CDC Memo

Lucy A. Peipins, Ph.D.

July 12, 2007

Risk Perception, Worry, and Use of Ovarian Cancer Screening Among Women at High, Elevated, and Average Risk of Ovarian Cancer

Response to OMB questions

We appreciate the opportunity to respond to questions raised by OMB staff. We have carefully considered the helpful comments provided and feel that we have addressed and incorporated all of the comments as follows:

1) Can we map the research questions outlined on page 3 and 4 in the supporting documentation with the questionnaire? Question numbers and page numbers are from Appendix E. The last research question (4) includes variables from the follow-up survey, Appendix F.

We will address the following research questions:

1. What are the predictors of a perception of being at high risk for ovarian cancer? Outcome variables are perceived lifetime risk of ovarian cancer and 10-year perceived risk of ovarian cancer, questions 19 and 20, page E-6.

We hypothesize that a strong family history of breast and ovarian cancer (question 28, page E-8 through E-10) having experienced cancer in one’s family or among friends (questions 38-47, page E-15), cancer knowledge (questions 71-73, pages E-25 through E-26), and an information-seeking style of coping with potentially threatening information (questions 5-6, pages E-4 and E-5) are positively associated with a perception of high risk of ovarian cancer.

2. What characteristics influence a woman’s likelihood of being screened for ovarian cancer through CA125 or transvaginal ultrasound? What are the strongest predictors of screening behavior? Outcome variables are CA125 testing (question 56, pages E-19 through E-21) and transvaginal ultrasound (question 58, pages E-21 through E-22).

We hypothesize that a positive family history of cancer (question 28, page E-8 through E-10), a perception of higher risk of ovarian cancer (questions 19 and 20, page E-6), a higher level of cancer worry (questions 25-27, page E-7), prior screening for other cancers (questions 49-54, pages E-17 through E-19), and an information-seeking style of coping with potentially threatening information (questions 5-6, pages E-4 and E-5) are associated with a higher likelihood of screening.

3. Is a perception of being at high risk for ovarian cancer also associated with increased cancer worry (questions 25-27, page E-7) and anxiety (questions 7-18, page E-5)? Does knowledge about ovarian cancer (questions 71-72, pages E-25 through E-26) affect that relationship?

We hypothesize that a high level of anxiety (questions 7-18, page E-5) and a higher perceived risk will be positively associated with a higher level of cancer worry (questions 25-27, page E-7).

4. What is the relationship between intent to undergo screening (question and actual screening behavior? Is this relationship moderated by change in family history or change in risk perception or other characteristics?

We hypothesize that the proportion of women reporting intent to undergo screening (question 57 for CA125, page E-21) and question 59 for transvaginal ultrasound, page E-23) will be smaller than those who actually undergo screening (question 8b, page F-4 Follow-up Questionnaire) and that this will be influenced by changes in perceived risk (questions 2 and 3, pages F-2 in Follow-up Questionnaire) and family history of cancer (question 12, page F-7 and F-8, Follow-up Questionnaire).

2) Will the results of this study be generalizable to the US population?

We are conducting this survey among HMO members in the Detroit metropolitan area. These women have access to health care and the ethnic/racial composition of enrollees in this HMO is 30% African American, 64% White and approximately 5% “other” racial designations. This ethnic/racial composition is not representative of, nor generalizable to, the US general population. In addition, insured women are not representative of all women in the US population. However, for the purposes of our study, it was important to recruit a population of women without major barriers to health care (such as screening) that you would find in a managed care organization.

3) Research shows an improved response rate with incentives. Response rates are increased when incentives are given prior to the administration of a questionnaire. Also receiving cash has been show to be preferable to gift cards.

We concur with this helpful recommendation. Although it is not feasible to provide cash incentives prior to administering the baseline questionnaire due to its linkage with the eligibility screener, a $10 cash incentive will be provided to the 1,600 baseline participants prior to participating in the follow-up survey. This cash incentive will be sent along with the reminder letter (Attachment I, Reminder Letter for Follow-up Survey).

4) What procedures or methods will we follow to improve response rates for the follow-up survey? Will we be taking any steps to assure that the contact information is correct prior to calling?

There are several methods we will use to ensure a good response rate for the follow-up survey. The following plan is intended to mitigate attrition and maintain the expected sample size of 1,600 at follow up. First, at the end of the baseline survey in the closing statements (pages E-26 through E-27), we ask for contacts who would know where the participant might have relocated should she relocate. The text is as follows:

In addition, since people often move within the span of a year, we would like to get the name of one or two other people who would likely know where we could reach you. Many people give their mother’s or sister’s name. Others give the name of a very close friend. Is there someone who would be a good person for us to contact in case we cannot reach you at this phone number?

If yes, what is their name? Would you spell that for me?
NAME:
And what is their relationship to you?
RELATIONSHIP:
Where would we contact him/her?
ADDRESS:				CITY:
STATE:				ZIPCODE:
Is there a phone number?			PHONE NUMBER:

A second approach to maximizing response rate is to ensure accuracy of participant contact information. We have an opportunity to ensure accuracy at several points between administration of baseline and follow-up surveys. A thank you letter for baseline survey participation will be sent (Attachment H). For any letters that are returned, we will follow-up via telephone with respondents to obtain more accurate address information, update the database, and res-send the letter. We will also match these addresses to the national change of address (NCOA) database and update any changes accordingly. At that point we will also investigate changed numbers and record the new changed number. We will also verify with the call center that interviewers are trained to disposition telephone numbers as ‘number changed’ if they encounter any new numbers.

We expect some proportion of the phone numbers to be no longer working when we call back for the follow-up survey. As a precaution, we will manually re-dial these numbers to make sure that the disposition of nonworking telephone numbers as assigned by the CATI interviewer, was in fact correct. If the nonworking rate is higher than anticipated, we will see whether it is concentrated within particular exchanges and find out whether there have been any changes to the area codes in that area since the baseline interview.

An additional step in assuring accurate information will be conducted through the Henry Ford Health System. Prior to sending out the reminder letter and $10 cash incentive, the Henry Ford Health System will check the addresses of the participants at the last contact against their HFHS Master Patient Index. Patient address information will also be checked through the HFHS Care Plus (the electronic medical records).

Finally, because contact between surveys is an effective way to garner cooperation in the follow-up, we will send a reminder letter and a cash incentive prior to administration of the follow-up survey. Timing of the reminder letter is important in maximizing response rate. A shorter time between reminder and telephone contact increases the impact of the incentive while a longer lag time increases the likelihood that participants may forget the source and purpose of the reminder and incentive. We will send the reminder letter 3 to 5 days prior to telephone contact and include the $10 cash incentive with this letter.

4