Attachment 6 MIHOPE-K FRN comment and response 01may18_unchanged

Mother and Infant Home Visiting Program Evaluation

COMMENT AND RESPONSE ON 60-DAY FEDERAL REGISTER NOTICE

April 2018

During the notice and comment period, we received one request for information and one comment. The comment and ACF’s response are below.

__________________

March 29, 2018

TO: Administration of Children and Families attn. Laura Nerenberg

FROM: Nurse-Family Partnership

RE: Response to FRN 2018-01683 proposed-information-collection-activity-comment-request

(https://www.federalregister.gov/documents/2018/01/30/2018-01683/proposed-information-collection-activity-comment-request

Thank you for this opportunity to review the plan for MIHOPE-LT. There is potentially valuable information about the participants to be discovered with this evaluation.

We strongly encourage the following considerations related to the planning of this evaluation:

• Adequate resources be deployed to engage all participants who were randomized.  Selective attrition is a huge problem with studies such as this.

• Focus scarce resources on outcomes of unquestionable public health importance measured directly.  Self-report is a challenge. An example of some administrative data would include:

  • Direct assessments of children’s records of hospitalizations for injury, including lengths of stay. 

  • A record infant and childhood deaths, with a focus on those that are preventable (note that this should specify deaths due to suffocation while sleeping as opposed to SIDS – changes in this category over time due to better autopsies today to sort out what is truly unexplained).  Other preventable categories include deaths due to injury.  These are indicators of child abuse and neglect.  This can and should be done through NDI

  • Assess maternal deaths, especially deaths due to more preventable causes, including overdose, suicide, injuries, and homicide.

  • Record whether each child is receiving SSI benefits due to their own disability (as opposed to parent death or disability).

  • Use of administrative data from Social Security Administration to assess earnings.

• We wonder whether and how “trauma” can be measured reliably and validly.  We recommend the focus of inquiry instead on directly measured outcomes.

• We would encourage that the evaluation includes the impact of maternal psychological resources on the long -term outcomes. This would require adequate baseline data on maternal psychological resources (IQ, mastery, and mental health) of the participants was collected as part of the original MIHOPE. Psychological resources will moderate outcomes in powerful ways, with childrearing and child outcomes being most pronounced among children born to mothers with the fewest psychological resources and pregnancy planning and employment outcomes most pronounced among those with initially higher psychological resources. 

• The times listed in the FRN seem less than will be needed across all areas of the assessments.

We are interested in reviewing the complete assessment package that will be used in this evaluation. Can you please provide details about when this information will be available for review?

Please contact Molly O'Fallon at [email protected] for any questions.

Sincerely,

President and CEO

Nurse-Family Partnership

Cc: Nancy Margie, Ph.D. Home Visiting Research Team Leader, ACF

David Olds, Ph.D. Prevention Research Center, University of Colorado

330 C Street, S.W., Washington, DC 20201 | www.acf.hhs.gov

DATE: April 30, 2018

TO: Frank Diadone, President and CEO, Nurse-Family Partnership

CC: Molly O’Fallon, Nurse-Family Partnership

David Olds, Prevention Research Center, University of Colorado

FROM: Nancy Geyelin Margie and Laura Nerenberg, Administration for Children and Families, Office of Planning, Research and Evaluation

RE: Responses to Federal Register Notice Comments on the Mother and Infant Home Visiting Program Evaluation Long-term Follow-up (MIHOPE-LT) Kindergarten Data Collection (FR Doc. 2018-01683)

Thank you for your thoughtful comments on the Administration for Children and Families’ (ACF) Federal Register Notice dated January 30, 2018 [FR Doc. 2018-01683]. Our responses to your comments are below.

We strongly encourage the following considerations related to the planning of this evaluation:

• Adequate resources be deployed to engage all participants who were randomized.  Selective attrition is a huge problem with studies such as this.

We agree that attrition in longitudinal studies such as MIHOPE-LT is a challenge. To minimize overall and differential attrition and maximize participation in future MIHOPE data collection activities, we are proposing to use a combination of locating methods that we found successful in earlier rounds of MIHOPE and incentives.

• Focus scarce resources on outcomes of unquestionable public health importance measured directly.  Self-report is a challenge. An example of some administrative data would include:

  • Direct assessments of children’s records of hospitalizations for injury, including lengths of stay. 

  • A record infant and childhood deaths, with a focus on those that are preventable (note that this should specify deaths due to suffocation while sleeping as opposed to SIDS – changes in this category over time due to better autopsies today to sort out what is truly unexplained).  Other preventable categories include deaths due to injury.  These are indicators of child abuse and neglect.  This can and should be done through NDI.

  • Assess maternal deaths, especially deaths due to more preventable causes, including overdose, suicide, injuries, and homicide.

  • Record whether each child is receiving SSI benefits due to their own disability (as opposed to parent death or disability).

  • Use of administrative data from Social Security Administration to assess earnings.

We plan to continue to collect and analyze administrative data, including child welfare, Medicaid, and the National Dataset of New Hires. These datasets will allow us to measure outcomes related to child abuse and neglect, hospitalizations (including reason for and length of stay), and earnings and employment, respectively. We also plan to pursue the acquisition of school records and National Death Index records in MIHOPE-LT. In addition, in response to this comment we added questions to the structured interview with caregivers to gather data about children’s receipt of SSI, and will be exploring the possibility of acquiring and analyzing SSA data.

• We wonder whether and how “trauma” can be measured reliably and validly.  We recommend the focus of inquiry instead on directly measured outcomes.

While we agree that directly measured trauma-related outcomes are important to capture, there are also trauma-related outcomes with implications for the future health and development of children and their families that can be reliably, validly, and most efficiently assessed through self-report. Outcomes such as intimate partner violence, maternal substance use, maternal mental health, and parental separation or divorce are measures of adverse childhood experiences (ACEs), and research that has shown that risk for a host of poor health and well-being outcomes increases as the number of ACEs increases.¹ Whenever available, we intend to use well-established measures of these outcomes, which were also used in earlier rounds of MIHOPE. For example, intimate partner violence will continue to be measured using the Conflict Tactics Scale and Women’s Experience with Battering Scale. In addition, we intend to ask the caregiver about their adverse childhood experiences, using questions from the Child Trends ACE Module (https://www.childtrends.org/wp-content/uploads/2014/07/Brief-adverse-childhood-experiences_FINAL.pdf).

• We would encourage that the evaluation includes the impact of maternal psychological resources on the long-term outcomes. This would require adequate baseline data on maternal psychological resources (IQ, mastery, and mental health) of the participants was collected as part of the original MIHOPE. Psychological resources will moderate outcomes in powerful ways, with childrearing and child outcomes being most pronounced among children born to mothers with the fewest psychological resources and pregnancy planning and employment outcomes most pronounced among those with initially higher psychological resources.

We agree that caregivers’ mental health and psychological resources are important potential mediators and outcomes. These measures were included in the baseline data collection (thanks to NFP comments on the previous OMB package for the MIHOPE baseline data collection), and we intend to examine their relation to long-term outcomes. In addition, we intend to continue to measure caregivers’ psychological resources, such as mastery, parental stress, depression, and short-term memory.

• The times listed in the FRN seem less than will be needed across all areas of the assessments.

The times estimated for each assessment are based on our previous experience administering the same or similar measures. In addition, we will pilot test the specific measure to make sure they can be completed in the amount of time specified in the OMB package. Also, for the structured interview with caregivers, we plan to use a technique called “planned missingness” to ensure that each respondent receives only 60 minutes of interview items. Specifically, for a construct with many items (such as social-emotional skills), each respondent will only be asked a subset of all of the items in the construct. In this way, groups of respondents will be assigned to answer only a portion of the items in this draft so that an individual’s total response time is 60 minutes or less.

• We are interested in reviewing the complete assessment package that will be used in this evaluation. Can you please provide details about when this information will be available for review?

The proposed measures are attached.

1 https://www.cdc.gov/violenceprevention/acestudy/about.html