VA Response to OMB Questions

Department of
Veterans Affairs
Date:

Memorandum

June 28, 2012

From: Aaron Schneiderman, Ph.D., Principal Investigator, and Acting Director, Epidemiology Program, Office of
Public Health, VA
Subj: “National Health Study for a New Generation of U.S. Veterans” (OMB No. 2900-0722) – Response to Office
of Management and Budget (OMB) Questions
To:

Bridget C.E. Dooling (OMB); Cynthia Harvey-Prior, Management Analyst (OMB Liaison), VH A; Denise
McLamb, Program Analyst, VA

1. How do the non-response bias results inform VA’s weighting
adjustments for the survey results that have already been collected?
VA has contracted with Mathematica Policy Research (MPR) for advice on
statistical approaches to address potential non-response bias. MPR will review
VA’s development of design weights that address non-response and provide
guidance on the development of a post-stratification plan. The post-stratification
plan is designed to account for differential response rates by stratification
variables and other important administrative variables, including age and
education.
2. What will VA do to bolster response rates of the next wave,
particularly for subgroups that have relatively low response rates?
VA will employ a multifaceted approach to enhance the response to future survey
efforts. The Office of Public Health (OPH) will conduct the following activities:
proactive outreach to Veteran Service Organizations to enhance the profile of
OPH projects in the Veteran community and development of a voluntary
Veterans Community Board with membership comprised of Veterans of different
conflict eras to advise OPH on outreach and communications. OPH will dialogue
with other VA researchers about the challenge of decreasing participation in VA
research and seek a way forward to support increased participation, especially
among younger Veterans. Any subsequent contracts for data collection will
include specific task language requiring contractors to provide plans for
maximizing response rates and investigation of new or alternative methods to
assure sufficient response rates. Contractors will be required to conduct a robust
literature review on the topic of non-response and employ the use of qualitative
research methods, such as focus groups, to develop plans to maximize response.
VA will also seek methods to better utilize incentives, such as cash or cash
equivalents, and other design elements to enhance the survey experience. The
VA will use sequential mailings, including an introductory letter from an
influential official, reminder postcards, and telephone follow-ups to encourage
response. A web-based survey will continue to be offered, which was a very
popular choice and enhanced response rates for the HSNGV.

Automated VA FORM 2105

3. Would VA be amenable to a term of clearance on the next approval
that directs VA, by the time of the next submission, to have a plan to
leverage its administrative data (i.e., EHRs) to perform more robust
non-response bias analysis?
VA will implement a more robust non-response bias analysis regardless of
whether there is a term of clearance associated with the next approval. We have
already started to include additional data elements in the non-response analysis
for the last survey iteration, including age and education, that are derived from
the sampling frame provided by DoD.
VA finds the suggestion to use the Electronic Health Record (EHR) intriguing;
however, this approach may not support a robust non-response analysis. The
potential pitfall is that our survey is not limited to users of VHA services, but
targets a population-based sample of all Veterans (of Operations Endurig
Freedom and Iraqi Freedom). Consequently, the survey responders, if
representative of the population, will consist of both VHA users and those who do
not receive care from VHA (and therefore no visibility in the EHR). A further
concern regarding the use of the EHR in non-response analysis is previous
analyses that suggest users of VHA services are different from the Veteran
population at large. There may be ways to use the EHR for non-response analysis
that are yet to be described. OPH will seek input from others and would welcome
OMB input toward describing a suitable method to leverage administrative data.
VA will adhere to the terms of clearance directed by OMB, but would prefer that
use of the EHR in non-response analysis be framed as a recommendation,
allowing VA the flexibility to further assess the feasibility of this approach.

Thursday, June 28, 2012
Office of Public Health, Post Deployment Group, Epidemiology Program