Occupational Injury and Illness Surveillance: Conceptual Filters Explain Underreporting

This earlier paper (2002) explored a range of reporting filters that take place against work-related injuries.

That is, a range of filters exist which influence whether incidents/injuries are reported, and if reported, influence how those events are categorised, notified, and acted upon.

Because of the paper’s age consider this more of an interesting snapshot, as likely a lot has changed since then [** however, numerous studies on my site and within the literature show that many of these factors still exist in some form or another.]

They note that the process of documenting and acting on work-related injuries and illnesses involves a detailed series of events. The term filter characterises partial barriers to these events.

Moreover, “unless the filters are totally permeable, the injuries detected at successive levels will be fewer and more severe” (p1421). The filters involve different factors at the individual, work environment, organisational and also to the broader economic, legislative and social contexts – all of which influence permeability.

The filters affecting the (2002) US reporting network is described below:

Note 1: This paper is very US-centric and particularly around workers compensation and insurance schemes. I’ve skipped over most of this.

Note 2: This isn’t a theoretical or opinion piece - virtually every key statement is supported by published evidence. There’s 116 references for only a 9 page paper.

Filters:

a. Filters to reporting to supervisors

Numerous barriers exist at this level. Workers may fear reporting injuries or illnesses as they fear disciplinary action, denial of overtime or promotions, drug testing, harassment or job loss.

Moreover, “Some safety incentive systems reward workers who do not report injuries with money, material goods, or recognition” (p1422).

While unionised members may have greater support for reporting, workers with insecure immigration status, limited permission to work or lack of marketable job skills are particularly vulnerable to dangers of underreporting.

Further, low wage and immigrant workers are vulnerable to threats of being fired or threatened for complaining.

Data indicated that contingent workers, including temporary and labour hire workers and “most construction workers” also risked future job assignments by reporting health problems.

Earlier data indicated that ~86 of 98 industrial employees experienced work-related illness or injury—leading to 50% experiencing this for over a week and 30% losing work time—yet fewer than 5% actually reported their condition as work-related.

Barriers to reporting in the above study were fear of discipline or being labelled a complainer. Importantly—and consistent with current research—other barriers to reporting included “a company goal of no reported injuries, reinforced by the presence of incentive programs that rewarded low levels of reported injuries” (emphasis added, p1422).

Further data indicated that 39% of injured workers in one sample hadn’t reported those events. Here the most common explanations was fear for being labelled as careless by their supervisor or considering the injury too minor to report.

The authors provide multiple other studies highlighting the widespread under-reporting of work-related injuries and illnesses (again, consistent with current research).

b. Filters to Lost Work Time Due to Work-Related Illness and Injuries

One demonstrable barrier here is that workers feel pressure not to report as they don’t believe they can afford lost time work.

Other barriers include missing out on work and overtime. This particularly affects low-wage, temporary or contingent workers who can face immediate poverty or job loss.

Also they highlight the very known issue that employers may seek to avoid lost time injuries by use of light duties and other types of bureaucratic juggling. However, they also importantly note that because of the efforts in avoiding the recording of a lost time injury, “injuries resulting in lost time may not represent underlying patterns” (p1422).

That is, as this earlier data and a large body of more recent data indicates – we, as a generality, have little reason to believe reported incident data as representative.

One study found ~30% of aircraft maintenance staff had reported low back pain severe enough to interfere with their daily activities but only 2.3% of the workers had lost any time due to that pain.

c. Filters to Medical Care for Work- Related Injuries and Illnesses

This section covers more issues – one including the insidious nature of work-related illnesses and disease and?particularly those with long latency periods (e.g. silicosis and the like).

I’ve skipped most of this section, but they highlight issues of medical insurance. One study found that workers with likely work-related musculoskeletal disorders with no health insurance were one third as likely to see a physician and have their condition diagnosed as work-related compared to those with insurance.

Other data found that some workers have no time or capacity to seek care. 30% of immigrant workers in one sample weren’t allowed to take sick or vacation days, paid or unpaid, without risk of being fired.

d. Filters to Recognition of Work- Related Injuries and Illnesses

A filter here may be that workers may not perceive a connection between their health condition and their work. Physicians may also fail to recognise work relatedness.

Lack of work relatedness may be especially prevalent for diseases with long latencies, symptoms common to nonoccupational disorders, or multiple causal factors (e.g. nonpneumoconiotic chronic respiratory illness, cancer, heart disease, renal disease, and neurological disorders).

An earlier study from California found that 2/3 of reported occupational diseases to physicians were skin and eye problems while “almost no occupational cancers were reported” (p1423).

Other data on new-onset and reactivated asthma was found to be significantly under-associated to work factors. Only 7% had been asked about occupational triggers by primary and urgent care physicians.

e. Filters to Charging Medical Care to Workers’ Compensation

As expected, barriers to reporting included concern about raising insurance premiums.

Moreover, within the construction industry (and elsewhere), contractors` records of compensation claims (** and incident data) also affect their competitiveness in contract bids.

Other barriers include the compensation investigation/assessment process. One dataset found 79% of patients with carpel tunnel syndrome were initially challenged or received no response from insurers and adjudication took an average of 429 days. Authorisation for surgery took an average of 318 days.

The authors also note that “Insurers were significantly more likely to challenge claims filed by non-Whites, low-wage workers, and union members” (p1424).

f. Filters to Recording Incidents in OSHA Logs

Much earlier data indicated that several of the largest corporations in the US “engaged in serious and willful underreporting of work-related injuries” (p1424).

Reasons for underreporting [or “safe-washing”, to coin O’Neill’s term], was the desire to avoid OSHA inspections as well as competition among companies, or to record low injury rates and enhance managerial performance evaluations.

Further, it’s noted that exemptions to OSHA inspection programs depended on low reported events – whereby they targeted employers with higher reported events; creating incentives to underreport.

Notably, “Employers may record occupational injuries and illnesses in ways that protect the business from liability, particularly in the identification of the source, causal event, and exposure leading to the injuries” (p1424).

One means to avoid reportable events was to provide on-site medical treatment; thereby classifying the injury as a first aid event to circumvent reporting requirements.

Another method was found whereby plant managers reduced reportable injuries by logging new injuries with old injuries in a single report.

One study argued that “the relatively low rates of reported injuries in establishments with fewer than 50 employees reflect lax record keeping rather than safer conditions,84 a position buttressed by the higher fatality rates in smaller companies employees” (p1424).

Much earlier data (1980s) found particular manufacturing establishments underreported total injuries and illnesses by between 10-25%. Other data found that organisations failed to document between 20-80% of occupational cumulative trauma disorders.

Another study found that large companies with “exceptional commitments to health and safety record keeping” failed to record 60% of applicable events.

Other data found that organisations underreported certain events by either 4-5 times (for acute trauma) and 69-93 times (!) for cumulative trauma disorders. Other data still found occupational skin diseases were underreported by a factor of 20-fold.

g. Filters to Filing First Reports of Injury to State Workers’ Compensation

A number of factors were highlighted here but I skipped a lot of it. Something I found interesting was that the workers’ compensation system detected around 26% of the 6k injuries in one sample, compared to the emergency department records detecting 81%.

i. Filters to Capture of Medical Records Data in Hospital Databases

Here they argue that even when injury treatment is correctly charged to workers compensation (given a sizeable percentage of work-related injury/illness cases aren’t linked to work), these same events can fall through cracks in the hospital reporting system.

In wrapping up, they note that numerous obstacles exist within the work-related incident reporting system. At each level, these filters “block documentation of health problems affecting populations especially vulnerable to workplace hazards, including immigrant and low-wage workers” (p1426).

Link in comments.

Authors: Azaroff, L. S., Levenstein, C., & Wegman, D. H. (2002). Occupational injury and illness surveillance: conceptual filters explain underreporting. American journal of public health, 92(9), 1421-1429.