Hospital Workers Compensation Benchmark Study

Transcription

1

2 Hospital Workers Compensation Benchmark Study P R E S E N T E D B Y B E E C H E R C A R L S O N I N S U R A N C E S E R V I C E S Beecher Carlson is pleased to present this fifth edition of the Hospital Workers Compensation Benchmark Study. We appreciate the opportunity and strive to continuously demonstrate value and support for the hospital industry through our efforts in developing this analysis. We applaud and appreciate the efforts of all participants and look forward to further discussions around the implementation and execution of strategies driven by the findings of this Study. Identifying the key elements that drive organizations total cost of risk and developing strategies to reduce those costs are shared objectives for the entire industry. We welcome the opportunity to discuss what this information may mean for your company specifically and to identify methods and opportunities for reducing claims. Claim frequency and severity are the key drivers of costs for workers compensation. This Study will provide a variety of methods to measure and review both. Simply put, if you can have fewer claims that cost less on average, you will reduce your organization s overall costs. Beyond these foundational measurements, we begin to consider the frequency of severe claims, as well as the other key performance indicators specifically related to overall workers compensation costs. SCOPE OF STUDY The 2014 Hospital Benchmark Study includes claim information from 2009 through 2013 and valued in 2014 representing: More than 500 hospitals in 36 states Nearly 600,000 employees 131,000 Non-Zero Claims over the five year period Just under 20,000 Lost Time Claims $856 million in incurred and $643 million in paid workers' compensation losses Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 1

3 The Data Frequency Frequency per $1 Million Payroll (Non-Zero) All Years All Years 28.5% reduction in frequency per $1million Payroll noted between 2009 and A key factor in reducing the overall cost of workers compensation claims is to reduce claim frequency. The chart above illustrates claim frequency relative to Payroll. Since 99% of claims are reported within the first 180 days, there is less of an impact on the newness of the data and the typical claims development as the accident-year matures. While there is a 28.5% reduction in claim frequency relative to Payroll over the five year period, it is important to take into consideration the increase in Payroll over that same period of time. As the rate of pay increases, the denominator would also increase thus possibly causing an apparent reduction in the frequency rate per Payroll. Frequency per 100,000 Manhours (Non-Zero) All Years All Years 21.1% reduction in frequency compared to Manhours reduces the likelihood the increase in pay has a significant impact and illustrates a reduction in the frequency of claims. In order to mitigate the impact of increased pay over the five year period, we measured frequency relative to Manhours worked. This measurement shows a 21% reduction over the five years in frequency of the Non- Zero Claims. This reduces the likelihood increased pay was responsible for the downward trend and supports a true reduction in frequency. The difference of approximately 7% over the five year period could be attributed to increased pay and is further supported by the reported national trends of about 2% increase in pay per year between 2009 and To clarify, $1.00 in wages in 2009 increased and compounded 2% each year would be worth $1.08 in Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 2

4 Frequency per $1,000,000 Payroll (Lost Time) All Years All Years Over the same five year period, the frequency of Lost Time claims is down 42%. Due to the variability between reserving philosophies, reporting requirements and the numerous methods for handling medical only losses, Non-Zero Claims can be a less than accurate measurement of frequency. While it is important to utilize Non-Zero Claims for each individual organization, the analysis of the frequency of Lost Time Claims may be more helpful in order to provide the greatest insight from the figures considered. During the same five years, we see a 42% reduction in the frequency of Lost Time Claims from 2009 to 2013 when measured against Payroll moving from 0.19 Lost Time Claims for every $1 million in Payroll to 0.11 claims. Again, it is essential in considering this indicator that increased payroll be factored into the analysis. Frequency per 100,000 Manhours (Lost Time) All Years All Years When compared to Manhours, the reduction is 36.7% thus removing the possible effect of increased pay during the same time period. Just as with the earlier analysis, it is important to consider the impact of increased pay during that same period of time and measure frequency against Manhours worked. Similar to the findings in the Non-Zero Claims, the rate of frequency for Lost Time Claims shows a reduction from 2009 to 2013 in the number of Lost Time claims for every 100,000 Manhours worked over the five year period. However, the reduction in Lost Time Claims is less evident than the Non-Zero claims. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 3

5 The Data Severity Medical inflation, an aging workforce, pharmaceutical costs, Medicare Set-asides, increased obesity; the list of challenges impacting our efforts to reduce costs is long and difficult. While we strive to reduce the frequency of claims, there is a long list of factors driving the costs for each claim higher and higher. It is important that all factors be considered when evaluating your organization s losses. Any one benchmark or measure considered alone can give a misconception of your loss prevention and mitigation efforts. For example, some organizations can experience an unexpected increase in severity with few claims but find their overall costs going down. Thus, none of the factors considered in this study or in your own organization s analysis should be evaluated in isolation. In measuring severity, we first analyze the average cost per Non-Zero Claim. Over the five year period, this figure is $6,499 per claim with 2013 being $4,127. We would expect the most recent year to be the lowest as these figures are undeveloped actual costs. This lower figure illustrates the recency or green nature of the newest loss year. Average Incurred (Non-Zero) $9,000 $8,000 $7,000 $6,000 $5,000 $4,000 $3,000 $2,000 $1,000 $0 $7,855 $7,102 $7,266 $2,668 $2,463 $2,375 $5,831 $1,743 $4,127 $1,004 $4,378 $3,929 $4,093 $3,467 $2,705 $710 $808 $799 $622 $ Expense Medical Indemnity It is also important to consider the cost allocation between the three claim components; expense, medical and indemnity. The ratio between these three components is rather consistent over the five years, except for the indemnity in the most recent loss year. As claims remain open, indemnity payments tend to increase over time, including any settlements or continued payments for lost wages. Thus, we can expect to see all components increase, but the larger increases will likely be in the indemnity and medical portions. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 4

6 To reduce the recency effect and compare loss years on an apples-to-apples basis, we have developed the losses using country-wide loss development factors as provided by The National Council on Compensation Insurance, Inc. (NCCI). These factors allow us to develop present day incurred figures to projected ultimates. Average Ultimate Severity Non-Zero Developed $9,000 $8,000 $7,000 $6,000 $5,000 $4,000 $3,000 $2,000 $1,000 $0 $7,778 $677 $8,818 $8,614 $963 $1,348 $7,825 $7,635 $1,994 $3, Expense Medical Indemnity IBNR Average ultimate severity for 2013 is actually projected to be down about 1.3% in comparison to However, 2010 and 2011 are still higher than forecasted in our prior report. While 2013 is projected to be nearly flat compared to 2009 for the average incurred developed loss amount per claim, 2010 and 2011 are trending even higher than forecast in our prior report. In comparison to 2009, 2010 is projected to run 12.9% higher and 2011 is developed to be 10.7% higher. Loss Rate per $100 Payroll by Year Limited to $500K $1.25 $1.05 $0.85 $0.65 $0.45 $0.25 $0.05 -$0.15 $1.01 $0.95 $0.08 $0.11 $0.90 $0.14 $0.73 $0.30 $0.31 $0.67 $0.25 $0.18 $0.16 $0.31 $0.48 $0.50 $0.43 $0.09 $0.32 $0.24 $0.09 $0.09 $0.08 $0.06 $ Expense Medical While 2010 is up slightly from 2009, the overall reduction over the five years is down 29.5%. Over the last five years, the Study shows a 29.5% reduction in the rate per Payroll to an ultimate developed rate of $0.67 per $100 Payroll for This is an even higher rate of reduction from our prior study which was just under 21% for the period 2007 to Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 5

7 Ultimate Costs Per 100,000 Manhours (Non-Zero Claims) $35,000 $30,000 $25,000 $20,000 $15,000 $10,000 $5,000 $0 $30,213 $32,700 $29,945 $24,531 $23, Expense Medical Indemnity IBNR There is a projected 22.7% reduction in Ultimate Severity comparing 2013 to 2009 and measured against Manhours. Again, to mitigate the impact of Payroll increases and the possible dilution of a severity pattern due to higher compensation, we analyzed a total of costs against 100,000 Manhours for Non-Zero Claims. In considering loss costs per 100,000 Manhours worked, we find the average cost has gone from $30,213 to $23,364 over the last five years when developed using NCCI loss development factors. This is a projected reduction in severity of 22.7%. However, 2010 spikes during this five year term and is actually forecasted to develop 21% higher than the projections from our prior analysis. While the data illustrates a reduction over the last five years in severity and loss rates, in looking back at our prior studies and estimated ultimates, there appears to be continued late development based on incurred, developed claims. This could indicate a pattern or concern around reserve adequacy and stair stepping reserves in the third and fourth year of claims lives. We will analyze historical results and report further in our Retrospective Scorecard Report which will be released first quarter We also recognize that the industry development patterns used may not properly reflect expected development for the participants in this Study. Actual expected developments for the participating organizations and their historical losses could impact the indications here. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 6

8 The Data Frequency of Severe Claims Percent of Total Claim Counts by Size of Incurred Loss (Non-Zero) 4.3% 1.3% Over 94% of all claims have incurred values less than $25, % 0-25K K >100K When addressing the overall cost of risk, it is essential to clarify and understand what claims are driving your organization s costs. In analyzing the losses for the Study, we found that 94.4% of all Non-Zero Claims had total incurred values of $25,000 or less. Incurred Percent of Total Incurred Loss by Size of Incurred Loss (Non-Zero) 26.5% 73.5% of incurred costs for the Non- Zero claims come from those valued at greater than $25, % 32.9% 0-25K K >100K This means 5.6% of all Non-Zero Claims having incurred values greater than $25,000 account for more than 73% of all costs! As we develop strategies to reduce costs, we focus our efforts to identify the factors that can impact and reduce the severity of that body of claims. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 7

9 Percent of Lost Time Claims to Non-Zero Claims 15% As we strive to further identify the body of claims responsible for the greatest costs and offer the opportunity to reduce the overall cost of risk, we look to several key performance indicators that assist in the analysis of an individual organization s loss experience. 85% Lost Time Non-Zero It is essential in measuring the frequency of severe cases to consider those that result in time lost from work relative to those that are medical only, or Non-Zero Claims without indemnity. For this year s Study, the ratio is down slightly from our prior report with 15% of all Non-Zero Claims being Lost Time Claims where some indemnity dollars were incurred. Percent of Lost Time Costs 17% Only 15% of the claims involve Lost Time, but they drive 83% of all incurred costs. Lost Time 83% Non-Zero It is just as important to recognize the impact 15% of all claims have on the overall costs. Approximately 83% of the costs from the five year period are tied to Lost Time Claims. This nearly inverse measure falls in line with the Rule and should assist in directing efforts for loss mitigation. Organizations seeing their ratio of Lost Time Claims increase should expect overall costs to increase. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 8

10 Average Incurred Losses by Reporting Lag $25,000 60% 53% $20,000 $19,436 50% 40% $15,000 28% $12,420 30% $10,000 $5,000 $5,569 $6,597 $7,987 $8,248 13% 20% 10% 4% 1% $0 1% >180 Average Costs $5,569 $6,597 $7,987 $8,248 $12,420 $19,436 Number of Claims 53% 28% 13% 4% 1% 1% 0% There is a clear trend over the prior four Studies that indicates the importance of prompt claims reporting. Every year provides evidence that claim costs increase significantly for losses with greater reporting lag time than those reported and addressed earlier. This result is in direct contrast to the expectation that the most catastrophic claims are known and reported almost immediately. The trend indicates a 19% higher incurred cost for claims reported after the first week than those reported in the first seven days. The interesting trend is that claims reported after the 8 th day but before the 31 st day have dropped as a percentage of the total claims. Claims reported between days 8-30 accounted for 18% of the losses in our prior Study. That has dropped to 13% in this report and actually appears to greatly account for the increase to 53% for the group of claims reported in the 0-2 day group which was at 45%. While 19% of the claims are being reported after the first week, resulting in at least 19% higher incurred costs than those reported earlier, this still represents a nearly 30% drop in the percentage of claims reported after the first week compared to our prior study. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 9

11 The Data Causation Cause Type Analysis Comparison of Total Incurred to Number of Claims by Causation Group TOTAL INCURRED 19% 19% 10% 1% 25% CONTACT (STRUCK BY/STRUCK EXPOSURE INCIDENT AGAINST) 19% COUNT 7% 27% FALL/SLIP/TRIP PATIENT INTERACTION 28% 10% 20% 15% STRAIN - OTHER ALL OTHER For this year s study, we performed an analysis of accident causation using claims with accident dates between 1/1/2009 and 12/31/2013. The workers compensation accident cause codes that are captured in Carrier/TPA claims systems are notorious for their lack of homogeneity, and our data quality assessment for this analysis revealed that reputation to be warranted. The complete set of claims data that we received contained tens of thousands of different causation codes, many of which were free-form text values. In order to convert this raw data into a more useable format, we employed a manual cause code grouping and reorganization strategy. We utilized the expertise of our loss control consulting team in the cause code grouping effort. This process resulted in the manual assignment of over 20,000 unique cause code values into one of the following groups: Contact (Struck By/Struck Against), Strain - Other, Patient Interaction, Fall/Slip/Trip, Exposure Incident, All Other. We have included a summary of the total and average incurred values for each of the observed cause groups in the table below. We can see from this exhibit that among these key groups, patient interaction has resulted in the highest average cost per claim. Please note that for this exhibit and for the graphical summary shown directly below the heading of this subsection, we used only claims with total incurred amounts greater than $0. Contact (Struck By/Struck Against) Exposure Incident Fall/Slip/Trip Patient Interaction Strain - Other All Other Count 31,922 8,388 26,128 12,266 24,133 24,272 Sum $79,912,689 $8,531,189 $213,232,592 $119,681,099 $225,283,852 $152,396,394 Mean $2,503 $1,017 $8,161 $9,757 $9,335 $6,279 Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 10

12 Employee Age Analysis: Average Claimant Age by Accident Year The raw data that we received for computing employee age (at the time of accident) was fairly complete, and appeared to be reasonably consistent. In order to perform the injured employee age section of our analysis, we eliminated from consideration any claims with missing age values as well as any claims which we felt were likely to have miscoded employee age values (in particular, any claims with an employee age less than 16 or greater than 89). The employee age analysis below contains both mean and median time series breakdowns by accident year. We defined employee age as being equal to: (date of loss employee date of birth)/ We also make the assumption that these values are decisively known at the time each claim is reported and therefore do not need to be developed in any manner to provide valid inferences. Interestingly, we can see from the chart below that the average age of injured employees was increasing until the year 2011, at which point it began to level off and drop back down. These findings run somewhat counter to our expectation that a continuously aging workforce effect would be discernable in the claims data. In order to validate them we broke down these results by individual data contributor. Specifically, we were interested in verifying that this pattern was not created solely by a data-related issue such as a survey participant with a disproportionately high average employee age leaving the study or having their data format change in such a way that age related variables are no longer available. Breaking the average injured employee age trend down by respondent, we found that the proportion of respondents with peak average age during each calendar year in the sample period were as follows: (2010; 13.3%), (2011; 33.3%), (2012; 40%), (2013; 13.3%). Based on these results we can see that the overall hump shaped pattern in the injured employee average age time series plot does not appear to be a discrepancy related only to data issues or sample composition, but rather to be indicative of an effect which is present in many respondents loss history on an individualized basis Mean Median Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 11

13 Employee Age and Cause Group Interactions To investigate the interaction between employee ages and accident cause types, we generated the following exhibit. This table shows the mean employee age by cause group (defined using the methodology described earlier), along with the 25th, 50th and 75th employee age percentiles for each of our accident cause groups. Cause 25 th Percentile 50 th Percentile (Median) Age 75 th Percentile Mean Contact (Struck By/Struck Against) Exposure Incident Fall/Slip/Trip Patient Interaction Strain - Other All Other **We can see from the exhibit above that the Fall/Slip/Trip accident cause affected a significantly older population of claimants than did any of the other accident cause types (50% of the Fall/Slip/Trip claimants were over 49.2 years of age, and 25% of the Fall/Slip/Trip claimants were over 56.8 years of age). **We can also see that the claimants suffering from injuries related to Exposure Incidents tended to be younger than other claimants, with 25% of them being younger than 28.2 years old. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 12

14 The Data Jurisdiction Ultimately, many organizations face challenges specific to their individual jurisdictions. Pay rates, compensation laws, compensability issues and the court system further impact results and costs. While there are opportunities to still address all the prior elements, they all must be tempered with the awareness and understanding of the laws and statutes of each state. To consider the possible impact of each state s laws, we calculated the average claim severity for all Non- Zero and Lost Time Claims. Each state s figures are broken down into paid and outstanding for a total incurred value and then ranked from highest to lowest by average severity. Average Non-Zero Incurred ( ) ALL CA AK OR IL MN LA OK NY NH SC AL GA VA MO ID MD PA NM MS FL UT CO KY IN AZ WV TN KS AR NV TX MI $3,671 $3,620 $3,264 $3,648 $3,727 $3,070 $2,616 $3,003 $2,831 $2,755 $2,558 $2,076 $4,932 $6,562 $8,170 $10,876 $8,097 $9,405 $9,766 $5,934 $7,230 $8,998 $6,483 $5,774 $4,403 $5,535 $5,336 $5,629 $5,234 $5,044 $5,067 $4,560 $4,358 $4,078 $4,179 $4,253 $4,008 $3,974 $3,959 $3,868 $3,497 $3,439 $3,427 $3,251 $3,230 $2,766 $2,578 $6,288 $5,243 $5,226 $5,136 $4,819 $4,482 $6,418 $6,326 $6,048 $5,997 $5,974 $5,434 $7,889 $7,233 $7,143 $7,036 $6,771 $11,274 $12,930 The countrywide average cost of Non-Zero Incurred claims is $6,562. Average Paid Average Incurred It is essential that severity, just like in prior discussions, not be considered alone or factored in as the only element. Further, due to the limited volume of claims data in certain states, we have only included states where at least 500 Non-Zero Claims and 50 Lost Time Claims were available. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 13

15 To eliminate the impact of reserving differences based on specific claims handling criteria in certain regions and organizations as well as to address those Lost Time Claims driving a majority of the costs, we reviewed and analyzed the average costs for Lost Time Claims over the five years. The average incurred amount is less than 1% lower than the average in our prior study. Average Incurred Lost Time Claim ( ) ALL VA SC AL PA IL AK KS IN GA OK LA NM CA TN NY MS CO KY ID UT NV FL MO OR MD WV AZ MI TX $11,678 $11,095 $23,247 $17,918 $18,122 $19,115 $14,987 $12,141 $26,745 $21,923 $20,817 $20,038 $32,439 $31,799 $29,796 $26,123 $37,136 $43,238 $42,262 $42,598 $37,571 $46,761 $40,094 $42,219 $39,965 $35,685 $34,851 $32,126 $25,649 $29,371 $28,765 $29,487 $27,385 $27,029 $27,451 $24,364 $24,083 $30,558 $35,348 $39,289 $38,620 $38,324 $36,892 $35,475 $34,836 $53,298 $47,030 $45,939 $45,581 $43,694 $42,403 $41,468 $40,664 $53,496 $53,370 $52,400 $50,254 $48,453 $47,326 The average incurred Lost Time claim value is $37,136. $64,461 Average Paid Average Incurred While there are many postive trends to consider from the 2014 Benchmark Study, the average cost of a Lost Time Claim is up from $32,871 to just over $37,000; representing a 13% increase. Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 14

16 Beecher Carlson Six Concourse Parkway, Suite 2300, Atlanta, GA beechercarlson.com Page 15