VA memo

See response below.

-OMB asked us about the Edwards et al. citation on p. 4 of our application.  I’ve attached a PDF of the reference to this e-mail. To further clarify the reference regarding up-front incentives increasing the likelihood of response by 61%, an OR (odds ratio) of 1.61 was presented by Edwards et al. (page 2 of the attached PDF) as an estimate of the increase in the odds of survey response when an unconditional/up-front incentive is provided.  We calculated the estimate of a 61% increase in survey response resulting from these incentives (as included in our application) as a relative effect, estimated as follows: Relative Effect = (OR - 1)x100.  Using the formula above, this is equivalent to an increase in odds of response of 61% [(1.61 - 1)x100].  If interested, further information about this calculation can be found in Essentials of Epidemiology in Public Health, Aschengrau and Seage (2003).

<<Edwards et al. (2007)_Methods to increase response rates.pdf>>

-OMB asked us to make some minor edits to survey question 53, assessing participant ethnicity.  The revised survey document with the requested revisions to this question (i.e., replacing “please check all that apply” with “please select one or more” in the question instructions; replacing “Alaskan” with “Alaska” in response choice 2; removing response choice 6 “Other”) is attached to this e-mail.

<<Survey Form 10-21089_REVISED OMB Version.pdf>>

-OMB asked us the status of our Certificate of Confidentiality application.  The final application has been under review at NIH since February 23, 2009.  Our contact at NIH, Brent Loomis, indicated that we should expect approval of the application within 4 – 6 weeks so we believe that final approval is imminent.

-The OMB reps asked that we consider additional variables that may be available to assess response bias if our survey results in a lower than expected response rate.  We will have access to a number of demographic variables (e.g., age, race, dates of service) from the database from which will be drawing the participant pool.  OMB asked us to consider the possibility of accessing relevant clinical status variables (e.g., diagnostic status, use of mental health services).  While it would be theoretically possible to access these data through linkages with the VA’s medical records system, there would be some substantial logistical issues and privacy concerns to be negotiated.  We will make a note of the suggestion, however, and undertake more significant exploration of the idea if our survey does in fact result in a lower than expected response rate.