Cancer Registries and Medical

Records
Rich Data Resources
Carol Lowenstein, MBA, CTR
Assistant Director
Survey and Data Management Core
Dana-Farber Cancer Institute

Co-Coordinator, Survey & Statistical Methods Core

 Objectives
Describe cancer registries

Review the history of cancer registries why we collect data

Review the types of registries and data available

Understand the varied uses of registry data where to go, what to use,
and the limitations

Explore the value added through linkages of registry data with other data
sources

Discuss medical record review vs. cancer registry data

Describe medical record review

 What Is a Cancer Registry?
Information system designed to:
Collect
Manage
Analyze
data on persons with the diagnosis of a malignant or neoplastic disease
(cancer).

Cancer registries can be classified into three general types:

Hospital based registries: maintain data on all patients diagnosed
and/or treated for cancer at their facility and report cancer cases to the
central or state cancer registry as required by law.
Population-based central registries: maintain data on all cancer patients
within certain geographical areas.
Special purpose registries: maintain data on a particular type of cancer,
such as brain tumors.

Source: National Cancer Registrars Association

 What Is a Cancer Registry?

Registry data is:

Used to make public health decisions

A valuable research tool for those interested
in the etiology, diagnosis and treatment of
cancer
Used in fundamental research on the
epidemiology of cancer

Source: National Cancer Registrars Association

 What Is a Cancer Registry?
Information maintained in a cancer registry:

Demographic Information: Age, gender, race/ethnicity, birthplace and

residence.

Medical History: Physical findings, screening information, occupation and

any
history of a previous cancer.

Diagnostic Findings: Types, dates and results of procedures used to

diagnose cancer.
Cancer information: Primary site, cell type and extent of disease.

Cancer Therapy: Surgery, radiation therapy, chemotherapy, hormone or

immunotherapy.

Follow-up: Annual information concerning treatment, recurrence, and

patient status is updated to maintain accurate surveillance information.

Source: National Cancer Registrars Association

 What Is a Cancer Registry?

How are these data used?

Evaluate patient outcome, quality of life, and satisfaction issues and
implement procedures for improvement
Provide follow-up information for cancer surveillance
Calculate survival rates by various data items
Provide information for cancer program activities
Analyze referral patterns
Allocate resources at the health care facility, the community, region
or state level
Develop educational programs for health care providers, patients
and the general public
Report cancer incidence as required under state law
Evaluate efficacy of treatment modalities

Source: National Cancer Registrars Association

 History of Cancer Registries
Key Events
2500 B.C.
Earliest known description of "cancer": the "Edwin Smith" and "George
Ebers" papyri which describe surgery, pharmacology, and mechanical and
magical treatments

400 B.C.
Hippocrates described a breast "cancer" as "karkinoma" (known now as
carcinoma) during surgical removal of a tumor

1629 A.D.
Cancer is first mentioned as a cause of death in the Bills of Mortality in
England

1728
London's "General Census of Cancer" - the first known systematic collection
of information on cancer is generated

Source: NCI SEER Program

 History of Cancer Registries
Key Events
1839
Implementation of death registration (what we now know as "death
certification") in the United States

1901
Earliest known population-based systematic collection of data on people
with leprosy in Norway (a population-based leprosy registry)

1926
A bone sarcoma registry established by Dr. Ernest Codman at
Massachusetts General Hospital, one of the earliest registries established
for a specific type of cancer

The first hospital-based cancer registry at Yale-New Haven Hospital was

organized in New Haven, Connecticut
Source: NCI SEER Program
 History of Cancer Registries
Key Events
Ernest Amory Codman, M.D

Kept track of his patients with End Results Cards

Tracked the outcomes of patient treatments

Identified areas of improvement in patient care

Believed that this information should be made public

and that this, not seniority, be the basis for
physician promotion

Help found the American College of Surgeons and

the Joint Commission on Accreditation of
Healthcare Organizations

Source: Countway Medical Library

 History of Cancer Registries
Key Events
1935
First population-based cancer registry in the United States established in
Connecticut

1956
The American College of Surgeons requires a cancer registry as a component
of an approved cancer program

1971
The U.S. National Cancer Act budgets monies to the National Cancer Institute
for research, detection, and treatment of cancer

1973
The Surveillance, Epidemiology and End Results (SEER) Program of NCI
establishes the first national cancer registry program

Source: NCI SEER Program

 History of Cancer Registries
Key Events

1992
U.S. Public Law 102-515 establishes the National Program of Cancer
Registries (NPCR) and is administered by the US Centers for Disease
Control and Prevention (CDC)

1993
Many state laws make cancer a reportable disease

Source: NCI SEER Program

 History of Cancer Registries
Cancer Registration
Late 16th Century
Prompted by the seemingly random geographic ravages of the
plague, the English Crown appointed elderly, epidemic-scarred
women to prowl the countryside in search of the dead and dying.
These 'Ancient Matrons' published weekly 'Bills of Mortality' for
each parish, tabulating deaths by causes such as 'the purples'
(probably leukemia), 'riting of the lights', 'consumption' (often an
effect of cancer), and of course, the plague. Just how this
information was used is not recorded. Perhaps the royalty found it
helpful to determine where the plague was active so they could be
somewhere else

Source: NCI SEER Program

 History of Cancer Registries
Cancer Registration
Mid 17th Century
Around 1665 a London businessman, John Graunt, created medical
history by subjecting decades of mortality data to critical and
mathematical analysis. He literally invented the science of medical
epidemiology and statistics, publishing a pamphlet with 108
conclusions. The list included such revolutionary observations as
the facts that women saw physicians twice as often as men yet lived
longer, and plague epidemics moved outward from swampy areas.
Graunt also was the first person to use mortality statistics to project
population survival, probably by crudely fitting data samples to a
logarithmic curve.

Source: NCI SEER Program

 History of Cancer Registries
Cancer Registration
Early 1900s

Bone sarcoma registry at MGH

Yale-New Haven Hospital Cancer Registry

Other hospitals began to develop registries as

physicians saw the benefits of comprehensive data
collection

Connecticut population-based cancer registry

 History of Cancer Registries
Cancer Registration
1937-39
First National Cancer Survey
Initiative of the newly formed NCI
Directed by Harold F. Dorn
Included only MD dxd cases, histolgic confirmation
3 regions of the US north, south, west
10 registries in metropolitan areas, MD abstractors
Included 10% of the overall population,
Resulted in several papers
Sparked questions about genetic factors,
racial, gender and economic disparities Harold F. Dorn

Source: Lilienfeld, Am J Public Health, Dec 2008

 History of Cancer Registries
Cancer Registration
1947-48
Second National Cancer Survey

1956
American College of Surgeons Commission on
Cancer (CoC) requires hospital cancer registries for
Commission approved cancer programs

1967-71
Third National Cancer Survey
 History of Cancer Registries
Cancer Registration

1973
SEER Registry

1992
NPCR funding for state cancer registries
 History of Cancer Registries
Cancer Registration

What were they trying to achieve?

A standard classification of disease

A systematic form of data collection
Analysis and use of the data
Sources of Cancer Registry Data

Levels of registries
Hospital
Central
National
International
 Sources of Cancer Registry Data

Data collected using ICD-O

Based on ICD-9 codes
Separate codes for site (T topography
codes) and histology (M morphology codes)
Coding manual
Topography numerical
Morphology numerical
Alphabetic index
 Sources of Cancer Registry Data

Reliance on data standards

Each data item collected has a very specific
set of rules
Edits
Developed by national organizations
Applied at hospital and central levels
 Sources of Cancer Registry Data

Hospital Cancer Registries

Massachusetts Cancer Registry (MCR)
North American Association of Central Cancer
Registries (NAACCR)
Surveillance Epidemiology and End Results
(SEER)
International Agency for Research o Cancer
(IARC)
 Hospital Cancer Registries

Cancer Reporting Law in MA

Variability in Size

Collect a large data set established by national

organizations

Sophisticated edits process

 Hospital Cancer Registries

Demographic Data
Address, age, gender, race, ethnicity, insurance, census
tract, birthplace, occ/ind, tobacco/alcohol hx, family hx
Tumor Data
Primary site, histology, grade, stage, laterality
Treatment Data
Surgery, chemo, XRT, hormone, BRM
Vital Status- Follow up Data
Report to the state/central registry
Within 6 months to 1 year after diagnosis
capture first course of treatment
 Hospital Cancer Registries

Quality of data
Varies
ACoS vs non-ACoS approved programs
Size of hospital
Teaching/research vs. community hospital
Involvement and interest of MDs and administration
Standardized data collection software
Standard edits developed by national organizations
 Hospital Cancer Registries
Data Produced

ACos required studies

Requirement of ACoS approval
Site specific
Quality and outcomes focused
Patient care improvements
Annual report
Administrative reports
Ad-hoc reports/requests for data
 Questions Answered With Hospital
Cancer Registry Data

Who are we serving?

Where are patients coming from?
How were they referred to our facility?
Was access a problem?
How sick are they when theyre diagnosed?
How are they treated?
How well do they fare compared to others?
Do we have adequate numbers to support
various clinical trials?
 American College of Surgeons
National Cancer Database (NCDB)
Joint program of the ACoS Commission on Cancer (CoC) and the
American Cancer Society
Nationwide oncology outcomes database for more than 1,500
Commission-accredited cancer programs in the United States and
Puerto Rico
70 percent of all newly diagnosed cases of cancer in the United
States are captured at the institutional level and reported to the
NCDB
Primarily used by CoC-accredited cancer programs as a means by
which to evaluate and compare the cancer care delivered to patients
diagnosed and/or treated at their facility with that provided at state,
regional, and national cancer facilities
 Massachusetts Cancer Registry
Population based state cancer registry
Data from 1982
Massachusetts Population 2010
Total: 6,547,629
White: 80.4%
Black: 6.6%
American Indian/Alaskan Native: 0.3%
Asian: 5.3%
Native Hawaiian-Other Pacific Islander: 0.0%
Reporting two or more races: 2.6%
Latino or Hispanic Origin: 9.6%
White, not Hispanic: 76.1%
 Massachusetts Cancer Registry
Race/Ethnicity Mass US

White 80.4% 72.4%

Black 6.6 12.6

American Indian and Alaska Native 0.3 0.9

Asian 5.3 4.8

Native Hawaiian and Other Pacific Islander 0.0 0.2

Persons reporting two or more races 2.6 2.9

Hispanic or Latino origin 9.6 16.3

White not Hispanic 76.1 63.7

 Massachusetts Cancer Registry

Data collected same data items as hospital data set

Provides the Abstracting and Coding Manual for
Hospitals
Consolidation of information from all reporting sources
Incidence data each tumor (occurrence of cancer) counted once
Primarily reported from hospitals
Some outpatient/path lab reporting
Data stored at multiple levels
Patient level data
Tumor level data
Treatment level data
 Massachusetts Cancer Registry
Patient A
Name
Address
Age etc.

Tumor 1 Tumor 2
Site Site
Histology Histology
Stage etc. Stage, etc.

Treatment A Treatment B Treatment A Treatment B

CTX Surgery CTX XRT
 Massachusetts Cancer Registry
Consolidation Process
Algorithms to dictate
Match
Non-match
Review
Consolidation rules
Multiple primary rules
Lymphomas and leukemias
Laterality
Benign vs. malignant
Massachusetts Cancer Registry
Consolidation Process

Patient A Patient A
Hospital 1 Hospital 2

Name: Jones Name: Jonnes

Primary Site: LLL Lung Primary Site: LUL Lung

Histology: Small cell Histology: Adenoca
 Massachusetts Cancer Registry
Consolidation Process

Consolidated Patient
Record
Hospital A
Name: Jones Hospital B
Name: Jones
Name: Jonnes
Primary site: LLL Lung
Consolidated Primary site: LUL Lung
Histology: Small Cell
Tumor Record Histology Adenoca
Site: LLL Lung
Histology: Adneoca
 Massachusetts Cancer Registry
Consolidation Process

Very time consuming!

Always looking for ways to further automate
Also involves real-time edit resolution
Males with endometrial ca
Impossible/rare site-histology combinations
Age flags
 Massachusetts Cancer Registry
Data - Reports

Two main reports:

Cancer Incidence and Mortality in Massachusetts
The City/Town Supplement
Special reports
Cervical and Uterine Cancers in MA
Cancer Incidence and Mortality in Boston
Neighborhoods
Childhood Cancers in MA
Cancer in MA by Race and Ethnicity
and more on their website
 Massachusetts Cancer Registry
Data - Reports

MassCHIP - Massachusetts Community Health

Information Profile
Online access to health and social indicators
http://www.mass.gov/dph/masschip
Linkages and data requests
Linkages with data sets provided to them
Approval for all data requests by the MDPH
Human Research Review Committee
Massachusetts Cancer Registry
Data - Reports

MCR website:

www.mass.gov/dph/mcr
 Questions Answered With
Massachusetts Cancer Registry Data

What is the race, gender, age breakdown for

cancers in the state?
What do certain areas look like in terms of
cancer type and demographic variables?
How does our hospital population compare to
the state in terms of stage at diagnosis?
Are there sufficient numbers of patients to justify
a new service or to open a trial?
 North American Association of Central
Cancer Registries

Promotes uniform data standards

Provides education and training; certifies
population-based registries
Gold and Silver status
Aggregates and publishes data from central
cancer registries
Promotes the use of cancer surveillance data
 North American Association of Central
Cancer Registries

All central cancer registries in the United States

and Canada are members
Cancer In North America (CINA)
Published annually
Contains: general demographic information about
each state, data quality indicators, site specific
incidence rates
Individual states
US Combined Report
 North American Association of Central
Cancer Registries - Reports

Cancer Incidence in North America (CINA) Monographs - four published

volumes of cancer statistics for a five-year period
CINA+ Online - an on-line query system utilizing CINA data
CINA Monograph Data in SAS Data Set - published data in CINA
Monograph (all four volumes) available for use in SAS
CINA Plus in SEER*Stat - data from CINA+ Online that are available in
SEER*Stat software, allowing more flexible queries and statistical testing.
Only available to NAACCR members or collaborators with NAACCR
members.
CINA Deluxe - a multi-registry NAACCR members analytical data file for
cancer researchers that includes more detailed data variables and
groupings. Only available to NAACCR members or collaborators with
NAACCR members.
 CDC National Program of Cancer
Registries (NPCR)

Established by Congress through the Cancer

Registries Amendment Act in 1992
Provided funding, supplemental and start up, to
state cancer registries
Contracted with NAACCR to provide guidance in
the areas of data standards and quality control
In collaboration with the NAACCR, produce the
United States Cancer Statistics: Incidence and
Mortality report
NPCR - Program Contacts by Funding
Status
 NCI SEER

Data from 18 population based central cancer registries

Covers approximately 26% of the US population
23.4% of Whites
22.7% of African Americans
40.4% of Hispanics
42.2% of American Indians and Alaska Natives
53.3% of Asians
69.8% Hawaiian/Pacific Islanders
NCI SEER
NCI SEER
 NCI SEER
Goals of the SEER Program
Collect data
Conduct quality control and quality improvement programs
Report on the cancer burden in selected segments of the
population
Identify unusual changes and differences in patterns of
occurrence
Describe temporal changes in the extent of disease at
diagnosis, trends in therapy and changes in survival
Monitor the occurrence of possible cancers that are
caused by cancer therapy
 NCI SEER
Goals of the SEER Program

Collaborate with other organizations on cancer

surveillance activities, data standards and training
Serve as a research resource to the NCI regarding
studies dealing with cancer prevention and control,
program and registry operations
Provide research resources to the general research
community
Provide training materials and web-based training
resources to the cancer registry community.
 NCI SEER
Data Available

SEER data are updated annually

Provided in print and electronic formats
SEER web site (http://seer.cancer.gov)
Reports on cancer statistics
Instructions on how to access the SEER public use file free of
charge on a CD-ROM or access provided through a server at
NCI.
Software provided free of charge by NCI that aids in the
analysis of the SEER database (SEER*Stat) and can be used to
analyze data from other cancer registries (SEER*Prep and
SEER*Stat).
 NCI SEER
Data Available
Three major categories of data:

1. Cancer Statistics
Web-based systems provide access to statistical tables, graphs, and
maps from various data sources.
2. Datasets and Software
Use SEER data to address multiple issues, such as looking at ne
stage at diagnosis by race/ethnicity, calculating survival by stage at
diagnosis, age at diagnosis, and grade or size of tumor, or
determining trends and incidence rates of cancers at various sites
over time.
3. Publications
The statistical publications provided were produced by the SEER
program or through collaborative efforts between SEER and other
organizations.
 NCI SEER
Data Available Cancer Statistics

Cancer Stat Fact Sheets

Printable summaries of the latest cancer statistics for major cancer
types.
Cancer Statistics Review (CSR), 1975-2008
Detailed report published annually to present a broad profile of
cancer statistics to the public.
Fast Stats
Build your own tables and graphs of key SEER and US cancer
statistics.
Cancer Query Systems
Provides more flexibility and a larger set of cancer statistics than
Fast Stats but requires more input from the user.
State Cancer Profiles
Dynamic maps and graphs enabling the investigation of cancer
trends at the county, state, and national levels
 NCI SEER
Data Available Datasets and Software

SEER*Stat version 7.0.5 (09/26/2011)

Statistical software that provides a convenient, intuitive
mechanism for the analysis of SEER and other cancer-related
databases. It is a powerful personal computer tool to produce
statistics for studying the impact of cancer on a population.

SEER*Prep software version 2.4.5 (2/7/2011)

Software that converts user-supplied ASCII text data files to the
SEER*Stat database format, allowing analysis of population-
based registry data using SEER*Stat.

Health Disparities Calculator (HD*Calc) version 1.1.0 (01/13/2010)

An extension of SEER*Stat designed to generate multiple
summary measures to evaluate and monitor health disparities.
 NCI SEER
Data Available Publications
Statistical Reports
SEER Cancer Statistics Review (CSR)
Annual Report to the Nation on the Status of Cancer
Cancer Incidence and Mortality Patterns among Specific Asian and Pacific
Islander Populations in the US
United States Cancer Statistics: Cancer Incidence and Mortality Data
Monographs (recent)
An Update on Cancer in American Indians and Alaska Natives, 1999-2004
Selected Comparisons of Measures of Health Disparities: A Review Using
Databases Relevant to Healthy People 2010 Cancer-Related Objectives
SEER Survival Monograph: Cancer Survival Among Adults: US SEER
Program, 1988-2001, Patient and Tumor Characteristics
 NCI SEER
Linked Databases
SEER-Medicare Linked Database
Links SEER and Medicare data.
SEER-Medicare Health Outcomes Survey Linked Database
Links SEER and the Medicare Health Outcomes Survey (MHOS)
that provides information about the health-related quality of life
(HRQOL) of Medicare Advantage Organization (MAO) enrollees.
National Longitudinal Mortality Study (NLMS) and Linked
SEER-NLMS Databases
SEER registry cancer patient records for 1973-2003 are linked
with the NLMS database, which was developed to examine the
relationships between self-reported demographic and
socioeconomic characteristics and mortality among a
representative sample of the US non-institutionalized population.
 NCI SEER
Linked Databases SEER-Medicare

SEER data Medicare claims data

Clinical, demographic and cause Covered health care services
of death information for from the time of a person's
persons with cancer Medicare eligibility until death

Unique population-based
source of information
Epidemiological and health
services research
 Why link the SEER-Medicare Data?
The linked data can be used for a number of analyses that span the
course of cancer control activities

Diagnosis/ Tx Survivorship Second Occurrence Terminal Care

Patterns of care Late effects of Rates of second Use of hospice
treatment primaries services
Peri-operative
complications
Post-diagnostic Relationship of
Volume outcomes surveillance second events to
initial treatment and Patterns of care
studies
ongoing during the last
surveillance year of life
Extent of staging Treatment of
prevalent cancers
Comorbidities

Survival

Source: Warren, NAACCR,NCI

 SEER-Medicare Data

SEER Data
Incidence, site, stage, initial tx, demographics and
vital status

Medicare claims
Short stay hospitals
Physician and lab services
Hospital outpatient claims
Home health and hospice bills

Source: Warren, NAACCR,NCI

 SEER-Medicare Data

Limitations
Non-covered Medicare services are not
included
Reasons for ordering tests and test results
are not included
Primarily over 65 population
 NCI SEER
SEER RTR Program

Residual Tissue Repository (RTR) Program

Established in 2003, the RTR maintains bio specimens
obtained from three of SEERs population-based cancer
registries: Iowa, Hawaii, and Los Angeles. Investigators
at government, academic, and nonprofit institutions may
apply to the program to obtain specimens to study
biomarkers, etiology, and other aspects with a
population-based sample of cancer cases.
Questions Answered With NAACCR and
SEER data

How is Massachusetts similar or different than

other parts of the country?
How does Massachusetts compare in terms of
age at diagnosis, stage and outcomes?
Where else might there be special populations
or rare cancers of interest?
Where might there be the potential for
collaboration regarding these populations?
 International Agency for Research on
Cancer - IARC

WHO
Data held and managed by the Section on
Cancer Information (CIN) of IARC
Cancer in Five Continents
Published every year since 1966
CIN Website CANCERMondial
http://www-dep.iarc.fr/
 IARC

CIN Databases
GLOBCAN
provides access to the most recent estimates (for 2008) of the
incidence of, and mortality from 27 major cancers worldwide.
CI5-Cancer Incidence in Five Continents
provides access to detailed information on the incidence of
cancer recorded by cancer registries (regional or national)
worldwide.
ACCIS (Automated Childhood Cancer Information System)
provides access to data on cancer incidence and survival of
children collected by European cancer registries.
 IARC

CIN Databases (cont.)

ECO (European Cancer Observatory)
provides access to the estimates (for 2008) of the incidence of, and
mortality from 25 major cancers in the countries of the European Union
(EU-27), together with a quick overview of on-going programmes of
screening for breast, cervical and colorectal cancers implemented in the
27 European countries.
NORDCAN
presents up-to-date long time series of cancer incidence, mortality,
prevalence and survival from 40 cancers recorded by the Nordic
countries, together with advanced prediction facilities.
SurvCan
presents cancer survival data from cancer registries in low and middle
income regions of the world. (Africa, Asia, the Caribbean and Central
America)
 Questions Answered With IARC data

How is the US, MA, Boston similar or

different than other parts of the country?
Where else might there be special
populations or rare cancers of interest?
Where might there be the potential for
collaboration regarding these populations?
 Rapid Case Ascertainment

Provides the ability to identify cancer cases

shortly after diagnosis

Central registry editing and consolidation can be

time consuming

Located within central cancer registries and/or at

academic medical centers
 Rapid Case Ascertainment

Value to researchers: provides immediate access

to patients
Interview patients prior to surgery
Obtain blood/DNA etc.
Initiate a series of questionnaires over time (i.e. 3, 6 and
9 months after diagnosis)
Interview patients about lifestyle elements that might tend
to change after diagnosis (alcohol consumption, smoking
history)
 SEER Rapid Response Surveillance
Studies
Began as Patterns of Care Studies

Evolved into a mechanism that allows studies to move from

the initial concept through completion in a relatively short
interval, usually within two years

Studies address new and emerging issues related to cancer

prevention and control

Conducted through he SEER Registries

Largely pathology based active case-finding (e-path or circuit

riding) vs. waiting for reports to be sent to the central registry
Limitations of Using Cancer Registry Data

Limited data set extensive, but limited

Strict rules for data collection good, but
restrictive
Time delays data can be 1 to 2 years old
Data collected and defined by the priorities of
others
Dataset designed by committee
Limitations of Using Cancer Registry Data

Reliance on primarily pathology based case-finding

image and lab based dxs may not be as readily picked
up (pancreas, brain, multiple myeloma)
Majority of cases reported by hospitals some lab and
outpatient reporting
Data can be limited by information captured only through
the hospital system
Outpatient information is often missing XRT at
freestanding centers, CTX in MD offices
Limitations of Using Cancer Registry Data

Lack of information about events leading up to dx

screenings, health behaviors
Limited information about co-morbid conditions
Follow-up information often limited to vital status
no detailed information on side effects to tx, tx
compliance
Very little information about recurrence of disease
 Medical Record Review

Advantages
Flexibility
Timeliness
Customized data set
Investigator control
 Medical Record Review Recent
Examples

Helping Ourselves Helping Others

PI: Ann Partridge, MD
Breast cancer in women 40 and under
Captured within 16 weeks of dx
Cases identified and reviewed at DFCI, MGH,
BWH, NWH, Faulkner, BIDMC, North Shore
Cancer Center, Lowell General
 Medical Record Review Recent
Examples

The Environment, Genes and Testicular Cancer

PI: Russ Hauser, MD
Newly diagnosed cases of testicular cancer
Identified shortly after dx at DFCI, MGH,
BWH, Faulkner, North Shore, BIDMC, Boston
Medical Center, Lowell General, Tufts-NEMC
 Medical Record Review Process

Plan data set with investigator/study team

Define data elements and data definitions
(coded 1, 2 responses vs. yes, no)
Think about the future
Will this data be linked or compared to other data
sets should race and ethnicity be collected using
cancer registry/census rules
What is being measured cant evaluate data that
hasnt been collected
 Medical Record Review Process

Design/program the data collection tool

Access, Excel, other programs
Consider output where is the data going-
other databases, are the variables the same
 Medical Record Review Process

Define rules for:

Missing data
Incomplete data
Illegible data
Keep in mind
Medical records are not designed for
research, there are limits to what they contain
 Medical Record Review Process

Develop case identification procedures

Review eligibility criteria
Identify facilities
Case sources
Pathology
Radiology
New patient admission/visits
Clinic logs
Access to case sources
Who has access
How does one obtain access
Is training required
 Medical Record Review Process

Consider the frequency of data collection

Is time to dx important
Are there other patient or patient flow events to
consider
How often do patients come in for treatment
Do they stay in the hospital
Is there a planned schedule of testing
Many cases will need to be reviewed to find eligible
cases this can be time consuming
 Medical Record Review Process

Work with investigator/study team on IRB issues

Provide background information on personnel
Provide information about methods for identifying
cases
Build in quality control measures
Computerized edits
Double abstraction (5-10%, inter rater reliability > 95%)
 Abstraction Process
Research Hypothesis

Define
Variables Exploration

Instrument Logic checks

Guidelines
Development Default values

Plans for Case

Staff Pilot testing
Identification

Record review Quality Mgmt

Revisions

Adapted from Keyzer, UC Davis

What Would They Think?
 Progress

Amount of work in Outcomes Research

Focus on End Results Reporting, particularly
Comparative Effectiveness Research
Accountability
Evidence based practices
Research based interventions
Questions?