Transcript Pharmaceutical Research Databases and Consulting Capabilities

Information Assets and Outcomes Research
Capabilities at
The Medstat Group, Inc.
Ron Ozminkowski, Ph.D.
Director, Health and Productivity Management Research
Agenda
• Medstat is ….
• Thomson Corporation
• MarketScan® Research Databases
• Pharmaceutical Research Capabilities
• How Do We Get Started?
• What Happens Next?
• Market Signals
2
Copyright 2002 Thomson Medstat
Medstat is ….
A healthcare information company that:
• Helps customers manage the cost and quality of health
services and benefits
–
–
–
–
Research services
Analytical consulting
Software applications
Databases
• Serves all major market segments
• Has longest track record in industry (founded 1981)
• Employs approximately 750 employees in nine locations
• Is part of The Thomson Corporation
3
Copyright 2002 Thomson Medstat
Medstat is Part of the Thomson Corporation
• Global e-information and solutions company with annual
revenues of approximately $7.2 billion
• Operations organized into four market groups:
–
–
–
–
4
Thomson Legal & Regulatory
Thomson Financial (includes First Call)
Thomson Learning
Thomson Scientific & Healthcare
• Thomson Scientific
• Thomson Healthcare
• Medstat
Copyright 2002 Thomson Medstat
Medstat Customer Profile
Market Sector
Customer Overview
• Pharmaceutical
Most U.S.-based manufacturers
• Employers
70+ large employers
• State Gov’t
20 state programs
• Federal Gov’t
CMS, AHRQ, CDC, SAMHSA, DOD
• Health Plans
100+ health plans
• Providers
1,000+ hospitals, IDNs, etc.
5
Copyright 2002 Thomson Medstat
Three Principal Product Lines
• Decision Support Systems and Solutions
– Adding value to internal information by developing databases,
applying methods to help understand the data, and developing the
tools necessary to organize the data for better decision making.
• Information Solutions
– Integrating external market intelligence and applying methods to
benchmark, forecast, and analyze for improved decision making.
• Research
– Helping the research community uncover important information for
better decision making and to fuel internal product development
and innovation
6
Copyright 2002 Thomson Medstat
MarketScan Database Development
Customer Internal Data
Data Management
Process
Eligibility

Inpatient claims
Outpatient claims



Prescription drugs


Privacy Protection
Integration
Standardization
Customization
Enhancement
QA/Improvement
Medstat
MarketScan
Databases
75 million facility and professional claim lines and 40 million drug claims annually
7
Copyright 2002 Thomson Medstat
MarketScan® Research Databases
• Large multi-source database of privately insured claim and
encounter data
– These sources ensure that MarketScan reflects the complexity
of the real world of healthcare delivery
– Databases represent true continuum of care, including carveouts
– 4 million+ annual covered lives
• Longitudinal data
– Databases support person-level analyses, with three to five
years of follow up
• The basis of more than 40 peer-reviewed journal articles in
the last three years
8
Copyright 2002 Thomson Medstat
How MarketScan Databases Are Built
• Raw data are collected from approximately 100 payers to build
individual client databases
– Variables standardized
– Clinical detail added
– Inpatient cases built
• Standard variables from individual databases are combined to form
the MarketScan Databases
–
–
–
–
Synthetic patient identifiers assigned, consistent across years
Clinical categories added (e.g., DRGs, DCGs)
Plan type identified (e.g., PPO, POS, comprehensive)
Place, service types, provider types, and industry classifications
standardized
– MSA and FIPS state-county codes added
9
Copyright 2002 Thomson Medstat
MarketScan Features
• Fully adjudicated claims data
• Consistent data definitions
• Inpatient and outpatient claims can be linked to enrollment and
drug data
• Consistent patient identifiers across years, for longitudinal analysis
• Linked drug files containing NDC, therapeutic class, copayment,
total payment, etc.
• Clinical classification systems such as MDCs, DRGs, DCGs, and
Disease Staging can be applied to the data
10
Copyright 2002 Thomson Medstat
MarketScan Database Organization
Standard Files
– Inpatient Admissions File (one record per admission)
– Inpatient Services File (one record for each service)
– Outpatient Services File
– Outpatient Pharmaceutical Claims File
– Person-Level Enrollment Files
11
Copyright 2002 Thomson Medstat
Medical Claims: Sample Data Elements
Demographic
Clinical
Provider
Patient ID
Admit Date
Provider ID
Age
Admit Type
Provider Type
Copay Amount
Gender
Length of Stay
Place of Service
Deductible Amount
Employment Class
DRG
Provider Specialty
Hospital Payment/Charge
Relationship to Benefic.
Principal Diagnosis
Geographic Location
Dx2-Dx14
Industry
MDC
Other Payment/Charge
Principal Procedure
Total Payment/Charge
Proc2-Proc14
Payment/Charge
Service Date
Service Type
12
Copyright 2002 Thomson Medstat
Financial
Physician Payment/Charge
Drug Claims: Sample Data Elements
Demographic
Drug
Provider
Financial
Patient ID
National Drug Code
Pharmacy ID
Payment
Age
Dispense As Written
Pharmacy Class Code
Average Wholesale Price
Gender
Date Filled
Copay Amount
Employment Class
Therapeutic Class
Ingredient Cost
Relationship of
Therapeutic Group
Dispensing Fee
Patient to Beneficiary
Geographic Location
Days Supply
Generic Product ID
Industry
13
Copyright 2002 Thomson Medstat
Commercial Claims and Encounters (CC&E)
• Inpatient and outpatient medical claims linked to drug data,
person-level enrollment data, and benefit plan design data
• Fee-for-service (FFS) with drug data available since 1992,
encounter with drug data available since 1994
• FFS and fully and partially capitated plans represented
–
–
–
–
–
14
Exclusive provider organizations
Preferred provider organizations
Point of service plans
Indemnity plans
Health maintenance organizations
Copyright 2002 Thomson Medstat
Commercial Covered Lives
Database/
Year
Total Covered
Lives*
With Drug Data*
With Drug and
Enrollment Data*
Commercial Claims and Encounters
1997
1998
1999
2000
2001
•
15
4,800,000
4,000,000
3,550,000
3,620,000
3,530,000
*Numbers are rounded
1,950,000 (41%)
1,900,000 (48%)
2,620,000 (74%)
2,790,000 (77%)
3,060,000 (87%)
1,190,000 (25%)
1,580,000 (40%)
2,500,000 (70%)
2,790,000 (77%)
3,060,000 (87%)
Note: Approximately 85 percent of covered lives
have at least one claim
Copyright 2002 Thomson Medstat
Medicare Supplemental and COB
• Inpatient and outpatient Medicare supplemental medical
claims linked to drug data, person-level enrollment data, and
benefit plan design data
• Available since 1995
• Both the employer-paid and Medicare-paid components of
care represented
16
Copyright 2002 Thomson Medstat
Medicare Covered Lives
Database/
Year
Total Covered
Lives*
With Drug Data*
With Drug and
Enrollment Data*
Medicare Supplemental and COB
1997
1998
1999
2000
2001
•
17
590,000
700,000
550,000
540,000
500,000
Numbers are rounded
Copyright 2002 Thomson Medstat
220,000 (37%)
490,000 (70%)
450,000 (82%)
480,000 (89%)
460,000 (92%)
N/A
290,000 (41%)
420,000 (76%)
480,000 (89%)
460,000 (92%)
MarketScan Databases
• Commercial Claims and Encounters (1992–2002)
– 2002 data to be complete in December
• MarketScan Early View Data Set (July, 2002 – June, 2003)
– A subset of the CCAE data set for 42 large employers
• Medicare Supplemental and COB (1995–2002)
– Ditto re: 2002 data
• Benefit Plan Design (1995–2002)
– 2002 data to be complete in January
• Health and Productivity Management (1997–2001)
– Contains information about illness-related absenteeism, STD, and WC
use, linkable to medical and drug claims and enrollment files
• Medicaid – three states (1999-2001)
– 2 more states to be added soon, with 2002 data
18
Copyright 2002 Thomson Medstat
How Customers Access MarketScan
• Database licensure
– Complete database, multiple years
– Subset of database, multiple years
• Custom reports
• Outcomes research studies
• Web-based products
• Clinical trial recruitment products
19
Copyright 2002 Thomson Medstat
Pharmaceutical Research Capabilities
20
Copyright 2002 Thomson Medstat
Pharmaceutical Research Overview
• Total staff of more than 40
– Many with advanced degrees, including
• 10 PhDs
• 2 MDs
– Widely published in peer-reviewed journals
– Multiple research disciplines, including economics, health
services research, psychology, and statistics
• Focus on pharmacoeconomics and clinical effectiveness
– Retrospective database studies
– Prospective observational studies
– Disease management and program evaluations
21
Copyright 2002 Thomson Medstat
Solutions for the Pharmaceutical Marketplace
• Burden of illness
• Product differentiation
• Market assessment
– Pricing strategy
– Market sizing
– Portfolio selection
• Economic protocol design
• Decision analytic and budget impact modeling
• Clinical trial recruitment
22
Copyright 2002 Thomson Medstat
Methods Often Used in Outcomes Research Studies
• Selection bias adjustments (Heckman, IV)
• Decision analytic models
• Treatment-effect models
• Multivariate techniques
–
–
–
–
–
–
–
–
23
Hazard regression
Linear regression (OLS)
Logit, probit, and propensity score models
Multinomial logit, ordered probit
Latent class analysis
Count models
Duration/time-to-event
Simultaneous equations
Copyright 2002 Thomson Medstat
Prospective Studies: The Medstat Advantage
• Assessment of the economic implications of clinical trials
• Development of economic protocols for clinical trials
• Evaluation of quality of life, productivity, and health status
• Support for multi-site clinical and economic studies
• Site selection, training, and management
24
Copyright 2002 Thomson Medstat
Sample Prospective Study
• Schizophrenia Care and Assessment Program (SCAP)
–
–
–
–
Multi-year study of 2,400 people with schizophrenia at six sites
Baseline and repeated assessments of functioning
Detailed medication tracking
Several research themes, including diffusion of new technologies,
outcomes, differential treatment, and outcomes by race
– Results described in several conference presentations and a few
publications
25
Copyright 2002 Thomson Medstat
Clinical Trials Examples
• Depression trial
– Used MarketScan data to derive prices for service utilization
observed during a clinical trial. Statistical bootstrapping was
applied to compare cost effectiveness for alternative measures of
effectiveness (responders and remitters) as well as direct and
indirect costs.
• Bowel motility trial
– Linked hospital discharge and UB-92 data to clinical trial data and
found length of stay reductions and subsequent cost savings for
study patients.
• Intermittent claudication trial
– Assessed the cost-effectiveness of a medication for in the
treatment of intermittent claudication (IC).
26
Copyright 2002 Thomson Medstat
Decision Analytic and Budget Impact Modeling
• Combine clinical trial efficacy data on patient/drug response
with cost information from naturalistic claims data
• Estimate real costs based on expected response using age,
gender, dose, medication switching
• Model impact of proposed drug by entering characteristics of
health plan or employer population into software
• Econometric modeling programs include those developed for
allergic rhinitis, GERD, cholesterol management, and health
risk reduction programs
27
Copyright 2002 Thomson Medstat
Clinical Trial Recruitment
• We can locate the best clinical investigators in the best locations
• We can find the best patient groups to whom you should advertise
• We can tell you the best way to reach them
This means much faster and more effective patient recruitment
28
Copyright 2002 Thomson Medstat
Depression Age 65+: Affective Psychoses
Clinical Investigator: Dr. Katz
Area: Philadelphia, PA DMA
19407
19001
19009
19401
19116
19075
19462
19031
19405
19038
19046
19406
19115
19444
19154
19095
19399
19428
19118
19150
19027
19087
19111
19126
19035
19136
19085
Dr. Katz is in a high area
(red) for depression
office visits.
19114
19152
19012
19138
19119
19333
19128
19149
19141
19144
19120
19127
19010
19135
19072
19041
19003
19140
19129
08065
19066
19137
19132
19096
08077
19133
19073
19134
19131
19121
19083
19125
19122
19151
19008
19028
08057
19130
19139
19082
He has a clear recruiting
advantage.
08075
19124
19004
19026
19123
Dr.
Dr.
19104Katz
Katz

19102
08110
08109
08102
08052
08105
19106
19143
19050
19146
19147
08103
19064
08002
19018
19142
19148
08104
19023
19063
19037
19145
19065
19070
19079
08108
08107
19153
08034
19043
19081
19052
08033
19033
19086
19032
08030
19112
19074
08106
08059
19078
08003
08035
19094
19014
19029
19015
19022
08063
19113
08099
08007
08045
08093
08086
08049
08078
19013
08043
08029
08083
08026
08066
19017
08096
08084
08027
08097
08021
08061
08014
08020
08056
08090
08051
08032
Pop 65+
5 Mile Radius
10 Mile Radius
15 Mile Radius
20 Mile Radius
Quintile 1
Quintile 2
Quintile 3
Quintile 4
Quintile 5
Total
Copyright 2002 Claritas/The MEDSTAT Group
29
Copyright 2002 Thomson Medstat
637,230
277,124
116,787
28,810
7,333
1,067,284
Hospital Visits
6,575
2,842
1,186
292
74
10,968
Physician Visits
16,159
6,989
2,931
720
183
26,981
Total Visits
22,733
9,831
4,116
1,012
257
37,950
Depression Age 65+: Top Cluster Locations
Clinical Investigator: Dr. Katz
Area: Philadelphia, PA DMA
Inner Cities (green)
and Single City
Blues (red) are the
best PRIZM clusters
near Dr. Katz.
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Dr.
Dr. Katz
Katz
These are the
perfect targets for
direct mail.

Delaware
Delaware
Delaware
Delaware
Delaware
Delaware
Camden
Camden
Camden
Camden
Camden
Camden
Pop 65+
5
10
15
20
Mile
Mile
Mile
Mile
Radius
Radius
Radius
Radius
Copyright 2002 Claritas/The MEDSTAT Group
30
Copyright 2002 Thomson Medstat
Single City Blues
New Empty Nests
Gray Collars
Inner Cities
Middle America
Total
39,791
57,741
41,722
43,638
42,210
225,102
% Pop
17.68
25.65
18.53
19.39
18.75
100
Depressed Pop 65+
3,975
3,796
3,261
3,195
2,481
16,708
% Pop 65+
10
6.6
7.8
7.3
5.9
7.4
How Do We Get Started?
• Sometimes prospective clients come to us with a question
• For example, do patients who use SSRIs for depression cost less than
patients who use TCAs?
– The issue is total cost, not just cost of the drug
– (Want to know the consequences of drug choice)
– Also want to know how it works in real life, not in a clinical trial in this case.
• We talk to the client, then write a proposal
–
–
–
–
–
–
–
–
31
Background on depression and pharmaceutical treatment
Description of study questions and hypotheses to test
Description of data assets that can be used to test those hypotheses
Statistical methods designed to provide most accurate answer to the
questions of interest (descriptive and multivariate approaches)
Description of project tasks, timeline, logistics
Pricing
Medstat experience and capabilities
Project staff
Copyright 2002 Thomson Medstat
What Happens Then?
• If we are selected to perform the work, a contract is negotiated.
• Then we do the work, and hopefully publish the results in a peerreviewed journal and make conference presentations
– In the past, we have done a lot of both.
• When clients are happy, they come back for more, and business
thrives.
• If clients are not happy, we try to make them happy.
• But credibility comes first and is never sacrificed.
32
Copyright 2002 Thomson Medstat
Market Signals
• The market suggests we are good at what we do.
– Business has grown from about $0.5 million in 1998 to over $5
million this year.
• We recruit good people and often hear compliments about their
work.
• We have fun!
33
Copyright 2002 Thomson Medstat