Friday, September 15, 2017

NHS 2011 - Commuter Flow Question

Question:
A couple of years ago I asked a question about some place of work information from the NHS that seemed to indicate a coding error.  People whose residence did not seem to make sense were seen to be working in the Village of Gagetown.  Of course, it turned out that there was a coding error based on respondents entries, and that the place of work was actually [CFB] Gagetown.  I have another such question from the same researcher. 

In looking through the 2011 commuter flow table (99-012-x2011032) for the Chipman/Minto area, he noticed that 45 people are listed as having their place of residence in Red Bank First Nation (Red Bank 4) and their place of work in Chipman Parish.  Given that Red Bank First Nation is more than an hour's drive to Chipman Parish, and given that there's also an area in Chipman Parish called Redbank, he's wondering what he can count on with respect to this data.  He's aware that "place of residence" is the location where respondents are enumerated and does not necessarily represent where respondents were at the time of the NHS; however, given his knowledge of the area he thought the numbers didn't seem realistic and that it might be possible that an error had occurred.  He's not sure, though, so he wanted to see if we could double check on this. 

Answer
Thank you for your feedback and observations of the 2011 National Household Survey (NHS) data on place of work (commuting flows).

Aboriginal reserves and settlements have always been a challenge for the automated and manual coding processes in place of work. Some have names that are similar, or identical except for numbers associated with them. We are also dependent on the respondent to fill out their place of work information as completely and accurately as possible. We try to do the best we can at getting them coded correctly.

After investigating the data, it appears as if the records in question should probably have been associated with Red Bank 4 and not Chipman. Unfortunately the processes in place for that cycle did not catch this particular issue.

We continue to work on improving our POW coding and verification processes, and hopefully for 2016 and future cycles you will see such improvements reflected in the data.

Questions about CIHI Discharge Abstract Database Coding

Question:
I was reviewing the 2015-2016 Status variables (ST_AT_1 through _20) on the clinical file, checking frequencies from before and after recoding those variables. I discovered that there were a large number of status codes which were missed by my recode syntax. I’ve looked at the online Canadian Coding Standards for ICD-10-CA and CCI (https://secure.cihi.ca/estore/productFamily.htm?pf=PFC2785&lang=en&media=0), the DAD abstracting manual (and for that matter, all of the other documentation files received with DAD), and haven’t been able to find a list of the possible field contents.

Is there a complete list of status attribute codes (I hope that I’ve not simply missed it)?

  • I’m missing meanings for codes: D1,  DX, PA, PB, PC, RA, RB, RC, and UN.
  • I have meanings for codes: 0, B, C, CR, D, DP, DS, E, I , M1, M3, M4, N, N2, N3, N4, N5, N6, P, Q, R, S, S2, TM, U, UR, V, and Z. (e.g., U – Initial consultation)
  • Are there still more status attribute codes that I haven’t encountered yet?
While looking for the intervention status attribute codes, could you find a list of intervention location attribute codes (LC_AT_1 through LC_AT_20) on the clinical file (or let me know where I should have found them)?

I’m not sure where to find explanations of codes FE, FY, JU, LY, ML, MR, PI, RY, SC, and UN.

Answer:
A CCI status code list that was provided by CIHI.  They have indicated it was pulled from Appendix C of the v2015 CCI folio product.

The Folios are essentially electronic, interactive versions of the Manuals so the information contained in them should be the same. Unfortunately we cannot provide the client with a copy of the CCI Folio as it’s a paid product and there are licencing agreements etc. Hospital medical record departments may have a copy they can borrow.

Follow-Up Question:
Would it be possible to put a copy of the Word document that you received from CIHI in the documentation folder for the DAD? It could (should) be annotated to indicate that:

  • the Status list corresponds to variables ST_AT_1, ST_AT_2, …, ST_AT_20 in the DAD excerpt files
  • the Location list corresponds to variables LC_AT_1, LC_AT_2, …, LC_AT_20 in the DAD excerpt files
  • the Extent list does not correspond to any variables in the DAD excerpt files
Follow-Up Answer:
I’ve looked into this further and the ICD-10-CA/CCI has four volumes in total:
  • Volume 1—International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Canada (ICD-10-CA) - Tabular List
  • Volume 2—International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Canada (ICD-10-CA) - Alphabetical Index
  • Volume 3—Canadian Classification of Health Interventions (CCI) - Tabular List
  • Volume 4—Canadian Classification of Health Interventions (CCI) – Alphabetical Index
Typically, for the DLI data cuts, we only include the Tabular lists (Volumes 1 & 3). As to why we only release the Tabular lists, I can’t say -  only that this is what was being done since I’ve been with the team.

Appendix C is provided in Volume 4, as well as appearing in the electronic Folio mentioned earlier, which is why it didn’t turn up in your search.

I just received volume 2 and 4 of the CCI/ICD documentation. They have been placed in the CIHI safe located here:  /MAD_CIHI_ICIS_DAM/Root/discharge-abstract-database-2015-16/Other_documents/ICD10CA CCI v2015/EN/

Tuesday, September 12, 2017

2011 and 2014 Public Service Employees Survey (PSES)

Question
Is it possible to get the Public Service Employees Survey (PSES) files for 2011 and 2014? At the moment the PSES files up to 2008 are available in CHASS.

Response
Thank you for contacting the Data Liberation Initiative (DLI). Public use files for the 2011 and 2014 PSES were not created. The 2008 cycle is the last reference period for which a PUMF was produced: http://www5.statcan.gc.ca/olc-cel/olc.action?ObjId=71M0016X&ObjType=2&lang=en&limit=0.

Public Service Employee Survey, 2014
Aggregate tables presenting the 2014 Public Service Employee Survey results for the entire public service and federal departments are now available upon request. Contact:
(toll-free)1-800-263-1136; or 514-283-8300
infostats@statcan.gc.ca
http://www.statcan.gc.ca/daily-quotidien/150205/dq150205f-eng.htm

The tables are also available on the DLI EFT’s collection to DLI Contacts. They are found here: /MAD_DLI_IDD_DAM/Root/other_autres/4438_PSES_SAFF

DAD File documentation

Question: I’m looking at processing the most recent (2015-2016) DAD files, but I have a question about the documentation.

The files Version 2012 ICD-10-CA Tabular List, Volume 1 PDF (4.6 MB).pdf released with a previous version of DAD, and ICD-10_Vol1_PDF_EN.pdf (released with this version of DAD) both say that they are the ICD-10-CA – tenth revision.
  • Do they have the same content, and are just showing different publishing dates (2012 and 2015, respectively)? I see that I had apparently missed category U (for emergent diseases of unknown cause (e.g., SARS)), and I’ve added a variable to capture that – anything else?
  •  A quick glance through the list didn’t pick up other substantial differences – but I didn’t compare the 1000+ pages of each release in great detail 
Similarly, are there substantive differences in the interventions recorded in the respective CCI classifications? I again did a quick scan: while new interventions may have been introduced, the high-level groups (e.g., Diagnostic Interventions on the Musculoskeletal System (2SA - 2WZ)) still appear to be the same.


Answer: The ICD-10-CA and CCI manuals are revised every 3 years and in this case it looks like your client is comparing two different versions (v2012 and v2015).

With every revision, there will be some changes in the document (new codes, retired codes, new coding methodology for certain conditions etc.), but from my experience, the majority of it will remain the same.

Because there are so many codes, and as your client has found out, a lot of ground to cover, I would recommend that they seek assistance from a medical coding professional (ex/ someone from a hospital medical records department) and have them validate the specific diagnosis codes in question for each year.

Friday, September 8, 2017

CanCHEC

Question
A researcher has expressed interest in the CanCHEC (Canadian Census Health and Environment Cohort) data. She would like to consult the documentation before deciding whether to submit a proposal to the RDC. When I look at the CRDCN website, I am not able to locate any documentation such as the data dictionary, codebook, etc. Any assistance you can provide is most appreciated.

Response
Apparently, there is no codebook for CanCHEC 2001. However, there are currently zero codebook available for CanCHEC 1991 which are usually provided to researchers interested in CanCHEC 2001. As per the SFP variables are roughly the same, except for the CCR – cancer information is not available in the 2001 dataset. However, the user guide is not part of the non-confidential tools located on the social surveys drive so, to my knowledge, it hasn’t been approved for release from the RDC.

Alternatively, researchers can get a comprehensive overview of the 2001 CanCHEC in the Health Analysis Division’s publication found here: http://www.statcan.gc.ca/pub/11-633-x/11-633-x2016003-eng.htm. I would recommend reviewing this in the meantime as it outlines almost everything that is found in the user guide. For variable information, they can use the 1991 CanCHEC non-confidential codebooks (attached to this email).

Friday, September 1, 2017

Data Wealth Distribution in Canada

Question
I have a question from a professor who would like the most recent data for wealth distribution in Canada - she is having trouble finding anything. All I've found is a reference to Survey of Financial Security from StatsCan and some reports.

She also said she would prefer decile although quantile will do if not available.

Response
The Survey of Financial Security would be your best bet. The CANSIM numbers are 205-0002 and 205-0003.

Survey of Financial Security (SFS), composition of assets (including Employer Pension Plans valued on a termination basis) and debts held by all family units, by age group, Canada and provinces, occasional (2012 constant dollars), 1999 to 2012:
Description http://www5.statcan.gc.ca/cansim/a34?lang=eng&mode=tableSummary&id=2050002&stByVal=1&&p1=1&p2=50
Table 205-0002 http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=2050002&tabMode=dataTable&p1=1&p2=50&srchLan=-1

Survey of Financial Security (SFS), composition of assets (including Employer Pension Plans valued on a termination basis) and debts held by all family units, by net worth quintiles, Canada and provinces, occasional (2012 constant dollars), 1999 to 2012:
Description http://www5.statcan.gc.ca/cansim/a34?lang=eng&mode=tableSummary&id=2050003&stByVal=1&&p1=1&p2=50
Table 205-0003 http://www5.statcan.gc.ca/cansim/a26?lang=eng&retrLang=eng&id=2050003&tabMode=dataTable&p1=1&p2=50&srchLan=-1

There could also be analytical reports or articles pertaining to wealth:
http://www5.statcan.gc.ca/olc-cel/olc.action?objId=13F0026M&ObjType=2&lang=en&limit=0
http://www.statcan.gc.ca/daily-quotidien/150603/dq150603b-eng.htm
http://www.statcan.gc.ca/pub/75-006-x/2015001/article/14134-eng.htm

The National Economic Accounts Division produces CANSIM 378-0121, 380-0072: http://www.statcan.gc.ca/daily-quotidien/161214/dq161214c-eng.htm