denials management solution using machine intelligence · case study 13 identified new patterns of...
TRANSCRIPT
Denials Management Solution using Machine Intelligence
1
Company Confidential & Proprietary 2
Company Confidential & Proprietary
The Revenue Opportunity
3
__ __ __
1 2 4 TYPEOF BILL
FROM THROUGH5 FED. TAX NO.
a
b
c
d
DX
ECI
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
A
B
C
A B C D E F G HI J K L M N O P Q
a b c a b c
a
b c d
ADMISSION CONDITION CODESDATE
OCCURRENCE OCCURRENCE OCCURRENCE OCCURRENCE SPAN OCCURRENCE SPANCODE DATE CODE CODE CODE DATE CODE THROUGH
VALUE CODES VALUE CODES VALUE CODESCODE AMOUNT CODE AMOUNT CODE AMOUNT
TOTALS
PRINCIPAL PROCEDURE a. OTHER PROCEDURE b. OTHER PROCEDURE NPICODE DATE CODE DATE CODE DATE
FIRST
c. d. e. OTHER PROCEDURE NPICODE DATE DATE
FIRST
NPI
b LAST FIRST
c NPI
d LAST FIRST
UB-04 CMS-1450
7
10 BIRTHDATE 11 SEX 12 13 HR 14 TYPE 15 SRC
DATE
16 DHR 18 19 20
FROM
21 2522 26 2823 27
CODE FROM
DATE
OTHER
PRV ID
THE CERTIFICATIONS ON THE REVERSE APPLY TO THIS BILL AND ARE MADE A PART HEREOF.
b
.INFO
BEN.
CODEOTHER PROCEDURE
THROUGH
29 ACDT 30
3231 33 34 35 36 37
38 39 40 41
42 REV. CD. 43 DESCRIPTION 45 SERV. DATE 46 SERV. UNITS 47 TOTAL CHARGES 48 NON-COVERED CHARGES 49
52 REL51 HEALTH PLAN ID 53 ASG. 54 PRIOR PAYMENTS 55 EST. AMOUNT DUE 56 NPI
57
58 INSURED’S NAME 59 P.REL 60 INSURED’S UNIQUE ID 61 GROUP NAME 62 INSURANCE GROUP NO.
64 DOCUMENT CONTROL NUMBER 65 EMPLOYER NAME
66 67 68
69 ADMIT 70 PATIENT 72 73
74 75 76 ATTENDING
80 REMARKS
OTHER PROCEDURE
a
77 OPERATING
78 OTHER
79 OTHER
81CC
CREATION DATE
3a PAT.CNTL #
24
b. MED.REC. #
44 HCPCS / RATE / HIPPS CODE
PAGE OF
APPROVED OMB NO. 0938-0997
e
a8 PATIENT NAME
50 PAYER NAME
63 TREATMENT AUTHORIZATION CODES
6 STATEMENT COVERS PERIOD
9 PATIENT ADDRESS
17 STAT STATE
DX REASON DX 71 PPS
CODE
QUAL
LAST
LAST
National UniformBilling CommitteeNUBC™
OCCURRENCE
QUAL
QUAL
QUAL
CODE DATE
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
a
b
a
b
Individual Claim
28%
11% 44%
17%
Macro Level Summary
Scope of Analysis Narrow Broad
Opportunity
Low
High
Company Confidential & Proprietary
The Revenue Opportunity
4
__ __ __
1 2 4 TYPEOF BILL
FROM THROUGH5 FED. TAX NO.
a
b
c
d
DX
ECI
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
A
B
C
A B C D E F G HI J K L M N O P Q
a b c a b c
a
b c d
ADMISSION CONDITION CODESDATE
OCCURRENCE OCCURRENCE OCCURRENCE OCCURRENCE SPAN OCCURRENCE SPANCODE DATE CODE CODE CODE DATE CODE THROUGH
VALUE CODES VALUE CODES VALUE CODESCODE AMOUNT CODE AMOUNT CODE AMOUNT
TOTALS
PRINCIPAL PROCEDURE a. OTHER PROCEDURE b. OTHER PROCEDURE NPICODE DATE CODE DATE CODE DATE
FIRST
c. d. e. OTHER PROCEDURE NPICODE DATE DATE
FIRST
NPI
b LAST FIRST
c NPI
d LAST FIRST
UB-04 CMS-1450
7
10 BIRTHDATE 11 SEX 12 13 HR 14 TYPE 15 SRC
DATE
16 DHR 18 19 20
FROM
21 2522 26 2823 27
CODE FROM
DATE
OTHER
PRV ID
THE CERTIFICATIONS ON THE REVERSE APPLY TO THIS BILL AND ARE MADE A PART HEREOF.
b
.INFO
BEN.
CODEOTHER PROCEDURE
THROUGH
29 ACDT 30
3231 33 34 35 36 37
38 39 40 41
42 REV. CD. 43 DESCRIPTION 45 SERV. DATE 46 SERV. UNITS 47 TOTAL CHARGES 48 NON-COVERED CHARGES 49
52 REL51 HEALTH PLAN ID 53 ASG. 54 PRIOR PAYMENTS 55 EST. AMOUNT DUE 56 NPI
57
58 INSURED’S NAME 59 P.REL 60 INSURED’S UNIQUE ID 61 GROUP NAME 62 INSURANCE GROUP NO.
64 DOCUMENT CONTROL NUMBER 65 EMPLOYER NAME
66 67 68
69 ADMIT 70 PATIENT 72 73
74 75 76 ATTENDING
80 REMARKS
OTHER PROCEDURE
a
77 OPERATING
78 OTHER
79 OTHER
81CC
CREATION DATE
3a PAT.CNTL #
24
b. MED.REC. #
44 HCPCS / RATE / HIPPS CODE
PAGE OF
APPROVED OMB NO. 0938-0997
e
a8 PATIENT NAME
50 PAYER NAME
63 TREATMENT AUTHORIZATION CODES
6 STATEMENT COVERS PERIOD
9 PATIENT ADDRESS
17 STAT STATE
DX REASON DX 71 PPS
CODE
QUAL
LAST
LAST
National UniformBilling CommitteeNUBC™
OCCURRENCE
QUAL
QUAL
QUAL
CODE DATE
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
a
b
a
b
Individual Claim
28%
11% 44%
17%
Macro Level Summary
Scope of Analysis Narrow Broad
Opportunity
Low
High
The Denials Opportunity
Company Confidential & Proprietary
Coding
Traditional Analytics Have Hit the Wall
5
Hypotheses Data Analysts Insight
Weeks/Months/Years/Never
Company Confidential & Proprietary
Complexity is the Challenge
6
c Dataset
# of possible insights in Dataset is exponential
function of size
# of possible insights in the Dataset after it grows another 40%
Company Confidential & Proprietary
The Solution is Machine Intelligence
7
Automated Discovery
Data Insights
Minutes
Decision Maker
Company Confidential & Proprietary
How It Works
8
• Apply algorithms to a data set
• Statistical • Geometric • Machine learning
• Construct a multi-faceted notion of similarity
• Cluster similar data points
• Create similarity maps that reveal mid-level patterns
Company Confidential & Proprietary
Similarity Maps Using Machine Intelligence
9
Nodes are groups of similar data points Edges connect similar nodes
Company Confidential & Proprietary
Surface Mid-Level Patterns of Denials
10
Discover distinct groups of similar claims that have high concentration of denials
Concentration of Denied Claims
Low High
Company Confidential & Proprietary
Identify Hot Spots
11
Uncover the unique attributes of each group to produce rich descriptions of denials patterns
Concentration of Denied Claims
Low High
Company Confidential & Proprietary
How Machine Intelligence Reduces Denials
1. Identifies drivers of rejections and denials for groups of claims
2. Prioritizes & streamlines work queue for claims resubmission
3. Informs upstream process changes to prevent future denials
12
Company Confidential & Proprietary
Case Study
13
Identified new patterns of denied claims. Identified specific drivers for known patterns of denials, suggesting new options for intervention.
SOLUTION
Despite years of effort, the provider continued to write off tens of millions of dollars annually.
PROBLEM
Solution is projected to address 70% of denied dollars and will continue to identify new denials patterns as they arise.
OUTCOME
Demo
Company Confidential & Proprietary
Data included: • UB-04 hospital billing data • Additional account data from provider’s
financial databases • Associated 835 remit transactions (denials,
rejections, underpayments, payments) • Primary payer data • Personal or family guarantors
__ __ __
1 2 4 TYPEOF BILL
FROM THROUGH5 FED. TAX NO.
a
b
c
d
DX
ECI
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
A
B
C
A B C D E F G HI J K L M N O P Q
a b c a b c
a
b c d
ADMISSION CONDITION CODESDATE
OCCURRENCE OCCURRENCE OCCURRENCE OCCURRENCE SPAN OCCURRENCE SPANCODE DATE CODE CODE CODE DATE CODE THROUGH
VALUE CODES VALUE CODES VALUE CODESCODE AMOUNT CODE AMOUNT CODE AMOUNT
TOTALS
PRINCIPAL PROCEDURE a. OTHER PROCEDURE b. OTHER PROCEDURE NPICODE DATE CODE DATE CODE DATE
FIRST
c. d. e. OTHER PROCEDURE NPICODE DATE DATE
FIRST
NPI
b LAST FIRST
c NPI
d LAST FIRST
UB-04 CMS-1450
7
10 BIRTHDATE 11 SEX 12 13 HR 14 TYPE 15 SRC
DATE
16 DHR 18 19 20
FROM
21 2522 26 2823 27
CODE FROM
DATE
OTHER
PRV ID
THE CERTIFICATIONS ON THE REVERSE APPLY TO THIS BILL AND ARE MADE A PART HEREOF.
b
.INFO
BEN.
CODEOTHER PROCEDURE
THROUGH
29 ACDT 30
3231 33 34 35 36 37
38 39 40 41
42 REV. CD. 43 DESCRIPTION 45 SERV. DATE 46 SERV. UNITS 47 TOTAL CHARGES 48 NON-COVERED CHARGES 49
52 REL51 HEALTH PLAN ID 53 ASG. 54 PRIOR PAYMENTS 55 EST. AMOUNT DUE 56 NPI
57
58 INSURED’S NAME 59 P.REL 60 INSURED’S UNIQUE ID 61 GROUP NAME 62 INSURANCE GROUP NO.
64 DOCUMENT CONTROL NUMBER 65 EMPLOYER NAME
66 67 68
69 ADMIT 70 PATIENT 72 73
74 75 76 ATTENDING
80 REMARKS
OTHER PROCEDURE
a
77 OPERATING
78 OTHER
79 OTHER
81CC
CREATION DATE
3a PAT.CNTL #
24
b. MED.REC. #
44 HCPCS / RATE / HIPPS CODE
PAGE OF
APPROVED OMB NO. 0938-0997
e
a8 PATIENT NAME
50 PAYER NAME
63 TREATMENT AUTHORIZATION CODES
6 STATEMENT COVERS PERIOD
9 PATIENT ADDRESS
17 STAT STATE
DX REASON DX 71 PPS
CODE
QUAL
LAST
LAST
National UniformBilling CommitteeNUBC™
OCCURRENCE
QUAL
QUAL
QUAL
CODE DATE
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
a
b
a
b
15
Example of data used for analysis Typical claim form
Company Confidential & Proprietary
Investigate Hot Spots
Staging DB, Data
Transforms
Generate Similarity Maps
Color by % of
Denials
Denials Management Workflow
16
Provider
Claims Data Output Denials Hot Spot
Report
Identify drivers of rejections Drive process modifications Reduce denials upstream
Company Confidential & Proprietary
How It Works
17
• Use a multi-faceted notion of similarity to cluster similar data points
• Create similarity maps that reveal mid-level
patterns
• Nodes are groups of similar data points • Edges connect similar nodes • Colors let you see values of interest • Position of a node on the screen doesn’t matter
Company Confidential & Proprietary
Uncovering Characteristics of a Hotspot
18
HOTSPOT ANALYSIS Dataset: denials_100K_wide_150429.tsv Network Parameters: Prin DX Codes + HCPCS + Payor1 (Red = under-‐represented) Summary Medical Necessity Rejec;ons Medical Necessity Denials Hotspot Claims Volume Percent Amount Volume Percent Amount Island A 2947 2590 87.89% $2,602,381 158 5.36% $34,357
Top Payors % Rejected Rejected $ Top Providers % Rejected Rejected $ Top Bill Types % in Group Rejected $ PAYOR 1 89.11% $2,318,982 Clinic 1 45.13% $1,174,455 **131 89.75% $2,335,637 PAYOR 2 1.97% $51,267 Hospital 1 12.76% $332,064 **851 9.30% $242,021 PAYOR 3 1.39% $36,173 Hospital 2 10.72% $278,975 TOTAL 99.05% $2,577,659 Hospital 3 4.21% $109,560 TOTAL 92.47% $2,406,422 TOTAL 72.82% $1,895,054
Revenue Codes Percent HCPCS Descrip;on Code Percent Medical/Surgical Supplies And Devices -‐ Sterile Supply 98.47% Normal saline soluYon infus J7040 54.22% Pharmacy -‐ Drugs With Detailed Coding 98.27% Tissue exam by pathologist 88305 36.34% Recovery Room -‐ General 97.93% Colonoscopy and biopsy 45380 26.87%
Gastro-‐IntesYnal Services -‐ General 94.03% Colonoscopy w/lesion removal 45385 22.87%
Pharmacy -‐ General 81.20% Colon ca scrn not hi rsk ind G0121 22.63% Anesthesia -‐ General 72.62% Colorectal scrn; hi risk ind G0105 17.92%
Laboratory Pathological -‐ Histology 36.41% Colonoscopy w/lesion removal 45384 11.03%
Medical/Surgical Supplies And Devices -‐ General 32.10% Anesthesia -‐ Anesthesia Incident To Other DiagnosYc Services 22.57% Laboratory -‐ Chemistry 18.49%
Company Confidential & Proprietary
Recap
1. Identifies drivers of rejections and denials for groups of claims
2. Prioritizes & streamlines work queue for claims resubmission
3. Informs upstream process changes to prevent future denials
19
Company Confidential & Proprietary
Facilitating the ICD-10 Transition
20
Code Agnostic
Understand the Long
Tail
Infer New Rules as they
Arise
Automated & Adaptive
Company Confidential & Proprietary
Re-Cap, Documents + Q&A
• The challenge of denials management is complexity – the low-hanging fruit is gone • Using Machine Intelligence, providers can find
actionable patterns of denials • This presentation and other docs in Handouts • To connect with me: [email protected]
or ayasdi.com/healthcare or [email protected]
21