Manuscript Detail | Home Visiting Evidence of Effectiveness

Paul, I. M., Savage, J. S., Anzman-Frasca, S., Marini, M. E., Beiler, J. S., Hess, L. B., ... & Birch, L. L. (2018). Effect of a responsive parenting educational intervention on childhood weight outcomes at 3 years of age: the INSIGHT randomized clinical trial. JAMA, 320(5), 461-468. https://doi.org/10.1001/jama.2018.9432

Manuscript screening details
Screening decision	Screening conclusion	HomVEE procedures and standards version
Passes screens	Eligible for review	Version 2

Study design details
Rating	Design	Attrition	Baseline equivalence	Compromised randomization	Confounding factors	Valid, reliable measure(s)
High	Randomized controlled trial	Low	Not assessed for randomized controlled trials with low attrition	No	No	Yes, details reported below for findings on valid, reliable outcomes that otherwise rate at least moderate

Notes:

Information on sample sizes and details about findings on body-mass index were based on correspondence with the author. The authors reported several findings that are not eligible for review because they do not examine the impact of the intervention on an eligible outcome.

Study characteristics
Study participants	Study participants were mother and infant dyads recruited after delivery in one Pennsylvania hospital. To be eligible, dyads had to include full-term, singleton births, with infants of normal birth weight. Mothers were English-speaking, primiparous, and at least 20 years old. A total of 291 dyads were randomly assigned to either the Intervention Nurses Start Infants Growing on Healthy Trajectories (INSIGHT) intervention (145 dyads) or a comparison intervention (146 dyads). Mother–infant dyads were randomly assigned to the INSIGHT intervention or the comparison intervention two weeks after birth, and outcomes were measured until the children’s third birthday. In the study, 89 percent of mothers were White, 6 percent were Black, 3 percent were Asian, and 1 percent reported another race. Seven percent of mothers reported Hispanic ethnicity.
Setting	The study took place in Hershey, Pennsylvania.
Intervention services	INSIGHT consisted of home visits conducted when infants were 3 to 4, 16, 28, and 40 weeks old, followed by annual clinic-based visits at 1, 2, and 3 years old. The curriculum taught parents to respond promptly and in developmentally appropriate ways to infant cues across four behavioral states (drowsy, sleepy, fussy, and alert/calm). Research nurses provided parents with developmentally appropriate sleep guidance during each visit. The guidance addressed bedtime routines, sleep location, and night waking. During and between visits, caregivers were provided with information and resources on responsive feeding, lactation support, soothing practices, and home safety (including crib safety and choking hazards). When infants were 2 weeks old, participants received a mailed packet with information on infant feeding.
Comparison conditions	Families assigned to the comparison condition were not eligible to receive intervention services through the INSIGHT program. However, these families received a similar number of home visits as the INSIGHT group (when infants were 3 to 4, 16, 28, and 40 weeks old) and annual clinic-based visits at 1, 2, and 3 years old. The home visits focused solely on home safety topics, including crib safety. When infants were 2 weeks old, participants received a mailed packet with information on infant feeding.
Subgroups examined	This field lists subgroups examined in the manuscript (even if they were not replicated in other samples and not reported on the summary page for this model’s report). • Boys • Girls • Intended feeding mode (breastfeeding) • Intended feeding mode (formula)
Funding sources	This research was supported by Award Numbers R01DK088244 and UL1TR000127 from the National Institutes of Health and 2011-67001-30117 from the U.S. Department of Agriculture. The research was also supported by the Children’s Miracle Network at Penn State Health Children's Hospital and the Penn State Clinical and Translational Science Institute.
Author affiliation	The authors are affiliated with several universities, including Pennsylvania State University, and are developers of the INSIGHT program.
Peer reviewed	Yes

Study Registration:

Clinicaltrials.gov Identifier: NCT01167270. SocialScienceRegistry.org Identifier: None found. Registry of Efficacy and Effectiveness Studies Identifier: None found. Study registration was assessed by HomVEE for Clinicaltrials.gov beginning with the 2014 review, and for other registries beginning with the 2021 review.

Findings that rate moderate or high

Child health
Rating	Outcome measure	Effect	Sample	Timing of follow-up	Sample size	Intervention group	Comparison group	Group difference	Effect size	Statistical significance	Notes
High	Accelerated weight gain	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	16 weeks old	269 infants	Unadjusted proportion = 0.09	Unadjusted proportion = 0.12	Difference = -0.03	HomVEE calculated = -0.19	Not statistically significant, p= 0.44	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	Accelerated weight gain	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	1 year old	253 children	Unadjusted proportion = 0.32	Unadjusted proportion = 0.34	Difference = -0.02	HomVEE calculated = -0.05	Not statistically significant, p= 0.97	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	1 year old	253 children	Unadjusted mean = 17.00	Unadjusted mean = 17.40	Mean difference = -0.40	HomVEE calculated = -0.28	Statistically significant, p = 0.02	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	2 years old	243 children	Unadjusted mean = 16.50	Unadjusted mean = 16.80	Mean difference = -0.30	HomVEE calculated = -0.21	Not statistically significant, p = 0.10	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	4 weeks old	279 children	Unadjusted mean = 13.80	Unadjusted mean = 14.30	Mean difference = -0.50	HomVEE calculated = -0.43	Statistically significant, p <.01	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	232 children	Unadjusted mean = 15.80	Unadjusted mean = 16.20	Mean difference = -0.40	HomVEE calculated = -0.32	Statistically significant, p = 0.02	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	16 weeks old	269 children	Unadjusted mean = 16.40	Unadjusted mean = 16.80	Mean difference = -0.40	HomVEE calculated = -0.27	Statistically significant, p = 0.03	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	28 weeks old	262 children	Unadjusted mean = 17.10	Unadjusted mean = 17.70	Mean difference = -0.60	HomVEE calculated = -0.39	Statistically significant, p <.01	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	40 weeks old	259 children	Unadjusted mean = 17.30	Unadjusted mean = 17.80	Mean difference = -0.50	HomVEE calculated = -0.33	Statistically significant, p = 0.01	Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	BMI (in standardized units)	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	2 years old	243 children	Unadjusted mean = -0.09	Unadjusted mean = 0.11	Mean difference = -0.20	Study reported = -0.21	Not statistically significant, p= 0.10	Author-reported effect size was calculated as mean difference between standardized scores.
High	BMI (in standardized units)	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	232 children	Unadjusted mean = -0.13	Unadjusted mean = 0.15	Mean difference = -0.28	Study reported = -0.28	Statistically significant, p= 0.04	Author-reported effect size was calculated as mean difference between standardized scores.
High	BMI percentile	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	232 children	Unadjusted mean = 46.90	Unadjusted mean = 53.80	Mean difference = -6.90	HomVEE calculated = -0.24	Not statistically significant, p= 0.07
High	BMI percentile	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	2 years old	243 children	Unadjusted mean = 48.10	Unadjusted mean = 52.40	Mean difference = -4.30	HomVEE calculated = -0.15	Not statistically significant, p= 0.26
High	Conditional weight gain	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	28 weeks old	262 children	Unadjusted mean = -0.18	Unadjusted mean = 0.18	Mean difference = -0.36	Not available	Statistically significant, p <.01	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	Downward crossing of 2 major percentile lines for height and weight	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	279 children	Unadjusted proportion = 0.02	Unadjusted proportion = 0.04	Difference = -0.01	HomVEE calculated = -0.32	Not statistically significant, p = 0.47	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention. Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.
High	Percent obese	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	2 years old	243 children	Unadjusted proportion = 0.01	Unadjusted proportion = 0.08	Difference = -0.08	HomVEE calculated = -1.47	Statistically significant, p= 0.01	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	Percent obese	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	232 children	Unadjusted proportion = 0.03	Unadjusted proportion = 0.08	Difference = -0.05	HomVEE calculated = -0.70	Not statistically significant, p= 0.08	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	Percent overweight	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	2 years old	243 children	Unadjusted proportion = 0.11	Unadjusted proportion = 0.21	Difference = -0.09	HomVEE calculated = -0.43	Statistically significant, p= 0.04	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	Percent overweight	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	232 children	Unadjusted proportion = 0.11	Unadjusted proportion = 0.20	Difference = -0.09	HomVEE calculated = -0.41	Not statistically significant, p= 0.07	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention.
High	Weight-for-age below 5th percentile	FavorableUnfavorable or ambiguousNo Effect	INSIGHT vs. comparison, Pennsylvania 2012-2014, full sample	3 years old	279 children	Unadjusted proportion = 0.08	Unadjusted proportion = 0.04	Difference = 0.04	HomVEE calculated = 0.39	Not statistically significant, p = 0.22	Submitted by user on Fri, 03/15/2019 - 14:29 Negative effect is favorable to the intervention. Submitted by nwu on Fri, 04/12/2019 - 18:31 Statistical significance is based on HomVEE calculations.