Health Plan Auditing

1,247 views

Published on

Auditing 100% of claims vs. random sample audits

Published in: Business, Economy & Finance
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,247
On SlideShare
0
From Embeds
0
Number of Embeds
231
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Health Plan Auditing

  1. 1. James Herrington Chief Marketing Officer Healthcare Data Management, Inc. A presentation by: April 20, 2010 George P. Sillup, Ph.D., M.S. Saint Joseph’s University Haub School of Business
  2. 2.
  3. 3. <ul><li># of Employees Cost Per Employee/Yr. Annual Plan Cost </li></ul>20,000 $10,000 $200,000,000 Note: An October 2009 study by Towers Perrin projected that the average annual health benefit spend will exceed $10,000 per employee for the first time in 2010. 10,000 $10,000 $100,000,000
  4. 4. <ul><li>Annual Plan Cost Percent Plan Waste Total Dollar Waste </li></ul>$100,000,000 8% $8,000,000 $200,000,000 8% $16,000,000 Note: HDM audits normally find from 5% – 8% wasted expense.
  5. 5. <ul><li>Stop the bleeding. </li></ul><ul><li>Hold Third-Party Administrators (TPAs) and Pharmacy Benefit Managers (PBMs) accountable for claim errors and fraud. </li></ul><ul><li>Bring about plan changes to reduce waste. </li></ul>
  6. 6. <ul><li>Audit based on 100-percent-of-claims analysis. </li></ul><ul><li>Focused audit of 100-percent-of-claims. </li></ul><ul><ul><li>Scans 100 percent of the claims for errors that fit a pre-determined profile. </li></ul></ul><ul><li>Random Sampling. </li></ul><ul><ul><li>Usually around 200 to 400 claims. </li></ul></ul>
  7. 7.
  8. 8. Ronald K. Klimberg, Ph.D. Saint Joseph’s University Haub School of Business Conducted by: The authors thank Healthcare Data Management, Inc. for funding this study. George P. Sillup, Ph.D., M.S. Saint Joseph’s University Haub School of Business
  9. 9. <ul><li>To examine the implications of two approaches to health plan auditing. </li></ul><ul><ul><li>Auditing based on a 100%-of-claims analysis, utilizing the HDM five-step protocol. </li></ul></ul><ul><ul><li>Random Sampling (300 and 400-claim samples). </li></ul></ul>
  10. 10. <ul><li>Steps 1 & 2 – Data Warehouse </li></ul><ul><ul><li>HDM downloads all of the data from steps 1 and 2 to data warehouse. </li></ul></ul><ul><ul><li>1. Paid claims data requested from administrator(s). </li></ul></ul><ul><ul><li>2. Client provides eligibility data, benefits covered and not covered, place of service, co-pays, etc. </li></ul></ul>
  11. 11. <ul><li>Step 3 – Benchmarks </li></ul><ul><ul><li>HDM integrates comparative logic created from ASO, SPD, formulary, industry standards (e.g. CMS, AMA, FDA rules, etc.) and proprietary logic to catch duplicates/fraud. </li></ul></ul><ul><li>Step 4 – Exception Analysis </li></ul><ul><ul><li>Analysis commences on 100% of claims and eligibility data compared to best business practices and administrator's procedures. Exception reports are generated with potential errors and issues identified. </li></ul></ul>
  12. 12. <ul><li>Step 5 – Claims Auditing </li></ul><ul><ul><li>After analyzing 100% of claims according to the ASO, SPD and industry standards, an onsite claims audit validates the logic and findings of the 100% analysis. </li></ul></ul><ul><li>Notes : </li></ul><ul><li>HDM refers to data that have not yet been subjected to an onsite audit as “pre-audit </li></ul><ul><li>data” and to data that have undergone the onsite logic check as “post-audit data.” </li></ul><ul><li>Exceptions = errors. </li></ul>
  13. 13. <ul><li>HDM provided medical claims data from two Fortune 100 corporations – company A and company B. </li></ul>
  14. 14. Entire Population of Claims Company Type $ Amount of Claims Paid # of Claim Records A Pre-Audit 2006 12,803,426 54,192 Pre-Audit 2007 12,544,893 54,371 Post- Audit 2006 12,803,426 54,192 Post-Audit 2007 12,544,893 54,371 B Post-Audit 118,368,625 463,919
  15. 15. $$ Exception claims—range from 4.6% to 24.7% ## Exception claims—range from 4.4% to 15.5% Entire Population of Claims Exception Claims Company Type $ Amount of Claims Paid # of Claim records $ Amount of Exception Claims # of Exception Claim Records % of $ Errors % # of Error Claims A Pre-Audit 2006 12,803,426 54192 $2,234,051 8299 17.4% 15.3%   Pre-Audit 2007 12,544,893 54371 $3,092,431 8428 24.7% 15.5%   Post-Audit 2006 12,803,426 54192 $1,327,346 3850 10.4% 7.1%   Post-Audit 2007 12,544,893 54371 $1,792,882 5465 14.3% 10.1% B Post-Audit 118,368,625 463919 $5,467,944 20395 4.6% 4.4%
  16. 16. Company Type $ Amount of Exception Claims Paid $ Amount of Over- Payment Records $ Amount of Under- Payment Claims $ Total Over- and Underpayment Claims A Pre-Audit 2006 $2,234,051 $229,012 $87,716 $316,728   Pre-Audit 2007 $3,092,431 $458,049 $62,498 $520,547   Post-Audit 2006 $1,327,346 $162,649 $26,654 $189,302   Post-Audit 2007 $1,792,882 $160,050 $25,459 $185,508 B Post-Audit $5,467,944 $702,556 $59,417 $761,973
  17. 17. Company A’s – Claims Histogram What claims are responsible for this expenditure?
  18. 18. Company A’s – Over- and Underpayment Claims Histogram
  19. 19. What claims are responsible for this expenditure? Company B’s – Claims Histogram
  20. 20. Company B’s – Over- and Underpayment Claims Histogram
  21. 21. <ul><li>The resulting set of medical claims record exceptions (errors) was used as the benchmark for comparison with random sampling results. </li></ul><ul><li>That is, this data set was assumed to be the total population of exceptions. </li></ul>
  22. 22. Population of Claims Random Sample Population of Exception Claims Random Sample
  23. 23. <ul><li>For each data set... </li></ul><ul><ul><li>Generated 100 random samples of sample size 300. </li></ul></ul><ul><ul><li>Generated 100 random samples of sample size 400. </li></ul></ul>
  24. 24. 300 Sample Size Amount Missed Company Type % of $ Amount of Exception Claims Paid % of $ Amount of Over- Payment Records % of $ Amount of Under- Payment Claims $ Amount of Claims Paid $ Amount of Overpayment Records $ Amount of Under- Payment Claims A Pre-Audit 2006 3.66% 3.97% 3.64% $2,152,257 $219,926 $84,522   Pre-Audit 2007 3.57% 3.72% 3.59% $2,982,092 $441,017 $60,253   Post-Audit 2006 7.87% 8.22% 7.73% $1,222,858 $149,282 $24,593   Post-Audit 2007 5.46% 5.56% 5.35% $1,695,056 $151,145 $24,096 B Post-Audit 1.48% 1.42% 1.45% $5,386,944 $692,615 $58,554
  25. 25. 400 Sample Size Amount Missed Company Type % of $ Amount of Claims Paid % if $ Amount of Over-Payment Records % of $ Amount of Under-Payment Claims $ Amount of Claims Paid $ Amount of Overpayment Records $ Amount of Under- Payment Claims A Pre-Audit 2006 4.85% 5.12% 4.85% $2,125,686 $217,294 $83,459   Pre-Audit 2007 4.85% 4.84% 4.78% $2,942,504 $435,859 $59,510   Post-Audit 2006 10.55% 10.78% 10.30% $1,187,249 $145,123 $23,908   Post-Audit 2007 7.44% 7.33% 7.24% $1,659,439 $148,326 $23,616 B Post-Audit 1.96% 2.02% 1.89% $5,360,753 $688,396 $58,297
  26. 26. <ul><li>Even the larger sample size missed over 90% of the claim errors identified by the 100% methodology. </li></ul><ul><ul><li>In terms of dollars, random sampling missed from about $200,000 to over $750,000 in over- and underpayment exceptions. </li></ul></ul><ul><li>100% auditing depicts plan utilization more accurately. </li></ul><ul><li>Random sampling missed the opportunity to identify the root causes of claim errors. </li></ul>
  27. 27. <ul><li>State of Wyoming: </li></ul><ul><li>A retrospective analysis of approximately two year’s worth of medical claims uncovered... </li></ul><ul><ul><li>$3,186,135 in exceptions among total paid claims of nearly $223,000,000. </li></ul></ul><ul><ul><li>More than one-half of the exceptions were due to billing and coding edits. </li></ul></ul>
  28. 28. <ul><li>State of Wyoming </li></ul><ul><li>Distribution of $3,186,135 in Exceptions </li></ul><ul><li>Random sampling would have failed to identify... </li></ul><ul><ul><li>A minimum of 90 percent – $2,867,522 – of the exceptions. </li></ul></ul><ul><ul><li>Order of magnitude per exception category. </li></ul></ul>
  29. 29. <ul><li>City of Fort Worth : </li></ul><ul><li>A retrospective audit of approximately two year’s worth of medical claims uncovered nearly $2,434,499 in errors among total paid claims of nearly $87,000,000. </li></ul><ul><li>Most of the errors were overpayments due to claims being paid in duplicate. </li></ul>
  30. 30. <ul><li>City of Fort Worth </li></ul><ul><li>Distribution of $2,434,499 in Exceptions </li></ul><ul><li>Random sampling would have failed to identify... </li></ul><ul><ul><li>A minimum of 90 percent – $2,191,049 – of the exceptions. </li></ul></ul><ul><ul><li>Order of magnitude per exception category. </li></ul></ul>
  31. 31. Population of Claims Population of Exception Claims -$$ Random Sample
  32. 32. <ul><li>Results indicate that random sampling missed at least 90 percent of the exceptions the “100 percent” methodology caught. </li></ul><ul><li>Identify a need for a paradigm shift from random sampling to goal of “zero defects” using 100% audit. </li></ul><ul><li>Begin to understand reason for claims and thereby eliminate them. </li></ul>

×