Primary Care Interventions to Promote Breastfeeding: Article

You Are Here: AHRQ Home > Clinical Information > U.S. Preventive Services Task Force > Topic Index > Counseling: Breastfeeding > Article

Appendix Table. Quality Assessment of Randomized, Controlled Trials

Study, Year (Reference)	Design	Method of Randomization^a	Allocation Concealment Adequate^b	Intentionto-Treat Analysis	Outcome Assessors Blinded	Loss to Follow-up,%^c	Results Adjusted for Confounding	Groups Similar at Baseline	Recruitment Method Appropriate^d	Statistical Analyses Appropriate^e	Overall Quality
Developed countries
Anderson et al., 2005⁹³	Parallel	Assigned by the study field coordinator	No	Yes	Unclear	15	None	Yes	Yes	Yes	Fair
Bonuck et al., 2005⁹⁴ and 2006⁹⁵	Parallel	Blocked and stratified according to center	Yes	Yes	No	21 (missing breastfeeding data)	Maternal age, ethnicity, Medicaid status, previous breastfeeding data	Yes	Yes	Yes	Fair
Carfoot et al., 2004^96,c	Parallel	Computer-generated randomization list, sequence of envelopes	No	No	No	7.1	None	Yes	Yes	No	Poor
Carfoot et al., 2005⁹⁷	Parallel	Computer-generated randomization list, sequence of envelopes	No	Yes	No	3.4	None	Yes	Yes	Yes	Fair
Chapman et al., 2004¹¹⁷	Parallel	Computer software program	Unclear	Yes	No	25	None	No	Yes	Yes	Fair
Dennis et al., 2002¹¹⁸	Parallel	Random number generated by a statistician	Yes	Unclear	Yes	1	None	No	Yes	No	Fair
Di Napoli et al., 2004¹²⁴	Parallel	Unclear	Unclear	Yes	No	10	Age, parental education, smoking, parity, participation in breastfeeding course, type of delivery	Yes	Yes	Yes	Fair
Ekström and Nissen, 2006;¹³⁰; Ekström et al., 2006¹³¹	Quasi	Randomized pairwise; centers matched in pairs that were similar in size and had similar breastfeeding duration	Yes	No	Unclear	Unclear (can be as high as 33)	None	Yes	Yes	Yes	Poor
Finch and Daniel, 2002⁹⁹	Parallel	No	No	No	Unclear	37	None	Yes(presumed)	Unclear	Yes	Poor
Forster et al., 2004¹⁰⁰	Parallel	A computerized system of biased urn randomization	No	No	Unclear	7	Income, smoking before pregnancy, education	Yes	Yes	Yes	Fair
Gagnon et al., 2002¹¹⁹	Parallel	Block randomization, using computergenerated blocks and stratified by parity	Unclear	Yes	Yes	15	None	Yes	Yes	Yes	Fair
Graffy et al., 2004¹²³	Parallel	Random permuted blocks by the statistical adviser	Yes	No	Yes	14	Decision about the feeding plan	Yes	Yes	Yes	Good
Henderson et al., 2001¹⁰¹	Parallel	Computer-generated balanced blocks of 20	Yes	No	No	6.3	None	Yes	Yes	Yes	Fair
Howard et al., 2003¹⁰²	Parallel	Computer-generated balanced blocks of 20	Yes	Yes	Yes	2	All predictors with P ≤ 0.10, including maternal race, previous births, and maternal education	Yes	Yes	Yes	Good
Kools et al., 2005¹²⁹	Cluster	Coin-flip for the center randomization, clusters matched by breastfeeding rates	Yes	Yes	Unclear	1.0	Variability of breastfeeding rates among the 10 centers	Yes	No	Yes	Fair
Kramer et al., 2001¹⁰³	Parallel	Computer-generated blocks of 4	Yes	No	Yes	8	Marital status, smoking	Yes	Yes	Yes	Good
Kronborg et al., 2007¹²⁸	Cluster	Computerized	Unclear	Unclear	No	~1.8	None	Yes	No	Yes	Fair
Labarere et al., 2003¹⁰⁴	Parallel	Computer-generated, random numbers in blocks of 8	Yes	Yes	Yes	9.5	None	Yes	Yes	Yes	Good
Labarere et al., 2005¹⁰⁵	Parallel	Random permuted blocks with a block size of 8	Yes	Yes	No	2	Age, education, white-collar worker, smoking, prenatal class attendance, primiparity, epidural anesthesia, infant birthweight and gestational age, breastfeeding <1 h after delivery, postpartum length of stay >4 d, expected breastfeeding duration >4 mo	Yes	Yes	Yes	Good
Lavender et al., 2005¹²⁷	Cluster	Wards were pair-matched; pairs were randomly allocated to the groups by a midwife, independent of the trial	Yes	Yes	Yes	5 to 7	None	Yes	Yes	Yes	Poor^f
Mattar et al., 2007¹¹⁴	Parallel	Computer-generated list	Yes	Yes	No	10	Adjust for multiple comparisons	Yes	Yes	Yes	Good
McKeever et al., 2002¹⁰⁶	Parallel	Central randomization procedures	Unclear	No	No	26	None	Yes	Yes	Yes	Poor
McLeod et al., 2004¹³²	Quasi	Random number	No	Yes	No	60	Yes (although unclear what variables were adjusted for)	No	Yes	Yes	Poor
Mizuno et al., 2004¹⁰⁷	Parallel	Unclear^g	Unclear	Unclear	No	10	Site of enrollment; age of infant at interview; maternal, paternal, and infant characteristics	Yes	Yes	Yes	Fair
Moore and Anderson, 2007¹¹⁵	Parallel	Randomization with algorithm to adjust for maternal age, education, marital status, race, smoking, breastfeeding intention, infant sex, and health care provider	No	No	Yes	5	None	Yes	Yes	Yes	Poor^h
Muirhead et al., 2006¹⁰⁸	Parallel	In a block of 10, separated for each of 4 strata (primigravidae, previous formula feeder, previously breastfed <6 wk, previously breastfed >6 wk)	Yes	Yes	No	2.3	None	Yes	Yes	Yes	Fair
Noel-Weiss et al., 2006¹²⁵	Parallel	Matching the sealed manila envelope with a sealed, sequentially numbered, opaque envelope containing the assignments	Yes	Yes	Yes	9	None	Yes	Yes	Yes	Good
Pugh et al., 2002¹²⁰	Parallel	Unclear	Unclear	Yes (no dropout)	Unclear	0	Yes (only on matching factors)	Yes	Yes	No	Poor
Quinlivan et al., 2003¹²¹	Parallel	Computer-generated	Yes	Yes	Unclear	9.4	Age, social class, baseline knowledge, factors that were unbalanced between the 2 groups (ethnic origin, social isolation, involvement of the father, homelessness)	Yes(unbalanced factors between the 2 groups were adjusted for)	Yes	Yes	Good
Ryser, 2004¹⁰⁹	Parallel	Participants select a sealed envelope	Unclear	Yes (no dropout)	No	0	No	Yes	Yes	No	Poor
Schlickau and Wilson, 2005¹¹⁰	Parallel	Unclear	Unclear	No	Unclear	17	None	Unclear	Yes	Yes	Poor
Su et al., 2007¹¹⁶	Parallel	Computer	No	Yes	Unclear	18	None	Yes	Yes	Yes	Fair
Wallace et al., 2006¹¹¹	Parallel	Telephone-balanced block and computer	Yes	Yes	Yes	6	None	Yes	Yes	Yes	Good
Wilhelm et al., 2006¹¹²	Parallel	Random number	Unclear	No	No	3	Baseline breastfeeding self-efficacy, length of time before returning to work	No	No	Yes	Poor
Winterburn et al., 2003¹²²	Parallel	Unclear	Unclear	Unclear	Unclear	Unclear	None	Unclear	Yes	Yes	Poor
Wolfberg et al., 2004¹¹³	Parallel	Unclear	No	No	Unclear	3	Breastfed previously, mother was breastfed as an infant, mother plans to breastfeed for first month, mother lives with father; mother’s breastfeeding beliefsⁱ	Yes	Yes	Yes	Poor
Developing countries (for BFHI only)
Coutinho et al., 2005⁹⁸	Parallel	Random-number table	Yes	Yes	Yes	6	None	Yes	Yes	Yes	Good
Kramer et al., 2001¹²⁶	Cluster	Random-number table	Unclear	Yes	Unclear	3	Birthweight, maternal age, previously breastfed infant for ≥3 mo, number of children in household, maternal smoking, family atopic history	Yes	Yes	Yes	Good

BFHI = Baby-Friendly Hospital Initiative.
a If cluster randomized, controlled trial (RCT), method used to generate the random allocation sequence, including details of any restriction (e.g., blocking, stratification, matching).
b If cluster RCT, method used to implement the random allocation sequence, specifying that allocation was based on clusters rather than individuals and clarifying whether the sequence was concealed until interventions were assigned.
c A good-quality RCT must have < 20% loss to follow-up.
d Appropriate consecutive or randomized.
e If cluster RCT, statistical methods used to compare groups for primary outcome indicating how clustering was taken into account; methods for additional analyses, such as subgroup analyses and adjusted analyses.
f Downgraded to poor quality because only 64.7% of women in intervention attended the workshop.
g Authors stated that it was not possible to randomize all sites because of constraints on willingness of different practices to provide different services and other reasons.
h Downgraded to poor quality because the RCT was underpowered to detect differences.
i Mother’s mother thinks that the baby should be breastfed, mother believes that her partner thinks her baby should be breastfed, or father would like the baby to be breastfed.

Return to Document