Skip to Main Content

An official website of the United States government

Contamination Survey

The Contamination Survey Health Status Questionnaire (HSQ) is designed to assess the rate of prostate, lung, colorectal and ovarian screening that occurred outside the trial protocol. Data from the HSQ provide an estimate of the contamination rate in the PLCO trial.

The survey contains questions about various types of exams, including the PLCO screening exam types. For each exam, the participant was asked three questions: whether they received the exam, how recently they received the exam, and the reason for the exam. For the exam types used in the PLCO screening protocol, intervention arm participants were explicitly asked to report an exam only if it occurred after their last PLCO screening exam of that type.

Initially, the HSQ asked each participant about eight gender-neutral exams; in addition, men were asked about two gender-specific exams and women were asked about four gender-specific exams. In 2004, the Spiral CT exam was added to the HSQ. In 2005, the blood in stool exam was added. As a result, starting in 2005, men were asked about 12 exams and women were asked about 14 exams.

The survey was administered to a randomly selected sample of participants in the years 1997, 1999 - 2005, 2007, 2009 and 2010. A new random sample was selected for each of the eleven survey years. Samples were stratified by gender, screening center, and study year. Additional selection criteria are addressed below.

The sampling criteria changed over time:

  • 1997 - 2003: During this time period, approximately 1,000 participants were selected each year. Only "uncontaminated controls" were eligible for sampling. "Uncontaminated controls" are defined as control arm participants who had not reported undergoing 2+ PSA exams or 1+ colorectal exams (flexible sigmoidoscopy / colonoscopy / barium enema) in the three years prior to trial entry. Information on screening prior to trial entry ("baseline contamination") was obtained from the baseline questionnaire (BQ).
  • 2004 - 2010: During this time period, approximately 2,000 participants were selected each year. Participants were distributed among the following groups:
    • ~1000 "uncontaminated controls" defined above.
    • ~200 "contaminated controls" defined as control arm participants who reported undergoing 2+ PSA exams or 1+ colorectal exams in the three years prior to trial entry.
    • ~800 intervention arm participants who were in study year T7 or later. These participants were beyond the screening period of the trial.

In total, approximately 13,000 participants were surveyed. Some participants happened to be selected in two or three separate years. The HSQ data contain information on only a subset of PLCO participants, some of whom have more than one record in the dataset, and therefore care should be taken when merging with other PLCO data.

HSQ data can be obtained through the HSQ dataset.

Summary
Available Data
  1. Health Status Questionnaire (HSQ) data
    • Main contamination variables
      • The HSQ contains questions about eight exams that were used to assess contamination, two for each of the PLCO cancer sites. For each of the exams, participants are asked whether they received an exam, when was their most recent exam, and what was the reason for the exam. (For the intervention arm, trial screens were not to be reported.) The answers to these questions are provided in summary variables ("[exam]_summary") for the following eight exams:
        • Prostate: PSA, DRE
        • Lung: X-ray, Spiral CT
        • Colorectal: Colonoscopy, FSG
        • Ovary: CA-125, TVU
      • In addition, there are summary variables that address whether either of two exams were performed. There are summary variables for the following screening combinations:
        • PSA and DRE
        • FSG and colonoscopy
        • CA-125 and TVU
    • Additional exam variables
      • A total of sixteen exams were included on the HSQ questionnaire forms, and for each of these exams there is a variable describing how recently an exam was performed ("[exam]_years_ago") and a variable on the reason for the exam ("[exam]_reason"). These variables exist for the eight exams listed above, as well as the following exams:
        • Non-PLCO exams for PLCO cancer
          • Barium enema
          • Blood in stool
          • Pelvic examination (females only)
        • Exams for other conditions
          • Eye examination for glaucoma / cataracts
          • Blood pressure check
          • Blood cholesterol level
          • Blood glucose level
          • Mammogram (females only)
  2. Baseline Questionnaire (BQ) data
    • The HSQ dataset contains baseline questionnaire data for participants who received an HSQ.
    • The baseline questionnaire (BQ) was administered to participants in both arms of the trial at baseline. Participants were asked whether they received PLCO screening exams in the three years prior to randomization and, if so, how many times. The HSQ dataset contains baseline questionnaire variables ("[exam]_bq") on the following screening exams:
      • PSA (males only)
      • DRE (males only)
      • Chest X-ray
      • Colorectal exams (FSG, Colonoscopy, or Barium Enema)
      • CA-125 (females only)
      • TVU (females only
    • Control arm participants are considered "contaminated controls" if they reported undergoing 2+ PSA exams or 1+ colorectal exams (flexible sigmoidoscopy / colonoscopy / barium enema) in the three years prior to trial entry. The HSQ dataset includes a target group variable that identifies control arm participants as "contaminated controls" or "uncontaminated controls".