If denials of primary liability  are improperly made, workers' compensation fails in its purpose of providing medical care, cash benefits and vocational rehabilitation to employees injured at work. Denials are also important because they frequently generate disputes. The denial rate among filed indemnity claims in Minnesota rose from 8 percent in 1984 to 14 percent in 1991, and has remained between 14 and 16 percent since that time. 
Research and Statistics has begun a project to study claim denials. The first phase, reported here, uses department administrative data to examine factors (e.g., claimant characteristics and injury type) associated with claim denial. It also examines how denial rates vary by insurer.
A future phase of the project will examine reasons for claim denial given by insurance companies on the denial notice. Other phases will look at what happens after claims are denied -- for example, disputes and their outcomes and employment and earnings after the claimed injury.
In the current study, the associations between different predictive factors and the denial rate were estimated with the other factors statistically controlled. Several factors were found to be associated with the likelihood that a lost-time claim  is denied. Among the findings:
the denial rate increases with age (except for those 65 and older).
the denial rate decreases with job tenure.
the denial rate is higher for lower-wage workers.
the denial rate is higher if the individual has had prior claims.
the denial rate is sharply higher the greater is the time taken to notify the employer and insurer of the injury and of lost time.
the denial rate is lower for natures and causes of injury that can readily be linked to the workplace.
the denial rate varies widely by individual insurer (including self-insured employers). Because this occurs with other predictive factors controlled, it suggests some insurers may deny claims at a higher rate than necessary.
The current study examines denials of primary liability for 179,000 lost-time claims for 1996 through 2000, from the DLI database. It uses a statistical model to estimate the effects of different factors on the likelihood of denial. The effect of each factor (e.g., injury type) is estimated with the other factors statistically controlled or "held constant."
Figure 1 presents findings from the statistical analysis. The figure shows estimates of what the denial rate in each category would be with the other factors in the model controlled, i.e., statistically held constant at their average levels for the overall population of claims. All results are statistically significant, meaning there is only a small chance that random variation could cause the association between the denial rate and each factor (e.g., claimant age) to be as great as it is estimated to be. (See note in Figure 1.) Although the model controls for several factors, other factors potentially important for predictive or control purposes are excluded because they are not in the DLI database. Some examples are the claimant's education, English proficiency and immigrant status. This matters for interpreting the results, because the excluded factors may be associated with some of the factors examined, such as the pre-injury wage.
Claimant age -- The denial rate is higher for 45- to 64-year-olds than for younger ages. One possible reason is that a longer healing period for older workers could increase the incentive for denial. Also, for an injury with an uncertain work connection, an older worker with more limited opportunities in the labor market (because of age discrimination or specialized skills) may have more incentive to file a claim than a younger person who may more readily shift to a different line of work (or return to school) where the injury does not matter. Another possibility is that even with controls for type and cause of injury, the model may not completely control for these factors. If so, the relatively high denial rates for 45- to 64-year-olds could arise partly from a relatively high rate of cumulative trauma injuries among these people. Interestingly, the denial rate drops for people 65 and older. For these people, the availability of Social Security and Medicare might reduce the incentive to file a claim in a borderline case, leading to a lower denial rate.
Gender and marital status -- Claimant gender does not have a significant association with the denial rate for married people. Single males, and to a larger degree single females, have somewhat greater denial rates than married people. The results for gender and marital status are an example of what happens with statistical controls for other factors. In a simple tabulation by gender and marital status, the denial rate is about four percentage points higher for women than for men, among both married and single people. With other factors statistically controlled in the current analysis, the gender differences largely disappear (they remain to a small degree for single people). These findings imply that other factors largely explain differences in denial rates between men and women.
Employment status -- Compared with full-time workers, part-timers have a slightly higher denial rate, while seasonal workers have a somewhat lower rate. Volunteers have a denial rate more than four points lower than full-time paid employees. This makes sense, because if volunteers are working for noneconomic reasons, it is less important to them to replace lost income, which would lead them to file claims for only the more obviously work-related injuries.
Job tenure -- The denial rate decreases with job tenure. Other things equal, employers and insurers may be more trusting of long-term employees or may be more concerned with retaining these employees if they have built-up skills specific to the job. Also, longer-term employees may be less likely to file questionable claims.
Pre-injury wage -- The denial rate is higher for workers with lower pre-injury wages. In a way, this is unexpected because in lower-wage cases, benefits are lower (other things equal) and so the insurer has less to gain from a denial. However, these results may be influenced by the absence of certain controls, such as health insurance coverage. Lower-wage workers are less likely to have health insurance and should, therefore, be more likely to file a claim for an injury with an uncertain connection to work, which would lead to a higher denial rate for these people. Also, employers and insurers may be more prone to deny claims of workers who are less likely to contest a denial, who are more prevalent among low-wage workers because of such factors as low education, limited English proficiency and immigrant status.
Employee residence -- Denial rates are somewhat higher in metropolitan Minnesota than in nonmetro areas.
Most recent claim -- Denial rates are higher for workers with prior claims, especially for those with a claim in the previous three months. This may reflect an insurer perception (correct or not) that multiple-filers are more prone to file questionable claims. Also, if there is a prior claim, the insurer may attempt to link the worker's current condition to the prior accident or exposure in order to shift cost to a different insurer (if the prior insurer was different) or reduce the benefit amount (if the prior pre-injury wage was lower).
Notification to employer and insurer -- Denial rates are higher the longer it takes to report the injury or lost time to the employer or insurer. This is entirely expected, because delays in notification are likely to arouse suspicion that the injury is not work-related.
First day of lost time -- Denial rates are lower when the first lost-time is delayed relative to the injury. Perhaps the insurer is more confident the claimed injury is truly work-related if the worker does not immediately claim cash benefits. Also, if lost time has not occurred by the time of the primary liability determination, the insurer at that point is uncertain whether it will eventually occur. This means a lower expected claim cost (as perceived at the time of the determination), which should produce a lower denial rate.
Insurance type -- The denial rate is somewhat higher for individual private-sector self-insured employers than for insured employers in the voluntary market. This may occur because individual self-insured employers directly bear the cost of their claims, while insurers can often transfer some of their claim costs back to the employer. Group private self-insurers have about the same denial rate as voluntary market insurers, which makes sense because of certain similarities between group self-insurance and voluntary market insurance. Two categories of insurance type -- the Assigned Risk Plan (ARP) and uninsured -- correspond to particular insurers. The ARP is administered by third-party administrators under contract with the Department of Commerce; uninsured claims are administered and paid by the DLI Special Compensation Fund (SCF). These two insurance types, especially uninsured claims, have elevated denial rates.
Insurer and self-insurer size -- Other things equal, the larger the insurer or self-insurer, the lower the denial rate. The reasons for this are uncertain.
Employer ownership -- Denial rates are relatively low among private nonprofit employers, which makes sense because economic incentives are probably less important for these employers. Interestingly, denial rates are higher in state and local government than in the for-profit sector, which is unexpected to the degree that economic incentives are less influential in the public sector.
Nature of injury -- As expected, the denial rate is clearly lower for types of injuries most readily linked to workplace accidents or exposures. It ranges from 3 percent for heat burns to 7 percent for upper-extremity fractures, 19 percent for asbestosis, 35 percent for infectious diseases and 91 percent for heart attacks.
Cause of injury -- Also as expected, the denial rate is lower when the claimed cause of injury is something readily verifiable. The denial rate is 5 percent for being caught in or between (machinery or objects), 9 percent for falls from height, 17 percent for reaching, turning or stretching, and 42 percent for exposure to dusts, gases, fumes or vapors.
Individual insurer -- As shown in Figure 2, denial rates vary widely by insurer (including self-insured employers). The overall denial rate during the study period was 15.2 percent. Out of the 134 largest insurers, 16 have "adjusted" denial rates (controlling for other factors) of less than 10 percent, while 19 have rates greater than 20 percent. (See note 3 in figure.) Two insurers have adjusted denial rates of more than 50 percent, while another five have rates from 28 to 42 percent. Of the seven insurers with rates greater than 28 percent, three are public-sector self-insurers, two are voluntary market insurers, one is an individual private self-insurer and one is the SCF (uninsured claims). Some of the difference among insurers is due to random variation. However, if all insurers had the same underlying tendency to deny claims with a 15.2 percent likelihood, there would be, on average, just one insurer (out of the 134) with a denial rate less than 10 percent and only two with a rate of more than 20 percent, as shown by the "expected" distribution of denial rates. Because these results occur with other predictive factors statistically controlled, they suggest that some insurers may deny claims more frequently than appropriate.
1. Primary liability is the overall liability of the insurer for any costs associated with a claim after the injury is determined to be compensable. An insurer may deny primary liability (deny the injury is compensable) if it has reason to believe the injury was not work-related, was intentionally self-inflicted, resulted from intoxication or happened during participation in a nonrequired recreational program.
2. Minnesota Workers' Compensation System Report, 2000, DLI Research and Statistics, January 2002. Indemnity claims are claims for indemnity benefits -- benefits to compensate for wage-loss or permanent functional impairment.
3. A lost-time claim is a claim in which the injured worker has been disabled for more than three days -- the threshold for wage-loss benefits in Minnesota. Some filed indemnity claims involve permanent functional impairment, but no wage loss. These claims are excluded from the study, because not all denied claims of this type are in the DLI database.
4. Many workers' compensation insurance policies carry deductible provisions, where the employer reimburses the insurer for claim costs below a deductible limit. Also, the insurer can often raise premiums in the next policy period, through a higher experience modification factor or the insurer's own pricing procedures. In some cases, the insurer can reduce policy dividends to the employer or increase the amount billed to the employer under a "retro" provision where the employer reimburses the insurer for claim costs at the end of the policy period.