DCI Data Validation and Quality Issues

Transcription

1 NCCI S DCI Data Validation and Quality Issues January 26 29, 2016 Palm Beach County Convention Center West Palm Beach, FL

2

3 DCI Data Validation and Quality Issues Presented by: James Bonk and Joanne Missry Objectives To understand why we collect Detailed Claim Information (DCI), the aggregate data quality observation process, and associated carrier outreach. 2 1

4 Agenda Why We Collect DCI Data Carrier Outreach Process Data Element Data Quality Life of a Claim Review 3 Why We Collect DCI Data 2

5 Why We Collect DCI Data Used extensively for a variety of purposes: State Insurance Department Reporting State Advisory Forums State Legislative Indemnity Pricing Temporary Partial Disability as a proportion of Temporary Disability benefits Payments by type of benefit for Permanent Partial Disability (PPD) claims Attorney involvement Research Studies 5 Carrier Outreach Process 3

6 Carrier Outreach Process NCCI has a three-step process to ensure that our data reporters provide quality DCI data Defined aggregate data quality observations run quarterly By Data Element Compared to prior results and classified Data quality observations sent to the carrier Trending with previous quarter results and review of prior carrier feedback 7 Aggregate Data Quality Observations What Are They? Observations look for patterns or anomalies and highlight carrier outliers from the industry Include measures of DCI data elements and comparative analysis between the elements in aggregate over at least six months 8 4

7 Aggregate Data Quality Observations Why Are They Necessary? All DCI data elements are required or conditionally required No optional elements DCI elements do not become optional because an edit does not require them Report all DCI elements as the claim develops Per claim values are reasonable but may be unreasonable in the aggregate 9 Aggregate Data Quality Observations This Should Make It Easier to Understand Claims are edited individually Most claimants return to work after recovering Some claimants do not return to work 10 5

8 Return To Work (RTW) Date Indicates the most recent date the claimant returned to work Carriers sometimes capture Initial RTW and Current RTW Impact to reporting Observations: For closed claims, the percentage of missing Accident Date to Return to Work Date is less than 4 days 11 Return To Work (RTW) Date The most recent date on which the claimant returned to work unless the claimant has not returned to work. Claimant is classified as Permanent Total Claim is closed with a lump sum payment prior to returning to work Closed Claim Claimant chooses not to return to work after maximum medical improvement date Claimant leaves company while on disability 12 6

9 Return To Work (RTW) Date The most recent date on which the claimant returned to work unless the claimant has not returned to work. Claimant is classified as Permanent Total Claim is closed with a lump sum payment prior to returning to work RTW Date should Closed be left Claim blank in these cases Claimant chooses not to return to work after maximum medical improvement date Claimant leaves company while on disability % Aggregate Data Quality Observation Return to Work Date Missing Percent RTW Date Missing 80% 60% 40% 20% 0% Coverage Provider Groups 14 7

10 Necessary Analysis Before Outreach Trending current quarterly results with prior results: Results trending with industry norms Still trending different than the industry Results have improved or no longer an outlier Review prior carrier feedback, if any, to see if an observation has been explained by the carrier An outlier is correct as reported 15 Monitor State Claim Volume Focused on individual states over an extended period of time (6 12 months) by Open and Closed Compare NCCI estimated claims based on Unit Statistical data with carrier supplied claims Review Unit Statistical data claim pool Review Sampling ratios 16 8

11 Outreach Monitor State Claim Volume NCCI can provide: Open or Closed Unit Statistical claims for a specific time period Compare the estimated Unit Statistical pool provided by NCCI with carrier s DCI pool of claims Common reasons for overreporting or underreporting: Subset of claims is not part of the pool Incorrect sampling ratio 17 Death and Permanent Total (PT) Claims Compare Unit Statistical Death and PT claims with DCI claims Claims with Injury Code 01 or 02 in Unit Statistical data not in DCI data Supply claim detail to the carrier Why was that claim not submitted to DCI? Verify logic for selection of Death and PT claims Common responses for missing DCI claims Only pulling 1st report level on the unit Not using Injury Codes to identify claims for DCI 18 9

12 DCI Match Game Questions Common underreporting issue Poor data quality can lead to Per claim data quality process Used to identify Death and Permanent Total claims Reasonableness data quality process Answers Edits Aggregate observations Pool of claims Unreliable usage Unit Statistical data 19 Data Element Data Quality 10

13 Data Element Data Quality Definition of the Data Element from claim handler perspective How it is tied to state statutes and possible legislative reform How the reporting of the element impacts usage of the data How NCCI determines outlier observations 21 Benefit Type Code Corresponds to the type of indemnity benefits that are being paid to the claimant at valuation Codes are state specific based on state statues Benefit Type Code May not be the same as the Injury Code reported on Unit Statistical data If a reserve has been established, report the code that reflects the anticipated indemnity payment 22 11

14 Benefit Type Code Code Benefit Type Code Used for: 01 Death 02 Permanent Total Disability 03 Scheduled Permanent Partial 04 Unscheduled Permanent Partial 05 Temporary Total Injury 09 Disfigurement 11 Temporary Partial 12 Employers Liability 15 Supplemental (LA, SD only) 50 Other Specified Indemnity Benefits 23 Benefit Type Code Code Benefit Type Code Used for: 01 Death 02 Permanent Total Disability 03 Scheduled Permanent Partial 04 Unscheduled Permanent Partial 05 Temporary Total Injury 09 Disfigurement 11 Temporary Partial 12 Employers Liability 15 Supplemental (LA, SD only) 50 Other Specified Indemnity Benefits 24 12

15 Permanent Partial Disability (PPD) Code Benefit Type Code Used for: 01 Death 02 Permanent Total Disability 03 Scheduled Permanent Partial 04 Unscheduled Permanent Partial 05 Temporary Total Injury 09 Disfigurement 11 Temporary Partial 12 Employers Liability 15 Supplemental (LA, SD only) 50 Other Specified Indemnity Benefits 25 Permanent Partial Disability (PPD) Code Benefit Type Code Used for: 01 Death 02 Permanent Total Disability 03 Scheduled Permanent Partial 04 Unscheduled Permanent Partial Edit Number Data Field Edit Message 05 Temporary Total Injury Benefit Type Code Must be valid for the 09 Disfigurement reported Jurisdiction 11 Temporary Partial State Code or equal 12 Employers Liability zero. 15 Supplemental (LA, SD only) 50 Other Specified Indemnity Benefits 26 13

16 Benefit Type Code Injury Code Benefit Type Code 27 State Statutes and Benefit Type State statutes determine a state s benefit types States that have both Scheduled and Unscheduled PPD are more problematic Example of Scheduled PPD Total Loss of Use: For the loss of a thumb, 62 weeks For the loss of a hand, 170 weeks For the loss of an arm, 222 weeks For the loss of a foot, 139 weeks 28 14

17 State Statutes and Benefit Type Partial Loss of Use as defined by a state may be Scheduled or Unscheduled Scheduled Partial Loss of Use; number of weeks of PPD is based on (percentage of loss) x (number of scheduled weeks for total loss) Example of Partial Loss of Arm: Total loss of an arm is 222 weeks Percentage of loss is 30% Claimant receives weekly PPD payment of (.30 x 222) or 66.6 weeks 29 State Statutes and Benefit Type Unscheduled PPD compensates the claimant for permanent injury not covered by the PPD schedule Examples of Unscheduled PPD: Difference between pre-injury wage and post-injury wage until claimant s earnings are back to pre-injury wage or exceed a state cap Indicates a specific number of weeks multiplied by a disability rating based on the whole body 30 15

18 Carrier Payment Codes Benefit Types are often mapped from a carrier s internal payment codes. Payment Code Lump Sum or Periodic Payment Payment Code Mapping Claim handler must use correct payment code Claim handler must determine if lump sum or periodic payment Payment Code must map to correct DCI Benefit Type 31 Carrier Payment Codes Example Claimant is receiving scheduled PPD benefits as a weekly payment: Payment Code Lump Sum or Periodic Payment Claim handler uses scheduled PPD payment code Since periodic, claim handler determines weekly rate and number of weeks Reported to DCI Mapped as periodic Benefit Type 03 with amount paid and weekly rate 32 16

19 Comparative Analysis on Benefit Types Within PPD Reported Industry 0.8% BT % BT 03 Group B 1.0% BT % BT % BT % BT Subsequent Valuations Benefit Type payments reported on prior valuations shouldn t go away or be reduced; expected in rare cases (e.g., a payment was voided and reclassified) Benefit Type was not correct on the prior valuation; fix those prior valuations The claim settlements process erased those prior payments; retain the payments that were made prior to the final settlement 34 17

20 Lump Sum Payments DCI Reporting Guidebook was updated in 2014 to define lump sum as either a lump sum payment or a lump sum settlement Use lump sum payments when the weekly rate that was paid for a benefit type isn t known Generally lump sum payments are still based on state statutes Two unique lump sum benefit type values 35 Benefits Covered by Lump Sum Benefit Code When to report medical dollars as a lump sum Benefits Covered Code 06 Medical Only 49 Indemnity and Medical Combined Description Lump sum benefit includes only medical payments Lump sum benefit includes indemnity and medical payments combined 36 18

21 Legal and Final Settlements Medical and Indemnity Comingled Benefits Covered by Lump-Sum Benefit Code Legal Settlements Allocate the appropriate dollars to each benefit type covered in the settlement. Start Finish 38 19

22 Use of Benefit Type 49 Percent Lump Sum Paid Set to % 80% 60% 40% 20% 0% Percent of Lump Sum Benefits Set to 49 Indemnity and Medical Combined Coverage Provider Groups 39 Pre-Injury and Post-Injury Wage Pre-Injury weekly wage is the wage the claimant was earning when injured Post-Injury weekly wage is the wage the claimant is earning when they return to work Post-Injury weekly wage is linked to claimant benefits determined by a wage differential: Temporary Partial Disability Wage Loss Rate of Pay Indicator is also tied to this situation 40 20

23 Return To Work Rate of Pay Indicator If the Claimant goes back to work at the same hours AND rate of pay Claimant has not returned to work Claimant returns to work with ANY of the following: Reduced hours Reduced rate of pay Restrictions Light duty State provides additional benefits Post injury wage doesn t impact post RTW benefits being paid Then report 41 Setting of RTW Indicator Percent RTW Rate Indicator Set to N 100% 80% 60% 40% 20% 0% Return to Work Rate of Pay Indicator Set to N (No) Coverage Provider Groups 42 21

24 Impairment Percentage and Basic Code Related Elements Some DCI elements are related Impairment Percentage Impairment Basis Code Body Part Impairment Basis Code relates to the Impairment Percentage, not the Body Part 10% or 25% Whole Body or Part of Body Arm or Multiple Keep Them in Sync 43 Maximum Medical Improvement (MMI) Date Provided by a physician when the claimant s injury is stabilized Permanent Disability rating can now be determined by the physician Often the claimant returns to work and PPD benefits are paid instead of Temporary Disability (TD) benefits Observations: MMI date is equal to closed date MMI date is within 30 days of Accident Date 44 22

25 Medical Extinguish Indicator Indicates if the medical payments are extinguish based on a lump sum settlement agreement Based on reasonable expectation that insured will not be obligated to make future medical payments Observation: Indicator is Y, claim is Open, and there is an outstanding medical reserve 45 Attorney Indicator Indicates whether the claimant has an attorney or authorized representative Claimant attorney representation varies by state and claim status Observation: Carrier percentage of Y compared with the industry on a per-state basis, by open and closed claims 46 23

26 Enhancement to DCI in 2016/2017 DCI Data Quality Views screens Observation results will be available to carriers System will assist carriers in monitoring data element observation outliers System will allow carriers to download claims associated with a data element observation E.g., claims where MMI Data is within 30 days of Accident Date 47 DCI Match Game Questions The Return to Work Rate of Pay Indicator Payment Codes What does Benefit Type Code 09 represent? In legal settlements When should the Return to Work Date be reported? Answers Disfigurement If the claimant has returned to work Will be blank if the claimant has not returned to work Must be mapped to correct Benefit Type Try to allocate to specific Benefit Type Codes 48 24

27 Life of a Claim Case Study: Complex Claim Development First Valuation (018) Claim is open Claimant was out of work for 25 weeks Claimant returned to work and was assigned light duty Claimant is still being treated by a physician 50 25

28 Case Study: Complex Claim Development First Valuation (018) Scenario: Claim is open Claimant was out of work for 25 weeks Claimant returned to work and was assigned light duty Claimant is still being treated by a physician What data should be provided? Benefit type Type while 05 as periodic claimant benefit with out of TTD work Amount Paid and TTD Weekly Rate Benefit type Type when 11 as periodic claimant benefit with TPD amount and TPD weekly returned rate to at work time of valuation Return to Work Date Data Post Injury Elements Wage to be reported Return to Work Rate of Pay Indicator No Should MMI be reported? 51 Case Study: Complex Claim Development Second Valuation (030) Claim is closed Claimant is back to full duty Attending physician has determined claimant will no longer benefit from further medical treatment and assigned a disability percentage of 25% for partial loss of use of the arm Carrier agrees with disability percentage and paying the PPD as a lump-sum payment 52 26

29 Case Study: Complex Claim Development Second Valuation (030) Scenario: Claim is closed Benefit Type that has ended Claimant is back to full duty Physician has assigned a disability percentage of 25% for partial loss of use of the arm Carrier agrees with disability percentage and paying the PPD as a lump sum payment T TPD has ended and BT-11 amount paid reflects full amount paid Physician has supplied two data elements related to PPD Maximum Medical Date and Disability Rating of 25% Carrier enters PPD data into their Impairment Percentage of 25% Impairment Basis system Code of 2 for Part of Body PPD Benefit Type that was paid PPD is Scheduled Benefit Type 03 paid as Lump Sum 53 Review 27

30 Review Name the four data elements tied to a claimant s work status: 55 Review What are the three steps NCCI uses in the carrier outreach process? 56 28

31 Review What are the 3 PPD Benefit Types and which one is often confused with Injury Code? 57 Review What are the 3 PPD Benefit Types and which one is often confused with Injury Code? 58 29

32 Review What are the 3 PPD Benefit Types and which one is often confused with Injury Code? Code Benefit Type Code Used for: 01 Death 02 Permanent Total Disability 05 Temporary Total Injury 11 Temporary Partial 12 Employers Liability 15 Supplemental (LA, SD only) 50 Other Specified Indemnity Benefits 59 Review Which of the following statements describe why aggregate data quality observations are necessary? A. Edits looking at individual claims cannot determine if the data reported is systemically influenced B. Observations look for patterns or anomalies and check for reasonableness 1. A only 2. B only 3. Both A and B 4. Neither A nor B 60 30

33 Review True or False? The Impairment Percentage Basis Code value of 2 for Part of Body should be used when the claimant has a permanent injury to a specific body part. 61 Review Related to Return to Work Date, which of the following is false? A. RTW Date should be left blank (zero filled) when the claimant has not returned to work B. NCCI evaluates RTW Date for closed claims C. When the claimant has multiple time-off-work occurrences, report the initial RTW Date 62 31

34 63 32

35 Supplemental Information

36

37 Presenter Biographies James Bonk is the director of NCCI s Medical and DCI Data Validation Department. Medical and DCI Data Validation encompasses the collection, editing, storage, and validation of data associated with NCCI s Medical Data Call and DCI Data Call, including the development and maintenance of NCCI s data compliance programs. Since joining NCCI in 2003, James has held numerous management and technical positions. Previously, he worked for HNC Software as director of software development, where he invented MIRA, an automated case reserves software tool. Prior to that, James worked at Medata and CompReview, two leading medical bill review software companies. James received a bachelor of science degree in computer science from Fullerton College in California. Joanne Missry is the data consultant for the Detailed Claim Information (DCI) Operations Team. She is primarily responsible for identifying, developing, and implementing solutions to data quality issues impacting DCI data used for NCCI s core products, research, and legislative analysis. This includes technical support to business partners and data users on data quality concerns. Joanne worked at Assurant Group and Courtesy Insurance Company in both the actuarial and financial analysis areas. She has a mathematics and statistics background, and has passed several actuarial exams. Prior to joining NCCI, her roles varied from analyzing and validating data and building models to conducting actuarial analysis for carrier filings and pricing. Additionally, she has strong technical skills as she created many process efficiencies using Excel, SAS, and SQL. Joanne earned an MBA from George Washington University and a BA degree in mathematics and economics from the University of Rhode Island.