How risky is your job… really?

There are hazards in every job and every workplace.  Despite barriers, safeguards, and defenses, exposure to those hazards can harm workers and others in the workplace.  The possibility of injury (including illness, disease and even death) is a reality of work but specific job or task risk data necessary to assess the risk are rare.    Data on the past frequency and impact of work injury that do exist are often industry-based rather than specific to a job or task.  Unfortunately, supervisors and workers may equate the lack of accurate, accessible, and appropriate risk data with minimal risk; they may believe, “If this job or task was risky, I am sure someone would warn me”. 

Why risk awareness is important

Whether it’s your investments, sports activities, or medical treatments, having access to and an understanding of risk data are essential to decision making.  Depending on your risk tolerance and  armed with an understanding of the risks, you can decide whether or not to invest in a particular mutual fund, take up a particular sport, or undergo a particular therapy or treatment.  Knowing the risks, you can also make choices to mitigate them (diversification in your investments, classroom training for your chosen sport, and performing specific stretching exercises between physical therapy treatments, for example).  

Knowing the risks in your work is no less important.  Few employment engagements include explicit and precise information about risks.  Yes, employers have a duty of care and a general duty to inform workers about workplace and job-specific risks but few job interviews cover your risk of occupational injury, illness, disease or death.  Your job does not come with a warning sticker outlining the risks.  Under-stated, misrepresented or incomplete risk data may lead to incorrect judgements about precautions you ought to take or dissuade you from exercising your right to refuse unsafe work. 

Assessing Risk

So, how risky is your job?  What tasks in my job are risky?  Is working in healthcare riskier than working in construction?  How would you know?

Some jurisdictions require formal "risk assessments”.  These often involve examining the likelihood and impact of harm from a given task.  (see WorkSafeBC, Assessing Risks, for example).  A risk assessment will include an analysis of who might be harmed, how that harm might occur, and what to do to eliminate, minimize or otherwise manage the risks, particularly those with the highest probability and impact.  

Risk Matrix.  Source:  WorkSafeBC, Assessing Risk

This type of risk matrix is typically applied to very specific job tasks and often relies on a subjective estimation of probability and impact.  At their best, words like "low", "unlikely" , "minor" are useful in relative terms but lack precision  and may, at worst, be misinterpreted as "not worth worrying about".  Even if the probability  is rare, the consequences may be extreme and warrant some form of mitigation.  This is particularly true for biological toxins where the probability of exposure is low but the consequences may be severe illness or death--a combination that may warrant a "high" rather than "medium" subjective risk rating. 

Quantifying and Comparing Risks

At the enterprise or sectoral level, performance data may exist to add objectivity to the risk analysis. Statistical measures help quantify both probability and impact in risk analysis.  Impact may be quantified by workers' compensation costs, average days away from work (calendar or working days), or thresholds that exceed a particular level or case definition such as "serious injury".  Probability may be quantified as a ratio based on exposure (cases per million hours worked or 100 employees).  

The lack of data at the jobsite, enterprise or sectoral level may be due to a number of factors.  Workplace injuries and deaths are (thankfully) relatively rare events.  With small numbers, it is often difficult to calculate an accurate frequency rate that adequately represents risk.  There are also counting issue.  Most risk data come from workers’ compensation administrative information or “OSHA Log” surveys but not all work injuries are recorded or result in workers’ compensation claims.  Poor record keeping, intentional under-reporting, claim suppression, high denial rates for some types of injuries and occupational diseases are among the main reasons that reported injury rates may not adequately reflect actual workplace risk.

Many sectors and employers use “injury rate”  or “incidence rate” data a way to quantify risk.  These are admittedly trailing indicators of safety and, (as we are always told when assessing risk in our investments), past performance may not be indicative of future results. 

The idea behind injury and incidence rates is to provide a standardized expression of risk in terms of injury (illness, disease or death) events relative to a quantity of exposure (a measure related to a quantity of employment such as “person years”).  The US Bureau of Labor Statistics publishes “incidence rates” (among other statistics) that provide data at an industry level.  Here are the top ten for 2016:

TABLE SNR02. Highest incidence rates1 of nonfatal occupational injury and illness cases with days away from work, restricted work activity, or job transfer, 2016  [Extracted from Supplemental News Release Tables, 2016]
Industry2	NAICS Code3	Incidence Rate
Nursing and residential care facilities (State government)	623	8.4
Other nonferrous metal foundries (except die-casting) (Private industry)	331529	6.0
Fire protection (Local government)	92216	5.9
Heavy and civil engineering construction (Local government)	237	5.8
Frozen cakes, pies, and other pastries manufacturing (Private industry)	311813	5.8
Couriers and express delivery services (Private industry)	4921	5.8
Scheduled passenger air transportation (Private industry)	481111	5.7
Truss manufacturing (Private industry)	321214	5.6
Amusement and theme parks (Private industry)	71311	5.5
Police protection (Local government)	92212	5.5

1 The incidence rates represent the number of injuries and illnesses per 100 full-time workers  and were calculated as: (N/EH) x 200,000, where
N = number of injuries and illnesses
EH = total hours worked by all employees during the calendar year
200,000 = base for 100 equivalent full-time workers (working 40 hours per week, 50 weeks per year)
2 High rate industries were those having the highest incidence rate of injury and illness cases with     days away from work, restricted work activity,
or job transfer and at least 500 total recordable cases at the most detailed level of publication,  based on the North American Industry Classification System -- United States, 2012.
3 North American Industry Classification System -- United States, 2012

Note the limitations of these data.  The “cases” relate to recorded cases; if record keeping is poor or cases are not reported, then the published incidence rate will under-represent the risk.  Also, note the calculation methodology; the specific calculation of the  100 full-time equivalents used as the denominator for this incidence rate is only one way to calculate risk.  Other sources may use different calculations and definitions. 

Rather than using an approximation for 100 full-time workers, SafeWork Australia use both Frequency rates (serious injuries per million hours worked) and Incidence rates (serious injuries per 1,000 employees).  Here are the top 10  from the 2016 tables [Extracted from Australian Workers’ Compensation Statistics 2015-2016] :

Table 22: Frequency rate (serious claims per million hours worked) by industry,
2000–01 and 2010‑11 to 2015–16p  [Top 10 extracted and re-ordered based on 2015-16 column]

Industry	2000-01	2010-11	2011-12	2012-13	2013-14	2014-15	% chg	2015-16p
Agriculture, forestry and fishing	14.3	10.5	10.8	10.7	9.1	9.9	-31%	8.9
Manufacturing	13.9	10.5	10.7	9.5	8.8	8.8	-37%	8.4
Construction	13.5	9	9	8.4	7.8	8.1	-40%	8
Transport, postal and warehousing	14.9	11.8	12.2	10.4	9.6	8.6	-42%	7.7
Health care and social assistance	12.1	10.7	10.5	10	9.1	8.7	-29%	7.4
Arts and recreation services	13.7	9.8	9.7	8	9.2	8	-41%	7.1
Wholesale trade	8.2	7.7	7.1	6.5	6.6	6.6	-20%	6.6
Public administration and safety	8.8	9.1	8.1	8.3	7.2	6.9	-22%	6.1
Accommodation and food services	8.9	7.2	7.5	7	6.6	6.1	-31%	5.9
Administrative and support services	11.6	9.4	8.3	7.4	6.7	5.6	-52%	5.8

Table 23: Incidence rate (serious claims per 1000 employees) by industry,
2000–01 and 2010–11 to 2015‑16p [Top 10 extracted and re-ordered based on 2015-16 column]

Industry	2000-01	2010-11	2011-12	2012-13	2013-14	2014-15	% chg	2015-16p
Agriculture, forestry and fishing	27.8	20.6	21.4	20.7	18.1	19.1	-31%	17.5
Construction	27.7	18	18	17.1	15.9	16.1	-42%	16
Manufacturing	27.2	20.2	20.7	18.1	16.4	16.6	-39%	15.5
Transport, postal and warehousing	29.3	22.4	23	19.7	18.1	16.3	-44%	14.4
Wholesale trade	16.1	14.7	13.4	12.2	12.5	12.7	-21%	12.3
Health care and social assistance	17.8	15.3	15.3	14.3	13	12.3	-31%	10.7
Public administration and safety	15.5	15.3	13.9	14.2	12.1	11.5	-25%	10.2
Arts and recreation services	18.6	12.4	12.8	10.7	12	10.1	-46%	9.7
Administrative and support services	19.1	15.2	13.5	12.1	10.8	9	-53%	9.2
Mining	25.1	12.5	12.2	11.9	11.1	9.9	-61%	9.2

Note these data relate to accepted workers’ compensation claims.  By definition, denied claims (and unreported injuries) are not included.  The definition of “Serious” is also important. In this context, only injuries resulting in absences of a working week or more are considered. The definition of serious is not standardized.

Note also that the rank order changes depending on the method of calculation.  Using both incidence and frequency rates provide a richer depiction of risk. 

Risks for males and females are more similar than injury counts might suggest

Frequency and incidence rates provide similar but different representations of risk.  A frequency rate may be more appropriate where there is wide variation in the hours worked by particular groups.  Men tend to work more hours in a work week than women.  On an incidence basis, the injury rate for women would under-represent risk.    The same Australian report notes the frequency rate (serious injury claims per million hours worked) for men and women differs:  4.9 for women and 6.2 for men.  

The injury frequency rates for men and women are much closer than conventional wisdom might suggest. One often quoted statement presents a wide variation in risk for men and women:

 "Women incurred less than one-tenth of the job-related fatal injuries and one-third of the nonfatal injuries and illnesses that required time off to recuperate in 1992-1996".  US Department of Labor,   "Women Experience Fewer Job-related Injuries and Deaths than Men",  Issues in Labor Statistics, Summary 98-8, July 1998

In the two decades since this analysis was published, women have increased their participation in the labor force.  Although most North American jurisdictions do not publish frequency or incidence rates specific to males and females, data representing risks by sex may provide valuable insights.  Women now account for about half the labour force in Canada, the US and Australia, although average hours worked per week are higher for men than women.  The apparent lower number of accepted workers’ compensation claims for women arises from a lesser exposure (the smaller pool of hours worked).  

I asked WorkSafeBC to apply the Lost Time Injury Frequency Rate (LTIFR) calculation to its data and data on work hours from Statistics Canada.  In this case, all accepted time loss injuries (rather than just serious injuries as used in the Australian study) were used in the calculation. 

WorkSafeBC  Unofficial Injury Rates and Estimated LTIFR for Males and Females - 2016

	2016  Injury Rate (Accepted time loss claims per 100 person years of employment)	2016 LTIFR (Accepted time-loss claims per 1,000,000 hours of employment)
Males	2.61	13.9
Females	1.75	11.6

Comparing LTIFR to the traditional Injury Rate (per 100 persons years of employment calculation) reinforces the point.  LTIFR may present a more accurate and compelling representation of work-injury risk for women. 

Risks and consequences

Risk calculations noted above may carry a level of consequence in the case definition.  The US analysis uses a definition of "work absence or restricted duties" while the Australian data includes "accepted workers’ compensation claims with a week or more away from work". Many state and provincial jurisdictions publish workers’ compensation injury rate statistics but the case definitions used to calculate the risk indicator will vary.    WorkSafeBC publishes an annual Statistics Report with subsector injury rates and claim durations.  Cases are accepted time-loss claims and this is a no-waiting period jurisdiction so claims cover wages lost beyond the day of injury.   Here are the top 10 from that jurisdiction. 

WorkSafeBC Top Ten Subsectors by injury rate and duration
[Based on data extracted and re-ordered from the WorkSafeBC Statistics 2016 edition, Table 2-11] 

Note that this analysis does not consider sectors that are “self-insured” (Deposit account employers including the provincial government).  

The duration part of the table is useful in considering conditional risk:  if you work in warehousing and  have an accepted time-loss workers’ compensation claim then, on average, you will miss 42 paid days from work (a bit more than eight calendar weeks).

Risk varies with age

The risk of work-related injury also varies with age.  Many studies point to the high risk associated with young male workers.  This table,extracted from an Australian study, reinforces this fact but it also demonstrates that risk varies with age.  Also, note the frequency of injury for females is essentially the same as males for ages 50 and above.  

Table 4: Frequency rate (serious claims per million hours worked) by injury or disease, sex and age group, 2015–16p    [Extracted from Australian Workers’ Compensation Statistics 2015-2016]  

Age group	Injury and musculoskeletal disorder claims [per million hours worked]
		Male	Female	Total
< 20 years		7.7	3.4	5.7
20-24 years		6.5	3.3	5.1
25-29 years		5.1	2.7	4.1
30-34 years		4.6	2.9	4.0
35-39 years		4.8	3.5	4.3
40-44 years		5.3	4.3	4.9
45-49 years		5.6	5.0	5.3
50-54 years		6.1	5.9	6.0
55-59 years		6.4	6.3	6.4
60-64 years		6.9	6.6	6.8
65 years+		5.1	5.7	5.3
Total		5.6	4.3	5.1

Demographic change in the US, Canada, Australia and many other countries is driving dramatic shifts in the age profile of labour force participants.  More women, more older workers, fewer younger workers are driving changes in the risk in the labour force.  Frequency and mix of injuries as well as the duration of disability are all likely to increase as a result.

Risk data may under-represent actual risk

All representations of risk using workers’ compensation administrative data are subject to several important caveats.  As pointed out in previous posts, exclusions from coverage, under-reporting (including  claim suppression), claim denial rates result many potential cases of work injury missing from the calculation .  As a result, statistics like those in this post may well understate the actual risk.  

Categories of classification also vary by jurisdiction making simple apples-to-apples comparisons very difficult.  That said, data from multiple sources may provide a better sense of the risks.  Just like data from multiple medical trials or rating agencies can better inform your decisions regarding your health and investments, work-risk data from multiple sources may help workplace participants better understand and mitigate work-related risk.   

Your workplace is unlikely to have a warning sticker on the entrance providing risk data.  Your risk in your job on your worksite is going to be very specific, subject to a lot of factors,  and difficult to estimate accurately.  Risks associated with your demographic characteristics and your industry’s (and maybe even your firm’s) experience are likely more available and may provide guidance on just how risky your job really is. 

Providing risk data may not be formal regulatory requirement but sharing the risk data that are available with supervisors and workers may lead to a more accurate appreciation for the risk of workplace injury, illness and disease.  And that may lead to safer and healthier workplaces.