Mc1

John:
<<Summary of emails related to CDC maternity care survey.pdf>> <<Appendix D Expert Panel Meeting Final Report.doc>> <<April 12 2007 IFC Letter (2).pdf>> <<Email correspondence received by CDC>>

Thank you for giving us the opportunity to discuss our OMB package (0920-05BU -- Assessment & Monitoring of Breastfeeding-Related Maternity Care Practices in Intrapartum Facilities in the U.S. and Territories) with you yesterday.  We are attaching the items we agreed upon.  These include the following:

1. A sample of 10 e-mail comments that we have received from hospital facilities, public health stakeholders, consumers, and other interested parties regarding the proposed survey.  Attached is an annotated table listing these comments along with the original e-mails.  [Unfortunately, because of system limitations on e-mail volume, some earlier e-mails are not retrievable.]

2. The Final Report from an Expert Panel Meeting convened in October 2003, to help us shape a system to monitor breastfeeding-related maternity care.  State Health department representatives comprised one-third of this panel.

3. The April 12, 2007, follow-up letter from the International Formula Council.

As we discussed yesterday, there are a number of compelling reasons for using a census rather than a sampling methodology:

• State health departments have voiced a strong desire to be able to identify hospitals whose breastfeeding support practices are particularly problematic.  We had originally considered keeping the hospital identifiers confidential, but state health departments commenting on the survey insisted that they need information on individual hospitals, not just state summary statistics.  If states took this on through independent surveys, it would be difficult to compare results across states.  Additionally, not all states would have the capacity (both staff and funding) to undertake such a survey.

• There is a large diversity among maternity care facilities in terms of size and type (urban/rural; profit/non-profit; teaching/non-teaching; private/public; serving economically disadvantaged populations/serving high SES).  There is a need to stratify results by these variables as breastfeeding practices likely differ by these characteristics.  Sample surveys in each state would be so small as to make this kind of analysis impossible. 

• According to the census plan outlined in our current methodology, each facility will benefit from receiving a benchmark report which it can use for rapid and localized assessment of its own internal issues.  Those facilities not selected for the sample would be disadvantaged as they would not receive a benchmark report as it is hoped that the benchmark report will be a catalyst for change within a facility.  Thus, the opportunity for feedback, acknowledgement, and understanding of practices would be lost for non-surveyed facilities.

• If a sampling methodology is employed, some facilities not selected for the survey would feel slighted or, even worse, not consider the survey results represent their facility but only “other” facilities because of the competitive nature of these types of facilities.

We heard your concerns about administration of a second fielding of this survey without additional OMB review.  Our long range intention (described more fully on p. 8 of the supporting statement) is to collect data every two years in order to monitor practices that support achievement of Healthy People 2010 Breastfeeding Objectives. The project will provide meaningful information to the facilities themselves and to CDC and States and will demonstrate our responsiveness to the voiced needs of states and health care facilities.  A specific concern of the Expert Panel (see p. 4 of the attached Expert Panel Report and p. 6 for list of attendees and their affiliations) was the risk to data quality associated with long intra-administration intervals.  Although panel members initially advocated for annual administration, they ultimately agreed that biannual administration would be acceptable, but that longer intervals between administrations would be problematic. 

Our plan is that the data from the 2007 survey will establish a baseline measure and the data from the second survey in 2009 will identify changes in practices over time. The survey instruments are designed to capture incremental changes that we anticipate will be taking place at the hospital level.  We plan to evaluate the two surveys before moving forward towards the establishment of such a system, but we think it would be inappropriate to re-evaluate a survey intended to examine trend data after the first iteration. 

Thank you again for your consideration of this OMB package.  We think this survey will provide critical information needed to more thoroughly understand maternity care practices taking place in hospitals and will help us and others design interventions that will improve these practices. 

Paulette E. Murphy, MLIS
Deputy Chief, Nutrition Branch
DNPA
NCCDPHP, CCHP, CDC