronchitis, acute or chronic not specified, infectious or noninfectious
    • bringing you in on existing discussion


    APACHE CHRONIC stuff related coding schemas

    • This does not trigger APACHE Acute Dxs in ICD10 codes while some other dxs that seem no more "intense" do. Is that right?
      • AG REPLY --- Leave it OUT, it's a wastebasket code and could be acute disorders. ALSO, in Feb we'll shift from tick boxes for the AP2 comorbs to identification via ICD10, for which I've made the coding already.
        • Julie has investigated the APACHE conversion further since this was brought up, so she should be included in any further conversation about this. Ttenbergen 14:45, 2018 August 6 (CDT)
    22 March 2019 20:27:14Michelle Lagadi
    Palliative Service
  • there was a question about palliation at beginning vs end of stay. It was discussed at task but never cleaned up. Could you have a look a this page? If this is all no longer an option, please delete the section. If it was resolved, then what did we decide? Or was that why we starte Comfort Care? Ttenbergen 00:01, 2018 November 27 (CST)
  • 28 November 2018 04:55:03Ttenbergen
    Transfer time rule
    • will we still want this now that we have Visit Admit DtTm field and will hopefully eventually move to using the EPR to glean arrive and dispo? Ttenbergen 17:06, 2016 May 25 (CDT)
      • deferring the question to after when Julie has done the new multiple encounter linking with Dispo. Ttenbergen 15:58, 2016 June 27 (CDT)

    with transfer tracker gone, what will be the official instructions for this? Ttenbergen 16:36, 2017 June 21 (CDT)
    1 February 2019 21:16:24Ttenbergen
    Query NDC Dxs vs TISS Dialysis
    • AG REPLY -- This is challenging since not everyone with either ARF or CRF gets dialysis AND not everybody who gets dialysis has renal failure -- e.g. dialysis is also occasionally used for drug overdoses, volume overload, and a very few other things.

    We are getting a lot of positives where a requiring dx does not have a TISS. Is our list of requiring dxs reasonable? ~ 118 patients like this.
    20 March 2019 16:41:35Pagasa Torres
    Check organ donors must be dead 24 December 2018 18:47:30Ttenbergen
    Check CRF vs ARF across multiple encounters
    • Using the ICD10 renal codes, we still need to know when the transition from acute to chronic occurs - so we can decide whether the multiple encounters consistency checking is still relevant. --JMojica 11:51, 2018 November 14 (CST)
      1. is the transition on the next hospital stay? Example in this hospital stay, patient is diagnosed with ARF and stayed continuously in both ICU and ward in same or different hospital. On the next hospital stay, he is now chronic renal patient.
      2. Or the transition is on the next ICU or ward stay? Ex. the first stay is ICU and diagnosed with ARF. then patient was transferred in a ward of same or diff hospital - is he now a chronic renal patient?
      • The data collection instructions are in the related pages, and additional info is in Renal Coding Considerations for ICD10, but they are a beast of a network of concepts. Those might tell you how we currently propose to collect the renal codes, but not necessarily what you or the users of the data would want. Usually these cross checks would be driven by what you need for data requests, so do our proposed instructions line up with how you want to use this? Or is this maybe too case-by-case of a concept to even make a cross check? Ttenbergen 18:59, 2019 January 6 (CST)

    I think this section is largely old rules that no longer apply with how we define Comorbid Diagnosis now, ie if something was clearly present before admission we can now code it, even if it had not been diagnosed. Are any of these still required, with that in mind? Ttenbergen 21:17, 2018 October 26 (CDT)
    1 March 2019 19:10:53Ttenbergen
    Eliminating distinction between different ward typesCTU vs NTU - we decided at task that we wanted to eliminate the distinction. A lot of things are part of their network of information, so we will need to work out the details above before we can move ahead.
    • I believe it was only related to transfer ready DtTm between CTU and NTU not eliminate the concept. Medicine program would have to weigh in. Trish Ostryzniuk 16:11, 2018 July 30 (CDT)
      • I seem to remember the reason to stop doing this also involved that we have more and more locations that don't fall into a clean place. For now we pretend we can give medicine this data, but is it true and meaningful? Ttenbergen 15:51, 2018 July 31 (CDT)
        • Allan will contact the medicine stakeholders Ttenbergen 12:18, 2019 January 24 (CST)
          • pinged again for Allan to Follow-up. Ttenbergen 13:44, 2019 February 25 (CST)
            Allan will contact the medicine stakeholders about eliminating this distinction. Will email Mary-Ann Lynch, VanAmeyde, Griffin.
            any of Julie's Reporting that use this concept?
    9 March 2019 22:27:38Ttenbergen
    STB Cardiac Care patientsCan you confirm new facts now that we are eliminating Registry Patient Type. (obviously this is a pretty old question...)10 March 2019 04:38:24Ttenbergen
    Charlson Comorbidity scoring in ICD10 codesCan you have a look at this page and the queries/tables it describes to see if it makes sense?2 January 2019 21:58:47Ttenbergen
    ICU Var 6 - AMADid we transition the following into tmp or otherwise? Ttenbergen 13:58, 2017 June 6 (CDT) If we did not then this question can just be removed, but if we did move this elsewhere we should explain where to.3 January 2019 21:31:41Ttenbergen
    S dispo.service typeDo you know what is the description? Especially in contrast to S dispo.loc_type. How do you use this? Please put the answer in the element_description above7 February 2019 04:47:51Ttenbergen
    Query NDC VAP no TISSFYI Maybe
  • Pneumonia,_ventilator-associated_(VAP)#VAPs_on_medicine_wards mentions to use the same rules as for other Acquired Diagnoses, ie the dx is acquired where it first manifests, so this check or rule no longer makes sense. If you are OK with this, please take this note out, else pls comment here.
  • 7 February 2019 14:08:04Ttenbergen
    Severe SepsisI don't think this was ever implemented, can't find any evidence of it. Do we need it? Ttenbergen 11:04, 2018 September 25 (CDT)31 December 2018 21:36:56Ttenbergen
    LOS Medicine per hospital admissionthis still talks about TMSX... what is the new status of this field?
    I think you made several of these at some point. Did we do anything even categorize them? If we annotate them right we can include them in the Data dictionary ...
    4 January 2019 02:34:53Ttenbergen
    AaDO2I wonder if "null" is actually intended here, or if it was supposed to be 0. Ttenbergen 14:44, 2017 January 8 (CST)
  • AaDO2 is null if there is no data for FIO2, PaO2 and PaCO2. However, no data was set to zero value instead of null. In addition, AaDO2 and the corresponding score are required only when FIO2 >= 50%, otherwise should be treated as null. (Similarly with PO2 score, it is required when FIO2 < 50% and otherwise should be treated as null. When FIO2 =0, both AaDO2 and its score and PO2 score should be null.) Are we setting null to zero to facilitate the programming calculation? Is it possible to differentiate null from valid zero in both L_LOG and created_variables_CC - will the work be big? In L_LOG, these are the FIO2, CO2, PO2 and SerCO2. 14:49, 2017 March 16 (CDT)
  • 3 January 2019 06:36:25Ttenbergen
    S dispo.loc typeThis value is not yet encoded on the wiki as Property:Collection Location Location Type for locations that have their own article, but maybe it should be. Should it be? Ttenbergen 09:46, 2017 November 9 (CST)
    Indeed, what is the description? Especially in contrast to S dispo.service type; when you give the answer, please put it behind "element_description" above and delete this question.
    25 October 2018 03:37:29Ttenbergen
    QA Septic ShockIs "Every entry for project QA Septic must have either a date or a time." a request for a cross check? or does that check exist already?3 January 2019 05:41:09Ttenbergen
    Transitional CareIs this part of any reports? Ttenbergen 10:03, 2017 November 9 (CST)10 March 2019 04:49:40Ttenbergen
    HSC IICU Collection GuideIs this still relevant after workload redistribution? Ttenbergen 11:47, 2015 May 20 (CDT) And is this how you want it?16 January 2019 16:08:27Ttenbergen
    Critical Care Vital Signs MonitoringIt says that CCVSM is in Quarterly report. If CCVSM is no longer, is it still in quarterly?21 January 2019 18:46:58Ttenbergen
    Bed holdsJulie seems to set the limit at 1 day - emailed Julie Ttenbergen 10:07, 2016 November 10 (CST)
    duplication on wiki needs to be cleaned up once we are on same page
    Laura and Tina discussed this and there clearly are different understandings about this. Need to review. LKolesar 14:43, 2017 March 1 (CST)
    3 January 2019 22:22:52Ttenbergen
    Transition to Database ServerJulie, can you confirm that SAS would be able to connect to an MS SQL Server via ODBC? Ttenbergen 22:02, 2018 March 14 (CDT)
    You were going to follow up with the new CHI person to make sure they are aware.
    11 March 2019 19:02:11Ttenbergen
    SAS Data Integrity ChecksNow that we have a structure for cross-checks we should add those you do in SAS to here as well, using the same structure as for those listed in Data Integrity Checks Ttenbergen 20:46, 2018 October 26 (CDT)30 October 2018 01:48:15Ttenbergen
    Query check ICD10 needs awaiting if TR Dt and Dispo Dt on diff daysThere are transfer ready reasons that would not result in an awaiting code. These resulted in false positives. If we want to check for date diff but no awaiting code we will need to enumerate these reasons, and all need to be present in data. I will put aside this half of the query until we address that. Details in wiki page.1 March 2019 19:10:56Ttenbergen
    Statistical AnalysisThis article will likely be one of the more common landing points for external users. What do we want to tell them? Do we have any project articles we want to link in that especially highlight what we can do? ALERT Scale?Ttenbergen 22:50, 2017 June 7 (CDT)7 February 2019 04:48:08Ttenbergen
    Chart Review ListsThis is linked from the front page and intended to give an idea of how one could use our data. Is there anything on Publications that would be a good example for how our DB was used for this? If not, should we take it out? With nothing here it doesn't look very good coming from front page.3 January 2019 22:31:09Ttenbergen
    Pre-OP Admit - Research Patient - CardiovascularThis is not coded under ICD10 or CCI. Do we report this, and therefore need to accommodate otherwise, or how will we treat this? We have Preparatory care (incl preop optimization) but that doesn't capture the full concept. Ttenbergen 22:45, 2018 November 27 (CST)31 December 2018 03:05:46Ttenbergen
    Pre op Admit-Cardiovasc PatientThis is not coded under ICD10 or CCI. Do we report this, and therefore need to accommodate otherwise, or how will we treat this? We have Preparatory care (incl preop optimization) but that doesn't capture the full concept. Ttenbergen 22:45, 2018 November 27 (CST)31 December 2018 03:06:12Ttenbergen
    Base Population for ResearchThis page is linked from the front page, so we should either make it good or get rid of that. Is there anything on Publications that would be a good example for how our DB was used for this?7 February 2019 04:47:59Ttenbergen
    Validation against Patient Registry DataThis page was started long ago to keep track of our attempt to get access to the registry. I think it would be good to re-convene on it so we have a central point where past efforts and current efforts can be tracked. That would also make it easier to take it to task or steering and have consistent info. Do you have a log of this somewhere? We can rename it if you want.3 January 2019 21:26:12Ttenbergen
    ADL General Collection InformationWhat else in addition to ALERT Scale Calculation uses this?4 January 2019 02:25:07Ttenbergen
    Query s ICD10 Chapter block dxsany other plans for these?11 January 2019 05:15:41Ttenbergen
    Project Borrow arrivedid they ever get back to us?23 January 2019 00:15:12Ttenbergen
    Query TISS Errors missing dayswhich report/s are these actually included in?11 February 2019 22:14:57Pagasa Torres
    QA Infection VAPwill we still need to collect this in ICD10, since I think all the data now lives in the dx codes as well. I am holding off on implementing Query s tmp QAInf tmp no dx until resolved. same reply as in QA CLI. --JMojica 12:04, 2018 December 27 (CST)3 January 2019 05:08:29Ttenbergen

    IT Access used

    Related Articles

    Related articles: