1. Portland State University
Testing the Effectiveness of Healthy Families Oregon: Summary of 1 & 2 Year Outcomes Best Beginnings Meeting 1/14/2016
Beth L. Green, Ph.D.
Portland State University
Jerod Tarte, M.A.
NPC Research, Inc.

2. Research Project Goals & Objectives
Conduct a large-scale randomized study of Oregon’s Healthy Families Oregon (HFO) program examining effects on substantiated maltreatment rates and other key outcomes in administrative data
Conduct a detailed cost-benefit analysis of HFO examining program investments & outcome costs
Develop & disseminate a web-based cost analysis tool to develop readiness & understanding of cost analysis in home visiting programs
11/4/2015

3. RCT: Target Population & Sample
Study implemented in 7 of Oregon’s HFO programs, selected for high implementation + many unserved eligible families
Random assignment happened at initial eligibility screening (Feb 2010-Feb 2012).
Eligibility = first time parent, 2+ Risk Factors on “New Baby Questionnaire” (or substance abuse, depression)
Most screenings done at birth
Final full sample n=2667 (1438 program, 1289 control)
11/4/2015

4. Methodology – Two Outcome Studies
Administrative Data: 2-year Intent To Treat design (n=2667):
Child welfare data, Medicaid billing records data, substance abuse treatment data, enrollment in self-sufficiency programs, criminal justice (arrests)
Interview Sample: 1 year follow up telephone interview with n=803 (403 HFO, 400 control):
Home visiting & parenting services received
Short term child and parent outcomes
11/4/2015

5. 12 Month Interview Outcomes
1. + Early School Readiness
+ Daily reading to children in HFO
+ More frequent developmentally supportive parent-child activities in HFO
2. + Healthy Child Development
HFO families were more likely to report that their child received a developmental screening in the past year
11/4/2015

6. 12 Month Outcomes + Parenting and Risk Reduction
HFO families had less total stress (measured on the short form Parenting Stress Index), and in particular, lower parenting-related stress
Some results were more positive for families more risk factors
11/4/2015

7. More Positive Results for Higher Risk?
11/4/2015

8. Administrative Data Approaches
Intent to Treat Group –everyone randomly assigned
BUT: 44% of randomized never got a first home visit, many never contacted/located after screening and initial interest..
Created 2 additional comparison groups to compare families with got a visit vs. controls/non visited families
Lose “gold standard” design; but statistically controlled for baseline differences between groups.
Visited families were more likely to be Hispanic (33% vs. 22%)
More likely to be depressed (23% vs. 13%)
More likely to report trouble paying for basic needs (82% vs. 78%)
11/4/2015

9. Effects on Child Welfare Reports Full Randomized Sample
367 (out of 2,667) children had a founded or unfounded report
Not Significantly Different
11/4/2015

10. In full sample, 101 children were placed out of home – 3.7% overall
Effects on Visited vs Non Visited Families – Placement Outcomes (N=40 children)
In full sample, 101 children were placed out of home – 3.7% overall
11/4/2015

11. Were There Differences for Any Subgroups of Families?
Hispanic vs. Non Hispanic
Specific Risk Factors present/not present

12. Subgroup Differences on Founded Reports
11/4/2015

13. Effects on Other Service Utilization Full Randomized Sample
More likely to be enrolled TANF 1st time (45.2% vs. 39.1%)
More days SNAP/food stamps (488 vs. 464 days)
Self Sufficiency
No differences in OHP enrollment or utilization, immunizations, well baby, ER
HFO more likely to receive AOD TX (4.9% vs. 3.8%)
Health & Well Being
11/4/2015

14. Service Outcomes for Visited Families
Controlling for baseline differences, visited families (compared to non-visited) had:
More days TANF (187 vs. 167)
More days employment services (111 vs. 89.2)
More days OHP coverage for mothers and babies
More medical claims for mothers and babies
More well baby visits (6.4 vs. 6.1)
Fewer arrests (2.4% vs. 4.3%)

15. Implementation Analysis
Low rate of HFO enrollment in randomized group
44% (636 of 1489) families got 1 visit
Fidelity concerns (examples):
42% of families received L1 for 6 months or more
Average duration was 15 months (range 0-24)
12% received < 90 days of home visiting
60% received 75% of expected visits
11/4/2015

16. Duration in HFO Makes a Difference
Families who remained in services longer were:
Less likely to be teen mothers
More likely to have been screened prenatally
More likely to have relationship problems at screening
Families who remained in services longer had:
Fewer days of TANF
Fewer days employment assistance
More SNAP
More OHP coverage
More Immunizations
More well baby visits
11/4/2015

17. Take Aways & Next Steps
Overall, findings consistent with other large-scale randomized studies of HFA – NY, MA
Early modest impacts on parenting, service utilization
Strong evidence for early surveillance and reporting by HFO visitors
89% of founded reports were when they were not receiving home visits
More founded reports on families with specific risks
81% of unfounded reports during enrollment
Need for ongoing follow up to track outcomes
NFP, EHS, + HFA studies show CW outcomes later, when children are 4, 5, + years old
Cost savings are seen 7-9+ years later

18. Identified Implementation Issues
“Drop off” between screening, initial acceptance, and enrollment
Intensity and duration ongoing challenges
Serving higher risk (4+), depressed, and prentally screened families may be particularly important
11/4/2015