Focus On Tomorrow
Focus on Tomorrow
ReseaRch funded by WoRksafebc
Who Works Where?
Building Occupational
Research Capacity in the
B.C. Linked Health Database
June 2005
Principal Investigator/Applicant
Dr. Mieke Koehoorn
RS2002/03-DG28
Who Works Where?
Building Occupational Research Capacity
in the BC Linked Health Database
June 30, 2005
Mieke Koehoorn 1 2
Kimberlyn McGrail 3
Kay Teschke 1 2
Jaafar Aghajanian 3
Fan Xu 1 3
1. Department of Health Care & Epidemiology, University of British Columbia, Vancouver BC
2. School of Occupational and Environmental Hygiene, University of British Columbia,
Vancouver, BC
3. Centre for Health Services and Policy Research, University of British Columbia, Vancouver BC
Direct correspondence to:
Mieke Koehoorn
Department of Health Care & Epidemiology
University of British Columbia
5804 Fairview Avenue, Mather Building
Vancouver BC, V6T 1Z3
604-822-5756
Who Works Where? Final Report
Acknowledgements
This project was funded by a Development Grant from the Research Secretariat,
Workers’ Compensation Board of British Columbia. The principal investigator,
Mieke Koehoorn, was supported by a Michael Smith Foundation for Health
Research Scholar Award.
A special thank you to the staff of the Workers’ Compensation Board of British
Columbia and the Centre for Health Services and Policy Research who provided
consultation on various aspects of the WCB employer registry and the BC Linked
Health Database respectively; and to Dennis Wong of Statistics Canada,
Western Region and Northern Territories Office for assistance with the North
American Industry Classification Coding.
2
Who Works Where? Final Report
TABLE OF CONTENTS
Executive
Summary
1. 0 Research Context
1.1
Rich Research Resource.................................................................. …….6
1.2
Centre for Health Services and Policy Research………………………..6
1.3 Research
Problem……………………………………………………………..7
1.4 Study
Purpose…………………………………………………………………..8
2.0 Methodology
2.1 Overview……………………………………………………………………….10
2.2
Premium Provider Codes……………………………………………..…….10
2.3
WCB of BC Industry Classification Codes…………….………………….11
2.4 Linkage
Strategy……………………………………………………………..12
3.0 Results
3.1 Key
Findings…………………………………………………………………………...21
3.2
Statistics Canada Results…………………………………………………………...22
3.3 Workforce
Coverage………………………………………………………………..22
4.0
Implications for Future Research on Occupational Health
4.1
Operating
Grants…………………………………………………………….24
4.2
Population Health and Learning Observatory…………………………24
4.3
WCB-CHSPR
Partnership…………………………………………………….25
5.0 Dissemination/Knowledge
Transfer
5.1 Occupational and Population Health Researchers…………………..26
5.2
WCB of BC Interactions……………………………………………………..27
6.0
Future Steps……………………………………………………………………………28
References…………………………………………………………………………………….30
3
Who Works Where? Final Report
EXECUTIVE SUMMARY – Key Points
The British Columbia Linked Health Database (BCLHD) is among the
richest data resources in the world for population health and health
services research;
This project developed an industry of employment variable in the
BCLHD to enhance the capacity of the database for future
research projects on the relationships between work and health;
For the 2001 BCLHD data file, we were able to assign an industry
code to 13,578 employers (86% success rate at the employer-level)
representing 908,430 workers (91% success rate at the employee-
level). (Note Approximately half of the BC workforce can be
identified as actively employed via variables in the BCLHD. Overall,
our success rate at the population level means there is an industry
code available for 47% of the BC workforce in the 2001 BCLHD data
file)
The industry of employment code is now available in the BCLHD
data holdings for researchers to include in population and
workforce studies.
4
Who Works Where? Final Report
EXECUTIVE SUMMARY
Introduction: The Centre for Health Services and Policy Research (CHSPR) at the
University of British Columbia is home to the British Columbia Linked Health Database
(BCLHD). The BCLHD is among the richest data resources in the world for population
health and health services research. It includes longitudinal, person-specific, health
data on the province’s four million residents, covering the period 1985 forward. The
data available includes a registration file of British Columbia (BC) residents, records
of services from physicians and alternative providers, hospitalizations, vital statistics,
cancer incidence, mental health services, pharmaceutical prescriptions, and
workers’ compensation claims. The purpose of this project was to create an
industry of employment variable in the BCLHD to enhance the capacity of the
database for future research projects on the relationships between work and health
at a population workforce level.
Methods: A ‘crosswalk’ was developed between an employer field (representing
the payer of the provincial health premium) in the BCLHD registration file and an
industry-coded employer field in the Workers’ Compensation Board (WCB) registry.
Both computer programming (deterministic and probabilistic linkages) and manual
searches (e.g. key words) were used to create the crosswalk. The crosswalk was
then applied to the 2001 registration file of BC residents to assign an industry of
employment code to workers. Overall, approximately 52% of the BC workforce can
be identified in the BCLHD using the employer paid health premium field. Not all
workers have an employer paid health premium and are therefore not ‘identifiable’
as a worker in the database.
Results: For 2001, there were 15,833 unique employers in the registration file
providing health premiums for 1,000,130 workers. Of these, we were able to find a
match with the WCB registry for 13,578 employers (86% success rate at the
employer-level) and assign an industry code to 908,430 BC workers in the
registration file (91% success rate at the employee-level for all employees with an
employer paid health premium). In summary, we are able to identify 52% of the BC
workforce in the 2001 BCLHD using employer-paid health premiums and assign an
industry of employment code to 47% of the BC workforce at the population level.
Conclusion: With the addition of industry codes, the BCLHD provides enhanced
capacity for investigating questions on the relationships between work and health
and on labour market and workers’ compensation experiences at the population
level. In particular, it enables surveillance of disease and injury by industry of
employment at the population workforce level using a variety of health and health
services data. We recommend the use of the employment and industry variables in
the BCLHD database for research purposes, recognizing that some employee
groups may be underrepresented by these variables.
5
Who Works Where? Final Report
WHO WORKS WHERE?
1. 0 Research Context
1.1 Rich Research Resource
The Centre for Health Services and Policy Research (CHSPR) at the
University of British Columbia is home to the British Columbia Linked Health
Database (BCLHD). The BCLHD is among the richest data resources in the world
for population health and health services research. It includes longitudinal,
person-specific, health data on the province’s four million residents, covering
the period from 1985 forward. The data available includes a registration file of
British Columbia (BC) residents, records of service from physicians and
alternative providers, hospitalizations, vital statistics, cancer incidence, mental
health services, pharmaceutical prescriptions, and workers’ compensation
claims.
1.2
Centre for Health Services and Policy Research
CHSPR has over 30 years of experience in handling data from the BC
Ministry of Health Services (MoHS) and other professional bodies, and now acts
as the access point for researchers wishing to use these data for research in the
public interest. CHSPR has comprehensive policies and procedures in place to
protect the confidentiality and security of data holdings. The foremost priority is
to respect the privacy of individual users and providers of health care and other
care services, and to protect data against loss, destruction, or unauthorized use,
6
Who Works Where? Final Report
while making data available to researchers to answer questions on the health of
British Columbians and the delivery of health services in the province.
CHSPR ensures data confidentiality and security in two main ways. The first
is by limiting access to personal identifying information and ensuring that data is
used for appropriate research purposes. The cornerstone of data protection at
CHSPR is an Access Policy (1), which has been successfully used to facilitate
research in the public interest using linkable data since 1996. The Access Policy
was developed to conform to the BC Freedom of Information and Protection of
Privacy Act. The second approach to ensuring data confidentiality and security
is through best practices in data protection. All data at CHSPR are managed in
a manner that is consistent with evolving international best practices. CHSPR’s
physical workspace is tightly controlled, and the security of data is further
ensured with sophisticated technical safeguards. Please see CHSPR’s web site
for further details on policies regarding privacy and security
(http://www.chspr.ubc.ca/privacy.htm).
1.3 Research
Problem
Prior to this project, there was limited data in the BCLHD that indicated
employment status, occupation or industry of employment of BC residents. This
data was only available for BC residents who incurred a workers’ compensation
claim, a small subset of the population and one defined by disability. The lack
of data on employment and employment characteristics limited the capacity of
7
Who Works Where? Final Report
the BCLHD for occupational and population workforce research and the ability
to answer important questions on occupational injury and disease, and on work
health programs and policies.
1.4
Study Purpose
The overall purpose of this development grant was to create an industry
of employment variable in the registration file of the BCLHD for residents in the
BC workforce by assigning an industry code to the employers who pay health
premiums; and to tabulate the proportion of the workforce whose industry of
employment is assigned using this technique.
The addition of an industry of employment variable to the BCLHD offers
the potential of vastly improved methods of occupational health research,
including: improved denominator data for calculating incidence and
prevalence rates of injuries and diseases by industry, stratified by demographic,
geographic and socioeconomic variables; enhanced capacity for surveillance
of injury and disease by industry across numerous databases that capture
multiple health outcomes; enhanced capacity for defining study and
comparison populations for occupational health studies; and the opportunity to
investigate questions about the relationship between work and health, labour
market experiences and workers’ compensation experiences at the population
health level. Ultimately the enhancement of the BCLHD is expected to
generate more opportunities for examining the health of British Columbia
8
Who Works Where? Final Report
workers, and in turn provide evidence for effective planning, implementation
and evaluation of workplace policies, legislation and prevention strategies.
Population-based, health research databases such as the BCLHD are
unique in the world, and are made possible because of Canada’s
comprehensive, public health care systems and research sensitive privacy
legislation. Similar databases are available in Manitoba, Quebec and Nova
Scotia. However, this will be the first database to add employment
characteristics at a population workforce level.
9
Who Works Where? Final Report
2.0 Methodology
2.1 Overview
A ‘crosswalk’ was developed between the employer field (representing
the payer of the provincial health premium) in the BCLHD registration file and an
industry-coded employer field (a ‘classification unit’ or CU) in the Workers’
Compensation Board (WCB) business registry. Approaches based on both
computer programming and manual review were used to create the crosswalk
between the two registries. Additional industry codes were added using
employer coding from Statistics Canada. The protocol was approved by the
Behavioural Research Ethics Committee of UBC (#B03-0579 03-1306).
The end product was a data matrix of unique codes for employers paying
health premiums linked to WCB of BC industry code(s) that can be applied to
individuals in the registration file whose employer pays their health premium via
their specific employer-paid premium provider code.
2.2
Premium Provider Codes
In British Columbia, al residents, with a few exceptions (e.g. federal
inmates, RCMP, foreign students) are covered by the provincial health care
system and pay a health care premium for their universal health care benefits.
The payer of the health premium is assigned a code (premium provider code or
PPC) and recorded for everyone in the registration file of the BCLHD. For many
residents this is their employer. In the remaining cases, premiums are self-paid,
10
Who Works Where? Final Report
paid by a spouse’s or parent’s employer, or paid by the government (e.g.
federal employees, individuals on social assistance, and those with native
status). The Ministry of Health Services provided a list of the health premium
provider codes and corresponding employer names effective December 31st,
2001 to the research team for this project. This file contained a total of 15,833
unique employer-paid premium provider codes.
The employer-paid health premium provider code allows for the
construction of an industry of employment code but not an occupation code.
An initiative is underway between the Workers’ Compensation Board of British
Columbia and the BC Ministry of Health Services to investigate recording
occupation data as part of medical services billing records, but this is not part of
this development grant.
2.3
WCB of BC Industry Classification Codes
The WCB of BC industry classification and coding system is used to group
employers into similar industries with similar work activities with similar levels of risk
for occupational injury and disease. By grouping employers in this way, the WCB
of BC is able to set equitable rates for workers’ compensation insurance.
Industrial activities are organized into a hierarchical structure consisting of
sectors (the greatest degree of aggregation), subsectors, and classification units
or CUs (the least degree of aggregation). It includes seven sectors, 24
subsectors, and approximately 600 classification units. The classification unit is
11
Who Works Where? Final Report
identified by a six-digit classification unit code. For example, ‘Residential Social
Service Facility’ is a unique classification unit (#766017) nested within the
subsector ‘Health Care and Social Assistance’, nested with the ‘Service Sector’.
Typically, each employer is assigned to the CU that best describes its primary
business activity. If an employer is engaged in two or more distinct activities, it is
assigned to two or more CUs, although this represents a small proportion of
employers (less than 3.5%). For more information and to view the industrial
classification system please refer to:
http://www.worksafebc.com/for_employers/premiums/classification/default.asp
Permission was obtained from the WCB of BC to receive a copy of their
registry of employer names and industry classification codes as of December
31st, 2001. This data file contained 181,085 records.
As a development grant, the work to create a crosswalk was limited to
one year (2001) to investigate the feasibility of creating an industry code. Based
on the results, the crosswalk can be used for the matched employers in other
years of the BCLHD registration files and additional work undertaken to identify
new (2004 forward) and in-active employers (1985 to 2000).
2.4
Linkage Strategy
2.4.1 Deterministic Programming
Several computer programming queries in Visual FoxPro 9.0 (2) and SAS (3)
were used to match employer names in the Ministry of Health Services (MoHS)
12
Who Works Where? Final Report
data file with names in the WCB of BC data file. The first was a ‘perfect’
matching query requiring the name in one file to be identical to the name in the
other file (Table 1) after removing ‘low’ weighted words such as ‘The’, ‘Inc’ or
‘BC’, and punctuation such as periods or hyphens from both lists.
Table 1: Examples of ‘Perfect’ Matches
MoHS Employer Name
WCB of BC Employer Name
Otis Canada Ltd
Otis Canada Ltd
The Landing Services Ltd
Landing Services
Nineties Sparerib House
Nineties Spare-rib House
2.4.2 FoxPro Programming
The second strategy employed a set of programming rules to identify matches
(Table 2) between the two lists excluding the perfect matches identified in the
first strategy. Removing low weighted words such as ‘The’ or ‘Ltd’, each word of
an employer name in the MoHS list was compared with each word of a name in
the WCB list and vice versa. A weight of ‘1’ was assigned to a word in the WCB
employer name if the whole word or the first 7 characters of the word matched
a whole word or the first 7 characters of a word in an MoHS employer name,
and vice versa. A weight of ‘2’ was assigned to a word in the WCB employer
name if the first 4 characters matched the first 4 characters of a word in a MoHS
employer name, and vice versa. A ‘0’ weight was given to words that did not
match on at least the first four characters. In this manner a string of weights was
generated for each name in the WCB list and the MoHS list (e.g. 1111 or 0210).
13
Who Works Where? Final Report
Based on visual inspection of a random sample of 200 programming matches,
we excluded matches where both the WCB string and the MoHS string included
at least one ‘0’ weight. We also excluded matches where less than half of both
strings were weighted with ‘1s’ (i.e. excluded ‘1222’ or ‘2’ or ‘02112’).
Table 2: Examples of Included Programming Matches
MoHS Employer Name (Weighted String)
WCB of BC Employer Name (Weighted String)
Examples where all whole words, first 7 digits or first 4 digits in one list match in the other list
Aaronson’s Pharmacy (11)
Aaronson’s Pharmacy Cook Street (1100)
Lufthansa (1)
Deutsche Lufthansa Aktiengesel schaft (010)
Brenda Mines Noranda (110)
Brenda Mines (11)
Pontiac Buick GMC– Kal ewski’s (1112)
Kallawsky Pontiac Buick GMC (2111)
Canadian Sulzer Pumps/Pompes Sulzer (21100) Sulzer Pumps Canada (112)
Chris Schenk MD (110)
Dr Christian Schenk (021)
Township of North Cowichin (211)
North Cowichan Town (111)
Kelowna Coop Assoc (111)
Kelowna Cooperative Association (122)
We investigated rerunning the programming strategy with the remaining
unmatched names allowing for weights of ‘3’ where only the first 3 characters of
a word in the WCB employer name matched the first 3 characters of a word in
the MoHS list and vice versa. However, a visual inspection of a random sample
of 10% of matches indicated a high rate of error (more than 30% mismatches)
and we eliminated this as a matching strategy.
14
Who Works Where? Final Report
2.4.3 SAS Programming
With the remaining unmatched list we used an alternative programming
strategy to find matches. The SAS SPEDIS function computes a ‘spelling
distance’ between characters in words. If the two words are identical, the
spelling distance is 0. A higher scores indicates that words are more dissimilar.
For each type of spelling error, SPEDIS assigns penalty points. For example, if the
first letters of the two words do not match, there is a relatively large penalty. If
two letters are reversed (e.g. dr verses rd) there is a smaller penalty. The final
spelling distance is also based on the length of the words being matched. A
wrong letter in a long word results in a smaller score than a wrong letter in a
shorter word. See Cody (5) for a list of SPEDIS operations and their penalties.
Table 3: Examples of SAS Programming Matches
MoHS Employer Name (SPEDIS Score)
WCB of BC Employer Name (SPEDIS SCORE)
Examples of Included Matches
E Archie Strimbold (27)
E A Strimbold (19)
Dicks Lumber & Bldg (30)
Dicks Lumber And Building Supplies (19)
Hetherington & Hooper Stores (1969) (26)
Hetherington Hooper (12)
Examples Of Excluded Matches
J Reid & Associates Ltd (25)
Dr J B Reid (22)
Checktech Services Ltd (26)
Caltech Tech Services (23)
Echo 8 Enterprises Inc (32).
Eh-Cho-Dene Enterprises (27)
15
Who Works Where? Final Report
Based on a review of a random 10% sample of matches, we included only those
matches were the SPEDIS Score was 20 or less for either the MoHS name or the
WCB of BC name.
2.4.3 Key Word Matching
Next, a ‘key word’ strategy was used to assign CUs to the remaining unmatched
employer names in the MoHS list. Using the total list of employer names
provided by the WCB of BC, frequencies of CUs were generated for 183 key
words representing different types of work activities, such as ‘chiropractor’ or
‘insurance’. The complete list of key words is provided in Appendix I. For each
key word with at least 20 employers, we tabulated the frequency of employers
with the identical CU. Where the same CU was assigned to 90% or more of the
employers, all employers in the MoHS list with that key word in their name were
assigned the same CU. For example, if 40 employers in the WCB of BC list had
the key business word ‘Chiropractor’ and 95% of these had the same CU, this
key word(s) was included in our matching strategy and the CU applied with
confidence to all remaining unmatched employers with the word ‘Chiropractor’
in the MoHS list (Table 3). We selected 20 employers and 90% CU frequencies to
try to minimize misclassification of industry coding in the matrix and subsequently
in the BCLHD registration file.
A total of 24 key business words met the inclusion criteria for application of
WCB CUs to the MoHS list (these are highlighted in bold in Appendix I).
16
Who Works Where? Final Report
Table 3:
Example of Key Word Strategy
Key Business Word(s)
Distribution of CUs
Key Words Included in Matching Strategy
Restaurant
1026 of 1084 employers with same CU (94.7%)
Day Care
96 of 96 employers with same CU (100%)
Taxi or Cabs
69 of 70 employers with same CU (98.6%)
Lawyers or Barristers
941 of 954 employers with the same CU (98.6%)
Pawn Broker
13 of 14 employers with the same CU (93%)
Key Words Excluded from Matching Strategy
College
107 of 179 employers with the same CU (59.8%)
Forest Products
60 of 244 employers with the same CU (24%)
Beef Farm
2 of 8 employers with the same CU (25%)
Jewelry
7 of 10 employers with the same CU (70%)
2.4.4 Manual Matching
A manual search was completed by the principal investigator (Koehoorn)
to identify any unresolved matches between the two lists based on general
knowledge of changes in business names over time or similar names that did not
meet the computer programming criteria but were deemed to be the same
employer upon further inspection (Table 4). This was completed by searching
the WCB list of employers for proper names such as ‘Hudson Bay’ or ‘Norske’, for
example, for all remaining unmatched MoHS employer names.
17
Who Works Where? Final Report
Table 4: Examples of Manual Matches
MoHS Employer Name
WCB of BC Employer Name
The Bay
The Hudson Bay Company
BCHydro
BC Hydro and Power Corp
Fletcher Challenge
Norske Skog Canada
Majdanski Chartered Accountants
Majdanski, MacLean and Associates CAs
Dr Aindrias O’Breasil Inc
A OBreasil MD
2.4.5 Statistics Canada Matching
Approval was obtained from Statistics Canada (Dennis Wong, January
2004) to submit a sample of unmatched employer names from the MoHS list for
linkage with their business registry, which includes a North American Industry
Classification System (NAICS) code, in an effort to increase the number of
employers with an industry code in the crosswalk and to confirm if this is a
feasible alternative for future industry coding. NAICS is an industry classification
system developed jointly by Statistics Canada, the US Economic Classification
Policy Committee and Mexico’s Instituto Nacional de Estadistica Geografia e
Informatica. NAICS is designed to provide common definitions of the industrial
structure of the three countries and a common statistical framework to facilitate
the analysis and comparison of labour statistics in North America (4). NAICS-
Canada coding is also hierarchical, with 20 sectors, 99 sub-sectors, and 921
national industries. The NAICS code in the Stats Can registry is assigned by
18
Who Works Where? Final Report
industrial activity and provides both a valid matching strategy and industry
code for our sample of unmatched business names.
Employer names were selected for submission to Statistics Canada for
industry coding on the basis of the number of residents by employer in the client
registry. Almost 2,000 of the largest unmatched employers were selected to see
if Statistics Canada could assign an industry code (NAICS) by matching with our
employer name via their business registry.
Finally, the principal investigator reviewed all of the matches across all of
the strategies and eliminated errors to arrive at the final list of approved
matches.
The industry matrix is held solely under the auspices of the Centre for
Health Services and Policy Research. Only the industry of employment code is
released to researchers, under the same protocols that govern the use of all
other variables in the BCLHD. The premium provider code and the employer
name are not part of the data holdings available for research purposes and will
not be released to researchers.
19
Who Works Where? Final Report
Figure 1 below provides an overview of the methodology.
MoHS
WCB
PPC
Employer Name
Employer Name
CU
Matching Processes
BCLHD Industry of Employment Matrix
PPC =767676
CU= 762022
CU added to residents in the BCLHD
BCLHD Registration File
registration file via PPC code match
PPC
BC Resident
CU
767676
Jane Doe
762022
Figure 1: Overview of Study Methods
20
Who Works Where? Final Report
3.0 Results
3.1
Summary of Key Findings
Approximately half of all employer names in the MoHS registry were
successfully assigned an industry code by a ‘perfect’ match with an employer
name in the WCB of BC registry (Table 5). This was followed by programming
matches (17.2%), key word matches (7.0%), manual matches (5.4%) and
Statistics Canada matches (5.3%). Just over 14% of the MoHS employer list
remained unmatched without an assigned industry code.
Table 5: Matching Results between MoHS and WCB of BC lists of employers
Matching Strategy
Number of Matches (% of Total
Employers or unique PPCs in
the MoHS registration file)
Perfect Matches*
8,069 (50.96%)
FoxPro Programming
2,474 (15.57%)
Matches*
2724 (17.21%)
SAS Programming
250 (1.56%)
Matches*
Key Word*
1,100 (6.95%)
Manual* 851
(5.38%)
Statistics Canada † 834
(5.27%)
Total Matches (% Employers with 13,578
(85.76%)
an Assigned Industry Code)
Total Non Matches (% Employers
2,255
(14.24%)
without an Industry Code)
Total Unique Employers
15,833 (100%)
* Assigned an WCB of BC Industry Classification Unit (CU) code
†Assigned a Statistics Canada North American Industry Classification System (NAICS) code only (No WCB of BC CU).
Statistics Canada was able to find 1,630 of our sample of unmatched 1,928 employers (86%) and assign a NAICS code.
After completing our key word and manual matching strategies, only 838 remained unmatched and we used the NAICS
code as a valid industry code in lieu of the WCB of BC CU code for these employers.
21
Who Works Where? Final Report
In summary, a total of 13,578 of 15,833 (86%) unique employers (a.k.a.
health premium providers) in the 2001 BCLHD registration file were assigned an
industry code (i.e. WCB of BC industry classification unit or Statistics Canada
NAICS code). A matrix of industry codes linkable to individuals via the employer
premium provider codes is now available for research purposes as part of the
BCLHD registration file, linkable to all other health data holdings.
3.3
Visual Inspection of Matches
A visual inspection of all perfect and programming matches by the
principal investigator resulted in the exclusion of 43 ‘perfect’ matches (0.5%),
arriving at the 8,070 figure in Table 5; and 160 programming matches (5.5%)
arriving at the final 2,728 figure in Table 5. Visual inspection was not appropriate
for key word and Statistics Canada matches (i.e. no WCB name to compare
with) or manual matches (i.e. already inspected by the PI).
4.0 Workforce
Coverage
In order to assess the assignment of industry codes to BC workers, the
matrix of industry codes was applied to BC residents in the 2001 BCLHD
registration file by premium provider code. In order to eliminate
misclassification, industry codes were only assigned to the person employed in
the industry (the ‘primary’) and not to ‘dependents’ such as spouses and
children, who can have their health benefits covered by a family member’s
22
Who Works Where? Final Report
employer. In the 2001 registration file, a total of 1,000,130 individuals with
primary status had an employer-paid premium provider code. Of these, 908,430
were assigned an industry of employment code (90.8% coverage).
According to Census statistics generated by BC Stats, there were a total
of 1,922,100 BC residents employed in the workforce in 2001 (4); whereas we can
identify as employed 1,000,130 individuals in the BCLHD registration file based on
a combination of the employer premium provider code and ‘primary’ status
(52.0% of the workforce). Industry of employment was assigned to 908,430 of
these representing 47.3% of the BC workforce. We were able to identify only
approximately 50% of the workforce in the BCLHD registration file for a few
reasons. First, ‘dependents’ (according to a registration file based on
responsibility for payment) were excluded from our analysis but may be in the
workforce. Second, because identification of an employer is based on that
employer paying for health care premiums, we will not capture employees of
smaller firms. Third, self-employed individuals are responsible for paying their
health premiums and are not identifiable in the registration by an employer-paid
premium. For these reasons, we hypothesize that industry of employment
coverage is lower for women, self employed individuals, and workers in small
worksites (See Section 6.0 for future steps to assess workforce statistics). Of those
we can identify as employed in the BC workforce, we are able to assign an
industry code to the vast majority.
23
Who Works Where? Final Report
4.0
Implications for Future Research on Occupational Health
4.1 Operating
Grants
Three research projects are either underway or proposed that will request
access to the industry of employment variable and will serve as pilot tests for the
new variable. In particular, the industry of employment variable is being used to
identify a cohort of workers in heavy industry for a study of work-related back
injury (Teschke & Koehoorn; WCB of BC funded project), and as a co-variate in a
multi-factor environmental study of cardiovascular disease associated with air
pollution in the Georgia Air Basin (Brauer, Demers and Koehoorn et al; Health
Canada and BC Centre for Disease Control funded project). It will also be used
in a study of asthma surveillance by industry of employment in British Columbia
(Koehoorn, Kennedy, McLeod, Bogyo; AllerGen NCE – CIHR funded project).
Pending recommendations from researchers on the refinement of the industry
variable, it will be made available with documentation as part of the regular
BCLHD holdings for Canadian researchers with approved research projects.
4.2
Population Health and Learning Observatory
Eventually, the BCLHD including the registration file with the industry codes
will become a component of the Population Health and Learning Observatory
(PHLO). PHLO is an umbrella organization funded by the Canada Foundation
for Innovation to expand and develop data resources and create crosswalks
between databases, to support population health research. PHLO wil broaden
24
Who Works Where? Final Report
the array of variables available for research and integrate them in a way that
will allow for innovative questions on occupational injury and disease, including
multi-factor studies (e.g. occupational exposures), detailed health outcomes
(combination of drug claims, outpatient visits, hospitalization, health registries,
self-reported surveys and deaths) and life course trajectories (from school, to
work to retirement). As such, the industry of employment variable will be a
cornerstone of the occupational research component of the Observatory for
the identification of study populations, the assignment of exposures, the
surveillance of health over the life course, and as an explanatory variable in
multi-factorial studies of health outcomes.
4.3
WCB-CHSPR Research Partnership
The industry variable will also be a focus for research projects under the
Partnership recently formed between the Workers' Compensation Board of British
Columbia (WCB) and the Centre for Health Services and Policy Research at
UBC. The intent is to conduct research that maximizes the use of the BCLHD for
evidence-based policies and prevention programs. The goal is to create safe
and secure workplaces and to provide the best opportunity for recovery after
injuries occur. For example, one project proposes to add health maps by
industry of employment to the BC Health Atlas for surveillance of injuries and
diseases (www.chspr.ubc.ca/cgi-bin/pub).
5.0
Dissemination/Knowledge Transfer and Policy Implications
25
Who Works Where? Final Report
5.1
Occupational and Population Health Researchers
The primary audiences for the industry variable in the BCLHD are
researchers and those interested in research findings. To reach these
audiences, several presentations have been completed and abstracts
submitted to local, national and international conferences in an effort to
communicate the findings from the Who Works Where? project and the
potential of the industry variable in the BCLHD for occupational and population
workforce research, as follows:
Koehoorn M. Who Works Where? Population Health and Learning Observatory
Seminar Series. October 21st, 2004. University of British Columbia,
Vancouver, BC.
Koehoorn M. Occupational Research using the BCLHD. OCCH 540 Research
Seminar. School of Occupational and Environmental Hygiene.
November, 2004. University of British Columbia, Vancouver BC.
Koehoorn M, McGrail K, Teschke K, Aghajanian A. Who Works Where?
Occupational Injury and Disease Surveil ance in British Columbia. 18th
International Symposium on Epidemiology in Occupational Health
(accepted). Bergen, Norway, September 11-14, 2005.
Koehoorn M, Aghajanian J, McGrail K, Teschke K, McLeod CB. Building Research
Capacity to Investigate Questions on Place and Health. Canadian Public
Health Association 96th Annual Conference (accepted). Ottawa, ON.
September 18-21, 2005.
The project has also been highlighted as part of presentations to the Workers’
Compensation Board on the partnership, including the Senior Executive
Committee. The researchers will continue to present on the Who Works Where?
project as requested.
26
Who Works Where? Final Report
5.1.1 BC Linked Health Database
Pending further pilot studies (as outlined above) and validation work (as
outlined below under Future Steps), the industry variable will be incorporated
into the regular BCLHD data holdings and application process, and available for
research purposes, with plans to include information on the variable on the
CSHPR, PHLO and WCB-CHSPR Research Partnership websites.
5.2
WCB of BC Interactions
Knowledge translation research has identified the development of
partnerships, based on regular exchange and contact, as the most effective
approach in facilitating the use of research evidence and in creating a
research-attuned culture among decision makers. Consistent with this
approach, the WCB-CHSPR Research Partnership meets every other month and,
in this way, senior WCB decision makers are engaged in the research process
from study design to implementation of findings. Surveillance projects using the
industry variable in the BCLHD as part of a future BC Health Atlas have been
identified as a priority for the WCB under this Partnership. Relationships with
senior policy makers and leaders in the compensation field are well-established,
and the ‘receptor’ capacity for the findings and mechanisms for interchange
and policy/program development are underway.
Findings from this project will also be posted on a public website
(www.chspr.ubc.ca/research/wcbpartnership.htm) that has been created to
27
Who Works Where? Final Report
profile the WCB-CHSPR partnership and to present the results of research
projects to a broader audience. Knowledge translation will also be facilitated
through our active collaboration with the UBC Centre for Health and
Environment Research, which has dedicated research personnel and resources
focusing specifically on knowledge translation in occupational health settings
(www.cher.ubc.ca).
6.0 Future
Steps
The research team is continuing to develop the industry variable in the
BCLHD under the WCB-CHSPR Research Partnership. There are four key
objectives for future work;
1. To generate descriptive statistics about the BC workforce in the 2001
registration file by demographic, socioeconomic, and geographic variables to
assess the coverage of the industry variable relative to Census and Labour Force
Survey statistics data. This step requires the submission of a Data Access Request
to the Ministry of Health Services to access the registration file at the individual
level, along with ethics approval at UBC (to be completed summer of 2005).
Findings from this step will be incorporated into data documentation for
researchers accessing the industry variable for approved projects to
communicate the limitations of the variable (e.g. under representation of self-
employed, certain industries and women)
28
Who Works Where? Final Report
2.
To improve upon the success rate by continuing with manual
matches between the MoHS and WCB of BC employer lists. We will also attempt
to assign industry codes to individuals in the registration file without an employer-
paid premium provider code or those with an unmatched employer name, by
using industry information from the WCB of BC database in the BCLHD for the
subset of the population with an accepted claim. This may help improve
assignment of industry codes for female workers and workers in small
workplaces.
3.
To continue to update the industry variable for new employers on
an ongoing basis and expand the coding beyond 2001. Employer paid
premium provider codes are associated with start and end dates of coverage
for the period 1985 to 2003 in BCLHD. It is therefore feasible to use our crosswalk
to retrospectively assign industry codes to individuals by PPC using start and end
dates.
4.
To investigate funding and methodological opportunities to assign
the NAICS code to employers and ultimately employees in the BCLHD via
Statistics Canada registry. Based on our success rate with Statistics Canada and
to improve comparability of research findings across jurisdictions, we have
submitted an operating grant to the Canadian Institutes for Health Research
(Koehoorn, June 2005) that proposes to link the entire MoHS employer list to the
Statistics Canada registry for NAICS coding.
29
Who Works Where? Final Report
REFERENCES
1. British Columbia Ministry of Health and Centre for Health Services and Policy Research
(1996). Access Policy for Research Uses of Linked Health Data. CHSPR: Vancouver, BC
(HIDU Report 1996-1).
2. Microsoft Visual Foxpro 9.0 (2005). Redmond, Washington.
3. Statistics for the Applied Sciences. SAS Procedures Manual. Cary, NC: SAS Institute Inc.
2002.
4. Statistics Canada (2003). North American Industry Classification System (Canada).
Ottawa, ON: Government of Canada.
http://www.statcan.ca/english/Subjects/Standard/naics/1997/ naics97-intro.htm.
Accessed March 31, 2003.
5. Cody R. (2004). SAS Functions by Example. Cary, NC: SAS Institute Inc. pp 84-88.
30
Who Works Where? Final Report
APPENDIX 1
1. ADVERTISING
2. AFTER HOURS SCHOOL
3. ANIMAL OR VETERINARY OR PET HOSPITAL
Bold indicates key words were
4. APPLIANCE(S)
included in the matching strategy
5. AQUA TOWING
6. ARTS ASSOCIATION (E.G. DRAMA)
based on the distribution of CUs
7. AUTO BROKER(S)
associated with employers with the
8. AUTO ELECTRIC
9. AUTO GLASS
key words in their name (>90% of
10. AUTOBODY OR AUTO BODY
employers with the key word had the
11. AUTOMOTIVE
12. AVIATION
same CU)
13. BAKERY
14. BANK
15. BANK BROKER(S) OR LOAN BROKER(S)
16. BARRISTERSOR LAWYER
17. BROKER (NOT BANK, AUTO OR FOOD)
18. BUILDING MAINTENANCE
19. BUSINESS OR MERCHANT ASSOCIATION
20. CAB(S) OR TAXICAB(S) OR TAXI(S)
21. CARPET CLEANER(S)
22. CARPET CLEANING
23. CATERING
24. CATHOLIC OR CHRISTIAN OR INDEPENDENT
SCHOOL
25. CHIROPRACTIC OR CHIROPRACTOR(S)
26. CHRISTIAN OR CHURCH ASSOCATION
27. CHURCH
28. CLEANERS (NOT CARPET)
29. COLLEGE
30. COLLEGE AND HEALTH PROFESSIONAL (E.G.
CHIROPRACTIC)
31. COMMUNICATIONS (NOT TELE)
32. COMMUNITY ASSOCIATION
33. COMMUNITY SCHOOL
34. COMPUTER(S)
35. CONSTRUCTION
36. CONSULTING OR CONSULTANT(S)
37. CONTRACTING
38. COURIER(S)
39. CREDIT UNION
40. CULINARY SCHOOL
41. DAY CARE
42. DDS
43. DELICATESSEN OR DELI
44. DENTAL
45. DENTAL ASSOCIATION
46. DENTAL CLINIC
47. DENTAL PRODUCTS OR DENTAL SERVICES
48. DEVELOPMENTS
49. DINER
50. DISTRIBUTOR OR DISTRIBUTING
51. DR OR DRS OR DOCTOR(S)
52. DRILLING
53. DRIVING SCHOOL
54. DRYWALL
55. ECLECTRONICS
56. ELECTRIC
57. ENGINEERING OR ENGINEER(S)
58. ENGINEERING MANAGEMENT
59. ENGLISH COLLEGE
60. ENTERPRISES
61. EXCAVATING
62. EXCAVATING OR EXCAVATION
63. FLOWERS
64. FOOD BROKER(S)
2
Who Works Where? Final Report
65. FOOD IMPORTERS
127. PENSION PLANS
66. FOODS
128. PHARMACEUTICALS
67. FOREST OR FORESTRY
129. PHARMACY OR DRUGS
68. FOREST PRODUCTS
130. PLUMBING
69. FURNITURE
131. PLUMBING AND ELECTRICAL
70. FURNITURE DELIVERY
132. PLUMBING AND HEATING
71. FURNITURE MANUFACTURING
133. PODIATRIC
72. GARAGE DOOR(S)
134. POOL MAINTENANCE
73. GARAGE OR SERVICE STATION
135. PRESCHOOL OR NURSERY SCHOOL OR PLAY
74. GARDENING
SCHOOL
75. GLASS
136. PRINTING
76. GLASS BLOWING OR GLASS WORKS OR GLASS
137. PRIVATE HOSPITAL
WARE
138. PROPERTIES
77. GOLF ASSOCIATION
139. PUB OR BAR
78. GOLF COURSE OR GOLF LINKS OR GOLF CLUB
140. PUBLISHING
79. GOLF SALES OR GOLF SHOP
141. REAL ESTATE
80. GRAPHICS
142. REALTY
81. GYMNASTICS OR KARATE OR TAEKWONDO OR
143. RECREATION ASSOCIATION (E.G. HOCKEY)
BIKING OR SOCCER OR HOCKEY OR KAYAKING OR
144. REFORESTATION
SAILING OR BASKETBALL SCHOOL
145. RESTAURANT
82. HAIR OR HAIRDRESSING OR COSMETOLOGY OR
146. RESTAURANT ASSOCIATION
BEAUTY
147. RESTAURANT SUPPLIES
83. HEALTH REGION OR HEALTH AUTHORITY
148. RESTAURANT AND MARINA
84. HELICOPETER OR AIRCRAFT MAINTENANCE
149. ROAD OR HIGHWAY OR PAVEMENT MAINTENANCE
85. HELICOPTERS
150. SCHOOL
86. HOLDINGS
151. SCHOOL ASSOCIATION
87. HOSPITAL AUXILLARY
152. SCHOOL BOARD
88. HOSPITAL LAUNDRY
153. SCHOOL DISTRICT
89. HOSPITALITY
154. SCHOOL SOCIETY
90. HOSPTIAL FOUNDATION
155. SCREEN PRINING
91. HOSPTIAL(S)
156. SEAFOODS
92. HOTEL
157. SHIPYARD(S)
93. HOTEL ASSOCIATION(S)
158. SOFTWARE
94. HOTEL SCHOOL
159. TECHNOLOGIES
95. HOUSING ASSOCIATION OR CO-OP
160. TELECOMMUNICATIONS
96. HYDRO
161. TIRE
97. IMPORTERS
162. TOURISM OR TOURIST
98. INN
163. TOWING
99. INSURANCE
164. TRADING
100. INSURANCE BROKER
165. TRANSPORT OR TRANSPORTATION
101. INTERIOR DESIGN
166. TRAVEL
102. INVESTMENTS
167. TRUCKING
103. LANDSCAPING
168. TRUCKING AND EXCAVATING
104. LAW CORPORATION
169. TRUST
105. LAW JOURNAL
170. UNION
106. LIVESTOCK OR CATTLE ASSOCIATON
171. UNION OR BROTHERHOOD
107. LODGE
172. UNIVESTIY
108. LOGGING
173. WELDING
109. LUMBER
174. WOOD PRODUCT(S)
110. MANGEMENT
175. WOODWORKERS
111. MARINE
176. WRECKING OR WRECKERS
112. MARINE INSURANCE
177. GREENHOUSE
113. MARKETING
178. MUSHROOM
114. MEDICAL CLINIC(S)
179. CRANBERRY
115. MEDICAL SUPPLIES
180. VINEYARD OR WINERY
116. MENTAL OR HANDICAPPED ASSOCIATION
181. JEWELRY
117. MINERALS
182. BEEF FARM
118. MONTESSORI
183. RANCH
119. MOTEL
120. MOVING AND STORAGE
121. MUSEUM OR GALLERY OR ARCHIVES
122. MUSIC OR DANCE OR ART OR LANGUAGE SCHOOL
123. NATUROPATHIC
124. NEWS
125. NEWSPAPER
126. OPTOMETRIC
3
All rights reserved. The Workers’ Compensation Board of B.C. encourages
the copying, reproduction, and distribution of this document to promote
health and safety in the workplace, provided that the Workers’
Compensation Board of B.C. is acknowledged. However, no part of this
publication may be copied, reproduced, or distributed for profit or other
commercial enterprise or may be incorporated into any other publication
without written permission of the Workers’ Compensation Board of B.C.
Additional copies of this publication may be obtained by contacting:
Research Secretariat
6951 Westminster Highway
Richmond, B.C. V7C 1C6
Phone (604) 244-6300 / Fax (604) 244-6295
Email: resquery@worksafebc.com
Document Outline
- DG28 Koehoorn Cover
- DG28 Koehoorn
- Back cover no year
