Intervention Review

You have free access to this content

Postnatal parental education for optimizing infant general health and parent-infant relationships

  1. Janet Bryanton1,*,
  2. Cheryl T Beck2,
  3. William Montelpare3

Editorial Group: Cochrane Pregnancy and Childbirth Group

Published Online: 28 NOV 2013

Assessed as up-to-date: 17 SEP 2013

DOI: 10.1002/14651858.CD004068.pub4


How to Cite

Bryanton J, Beck CT, Montelpare W. Postnatal parental education for optimizing infant general health and parent-infant relationships. Cochrane Database of Systematic Reviews 2013, Issue 11. Art. No.: CD004068. DOI: 10.1002/14651858.CD004068.pub4.

Author Information

  1. 1

    University of Prince Edward Island, School of Nursing, Charlottetown, Canada

  2. 2

    University of Connecticut, School of Nursing, Storrs, Connecticut, USA

  3. 3

    University of Prince Edward Island, Department of Applied Human Sciences, Faculty of Science, Charlottetown, Prince Edward Island, Canada

*Janet Bryanton, School of Nursing, University of Prince Edward Island, 550 University Avenue, Charlottetown, C1A 4P3, Canada. jbryanton@upei.ca.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 28 NOV 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Anderson 1981

MethodsRandomization to 3 groups: 1 control and 2 intervention. Method of randomization not stated. Blinding of outcome assessors. Single-centre trial. 


ParticipantsIncluded 30 mothers and newborns. Inclusion criteria were: primiparas having uncomplicated pregnancies, labours, and births, local or regional anaesthesia, full term, healthy, female newborns, no post-birth complications. All breastfeeding. Trial conducted in a 500-bed community hospital likely in Chicago, Illinois, USA.


InterventionsGroup 1 (Control) received no instruction. Brazelton Assessment performed in nursery with mom not present, no feedback to moms, offered a class on infant furnishings during postpartum hospital stay. Group 2 (Explanation only) Brazelton Assessment performed in the nursery without mom present. Following assessment the investigator met with the mother in her postpartum room and explained the assessment done, her infant’s performance and positive aspects of performance stressed. Group 3 (Demonstration and Explanation) Investigator met with mother in her room, performed the Brazelton Assessment in the mother’s presence and gave an explanation of the assessment throughout. 10 dyads randomized per group.


OutcomesAn independent assessor (graduate student) assessed mother-infant dyads observed during a feeding at 24-48 hours (pre-test) and 10 to 12 days postpartum (post-test) (not clear where data were collected at 10 days). Measured Maternal Reciprocity (13 items), Infant Reciprocity (5 items) and Maternal-Infant Interaction (3 items). Items scored 1 to 5 with 1 representing low reciprocity and 5 representing high reciprocity. 


NotesSmall sample, no power analysis noted. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not stated.

Allocation concealment (selection bias)Unclear riskAllocation concealment not discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo losses discussed.

Selective reporting (reporting bias)High riskFindings reported but no data on maternal-infant interaction. Infant reciprocity data incomplete for Group 3.

Other biasLow riskGroups equivalent at baseline.

Bolam 1998

MethodsBlock randomization to 4 groups: 3 intervention and 1 control. Blinding of outcome assessors (different assessors at 3 and 6 months). Single-centre trial.


ParticipantsIncluded 540 Nepalese participants, all pregnant women admitted to Prasute Griha Hospital. Exclusion criteria were: women who had stillbirths or discharged from hospital before randomization occurred.


InterventionsIncluded 1-on-1 teaching for mothers in treatment groups. None for control (Group A - teaching at birth and 3 months, Group B - birth only, Group C - 3 months only, Group D - none) (n = 135 women randomized to each group). Intervention occurred in hospital and at home and included topics such as breastfeeding, treatment of diarrhoea, recognition of and response to upper respiratory tract infections, immunization, and family planning.


OutcomesMothers' knowledge of pneumonia, indrawing, tachypnoea; knowledge of diarrhoea; give rehydration solution; immunized infant. Infants' weight gain, length, and head circumference. Comparisons presented in analyses are for Group B vs Group D as only these groups met review inclusion criteria of 2 months post-birth.


NotesMay have included teens and mothers of NICU babies. Power analysis on all outcomes.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskBlock randomization using blocks of 20, random ordering of numbers from 0 to 19. Numbers 0-4, 5-9, 10-14, and 15-19 were assigned to groups A to D respectively.

Allocation concealment (selection bias)Low riskSequentially numbered, sealed, opaque envelopes. Generator of assignment was not involved in execution of allocation.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskSingle blinding of outcome assessors (different assessors at 3 and 6 months).

Incomplete outcome data (attrition bias)
All outcomes
High riskReported losses for all groups at 3 and 6 months due to stillbirth, infant death, and loss to follow-up. Total losses Group A (30%), Group B (22%) Group C (29%), Group D (27%). No imbalance across outcomes. No intention to treat. Greater than 20% attrition.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasLow riskGroups equivalent at baseline except 15% difference in illiteracy between groups B and C.

Christophersen 1982

MethodsMothers were selected sequentially from the obstetrics ward and randomly assigned by coin toss, in pairs to either the restraint seat group or usual care. There was no blinding of patients, caregivers, or outcome assessors. Single-centre trial.


ParticipantsIncluded 30 mothers who gave birth at Shawnee Mission Medical Center, a private non-profit hospital in suburban Kansas City, USA to a single live infant with baby's doctor practising within a 10-mile radius.


InterventionsIntervention (restraint seat): mothers (n = 15) were approached in their rooms prior to discharge by trained staff and offered a demonstration of the proper placement of the infant in the infant restraint seat, how to carry the seat with the infant in it, and the correct restraining of the seat with the automobile lap belt. They also placed the infant in the seat and securely in the car. Control group mothers (n = 15) were discharged in the usual manner.


OutcomesCorrect use of infant restraint seat at discharge and at 4 to 6 week follow-up.


NotesMothers were not asked to participate in the study; however their obstetricians, paediatricians, and family practitioners were asked for informed consent to observe their patients. The study was reviewed by the Medical Center Committee on human subjects. Interobserver reliability 100% at both times. Although education was given in the experimental group, the intervention included placing the newborn correctly in the car. Thus the observed outcome was either the result of the parents physically removing the infant or the staff member putting the infant in correctly.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomly assigned by coin toss to "restraint seat" or "no restraint seat" group.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants were not aware they were in a study so they were essentially blinded.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Low riskAt discharge no loss. At 4 to 6 weeks, .5% loss in treatment (n = 1) and 1% in control (n = 2). No intention-to-treat.

Selective reporting (reporting bias)Low riskMain outcome reported for both groups at both times.

Other biasHigh riskNo baseline data collected so unclear if groups equivalent at baseline.

Flagler 1988

MethodsRandomization to intervention and control group. Method of randomization not stated. No blinding. Likely a single-centre trial.


ParticipantsIncluded 74 American participants. Inclusion criteria were: vaginal birth, primipara, 20-30 years of age, living with father, normal full-term infant, and uncomplicated postpartum.Trial conducted in San Francisco, CA, USA.


InterventionsA 20-minute Brazelton assessment, individual demonstration of normal newborn behaviour with mother on second or third day postpartum in mother's room. Nothing stated regarding control group. The article did not state the number of participants randomized to each group (at analysis n = 31 in intervention n = 30 in control).


OutcomesIncluded maternal role competence as measured by: (1) the mother's perception of herself as a mother as compared against an ideal mother (15 items on a 7-point scale), with smaller distances between the 2 indicating greater role competence; (2) mutuality - the mother's ability to perceive her infant's cues, using the 10-item Issue One Subscale of the Maternal Attitude Scale; and (3) maternal anxiety related to child-rearing using the 14-item MAS subscale - a higher score reflects less anxiety. Data collection occurred at baseline and in the home 4 to 6 weeks' postpartum.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not stated.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding.

Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNumbers not given for initial randomization to groups. 13 lost to follow-up (18%) but not stated which group. No intention-to-treat.

Selective reporting (reporting bias)Low riskReported on all outcomes

Other biasLow riskGroups at baseline comparable except for gender of infant.

Gibson 1995

MethodsRandomization done by having mother select a number from a box that contained numbered wooden squares. Mothers selecting even numbered squares were placed in the intervention groups and those selecting odd-numbered squares were placed in the control group. No blinding evident. Single-centre trial.


ParticipantsIncluded 40 mothers on the postpartum unit of a metropolitan hospital in Arlington, Texas, USA. Inclusion criteria were not stated but all newborns were full term and were in the normal newborn nursery. 


InterventionsThe control group (n = 20) received 20 minutes of education related to first year infant feeding practices. The intervention group (n = 20) received the same 20 minutes of education as well as an additional 10 minutes of teaching with the same information  in the form of a video but condensed (50% over learning beyond the mastery level). Teaching was done in group format on the postpartum unit. Did not state number of days postpartum. 4 lost to follow-up at 2 weeks.


OutcomesIncluded maternal knowledge of infant feeding practices measured by the Infant Feeding Questionnaire which contained 16 multiple choice and 8 true and false questions. Pretest given to all mothers. If scored less than 95% (less than mastery), they were given correct answers and brief explanation. Then they answered questions they missed and process repeated until they had 95% or above. Post-test at 2 weeks during a visit to paediatrician or home visit.


NotesContent validity and internal consistency of instrument satisfactory. Ratings done by 2 nurses on similarity between verbal directions and film (4.6 and 4.8 on a scale of 5). No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomization done by having mother select a number from a box that contained numbered wooden squares.

Allocation concealment (selection bias)High riskNo allocation concealment.

Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding evident.

Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding evident.

Incomplete outcome data (attrition bias)
All outcomes
Low risk4 lost to follow-up at 2 weeks were not included in analysis.

Selective reporting (reporting bias)Low riskOutcomes reported but data not useable as only reported t-test and P values.

Other biasLow riskStated extraneous variables controlled by randomization but no baseline characteristics reported.

Goetter 2005

MethodsRandomization to intervention and control group. Method of randomization not stated and no blinding occurred. Single-centre trial.


ParticipantsIncluded 61 American participants. Inclusion criteria were: primiparas, 18-35, English-speaking. Exclusion criteria were: NICU infants and women who knew researcher. Trial conducted in a mountain community hospital, possibly Wyoming.


InterventionsIntervention (n = 32) included positioning for SIDS prevention, individual instruction with mother in hospital postpartum, explanation and demonstration of positioning, and how to maintain position, definition and causes of SIDS. Control (n = 29) received usual care (inconsistent individual or group teaching SIDS might or might not be mentioned).


OutcomesIncluded first week positioning (supine/not supine). Also, at 6 weeks, current position, last night's position, and today's nap position.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not stated.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding.

Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses to follow-up.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasUnclear riskGroups not equivalent at baseline. Intervention group was older and included more Caucasians than the control group.

Golas 1986

MethodsRandomization to intervention, contrast, and control groups. Method of randomization not described. Groups equivalent at baseline. No blinding. Single-centre trial.


ParticipantsIncluded 54 mothers and newborns. Inclusion criteria were: primiparas, uncomplicated births, newborns received examinations by private paediatrician group practice, term infants, no congenital anomalies  or medical complications in first 2 weeks of life. Exclusion criterion was: multiple births. Mothers were well educated, adults, married, white, middle-class Americans from Baltimore, MD, USA.


Interventions19 mother-infant dyads randomized to intervention: n = 19 to contrast and n = 16 to control group. Intervention included individual 2-hour session by nurse practitioner with each mother-infant dyad within 5 days of newborn being 2 weeks old. Setting: examination room of paediatrician’s office. Teaching plan with specific goals but delivered in a flexible order based on infant behaviour and mother’s responses and questions. (1) viewed a 25-minute film (The Amazing Newborn); (2) oral and visual presentation of different states of infant behaviour and appropriate maternal response; (3) demonstration of selected items from BNBAS and return demonstration by mothers, individually based on newborn. Control received no contact with paediatric nurse practitioner at 2 weeks. Contrast received no contact with nurse practitioner at 2 weeks. Completed Newborn Information Checklist at 2 weeks. Teaching provided to control and contrast mothers following study.


OutcomesMeasurements taken at 4 weeks postpartum lasting 20-25 minutes. Included (1) maternal knowledge of newborn characteristics and capabilities from birth to 4 weeks: consolability; interpretation of physical cues, crying, visual auditory responses. 20 items worth 1 point. Total score 0-20; (2) maternal confidence in interpreting infant behaviour. 12 items rated 1-5. Total score 12-60; (3) maternal satisfaction with teaching intervention-mean response (only intervention group). 


NotesSmall sample. No mention of power analysis.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not discussed.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding.

Blinding of outcome assessment (detection bias)
All outcomes
High riskResearcher conducted intervention and collected data.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskOverall a 14.8% loss (10.5% intervention; 15.8% contrast; 18.8% control). 5 withdrew after randomization (n = 2 intervention, n = 1 contrast, n = 2 control), and 3 were excluded from analysis because 1 had twins and 2 were enrolled in a similar class (n = 2 contrast and n = 1 control). Did not use intention-to-treat.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasLow riskChi2 analysis to rule out confounding by telephone information that was included in the teaching intervention given to mothers across groups. No significant difference in occurrence of calls to office between groups.  Groups were comparable at baseline.

Hall 1980

MethodsRandomization to intervention and control group. Method of randomization not stated and no blinding was described. Single-centre trial.


ParticipantsIncluded 30 participants recruited in the US, possibly in South Carolina. Inclusion criteria were: 18-30 years, primiparas, married, no chronic disease, an uncomplicated pregnancy, vaginal birth, full-term, no complications during birth and postpartum.


InterventionsIncluded individual teaching for mothers at home in 1 session 2 to 4 days postpartum. They were assessed individually based on learning needs, taught normal newborn behaviour, crying, feeding, spitting up, elimination, and predictability. 15 each randomized to intervention and control group. No mention if control received usual care.


OutcomesData collected at 1 month and included perception of infant and mean scores on neonatal perception. Results are presented as differences between the average baby and the mother's baby; lower values indicate more positive perceptions.


NotesSample chosen deliberately. No power analysis mentioned.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not discussed.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo mention of losses. Sample sizes at analysis were not given.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasLow riskGroups similar at baseline.

Issler 2009

MethodsBlock randomization Monday to Friday (excluded weekends to prevent contamination between groups). Blocks of 4 consecutive weeks at a time. Outcome assessor blinded. Single-centre trial.


ParticipantsIncluded 228 mother-infant pairs "living in a previously selected area of Porto Alegre" from September 2005 to September 2006. Pairs were excluded if mothers had a "severe physical handicap or mental health problems" such as "profound depression or overt schizophrenia". Study took place on the maternity ward of a large teaching hospital in a large city in Brazil.


InterventionsIntervention: 112 mothers received a 1-on-1 oral orientation and demonstration using a baby doll model regarding sleep position. They were also given a folder with information about infant sleep positioning and received the routine oral orientation from hospital personnel. The education occurred in the mother's hospital room. Control: 116 mothers received usual care which included routine oral orientation from hospital personnel.


OutcomesInfant sleep position at 3 and 6 months observed by an interviewer in the home or demonstrated by the mother if infant was awake during visit.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskTossed a coin to randomize in blocks of 4 consecutive weeks. Women giving birth on weekends were not included but this should not result in a selection bias.

Allocation concealment (selection bias)Low riskAll women who gave birth during the week assigned by randomization were eligible. All received the same condition for that week with no recruitment to minimize possible contamination effects.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot clear if women were blinded to study hypotheses. Otherwise clinicians nor mothers were blinded.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome assessors were blinded to group assignment. Once initial data were collected, group assignment was sealed in an envelope and not opened until the study was completed.

Incomplete outcome data (attrition bias)
All outcomes
High riskFollowing randomization, loss to follow-up in intervention group n = 21 (2 refused to participate, 10 moved to another town, 9 were not found at the stated address) and n = 16 in the control group (1 refused to participate, 8 moved to another town, 7 were not found at the stated address). The control group lost another 6 by 6 months. Attrition was comparable in both groups. (Intervention = 19% and control = 14% at 3 months and 19% for both at 6 months). No intention-to-treat analysis. At 3 months 91 were analyzed in the intervention group and 100 in the control and at 6 months 91 and 94 were analyzed respectively.

Selective reporting (reporting bias)Low risk2 outcomes reported.

Other biasLow riskBaseline equivalence except for marital status. In intervention group more women were married or living with a partner. This was not associated with sleeping position in bivariate analysis.

Jones 1977

MethodsRandomization to 4 groups: 3 intervention and 1 control group. Method of randomization not stated. Experimenter and outcome assessor blinded. No other blinding mentioned. Single-centre trial.


ParticipantsIncluded 40 mothers from Oklahoma, USA. Inclusion criteria were: well, primiparous women of a clinic population of 1 hospital who had a vaginal birth and planned to keep their infants; well, term infants without evidence of IUGR or anomalies, with a 1-minute Apgar greater than 4 and a 5-minute Apgar greater than 7. Exclusion criterion was: multiple births.


InterventionsIntervention and data collection occurred on the postpartum unit of 1 hospital. 10 mothers received “Additional Information (AI)” by the neonatologist within 24 hours post-birth. They were informed about sensory capabilities of newborns and encouraged to be aware of their infant’s capabilities. 10 mothers received “Additional Contact (AC)”. They were instructed by a nurse within 24 hours of birth to stroke their nude infants for 10 minutes, dress and hold or rock their infants for 5 minutes after 4 daily feedings. Benefits of skin to skin also given. 10 mothers received “Additional Contact and Additional Information (AI-AC). 10 control mothers received traditional contact which was a glimpse of the infant after birth, brief contact at 6 to 12 hours and visits of 20 to 30 minutes every 4 hours for feeding. Data collection occurred before discharge in the nursery on days 3 to 6.


OutcomesMaternal behaviours were observed from video-taped sessions of mothers watching their infants during the discharge exam (distance from infant, questions about care, questions about condition, contact with infant, behavioural responsiveness, verbal responsiveness to infant, verbal responsiveness to physician, distress of infant) and during a feeding (enface position, lateral trunk contact, fondling of infant, vocalizations). Behaviours were given a total score representing the presence of the behaviour at any time within a 30-second film clip summed over a series of 5 film clips. Left Sided Preference to Hold Infant presented from midline also observed. 


NotesInter-rater reliability 70% to 100% on 10 observations with an average of 91.6%. 10% refusal rate. Teenagers included in sample. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not stated.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskExperimenter blinded, otherwise not discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome assessor blinded.

Incomplete outcome data (attrition bias)
All outcomes
Low risk3 moms discharged before data collected were replaced by next eligible subject assigned to condition.

Selective reporting (reporting bias)High riskAll outcomes discussed but ANOVA tables for each outcome present only mean square and f ratio and data not useable.

Other biasLow riskGroups equivalent at baseline.

Keefe 2005

MethodsRandomized by project co-ordinator using central computer randomization. 3 groups: intervention, control, and post-test only group. Post-test only group not randomized (used to compare to control on outcome data to illustrate any changes due to attention from evaluation team or developmental changes alone). Evaluation team blinded, separate from intervention team. 2 centres.


ParticipantsIncluded 180 newborns and their parents. Inclusion criteria were: healthy, full-term, low risk infants between 2 to 6 weeks of age, living within a 2-hour radius of the city, and having unexplained crying more than 2.5 hours per day over the past week. Conducted in Charleston, South Carolina and Denver, Colorado, USA.


InterventionsControl group received routine care (not described). 121 in intervention and control. Doesn’t state number per group except 43 in non-randomized post-test only. Intervention parents received REST program. Intervention nurses formed specific recommendations and care plans for each infant. 4 home visits, 1 per week. Both parents received individual instruction, demonstration, written material, video in the home over the 4 visits. 16 lost to follow-up, not included in analysis.


OutcomesIncluded hours of infant crying using a crying log and infant fussiness using Fussiness Rating Scale (Likert-type scale). Collected at baseline, 4 weeks and 8 weeks.


NotesPower analysis done. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomized by project coordinator using central computer randomization.

Allocation concealment (selection bias)Low riskAdequate concealment.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskEvaluation team blinded, separate from intervention team.

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk16 withdrew or were lost to follow-up after randomization not included in analysis. Not explicit about numbers per group and losses per group not clear. States no differences between those who left and completed. No intention-to-treat.

Selective reporting (reporting bias)Unclear riskAll outcomes reported but not able to use data as unable to determine numbers per group and only P values or percentages given.

Other biasLow riskNo baseline data given but states groups are comparable on race, gender, birth outcome, parental characteristics.

Liptak 1983

MethodsInfants randomly assigned following Brazelton Assessment to 1 intervention and 2 control groups. Method of randomization not stated. Blinding of outcome assessor. Single-centre trial.


ParticipantsIncluded 75 mother-infant dyads. Inclusion criteria were: full term, first-born, healthy, white newborns of English speaking, middle class parents from Strong Memorial Hospital, Rochester, New York, USA.


InterventionsControl Group 1 infants received usual care. Physical exams were performed in the absence of parents. Mothers were visited by 1 of the researchers every day and any concerns raised by the mothers were addressed. Routine anticipatory guidance given. Control Group 2 infants received the same as Group 1 but in addition received complete physical exams in front of the mothers on the morning of discharge from hospital as well as a discussion of normal findings. This group was selected to control for the Hawthorne effect of an investigator intervention. Intervention Group 3 infants received the same care as those in Control Group 1 but also had parts of the Brazelton Neonatal Assessment Scale demonstrated to their mothers on the morning of discharge approximately 30 minutes following a feeding. A 20-minute semi-structured discussion, modified for the individuality of each infant, was held with the mother. Number randomized to each group not given; however it may have been 25 per group. 3 lost to follow-up; 2 before the first home visit and 1 after. Did not report on which group losses occurred.


OutcomesOutcomes measured at 1 and 3 months postpartum on home visits. Included observations and measurement of (1) gross physical contact, (2) feeding additional food, (3) positive contact, (4) overall Ainsworth rating, (5) mothers asking developmentally related questions, (6) mothers continuing to seek medical advice as well as self-reports of degree of bother, infant temperament. Following feeding, 35- and 45-minute observations during play. Observer recorded infant and maternal behaviour every 15 seconds using a behaviour checklist. After leaving home, observers rated mother-infant interactions using 21 9-point interaction scales by Ainsworth. At 1 month completed Broussard Perception Inventory and modified Degree of Bother Scale, and Carey Infant Questionnaire at 3 months. 


NotesPower analysis done. Inter-rater reliabilities ranged from 83% to 96%. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo method of randomization discussed.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk3 lost to follow-up but do not state to which group. No intention-to-treat.

Selective reporting (reporting bias)High riskAll outcomes addressed but they do not give the number randomized per group. Not able to use data because of this as only percentages and P values reported.

Other biasUnclear riskDemographic baseline characteristics appear comparable although not discussed.

McRury 2010

MethodsBasic randomized design. Mother-infant dyads assigned using a random numbers table. Outcome assessor blinded. Single-centre trial.


ParticipantsIncluded 51 mother-infant dyads. Infants were singletons, 37-41 weeks' gestation, and admitted to the normal newborn nursery. Mothers were required to have resources to view VHS tape. Exclusion criteria: mothers who did not speak English or were unable to fill out a diary. Conducted in Columbus, Ohio in a large community hospital.


InterventionsIntervention: 27 mothers were given a 30-minute video to view in hospital and then take home to view as well. Specific instructions about 5 steps to use to soothe infant during crying. Control: 24 mothers watched a 30-minute video "Begin with Love" that was offered on the newborn channel with standard anticipatory guidance about infant needs and care. Both groups received a lightweight blanket for swaddling.


OutcomesMothers documented sleep and crying in a diary 3 days per week during the 1st, 4th, 6th, 8th, and 12th week. Diaries were analyzed as 1) mean hours of total infant crying per day (time fussing, crying, inconsolable was calculated for each day; the mean of the daily sums at each time was calculated) and 2) mean daily total sleep time was calculated similarly to crying time. "Diary data that had more than 3 hours not recorded or recorded as 'can't remember' were excluded from the analysis". 3) Mother's level of stress was measured by the Parenting Stress Index at 6 and 12 weeks. "Total stress scores can range from 131 to 320 (less stress to more), with a score of 222 being the 50th percentile".


Notes1408 flyers distributed and 51 recruited. Nothing mentioned re validity or reliability of instrument but stated that it was "scored according to the manual". Sample small and under powered. Needed 42 per group to "detect a minimally intense intervention to reduce crying time by 40%".


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandom numbers table used.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskMothers blinded. Given an unmarked jacket on video so did not know which video they received.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskResearch assistant was blinded during baseline data collection but not during follow-up data collection as intervention fidelity was assessed.

Incomplete outcome data (attrition bias)
All outcomes
High risk9 pairs were lost to the intervention group (33%) and 7 to the control group (29%). Lower levels of education in those who left but about the same number in each group. No intention-to-treat. 18 analyzed in intervention group and 17 in control. Some diaries were incomplete.

Selective reporting (reporting bias)Low riskAppears to have provided data on all outcomes.

Other biasLow riskBoth groups equivalent at baseline.

Moore 1987

MethodsRandomization to intervention and control group. Method of randomization not stated. No mention of blinding. May have been a multi-centre trial.


ParticipantsIncluded 159 mothers who gave birth at the University of Texas Medical Branch Hospitals, Galveston Texas, USA. Inclusion criteria were: mothers who had a cesarean birth, were literate, spoke English, 17 years of age or older and singleton infants born after 37 weeks. Also, mother had to be able to hold and feed infant in her room.


InterventionsOn the third postpartum day, mothers in intervention group saw a video focusing on the manner in which infants from birth to 9 months exhibit multi-sensory capabilities for communication and how with interaction these capabilities can be enhanced. No mention if intervention was individual or group. No description of control group given. Number randomized to each group not stated.


OutcomesMaternal behaviour was assessed on the fourth postpartum day during a feeding. Maternal attachment behaviour assessment total and composite factor scores (e.g., maternal attentiveness, tactile response, body contact) were used. 


NotesNo useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo method of randomization discussed.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo mention of losses. Number per group not given.

Selective reporting (reporting bias)High riskAbstract only. No data presented.

Other biasUnclear riskNo baseline equivalence discussed.

Myers 1982

MethodsRandomization to 3 groups: (1) mother intervention, none for father, (2) father intervention, none for mother, (3) control was made post recruitment by the throw of a die. Single-centre trial.


ParticipantsIncluded 42 middle-class married couples in the US who had just had their first baby. No inclusion/exclusion criteria were specified.


InterventionsThe 2 intervention groups (n = 14 fathers in 1 group and n = 14 mothers in another) were taught individually how to administer most of the items of the Brazelton exam. Throughout the session, the experimenter gave information about infant development related to the infant's performance. The experimenter aimed at being supportive. Information learned was encouraged to be given to the other partner and to be used at home. The intervention was given after second day post-birth but before departure for home at day 4. Sessions were held in mother's room for the mothers and in a small room near the nursery for the fathers. Sessions lasted 45-60 minutes.


OutcomesIncluded (1) knowledge of infant behaviour as measured by a set of 15 multiple choice factual questions related to infants' physical capacities, including reflexes and senses (higher scores out of a maximum of 15 indicate more knowledge); (2) feelings of confidence as a caregiver; (3) feeling of affection and satisfaction with the infant; (4) behaviour with the infant. Measures were made greater than 6 hours after intervention.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandom assignment by the throw of a die.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNot discussed.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasUnclear riskAcceptance rates post randomization were about 66% overall but greater (no % given) in father intervention and control groups.Timing of outcome assessment an issue, since some were tested as soon as 6 hours after the intervention but no data are presented by group on this issue. No data on group comparisons at baseline were given.

Paradis 2011

MethodsBlock randomization using random numbers table to 1 intervention and 1 control group. Randomization occurred by week in blocks of 8 weeks to account for seasonal variation in birth rate. Blinding of outcome assessors. Single-centre trial.


ParticipantsIncluded 126 mothers and 11 fathers who were: 18 years of age or older, able to read and speak English, had access to a working telephone, and whose infant was less than 1 month old. Exclusion criteria: parents of premature or medically complex infants. Trial conducted in Rochester, New York, USA in a large hospital with a hospital-based primary care paediatric practice.


InterventionsIntervention: a media-based learning intervention of a 15 minute locally produced DVD was given to 70 parents during the baby's first visit to the paediatrician's office. The video depicted basic aspects of newborn care that were endorsed by the American Academy of Pediatrics and Brighter Futures guidelines. Content included: normal newborn breathing patterns, bathing, feeding, safe sleep, dealing with crying, promoting development. A staff member started the DVD in the exam room and it was stopped if it was still running when the provider came in. The DVD was then given to parent to take home. Control: 67 parents received an enhanced standard of care. They were given a packet of written handouts (written at the 4th grade reading level) that were already available at the clinic and covered similar but not identical material to the DVD. Parents were given the packet during the first clinic visit.


OutcomesParent: 1) knowledge of infant development using 14 questions from a 58-item Knowledge of Infant Development Inventory (chose questions pertaining most to newborns, answered agree, disagree, not sure); 2) self-efficacy using 20 items from a 52 item Infant Care Survey (rated on a 5-point scale from 1 (little confidence) to 5 (quite a lot of confidence); 3) problem solving competence using the How I Deal with Problem Regarding Care of My Baby Questionnaire (rated from 1 (never) to 9 (always) for a maximum score of 90). 3 scales administered at baseline and at 2 weeks post visit by telephone. Did not separate out mothers and fathers. Infant: health care utilization for infant assessed by chart review at 2 months. Utilization included: parent initiated clinic visits, phone calls, emergency department visits.


NotesNo validity or reliability information was given on the 2 adapted scales. Stated the third scale was validated but no further information. No useable data. Attempted to contact author with no response.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low risk"Randomization occurred by week (rather than by subject) to simplify administration of the intervention for the nursing staff". "Used a random-numbers table to generate the assignment of weeks to treatment or control in blocks of 8 to account for seasonal variations in birthrate".

Allocation concealment (selection bias)Low risk"At the beginning of each week, the principal investigator informed the nursing staff of whether" it was a treatment or control week. It is unlikely that a parent would not come for a visit knowing it was a control week or a treatment week.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot clear if participants were blinded. Personnal were not blinded. Recruiter was blinded.

Blinding of outcome assessment (detection bias)
All outcomes
Low risk"Outcomes were assessed by research associates who were blinded to group allocation".

Incomplete outcome data (attrition bias)
All outcomes
Low riskLoss of 6 participants, 3 per group: 1 refused and 2 lost to follow-up per group. Intervention = 4.3% loss and control = 4.4% loss. Performed an intention-to-treat analysis.

Selective reporting (reporting bias)Low riskDescribed findings for primary and secondary outcomes.No useable data, as data are reported in pre- and post-mean within-group differences.

Other biasUnclear riskIntervention group had more mothers and infants of Hispanic ethnicity and control group had more babies born outside hospital and being exclusively breastfed. Control group had better knowledge of infant development. Researchers adjusted for these baseline differences in analyses.

Paul 2011

MethodsStratified, randomized factorial 2X2 design. Mother baby dyad unit of randomization. Randomization method not described. Single-centre trial. It appears no blinding was used.


ParticipantsFull study included 160 mother baby dyads. Focus for this review is 80 mother baby dyads: 39 in intervention and 41 in control. Infants were included if they were singleton, 34 weeks' gestation or greater, without morbidities that would affect sleeping or feeding. Mothers were English speaking primiparas, who intended to breastfeed and follow-up care with a university affiliated primary care provider. Pairs were excluded if either newborn or mother had a hospital stay 7days or more and if mothers had a major pre-existing morbidity or condition that would affect postpartum care or study participation such as cancer, MS, or lupus. Study took place in Hershey, Pennsylvania, USA.


InterventionsThe "Soothe/Sleep" intervention was delivered by nurses during the first home visit at 2 to 3 weeks. Intervention included 1-on-1 verbal information, an instructional handout, and a commercially prepared video. Parents were taught "alternative strategies to feeding as an indiscriminate first response to infant distress" in an attempt to have parents understand the difference between crying from hunger versus other causes of crying. The video included strategies to calm and soothe the infant during the day and night when it was time for sleep. They also received a standard infant parenting book. The control parents received the standard infant parenting book. All parents had their questions answered.


OutcomesIncluded: infant weight-for-length percentile at 1 year; total daily crying and fussing; total daily sleep; total nocturnal sleep from 9pm to 6 am; timing and content of feedings measured by 96 hour diary cards at 3, 4, 8, 16, 24, 36, and 48 weeks post-birth. These were detailed at 15-minute intervals. Relative percentage of breast milk versus formula was documented using a visual analogue scale.


NotesMay have included infants in NICU. No useable data. Contacted author who chose not to provide data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNot addressed other than being stratified based on maternal prepregnancy BMI < 25 or >= 25.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
High riskAppears no blinding was used. Nurses providing the intervention could not be blinded.

Blinding of outcome assessment (detection bias)
All outcomes
High riskNurses measuring infants were not blinded.

Incomplete outcome data (attrition bias)
All outcomes
High riskGreater than 20% loss to follow-up in both groups. Losses equal across groups. Mothers not completing were significantly younger, less educated, single, non-white, and medicaid insured. No intention-to-treat for those who were randomized but did not receive the intervention. Some for non-completers who had growth data from second home visit.

Selective reporting (reporting bias)Low riskAppears to have reported on all outcomes but data not useable.

Other biasLow riskBaseline equivalence across all groups.

Petrowski 1981

MethodsRandomization to 4 groups: 3 intervention and 1 control. No method of randomization stated. No blinding noted. 2 hospital sites.


ParticipantsIncluded 56 mothers. Inclusion criteria were: primiparas who were staff patients of 2 inner-city hospital clinics in Washington, DC, USA, were educated in the USA, and who gave birth to healthy, normal newborns not less than 5 pounds.


InterventionsGroup 1 (n = 10) received the Instructional Package Form 1 in the last trimester of pregnancy. (Instructional Package included 4 cassette recordings: Care of the Umbilical Cord and Navel, Burping or Bubbling a Baby, Perineal Care, and Rest, Activity, and Exercise. Each tape accompanied by 13 pictures. 2 alternate forms). This group was also re-instructed postpartum in hospital using Form 2. Group 2 (n = 10) received the Instructional Package Form 1 in the last trimester of pregnancy. Group 3 (n = 10) received the Instructional Package, Form 1, from the first to fourth day postpartum in hospital. Group 4 (n = 10) (control group) received no instruction unless they requested it after the outcome measure was taken. 


OutcomesIncluded maternal acquisition of knowledge related to newborn and postpartum care measured using a multiple choice test of 20 questions based on the information in the package. Data were collected in the home from the fifth to thirteenth day postpartum.


NotesAntenatal and postpartum intervention for Group 1 and antenatal only for Group 2. Initially wanted sample size of 56 and then reduced it to 40. No mention of power analysis or basis for reduction. Controlled for participants’ educational level. Content validity of instructional packages assessed. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo method of randomization described.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
High risk20 of initial 56 randomized lost to follow-up (36%). No reasons given. Did not give losses per group. Recruited 4 new participants after initial randomization. No intention-to-treat.

Selective reporting (reporting bias)Unclear riskOutcome reported for all groups but data not useable, as no means or standard deviations presented.

Other biasUnclear risk No mention of baseline equivalence of groups.

Regan 1995

MethodsRandomization to intervention and control group was performed using a table of random numbers. Groups were similar on age and education but differed on 'other ethnicity'. Single-centre trial.


ParticipantsIncluded 100 women who were included in the caseloads of CNMs at the University of New Mexico Hospital (USA). Inclusion criteria were: English-speaking primiparas providing written consent.


InterventionsBoth groups attended a daily nurse-conducted class and received a packet of handouts of postpartum information. The experimental group (n = 45) also received individual education by the CNM. The content and length of intervention not given. Average age of baby was 17.6 hours at intervention. The control group (n = 55) did not receive individual education.


OutcomesKnowledge of postpartum mother-infant care based on written instructions developed by nurse-midwives and measured by a 20-item true/false instrument developed by the researchers; maximum score was 20.


NotesTo control for the confounding variable of nurses' teaching, the post-test was administered prior to the usual daily nurse-conducted postpartum class.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomization using a table of random numbers.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses.

Selective reporting (reporting bias)Low riskReported on outcome.

Other biasUnclear riskGroups were similar on age and education but differed on ethnicity.

Riesch 1984

MethodsRandomization to intervention and control group. Method of randomization not stated. No blinding mentioned. 2 hospital sites.


Participants2 studies were described. The first study included 108 mother-infant dyads. Inclusion criteria were: term, infants and women from 2 urban medical centres in a mid-western city, USA. The second study included 32 mother-infant dyads. Inclusion criteria were: preterm infants (34 to 37 weeks' gestation), singleton births, absence of physical malformations, absence of serious medical complications at 72 hours of age such as RDS, sepsis, convulsions, severe acid base imbalance and women from 2 urban medical centres in a mid-western city, USA. Exclusion criteria were: high-risk or ill mothers or infants including mothers with chronic illnesses, teen mothers, or mothers who required a cesarean birth.


InterventionsTerm study: 54 participants in intervention and control groups. Preterm study: 16 participants in both groups. The intervention groups received an audiotape with accompanying text which the mother listened to in her own room during a time mutually agreed upon. Visitors, calls, etc. were postponed during the listening period. The tape described neonate’s ability to demonstrate protective reflexes; to habituate to sound, light or motion; to self quiet; and to attend to objects. Appropriate maternal actions were described. Printed text was left with mother. In the term study the intervention occurred within 48 hours of infant’s birth. In the preterm study the intervention occurred within 5 days post-birth. The control groups received usual care (e.g., physical care, infant feeding, bathing).


OutcomesA 10-minute continuous observation using a checklist of 22 neonatal social behaviours and 16 maternal social behaviours, designed by researchers. Maternal self-report of her own and infant behaviours. Term study: maternal-infant interaction during an evening feeding, in hospital 24 to 72 hours post-birth. Preterm study: maternal-infant interaction during a feeding, in the home 2 to 4 days after discharge.


NotesInter-rater reliability .96 to 1.00. Content validity of instrument. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not described.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo losses described.

Selective reporting (reporting bias)Unclear riskOutcomes reported but no usable data.

Other biasLow riskGroups equivalent at baseline for both studies.

Shapiro 1987

MethodsRandomization to intervention and control group. Method of randomization not stated. No blinding described. Multi-centre study.


ParticipantsIncluded 696 women admitted to 3 maternity wards in US hospitals.


InterventionsBoth groups received a pamphlet about burns from hot water taps and a thermometer for testing maximum water temperature. The intervention group (n = 302 at analysis) also received a 1-minute verbal summary of the dangers of hot tap water, calling the pamphlet and thermometer to their attention. The control group (n = 302 at analysis) did not receive the verbal summary. The intervention was provided on the postpartum unit. Numbers randomized to both groups were not given.


OutcomesMaternal awareness of tap water burns and greater use of temperature testing; reported as percentages.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not described.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk92 participants were lost post randomization (13%); however, no information was given on numbers lost per group. No intention-to-treat.

Selective reporting (reporting bias)Low riskAppears that all outcomes are reported.

Other biasLow riskStated major socio-demographic attributes of the 2 groups were comparable but no data given.

St James-Roberts 2001

MethodsRandomization to 2 intervention and 1 control group was determined by a computer-generated randomization schedule and placed in a sealed envelope, which was opened between days 8 and 14 in their homes. No blinding described. Single-centre trial.


ParticipantsIncluded 610 consenting women who consecutively gave birth to a live singleton infant of greater than 37 weeks' gestation on postnatal wards of a large general hospital in the UK. Exclusion criteria were: non-fluency in English, no telephone, infants with congenital anomalies or admitted to the NICU.


InterventionsBehavioural group (n = 205) received a leaflet describing a 9-point program. The educational intervention (n = 202) consisted of a 10-page guide to baby crying and sleeping that was developed with local health professionals and a telephone number for CRYSIS (a voluntary organization for parents with young babies). Guide included a question and answer section on common problems and how to deal with them and a step-by-step guide to preventing crying and sleeping problems. This highlighted the need for a regular and structured approach to care that emphasized day-night differences, recommended settling baby in a crib or other designated place and sanctioned leaving baby to fret for a few moments when judged to be suitable. It did not prescribe a focal feed. It provided written advice and suggestions that could be adapted (not prescriptions, as was the case in the behavioural intervention). The control group (n = 203) received "normal services" (also received by 2 other groups greater than 1 home visit by a health visitor nurse, clinic visits, and access to a family doctor).


OutcomesInfant crying and sleeping measured using 24-hour behaviour diaries which gave a continuous record of starting time, end time, and duration of each behaviour. These were completed at 1 week of age (baseline) and at 3, 6, 9 and 12 weeks of age.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskA computer-generated randomization schedule was used for allocation to groups.

Allocation concealment (selection bias)Low riskGroup assignment was placed in a sealed envelope, which was opened between days 8 and 14 in their homes.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
High riskAt 1 week, losses were 3%, 2% and 6% respectively for diary completion. At 12 weeks, losses were 21%, 19%, and 22% respectively for diary completion. At 9 months, there was a 19% loss in behavioural group, 18% in educational group, and 19% in control group for questionnaire completion. Intention to treat not mentioned.

Selective reporting (reporting bias)High riskSpecific comparisons of sleeping and fuss/crying not reported for 1 week and 9 months "because of large numbers of measures collected"; however no significant differences noted at these times.

Other biasLow riskGroups comparable on demographics at baseline.

Stremler 2006

MethodsRandomization centrally by computer-generated random number sequence.The groups appear comparable at baseline.The outcome assessor was blinded. Single-centre trial.


ParticipantsIncluded 30 Canadian primiparous women who had a healthy singleton baby born greater than or equal to 37 weeks' gestation, lived in greater Toronto area, planning to provide care to their infant for first 6 weeks postdischarge. Exclusion criteria were: experienced complications requiring lengthy hospital stay, previous stillbirth/neonatal death, chronic poorly controlled maternal illness, mother used medications affecting sleep, history of drug/alcohol abuse, diagnosed sleep disorder, mother's partner worked night shifts, mother unable to read or understand English, no telephone.


InterventionsIntervention group (n = 15) received TIPS (Tips for Infant and Parent Sleep). The sleep intervention included a 45-minute meeting with a nurse to discuss sleep information and strategies, an 11-page booklet, and weekly phone contact to reinforce information and problem solve. The control group (n = 15) received a 10-minute meeting during which only maternal sleep hygiene and basic information about infant sleep were discussed, a 1-page pamphlet, and calls at weeks 3 and 5 to maintain contact without provision of advice.


OutcomesIncluded 6 infant sleep measurements determined by sleep diaries and Actigraph at 6 weeks. The Actigraph detects and records continuous motion data by use of a microprocessor. Detected movements are translated into digital counts across investigator-designated recording parameters. Brief movements in the middle of sleep periods are recorded as sleep, and brief periods of no activity within time intervals of extensive wakeful movement are recorded as awake.


NotesSame research nurse provided both TIPS and control intervention. Pilot study limited to 6 weeks. Small sample. Only 6 randomized per week due to limited number of Actigraphs.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomization of groups was done using a computer-generated random-number sequence.

Allocation concealment (selection bias)Low riskRandom-number sequence was held by a research associate outside of the immediate research team.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskContamination between groups was unlikely because "women were randomly assigned before discharge with little opportunity to share information before discharge". Also 1 woman per room was enrolled at a time.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskThe outcome assessor was blinded.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo losses.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasLow riskGroups were similar at baseline.

Sullivan 1980

MethodsRandomization to 4 groups: 3 intervention and 1 control. Method of randomization not stated. Outcome assessors blinded. Likely single-centre trial although not clear. 


ParticipantsIncluded 53 mother-infant dyads. Inclusion criteria were: mothers were healthy, had a vaginal birth of their first or second infant. Infants were healthy and full-term. Trial conducted in Torrance, California, USA. 


InterventionsIntervention Group 1 (n = 14) received 30 minutes of unstructured early contact during the second hour postpartum plus 30 minutes of special training in infant care based on BNBAS results for their infant during the postpartum period. Intervention Group 2 (n = 12) had early contact but not training. Intervention Group 3 (n = 14) had training without early contact. Control Group mothers (n = 13) had neither early contact nor training. Intervention occurred in hospital but not clear where training was provided.


OutcomesIncluded (1) levels of attachment behaviour, (2) self-confidence, (3) competency, and (4) mutuality of interaction at 1 and 2 months based on infant physical exam, 2 structured interviews, and scoring by independent raters of videotaped feelings and interactions. Not stated where data collection occurred.


NotesNo useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo method of randomization stated.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome assessors blinded.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo mention of losses.

Selective reporting (reporting bias)High riskNot all outcomes reported. Only P values presented. Abstract only.

Other biasUnclear riskNo baseline equivalence mentioned.

Symon 2005

MethodsBlind, randomization to intervention and control groups within block sizes of 8 was performed by a specific staff member. Single-centre trial.


ParticipantsIncluded 346 families recruited from birth notification published in the only South Australian newspaper with contact details from the telephone directory. They were contacted by telephone within 2 weeks of birth. Inclusion criteria were: gave birth at 36-42 weeks' gestation; English speaking; mother planning to provide full-time care to infant for greater than 12 weeks post-birth. Exclusion criteria were: infant was admitted to the NICU within the first 2 weeks post-birth.


InterventionsIntervention group (n = 171 families) received usual care plus both parents were invited to attend a 45-minute consultation with a nurse at 2 to 3 weeks on normal sleep patterns in newborn infants and a 50-page book reinforcing the information. If infant weight gain was less than 30g/d, parents were encouraged to see their usual postnatal care provider. Control care (n = 175 families) not described.


OutcomesIncluded greater than or equal to 15 hours of total sleep per 24 hours as per sleep diary for 7 consecutive days at 6 and 12 weeks, reported as percentages. Group differences in mean number of hours for sleep time were also reported at 6 and 12 weeks, as well as mean crying time per 24 hours.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskBlind, randomization to intervention and control groups within block sizes of 8 was performed by a specific staff member. Actual method not stated.

Allocation concealment (selection bias)Low riskBlind, randomization was performed by a specific staff member.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding discussed.

Incomplete outcome data (attrition bias)
All outcomes
High riskLoss at 6 weeks was 23% in intervention group and 31% in control group. At 12 weeks, loss was 36% and 39%, respectively. Losses greater than 20%. Intention-to-treat.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasHigh riskHigh refusal rate before randomization (65%).Groups appear similar on the only baseline information provided: infant sex and the socio-economic status of families.

Wendland-Carro 1999

MethodsRandomization to intervention and control groups. Method of randomization not stated. Outcome coders were blind to objectives and group assignment. Single-centre trial.


ParticipantsIncluded 38 primiparous mothers and infants from Poxto Alegre, Brazil. Inclusion/exclusion criteria not given.


InterventionsThe intervention (n = 17) was designed to enhance the mother-infant interaction. On day 2 or 3 postpartum, a video covering 5 areas was viewed and discussed with the mother individually to determine what she knew, whether she had seen these infant behaviours, and what her expectations were vis-a-vis her own infant. The investigator directed the mother's attention toward the infant's potential to interact as well as the importance of affectionate handling and sensitivity toward the infant and discovering the infant's individuality. A written list describing each of these items was then given and they were encouraged to identify the behaviours in their infants. The video lasted 15 minutes, the session lasted 50 minutes. 19 dyads participated in the control group and received an intervention that emphasized basic care giving skills.


OutcomesIncluded sensitive maternal responsiveness to the infant at the end of the first month. Video tapes during free play and bathing were used to capture this. Subsequently coded as synchronous (i.e., responsiveness to the behaviour of the other in the dyad). 13 categories used. Each observation period was divided into 15-second intervals and co-occurrence of synchronous categories were determined. Reported as mean frequency of co-occurrences.


NotesSmall sample. Intercoder reliabilities exceeded 80% for all categories.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not described.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo blinding discussed.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome coders were blinded to objectives and group assignment.

Incomplete outcome data (attrition bias)
All outcomes
Low risk2 mothers (11%) dropped out of the intervention group after being randomized, no loss from control group. No intention-to-treat.

Selective reporting (reporting bias)Low riskAll outcomes reported.

Other biasLow riskGroups comparable at baseline except more control fathers had unskilled occupations.

Worobey 1982

MethodsRandomization to 3 groups: intervention, contrast, and control. Method of randomization not described. Mothers were blinded to the hypotheses. Outcome assessors blinded to experimental design and purpose of study. Single-centre trial.


ParticipantsIncluded 48 mother-infant dyads from the Centre Community Hospital, State College, Pennsylvania, USA. Inclusion criteria were: healthy, full-term newborns.


Interventions16 dyads randomized per group. The 1-on-1 intervention consisted of 3 methods of relaying information from BNBAS. The control group (minimal information through auditory feedback) received usual care which consisted of a verbal description of their newborn’s performance on the BNBAS immediately following the assessment lasting approximately 15 minutes. The contrast group (passive observation with auditory and visual feedback) watched the examiner administer the BNBAS on their infants and received a verbal description of their infant behaviour as well as a 1-page summary 2 days later. The assessment lasted approximately 45 minutes. It occurred on the second or third day post-birth. The intervention group (active information through auditory, visual, and tactile interaction) were guided through an interaction with their newborns in which they administered the BNBAS. The facilitator did not touch the infant but rather guided the mother through the procedure. The intervention lasted approximately 45 minutes on the infant’s second or third day. These mothers also received a 1-page summary


OutcomesAt 4 to 6 weeks post-birth, dyads were observed for an hour in the home before, during, and after a bathing session. Maternal-infant behaviours were observed and documented on a pre-coded check list each time a behaviour from 16 behaviour categories occurred in a 15-second period. 5 a priori clusters of behaviours were used to create scores: (1) contingent interaction, (2) embellished maternal involvement, (3) simple maternal attention, (4) basic maternal care, and (5) infant behaviour. 


NotesInter-rater reliability between observers .87 to .99. No useable data.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMethod of randomization not described.

Allocation concealment (selection bias)Unclear riskNot discussed.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskMothers were blinded to the hypotheses.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome assessors blinded to experimental design and purpose of study.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo mention of losses.

Selective reporting (reporting bias)Unclear riskAll outcomes reported but no useable data.

Other biasLow riskGroups equivalent at baseline

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Adam 1985Study is not a RCT.

Akai 2008Not a structured form of postnatal education. Home-based support program. On average infants were 4.3 months old at start of intervention. Mother-infant dyads participated for about 4.5 months and then outcomes measured when infants were approximately 9 months of age.

Baqui 2008Part of an antenatal education and postnatal preventative and curative newborn care intervention for community health workers not parents.

Barlow 2013Study of teen parents.

Barr 2009Not a structured form of postnatal education. Study is not exclusively a postpartum intervention. Education started prenatally.

Bashour 2008Home visitation program, not a clear structured educational intervention.

Beal 1989Study is not a RCT - alteration of assignment to groups, no random start.

Beiler 2011Not a structured form of postnatal education. Office-based versus home-nurse visits.

Bristor 1984Study is not a RCT, no randomization of groups.

Cevasco 2008Singing songs is not a structured form of postnatal education.

Cohen 1980Study is not a RCT, although the abstract indicated it was.

Cupples 2011Study is not exclusively a postpartum intervention. Education began twice monthly during pregnancy.

Dihigo 1998Study is not a RCT - randomization to 2 treatment groups only, not control group.

Doherty 2006Study is not exclusively a postpartum intervention and results were not presented separately.

French 20125 time periods for intervention up to 12 months postpartum. Not able to separate out the 2-week and 2-month interventions.

Glazebrook 2007Parents of infants in NICU.

Goyal 2009Not a RCT. Also not an educational intervention. Support for postpartum depression.

Hansen 1990Study is not exclusively a postpartum intervention and results were not presented separately for intervention received at 0 to 2 months.

Hawkins 2008Couples had already enrolled in Welcome Baby program prior to being randomly assigned to 1 of the 2 treatment groups. Assignment to control group was not randomized. First 40 couples who were not interested in Welcome Baby program but were interested in being in study were enrolled in control group. Postnatal education did not start until 3 months after birth. Data collected for post-tests at 4-5 months and again 9-10 months postpartum.

Ho 2009A support-based intervention for postpartum depression not a structured form of postnatal education.

Kabakian-Khasholian 2005Study outcomes were related to women's use of postpartum services.

Kemp 2011Not a structured form of postnatal education. A home-based support intervention. Started in the prenatal period.

Kistin 2011Not a structured form of postnatal education.

Lee 2012Intervention not related to infant health/care or infant-parent relationships. Intervention related to women's postpartum sexual health behaviours.

Leff 1988Study is not a RCT - 2 intervention groups and no control group.

Morrell 2000Study does not test an educational intervention.

Niccols 2008Parenting program not within 2 months postpartum; infants from 1-24 months.

Petch 2012Study is not exclusively a postpartum intervention. Education began in the prenatal period.

Reich 2010Not a structured form of postnatal education. Study is not exclusively a postpartum intervention. Education began in the third trimester of pregnancy.

Rotheram-Borus 2011Not a structured form of postnatal education. Study is not exclusively a postpartum intervention. Home-based support program, intervention began with 4 antenatal visits.

Santelices 2011Study is not exclusively a postpartum intervention. There were 6 prenatal group sessions.

Scott 1990Infants recruited at less than 18 months; no data presented separately for infants receiving the intervention prior to 2 months old.

Shapiro 2011Study is not exclusively a postpartum intervention. Workshop was completed during pregnancy.

Simons 2001Study is not a structured teaching intervention and effect of education not isolated from other intervention components.

Spigelblatt 1991Study is not a RCT although the study title indicated that it was.

Wagner 1999Study does not focus on infants less than 2 months old.

Waterston 2009Postnatal education was beyond 2 months postpartum. The newsletter went through the first 12 months postpartum.

Wen 2011Home-based support program, intervention started in the prenatal period.

Wolfson 1992Study is not exclusively a postpartum intervention. Education started prenatally.

 
Characteristics of studies awaiting assessment [ordered by study ID]
Kim 2004

MethodsNot clear if this study is a RCT. It is written in Korean except for the tables. Attemping to secure a translation.

Participants35 mothers and their infants.

InterventionsEducation related to sensory stimulation administered twice a day.

OutcomesInfant growth, mother-infant interaction.

Notes

 
Characteristics of ongoing studies [ordered by study ID]
Cook 2012

Trial name or titleBaby business: A randomized controlled trial of a universal parenting program that aims to prevent early infant sleep and cry problems and associated parental depression.

MethodsRCT

Participants750 English speaking parents in Australia with healthy newborns greater than 32 weeks' gestation.

InterventionsExperimental group receives Baby Business program providing information to parents about infant sleep and crying via DVD and booklet mailed soon after birth and telephone consultation at 6-8 weeks postpartum. Control groups receives usual care.

OutcomesParent report of infant night time sleep as a problem at 4 months, parent report of infant day time sleep or crying as a problem, mean duration of infant sleep and crying/24 hours, parental depression symptoms, parent sleep quality and quantity and health service use.

Starting date2011

Contact informationFallon Cook at fcook@parentingrc.org.au

Notes

 
Comparison 1. Education on sleep enhancement versus usual care

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Total infant sleep >= 15 hrs per 24 hrs1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    1.1 At 6 weeks
11749Risk Ratio (M-H, Fixed, 95% CI)1.72 [1.56, 1.90]

    1.2 At 12 weeks
11497Risk Ratio (M-H, Fixed, 95% CI)1.73 [1.54, 1.95]

 2 Total minutes of infant sleep in 24 hrs (mean diff)3Mean difference (Fixed, 95% CI)Subtotals only

    2.1 At 4 weeks
1Mean difference (Fixed, 95% CI)60.0 [-24.02, 144.02]

    2.2 At 6 weeks
3Mean difference (Fixed, 95% CI)62.08 [42.88, 81.29]

    2.3 At 8 weeks
1Mean difference (Fixed, 95% CI)-12.0 [-78.58, 54.58]

    2.4 At 12 weeks
2Mean difference (Fixed, 95% CI)61.41 [28.08, 94.73]

 3 Night-time minutes of infant sleep in 24 hrs (mean diff)3Mean difference (Fixed, 95% CI)Subtotals only

    3.1 At 6 weeks
2Mean difference (Fixed, 95% CI)29.13 [18.53, 39.73]

    3.2 At 12 weeks
2Mean difference (Fixed, 95% CI)16.18 [4.41, 27.95]

 4 Longest uninterrupted night-time minutes of infant sleep in 24 hrs (mean diff)3Mean difference (Fixed, 95% CI)Subtotals only

    4.1 At 6 weeks
2Mean difference (Fixed, 95% CI)13.74 [-1.11, 28.58]

    4.2 At 12 weeks
2Mean difference (Fixed, 95% CI)11.45 [-5.40, 28.30]

 5 Day-time minutes of infant sleep in 24 hrs (mean diff)3Mean difference (Fixed, 95% CI)Subtotals only

    5.1 At 6 weeks
2Mean difference (Fixed, 95% CI)39.59 [25.01, 54.17]

    5.2 At 12 weeks
2Mean difference (Fixed, 95% CI)9.92 [-1.83, 21.66]

 6 Longest uninterrupted day-time minutes of infant sleep in 24 hrs (mean diff)3Mean difference (Fixed, 95% CI)Subtotals only

    6.1 At 6 weeks
2Mean difference (Fixed, 95% CI)5.57 [-2.31, 13.45]

    6.2 At 12 weeks
2Mean difference (Fixed, 95% CI)0.60 [-3.89, 5.09]

 7 Infant crying time in 24 hrs (mean diff)2Mean difference (Fixed, 95% CI)Subtotals only

    7.1 At 4 weeks
1Mean difference (Fixed, 95% CI)18.0 [-25.63, 61.63]

    7.2 At 6 weeks
2Mean difference (Fixed, 95% CI)4.36 [-6.44, 15.16]

    7.3 At 8 weeks
1Mean difference (Fixed, 95% CI)42.0 [-6.41, 90.41]

    7.4 At 12 weeks
2Mean difference (Fixed, 95% CI)0.55 [-8.38, 9.47]

 8 Night-time minutes of infant sleep in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    8.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)6.0 [-8.21, 20.21]

 9 Longest uninterrupted night-time minutes of infant sleep in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    9.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)-5.0 [-27.66, 17.66]

 10 No. of infant night-time sleeping episodes1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    10.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)0.10 [-0.18, 0.38]

 11 Day-time minutes of infant sleep in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    11.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)-5.0 [-19.86, 9.86]

 12 Longest uninterrupted day-time minutes of infant sleep in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    12.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)-5.0 [-11.40, 1.40]

 13 No. of infant day-time sleeping episodes1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    13.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)0.10 [-0.12, 0.32]

 14 Night-time minutes of infant fussing/crying in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    14.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)2.0 [-2.24, 6.24]

 15 Longest uninterrupted night-time minutes of infant fuss/cry in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    15.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)0.0 [-2.24, 2.24]

 16 No. of infant night-time fussing/crying episodes1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    16.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)0.20 [-0.08, 0.48]

 17 Day-time minutes of infant fussing/crying in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    17.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)3.0 [-4.07, 10.07]

 18 Longest uninterrupted day-time minutes of infant fuss/cry in 24 hrs1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    18.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)-1.0 [-3.24, 1.24]

 19 No. of infant day-time fussing/crying episodes1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    19.1 At 12 weeks
1316Mean Difference (IV, Fixed, 95% CI)0.40 [-0.10, 0.90]

 20 Maternal stress score170Mean Difference (IV, Fixed, 95% CI)30.31 [15.20, 45.41]

    20.1 At 6 weeks postpartum
135Mean Difference (IV, Fixed, 95% CI)23.80 [2.08, 45.52]

    20.2 At 12 weeks postpartum
135Mean Difference (IV, Fixed, 95% CI)36.40 [15.38, 57.42]

 
Comparison 2. Education on infant behaviour versus usual care

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Knowledge of infant behaviour2Mean Difference (IV, Fixed, 95% CI)Subtotals only

    1.1 Maternal at 1-4 days postpartum
128Mean Difference (IV, Fixed, 95% CI)4.6 [2.60, 6.60]

    1.2 Maternal at 4 weeks postpartum
256Mean Difference (IV, Fixed, 95% CI)2.85 [1.78, 3.91]

    1.3 Paternal at 1-4 days postpartum
128Mean Difference (IV, Fixed, 95% CI)6.5 [4.88, 8.12]

    1.4 Paternal at 4 weeks postpartum
126Mean Difference (IV, Fixed, 95% CI)3.70 [1.93, 5.47]

 2 Child-rearing anxiety161Mean Difference (IV, Fixed, 95% CI)3.70 [-1.33, 8.73]

 3 Mother distance161Mean Difference (IV, Fixed, 95% CI)0.20 [-0.11, 0.51]

 4 Mutuality161Mean Difference (IV, Fixed, 95% CI)1.10 [-3.07, 5.27]

 5 Synchronous co-occurrences during free play - visual136Mean Difference (IV, Fixed, 95% CI)10.10 [5.96, 14.24]

 6 Synchronous co-occurrences during free play - vocal136Mean Difference (IV, Fixed, 95% CI)6.73 [3.64, 9.82]

 7 Mothers' perception of their infants130Mean Difference (IV, Fixed, 95% CI)0.66 [-8.44, 9.76]

    7.1 At 1 month postpartum
130Mean Difference (IV, Fixed, 95% CI)0.66 [-8.44, 9.76]

 8 Maternal confidence interpreting infant behaviour130Mean Difference (IV, Fixed, 95% CI)3.70 [-1.16, 8.56]

    8.1 At 4 weeks postpartum
130Mean Difference (IV, Fixed, 95% CI)3.70 [-1.16, 8.56]

 
Comparison 3. Education on general post-birth infant health or care versus usual care

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Maternal general post-birth knowledge score (mean)1100Mean Difference (IV, Fixed, 95% CI)0.40 [-0.27, 1.07]

 2 Maternal general post-birth knowledge score (per cent correct)1100Risk Ratio (M-H, Fixed, 95% CI)1.07 [0.92, 1.23]

 3 Infant weight (kg)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    3.1 At 6 months post-birth
1203Mean Difference (IV, Fixed, 95% CI)0.10 [-0.19, 0.39]

 4 Infant length (cm)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    4.1 At 6 months post-birth
1202Mean Difference (IV, Fixed, 95% CI)0.30 [-0.88, 1.48]

 5 Head circumference (cm)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    5.1 At 6 months post-birth
1203Mean Difference (IV, Fixed, 95% CI)-0.20 [-0.76, 0.36]

 6 Appropriate immunization1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    6.1 At 6 months post-birth
1202Risk Ratio (M-H, Fixed, 95% CI)1.04 [0.97, 1.11]

 7 Knowledge of signs of infant pneumonia1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    7.1 Indrawing
1203Risk Ratio (M-H, Fixed, 95% CI)1.21 [0.73, 2.03]

    7.2 Tachypnea
1203Risk Ratio (M-H, Fixed, 95% CI)1.19 [0.90, 1.58]

 8 Knowledge of action to take in the case of infant diarrhoea1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    8.1 Continue breastfeeding
1203Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.75, 1.31]

    8.2 Give oral rehydration solution
1203Risk Ratio (M-H, Fixed, 95% CI)1.00 [0.92, 1.08]

 
Comparison 4. Education on infant safety versus usual care

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Infant restraint seat fastened by lap belt1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    1.1 At hospital discharge
130Risk Ratio (M-H, Fixed, 95% CI)21.0 [1.34, 328.86]

    1.2 At 4-6 weeks
127Risk Ratio (M-H, Fixed, 95% CI)1.24 [0.34, 4.51]

 2 Post-hospitalization awareness of tap water burns1604Risk Ratio (M-H, Fixed, 95% CI)1.07 [1.04, 1.11]

 3 Greater use of temperature testing1604Risk Ratio (M-H, Fixed, 95% CI)1.76 [1.43, 2.17]

 4 Supine infant sleep position2Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    4.1 At 1 week
161Risk Ratio (M-H, Fixed, 95% CI)1.31 [1.00, 1.72]

    4.2 At 6 weeks
161Risk Ratio (M-H, Fixed, 95% CI)1.21 [0.85, 1.71]

    4.3 At 3 months
1191Risk Ratio (M-H, Fixed, 95% CI)1.79 [1.17, 2.72]

    4.4 At 6 months
1185Risk Ratio (M-H, Fixed, 95% CI)2.18 [1.35, 3.53]