Rosemary D' Cress, DrPH

1
Improving Occupation Information in Central
Cancer Registries for Use in Occupational Cancer
Surveillance
Rosemary D. Cress, DrPH Research Program Director
NAACCR June 13-19, 2009, San Diego, CA
2
Introduction

• Occupational exposures are an important cancer
  risk factor, and may contribute to 4-20 of
  cancer deaths.
• CDC/NPCR requires its registries to collect
  industry and occupation (I/O) on patients
• Thus population based cancer registries provide a
  rich data source for studies of cancer and
  occupation

3
Introduction

• However, I/O information is collected in
  narrative form and has not been consistently
  collected or coded
• Although registrars are asked to record usual
  occupation, nearly half of cancer patients have
  an occupation listed as unknown or retired

4
Introduction

• Despite these limitations, a review of studies by
  Dennis Deapen in LA demonstrated that registry
  based studies of occupation and cancer obtained
  results consistent with more expensive studies,
  even with incomplete data.
• Such studies are useful for hypothesis
  generation.

5
Introduction

• Both the Los Angeles Cancer Surveillance Program
  and the California Cancer Registry have a long
  interest in collection of I/O information
• Both registries employed I/O coders in the past,
  and both developed automatic coding programs that
  were used until 2002

6
Introduction

• However, coding of I/O is highly labor intensive
  and expensive
• With budget cuts and competing priorities, I/O
  coding of statewide CCR data was abandoned
  several years ago

7
Introduction

• The CDC National Institute for Occupational
  Safety and Health (NIOSH) also has a keen
  interest in identifying occupations that increase
  cancer risk
• In 2007, NIOSH funded the CCR to conduct a pilot
  study to improve coding of registry I/O data

8
NIOSH/CCR Pilot Study

• The purpose of this study was to assess the
  feasibility of coding I/O information in the
  cancer registry, and of using this information as
  a surveillance tool
• The focus of the pilot was on coding construction
  occupations and assessing cancer incidence among
  construction workers

9
NIOSH/CCR Pilot Study

• During the first year, NIOSH staff translated the
  older CCR autocoding program to SAS. The program
  utilizes a look-up table with I/O codes and
  text
• CCR staff utilized the program to evaluate
  completeness of I/O coding in the CCR database
  and to identify uncoded text strings

10
NIOSH/CCR Pilot Study

• Uncoded I/O text strings were prioritized by
  frequency
• USC and NIOSH staff coded text strings to the
  Bureau of the Census 1990 I/O codes
• These new codes and text strings were added to
  the SAS program, and the process was repeated

11
Preliminary Results

• Of 2,258,857 records in the statewide CCR
  database, 1,114,620 (49.3) had IO coded at the
  beginning of the study.
• As of May 2009, 76.4 of records were coded

12
Industry Information on CCR cases,
1998-2008(n2,843,003)
13
Preliminary Results

• Over 50,000 cancer patients with a construction
  occupation have been identified.
• Over 387,000 patients are coded as retired, and
  over 743,000 as unemployed or unknown making
  their records unusable for I/O research
• Over 600,000 remain uncoded

14
Next steps

• NIOSH, CCR and USC will continue to work together
  to code as many I/O text strings as possible
• Standardized rate ratios will be calculated for
  major cancer sites for workers in the
  construction industry as well as those in other
  industries.

15
Conclusions

• This study is demonstrating that it is feasible
  to code I/O text fields and use this information
  for research into occupational risk factors for
  cancer
• However, inconsistent and incomplete information
  recorded for I/O in the medical record will limit
  this activity

16
Conclusions

• Additional education of abstractors on what to
  look for and how to code usual occupation and
  industry will improve our ability to collect
  these data for research.

17
Acknowledgements

• CDC/NIOSH
• Geoffrey Calvert
• Sara Luckhaupt
• Pamela Schumacher
• Rui Shen
• California Cancer Registry, Sacramento
• Katrina Bauer
• Mark Allen
• Los Angeles Cancer Surveillance Program, USC
• Dennis Deapen
• Shirley Miyashiro