A Systematic review of the effectiveness of infant massage programmes in improving mother and infant attachment and dyadic attunement.

06 June 2019

Rebecca Balakrishna, Melanie Teixeira, Jane Meyrick and Roxanne Hart sought to answer this question with a systematic review of the evidence from studies across 36 years.


Melanie Teixeira Bsc (Hons) MBPsS
Assistant Psychologist/Family Support Key Worker (CAMHS & North Somerset Council)

Rebecca Balakrishna (formerly Milford)
Consultant Nurse Manager, CAMHS, Children's Services, Weston Area Health Trust

Research summary 

  • A secure attachment relationship between mother and infant is recognised as a protective factor in the likelihood of children having positive mental health in adulthood.
  • National, regional and local drives to enhance perinatal and infant mental health, some directly reference infant massage.  International studies featuring infant massage from January 1980 to May 2016 were reviewed to find evidence that massage groups are an effective intervention to improve mother and infant attachment.
  • Of 732 papers identified, 10 met the criteria for inclusion.
  • Some evidence was found of improvement in mother-infant interaction and dyadic attunement, and an increased sense of wellbeing and confidence in the mothers.
  • In some studies a higher level of improvement was seen in women who had previously experienced moderate mental health difficulties.
  • All studies were too weak to make generalisations from findings. Some significant evidence still emerged.
  • More robust research is needed to support the current findings. It should focus on using infant massage programmes to improve the mental wellbeing of mothers with previously recognised moderate mental health concerns.


A secure attachment relationship between mother and infant is now recognised as a protective factor and key component in increasing the likelihood of children developing positive mental health in adulthood.

Research over 10 years suggests 23% of mothers, with no known mental health difficulties, have insecure-dismissive attachment, 19% insecure-preoccupied attachment and 58% secure attachment. Acknowledging these links between maternal attachment style and the dyadic relationship with the infant, this review looked at the potential effectiveness of infant massage in promoting improved mother-infant attachment and dyadic attunement. 


NICE guidance for planning and commissioning children’s services suggests that health visitors and midwives should consider evidence-based interventions, such as infant massage, as part of provision for the social and emotional well-being of new mothers and the under 5’s. The aim of this study was to establish the evidence of the effectiveness of infant massage interventions in order to support decision makers about funding these resources within perinatal services.


A Systematic review was conducted using 4 databases including publications from January 1980 to May 2016. The criteria included massage or touch, when used as an intervention on more than one occasion. Of 732 papers identified, 10 papers met the criteria for inclusion in the systematic review.


Synthesis of papers was difficult, as they varied in many criteria. They were grouped into those considering improvement in mother-infant Interaction, improvement specifically in mothers presenting with mental health difficulties and those focusing on other improvements. Moderate evidence was found of improvement in all three areas following infant massage intervention, indicating that it has a positive effect on the dyadic relationship.


Although there was moderate evidence for short-term improvement in infant attachment, following infant massage interventions, no long-term conclusions could be made from the review. More robust and longitudinal research would be needed to support current findings and careful consideration given to the use of appropriate outcome measures. Two papers suggested that further research should focus on using infant massage programmes to improve the mental wellbeing of mothers with previously recognised moderate mental health concerns.

Introduction and Background

The practice of mothers massaging their babies from birth is seen in many cultures worldwide and has been traditionally used to calm, sooth and stimulate babies. In the 1970s Vimala McClure was inspired by these cultural practises to establish the International Association of Infant Massage (IAIM), developing protocols and methods for a standardised infant massage programme. Her key aim was to promote and enhance communication and attunement between mother and infant. In the field of attachment, attunement refers to the mother’s sensitivity to the nonverbal cues from their infant relating to feelings and emotions. Learning to read these cues allows the attuned mother to respond appropriately to the infant’s physical and emotional needs, thereby enhancing the infant’s feelings of security and containment (Stern, 1985).  To promote a feeling of regulation between mother and infant, the mother will enhance the interaction with exaggerated facial expressions, body gestures, tone of voice and range of sounds (Fonagy, Gergely, Jurist, & Target, 2002). Infants experiencing an attuned relationship can then learn about regulating their own emotions by observing their mothers’ emotional responses. This co-regulated relationship builds an important understanding of each other (Winnicott, 1987).

Positive attunement, as part of a secure attachment relationship between mother and infant, is now recognised as a key component in increasing the likelihood of children developing positive mental health in adulthood and protecting against the effects of negative risk factors (National Institute for Health and Care Excellence [NICE], 2016; Oates et al., 2007; Bowlby, 1969). Infants exposed to significant risk factors (e.g. parents with relationship and attachment difficulties, parental drug abuse, parents who have mental or physical health problems, teenage parents etc) face a 90-100% likelihood of having one or more delays in their cognitive, language or emotional development (NICE, 2016; Barth et al., 2008; Runyan, Wattam, Ikeda, Hassam, & Ramiro, 2002). Rees (2007) concluded that positive attachment is fundamentally important to child protection.

Mothers who have an insecure attachment style often find it hard to contain their own emotions, struggle to read their babies nonverbal signals and can feel a sense of being dominated by their baby (Oates et al, 2007; Gerhardt, 2004). Research by Bakermans-Kranenburg and Van IJzendoorn (2009) suggested 23% mothers, with no known mental health difficulties, have insecure-dismissive attachment, 19%  insecure-preoccupied attachment and 58% secure attachment. 

NICE guidance for planning and commissioning children’s services suggests that health visitors and midwives should consider evidence-based interventions, such as group based baby massage, as part of provision for the social and emotional well-being of under 5’s. It also suggests that infant massage might be effective for use with depressed mothers. Robust evidence of the effectiveness of these courses in promoting mother-infant attachment would give confidence to decision makers about funding these resources within perinatal services, as a cost-effective and efficient way of increasing health outcomes.


Summary of policy framework driving the focus on maternal mental health and infant wellbeing

There are national, regional and local drives for developing services that enhance perinatal mental health and infant mental health, some of which directly reference infant massage.

Future in Mind (NHS England & Department of Health [DOH], 2015) focuses on promoting, protecting and improving children and young people’s mental health and wellbeing, particularly focusing on enhancing existing maternal, perinatal and early years health services to strengthen attachment between parent and child. Avoiding early trauma and building resilience could be achieved by ensuring parental access to evidence-based programmes of intervention and support. In addition, the Five Year Forward paper (NHS England, 2016) – to be achieved by 2020 – also discusses promoting resilience, prevention and early intervention by improving access to evidence-based programmes of intervention and support for the most vulnerable families.

NHS England report (2015) highlights the largest population of mothers affected by perinatal mental health issues are those with moderate illness and anxiety states, post-traumatic stress (following difficult birth) and adjustment disorder. It recommends follow-up and support for these women, should include promotion of ‘sound infant-parent attachment’. The figures for this group in England 2012-2013, based on the Office for National Statistics live births, was up to 318966 (480 per 1000 live births), whilst mothers with postpartum psychosis or chronic serious mental illness numbered 2650 (four per 1000 live births). This estimated rate shows the greatest number of mothers with mental health difficulties are cared for in the primary care and public health arena, where support and interventions are limited and cuts are being made.

The Healthy Child Programme outlines a programme of interventions for Early Years Workers, specifically Health Visitors (DOH, 2009), putting heavy emphasis on supporting mothers and fathers to provide sensitive and attuned parenting, in particular during the first months and years of life. Infant Massage is mentioned as an example of a dyadic therapy to increase maternal sensitivity.

In response to increasing referral numbers to specialist Child and Adolescent Mental Health Services (CAMHS), NHS England developed Children and Young People-Increasing Access to Psychological Therapy (CYP-IAPT) and in 2016 the 0-5 curriculum was devised to train workers in the early year’s field in perinatal and infant mental health care. The course outline includes development of early access universal and evidence-based interventions, including baby massage.


Criteria for considering studies for the review

Types of Study Considered

Studies using a randomised control trial, were included in the review. Studies which used outcome measures pre and post intervention were also included. International studies were included, but studies that simply expressed opinion, or focused only on describing an intervention, were not included.


Mothers and full-term infants, up to the age of one year were included. Teenage mothers (under 18), preterm infants and infants with developmental delay or significant health problems were not included, to allow generalisation of any findings for the general population.

Types of Interventions Considered

All infant massage approaches were included, regardless of different training approaches. Studies where the practice was culturally imbedded were also included. Note was taken of studies where other interventions were offered alongside infant massage

Search Criteria.

The electronic search strategy was replicated across the four databases, with syntax and subject headings changed slightly to account for different databases and platforms. See appendix 1.

Selection of papers

Combined searches, minus duplicates, produced 906 papers. Foreign language only studies were removed, reducing papers to 776. Animal studies were removed, reducing papers to 732.

See appendix 2 flow diagram.

The 732 abstracts were read independently by the researchers (RB) and (MT), with outcomes recorded on a database with additional comments. This was reviewed and disagreements discussed. A third party was available if needed.

Data Extraction including resolution of disagreement

By following this process, ten papers were included in the review for synthesise.

See Appendix 3.


Study Design 

The research methods, outcome measures, data items collected and collection environment for each study were assessed following the structure of PRISMA to inform determination of quality of research design.

Research method

The methods used in the ten final studies were categorised into groups to enable comparison of results:

  • Four studies used non-equivalent control groups with a pre-test/ post-test design
  • Three studies used a random control design
  • One study used a pre-test/ post-test design without a control group
  • One study used a post intervention evaluation only
  • One study completed a critical realist research design with mixed method

Outcome measures

A variety of outcome measures and questionnaires were used and in the majority of studies they were completed pre and post intervention. Each study collected and extracted differing data:

Data items

 The range of data focus of the studies considered:

Maternal behaviours and emotional responses,

Child behaviours and emotional responses,

and Dyadic Interactions.

Data Collection process

Data collection times varied between studies as did Data collection environments (including home, clinic and by mail) which could have influenced the data.

Synthesis Results

Comparison between studies was inadvisable, due to variety of criteria (recruitment methods, interventions offered, focus on differing benefits of infant massage). Therefore, a narrative synthesis was carried out, structured around those which showed greatest effect in three areas: mother-infant interaction, mother’s with mental health difficulties, and other improvements highlighted by the studies.

Highest improvement in mother-infant interaction

Two of the studies showed a specific improvement in mother-infant interaction, following the IAIM massage intervention. The Onozowa et al. (2001) study showed improvement in both control and experimental groups, but significantly higher levels of improvement in the infant massage group in the overall dyadic quality of interaction and feelings of closeness toward the infant. Lee (2006) also showed significant improvement in mother-infant interaction following intervention. Both Onozowa and Lee reported small sample size and short study period, rendering indication of long-term benefits impossible. Lee believed lack of participant similarity and the difference in time and quality of daily massage routines was also likely to influence results.

Gurol and Polat (2012) found a significant difference between groups post intervention in their small-scale study, with the infant massage group achieving a greater increase in attachment scores using the Maternal Attachment Inventory. However, they identified lack of randomised design in their trial and potential for control group mothers to massage their babies weakened their findings.

No significant difference was found by Watanabe et al. (2012) study, despite minor correlations being identified. Quality of results was limited by length of study.

The Fujita et al (2006) study was also limited by length of study, and possible bias due collection timings of cortisol samples.

All of the above studies looked at aspects of the mother-infant attachment relationship and drew broad conclusions that there was enough evidence of the effectiveness of infant massage programmes to warrant further research.

Highest improvement specifically in mothers presenting with mental health difficulties:

The main focus of all studies included was the mother-infant relationship. However, some included outcome measures assessing the mother’s mood state. Interestingly, Underdown (2013) found that low-risk mothers showed no significant change in mother-infant interaction, regardless of the programme they followed, but moderate-risk mothers showed a significant improvement, when using a ‘high quality programme’. High-risk mothers showed no improvement in interaction, regardless of quality of programme, and also showed some increased intrusive behaviours. The International Association of Infant Massage baby massage course, was considered to be a ‘high quality programme’, but ‘low quality programmes’ were unspecified. They suggest that the high level of low-risk mothers recruited may have affected results. They suggest mothers are specifically recruited with moderate levels of need in future research.

O’Higgins et al. (2008), showed mothers identified pre intervention as having moderate mood disorder, had similar levels of sensitivity to their infant post intervention as the non-depressed mother group. This improvement was not seen in the control group. They proposed future research should consider whether this change could be attributed to anticipated support, as this could be a confounding variable.

Onozawa et al. (2001) found that levels of Depression using the Edinburgh Postnatal Depression Scale decreased in both groups, however, this was more significant in the infant massage group.

Fujita, et al (2006) found that mothers who experienced the infant massage intervention showed a reduction in depression and increase in vigour at three months. They concluded that infant massage had a positive effect on the mood status of mothers, but their study was limited by being conducted in a short timescale.

In summary, despite the acknowledged limitations of the studies, there was moderate evidence that infant massage interventions had a positive effect on maternal mood, particularly, (and in some cases only) in mothers who are experiencing moderate mental health difficulties. None of the studies in the systematic review recruited women specifically who were experiencing moderate mental health difficulties. This would be an interesting future area of research.

Other improvements identified by the studies:

Hart (2003) and Beyer and Strauss (2003) both describe an increase in the mothers’ perceived sense of wellbeing. In Hart’s (2003) study 90% scored the massage programme as ‘very good’ or ‘excellent’, and rated the massage programme highly in its usefulness. Decreased scores in perceived levels of stress were identified in Beyer and Strauss’ study and an improved perception of competence. Hart reported bias due to usage of one trainer, and believed size and timescale of study also prohibited drawing significant conclusions. Beyer and Strauss were concerned that all participants had considerable social support outside the intervention. They suggest future research controls for the issue of time spent with the infant in other activities versus massage.

Clarke (2002), found that the intervention encouraged protected time with their baby and promoted a sense of parenting competency, including an increase in perceived levels of responsiveness and closeness, competency in communication and calming the infant’s physical needs. However, no statistically significant impact was found using outcome measures. Limitations were noted, due to study length, and participant similarity.


Results showed, despite limitations identified, moderate evidence that, following an infant massage intervention, some mothers showed improved mother-infant interaction, improved mood and a higher level of perceived wellbeing and parenting competency. Further research of improved quality is needed to enable conclusions to be drawn. Some studies indicated that improvements were more significant in women experiencing moderate mental health difficulties. Further research specifically targeting these women would be useful.

The process of ascertaining any impact of infant massage intervention from the studies considered has been complicated by the heterogeneity of the studies. However, some conclusions can be drawn from studies where relevant similarities have been identified in intent or method. 

Several studies identified improvement in the self-reported maternal level of warmth felt towards their baby and an increased parenting confidence. Other studies found improvement in levels of mother-infant interaction, in the overall dyadic quality and feelings of closeness toward the infant. These findings appear to support the idea that infant massage promotes improvement in the perceived and actual mother-infant dyadic relationship.

Key problems with existing studies:

Lack of detailed information in the studies constrains interpretation of results. Information on participant compliance is limited, particularly in regard to the daily administration of massage at home. Eighty percent of the studies do not specify whether one or more trainers were used, and the elements of social interaction and peer support have not been fully explored. Seven studies used the International Association of Infant Massage course (Field, Diego and Hernandez-Reif, 2007; Pigeon-Owen, 2007) as an intervention. The frequency and dosage of massage to be applied by parents at home was not specified in nine of the ten studies, and all relied on participants self-reporting compliance. It is therefore unlikely that the massage was completed in equal dosage and frequency for each infant, reducing the quality of the studies.

The control group intervention was not described in ninety percent of studies, reporting it as normal routine post-natal support care. It is unknown which interventions were offered and whether these were also available to the intervention groups. Future studies need to clarify specific details of routine care and to whom this was available.

Caution must be exercised when generalising from results as most of the studies were conducted in a western environment with predominantly white, educated women. Participants in all studies were volunteers, recruited either though health clinics or postnatal groups.

Implications for Practice:

There is growing belief by early years practitioners that infant massage is an easily accessible and effective approach to supporting good attachment between infants and vulnerable mothers (Day, 2014; Lorenz, Moyse, and Surguy 2005).  However, Valerie Smith (2014) identified that there appears to be a lack of high-quality research, focusing on the specific effect of infant massage on strengthening mother-infant attachment. This review concluded, that when combining all the outcomes and clinical observations of the ten studies, there is some evidence that baby massage has a positive effect on the dyadic relationship, and when added to previously conducted research showing good outcomes for physical growth and cognitive development (Lihong, Weihong, & Fentao, 2002), it would be reasonable to argue that infant massage is likely to have a positive effect on the overall wellbeing of the infant. There is also so incidental evidence that infant massage may have a particularly positive effect on mothers experiencing moderate mental health difficulties.

Implications for Research:

The two studies using a more robust approach, by combining outcome measures and clinical observations, showed more depth to their results and therefore more significant findings. Three studies also assessed the correlation and interconnectivity between maternal mental health and infant interaction. These studies all concluded that more research in this area was needed.

The studies in this review showed design limitations. The most common limitations included: intervention length, not fully randomised, one trainer, differing intervention and outcome measures, timings and environments and participant similarity. Most studies were short term and did not consider the longer-term effects of infant massage.

Using International Association of Infant Massage (used in 60% of the trials) as an intervention would be a way of guaranteeing standardisation of delivery and enabling comparisons to be made between studies.

High levels of bias were reported by the authors of the studies, particularly in reference to participants remaining in the study. These were often socially stable, well-educated and motivated to take part. They may have been well attuned to their babies and mentally well. This raises the question of whether infant massage is accessible or attractive to women who are struggling to attune to their babies. 

There was little evidence of service user participation in study design, focus and choice of outcome measures. This is likely to be a useful strategy in future. Service users could also increase participation maintenance, give useful feedback and be helpful in reviewing final results and conclusions.


The review focused on identifying evidence of the effectiveness of infant massage interventions, to inform decision makers about funding these resources within perinatal services. The review focused specifically on the effects of infant massage on mother-infant attachment and attunement and the dyadic relationship. Some moderate evidence was identified that this intervention does have a positive effect on both. However, evidence was weakened by the methodological quality of the studies. The variety of outcome measures and data items used made comparisons between studies inadvisable, but a narrative synthesis of results was undertaken. In future, identification of appropriate outcome measures, with service user participation in research design, are likely to benefit this process.

In addition to the main focus, several studies considered the effects of baby massage on the mother’s wellbeing and mood state, concluding that the highest level of benefit was seen in women with existing moderate mental health difficulties. Further research focusing on this particular group of women is likely to be useful, considering any correlation between the improvement in the mother mental health and the mother-infant attunement.


Bakermans-Kranenburg M J & Van IJzendoorn MH. (2009). The first 10,000 adult attachment interviews: distribution of adult attachment representations in clinical and non-clinical groups. Attachment & Human Development, 11(3), 223-263.

Barth RP, Scarborough AA, Lloyd EC,  Losby JL, Casanueva C & Mann T. (2008). Developmental status and early intervention service needs of maltreated children. Washington, DC: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.

Bennett C, Underdown A & Barlow J. (2013). Massage for promoting mental and physical health in typically developing infants under the age of six months. Cochrane Database of Systematic Reviews, 2013 (4), 1-158. doi: 10.1002/14651858.CD005038.pub3.

Beyer K & Strauss L. (2003). Infant massage programs may assist in decreasing parental perceived stress levels in new parents. Occupational Therapy in Health Care, 16(4), 53-68.

Bowlby J. (1969). Attachment and loss: Volume 1 attachment (2nd ed). New York: Basic books.

Brandon M, Sidebotham P, Bailey S, Belderson P, Hawley C, Ellis C & Megson M. (2012). New learning from serious case reviews: A two year report for 2009-2011. Department of Education. (DFE-RR037).

Bristol Health Partners. (2016). Improving Perinatal Mental Health (IMPROVE). See: (accessed 27 May 2019).

Clarke CL, Gibb C, Hart J & Davidson A. (2002). Infant massage: Developing an evidence base for health visiting practice. Clinical Effectiveness in Nursing, 6, 121-128.

Cooke A. (2015). Infant Massage: The practice and evidence base to support it. British Journal of Midwifery, 23(3), 166-170.

Crittenden PM. (2008). Raising parents: Attachment, parenting, and child safety. Collumpton: Willan Publishing.

Day J. (2014). Benefits of infant massage. The Practising Midwife, 17(5), 18-21.

Department of Health. (2009). Healthy Child Programme [online]. London: Department of Health. See: (accessed 27 May 2019).

Feijo L, Hernandez-Reif M, Field T, Burns W, Valley-Gray S & Simco, E. (2006). Mothers' depressed mood and anxiety levels are reduced after massaging their preterm infants. Infant Behavior and Development, 29(3), 476-480.

Field T, Diego M & Hernandez-Reif M. (2007). Massage therapy research. Developmental Review, 27(1), 75-89. 

Fonagy P, Gergely G, Jurist EL & Target M. (2002). Affect regulation, mentalization, and the development of the self. New York: Other Press.

Fujita M, Endoh Y, Saimon N & Yamaguchi S. (2006). Effect of massaging babies on mothers: pilot study on the changes in mood states and salivary cortisol level.  Complementary Therapies in Clinical Practice, 12(3), 181-185.

Gerhardt S. (2004). Why love matters: How affection shapes a baby’s brain. New York: Brunner-Routledge.

Gurol A & Polat S. (2012). The effects of baby massage on attachment between mother and their infants. Asian Nursing Research, 6(1), 35-41.

Hart J, Davidson A, Clarke C & Gibb C. (2003). Health visitor run baby massage classes: investigating the effects. Community Practitioner, 76(4), 138-142.

Lee HK. (2006). The effects of infant massage on weight, height, and mother-infant interaction. Journal of Korean Academy of Nursing, 36(8), 1331-1339.

Lihong D, Weihong L & Fentao S. (2002). Studying impact of touch on growth and development of infants. Chinese Nursing Research, 16(5).

Lorenz L, Moyse K, & Surguy H. (2005). The benefits of baby massage. Paediatric Nursing, 17(2),15-18.

Milford R & Oates J. (2009). Universal screening and early intervention for maternal mental health and attachment difficulties. Community Practitioner, 82(8), 30-33.

Moher D, Liberati A, Tetzlaff J, Altman DG & The PRISMA Group. (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med, 6(7).

NHS England. (2014). CYP- IAPT principles in child and adolescent mental health services values and standards.  Delivering With and Delivering Well. CAMHS Press. See: (accessed 27 May 2019).

NHS England & Department of Health.(2015). Future in Mind; Promoting, protecting and improving our children and young people’s mental health and wellbeing. See: (accessed 27 May 2019).

National Institute for Health and Care Excellence [NICE]. (2012). PH40 Social and emotional wellbeing: early years. See: (accessed 27 May 2019).

National Institute for Health and Care Excellence [NICE]. (2016). QS128 Early years: promoting health and wellbeing in under 5s. See: (accessed 27 May 2019).

NHS England. (2016). The Five Year Forward view for mental health. See: (accessed 27 May 2019).

Oates J, Gervai J, Danis I & Tsaroucha A. (2007). Validation studies of the mothers object relations scales short form (MORS-SF). Poster presentation in 12th European Conference on Developmental Psychology. La Laguna, Tenerife 2007.

O’Higgins M, St James Roberts I & Glover V. (2008). Postnatal depression and mother and infant outcomes after infant massage. Journal of Affective Disorders, 109(1-2), 189-192.

Onozawa K, Glover V, Adams D, Modi N & Kumar RC. (2001). Infant massage improves mother–infant interaction for mothers with postnatal depression. Journal of Affective Disorders, 63(1-3), 201–207.

Pigeon-Owen, K. (2007). The benefits of baby massage. Independent Nurse, 8(1).

Pugh K, O’Herlihy A & Fonagy P. (2016). What next for CYP IAPT? Update from NHS England. February 2016. See: (accessed 27 May 2019).

Rees C. (2007). Childhood Attachment. The British Journal of General Practice, 57(544), 920-922.

Runyan D, Wattam C, Ikeda R, Hassan F & Ramiro L. (2002). Child abuse and neglect by parents and other caregivers. In E. G. Krug, L. L. Dahlberg, J. A. Mercy, A. B. Zwi, & R. Lozano (Eds.), World Report on Violence and Health (pp. 57-86). Geneva, Switzerland: World Health Organization.

Schneider McClure V. (1989). A handbook for loving parents. New York: Bantam Books.

Schneider McClure V. (2001). Infant massage: A hand book for loving parents. London: Souvenir Press Ltd.

Smith V. (2014). Cochrane Corner: Massage for promoting mental and physical health in infants under six months. The Practicing Midwife, 42-43.

South West Strategic Clinical Network & NHS England. (2015). Perinatal and infant mental health care in the South West: Improving care pathways. See: (accessed 27 May 2019).

Stern DN. (1985). The international world of the Infant: A view from psychoanalysis and developmental psychology. New York: Basic Books.

Underdown A. (2013). Infant massage as a community intervention for infants aged under 6 month. Institute for health visiting. Excellence in practice - Evidence Digest (infant massage) Nov, 1-3.

Watanabe H, Tanaka Y & Nomachi S. (2012). Effects of full-term infant massage on mother’s emotional state. Journal of Obstetric, Gynaecology and Neonatal Nursing, 41(Supplement 1), S134. 

Winnicott DW. (1987). Babies and their mothers. New York: Addison-Wesley.

York A & SW Collaborative. (2016). National Updates for the SW Collaborative. September 2016.

Zeanah C & Benoit D. (1995). Clinical applications of a parent perception interview. In K. Minde (Ed.), Infant psychiatry: Child psychiatric clinics of North America (pp.539-554). Philadelphia: Saunders.

Appendix I

Search terms strategy use as follows


  1. MH Massage+
  2. Massage N1 therap*
  3. MH Touch or MH Therapeutic Touch
  4. Therap* N1 touch*
  5. Tactile N1 stimu*
  6. 1 or 2 or 3 or 4 or 5
  7. MH Infant+
  8. Baby or babies
  9. 7 or 8
  10. MH Parent-child relations or MH Mother-Infant Relations or MH Parent-Infant relations or MH Mother-child relations or MH Parent-Infant bonding
  11. Attachment*
  12. Bond*
  13. Reciproc*
  14. Interact*
  15. Containment
  16. 10 or 11 or 12 or 13 or 14 or 15
  17. 6 and 9 and 16.

Appendix 2
Flow Diagram ( following the PRISMA model)


Appendix 3   

Data Extraction including resolution of disagreement







Outcome Measures

Data Collection Process

Data Items

Author's Judgment of Bias/Limitations

Summary of Measures



Randomised controlled

39 (18 dropped out) full term, singleton Mothers @ 6 weeks postnatal with no complications

Field's Method of Baby Massage (did not attend a course). Applied 10 mins per day until 3 months

Cortisol in saliva, Profile of Mood States

Cortisol taken between 10am and 3 pm at the start and end of the study (in hospital and at home),  Questionairres were completed at home before and after intervention. All mothers were mailed the questionairres.

 Tension, Depression, Anger, Vigour, Fatigue, Confusion,Concentration of Cortisol Levels determined by enzyme-linked immnuosorbent assay

Small sample size. Short period of intervention,  requiring participants to mail back responses, The second Cortisol samples were delayed until 6 weeks after the intervention ended and taken in different environments.

SPSS. Wilcoxon rank sum test, Student t-test.. Cortisol Levels determined by enzyme-linked immnuosorbent assay.



Non-equivalent Control Group Post-test OnlyDesign

94 (52 didn't respond) Parents usually presenting with a perceived problem (eg. Crying). 60 (140 didn't respond) Parents with no perceived problems. No selection - voluntary participation.

IAIM 5 week course. 15 courses in a variety of settings. 12 parents at each course.

Infant Social Interaction Questionairre, Programme evaluation Questionairre. Focus group interviews

Questionairres mailed to participants after intervention. Interviews conducted at focus group.

Parenting sense of competancy, Self Esteem, Breast Feeding, Reading to child, Chatting, General health of parent and child,

Study focuses on one trainer. Short period of intervention  Not fully randomised in design. requiring participants to mail back questionairres.

Descriptive Statistics. Open coding to identify common themes. .



Non-equivalent Control Group Pretest-Posttest Design

129 (13 dropped out) Mothers of Full term, healthy 2-6 month old babies, recruited at  Massage programme or Well Baby Clinic. 26 Mothers agreed to the massage programme and 103 took part as the control group.

4 week IAIM Baby Massage course attendance and written checklist. To be applied 4 x weekly at home for 4 weeks.

Mother's Perception of Infant's Temperament Scale (modified), Weight, Height, Mother-Infant Interaction (MIPIS Walker and Thompson 1982)

Questionairre - not specified how given. Video recordings of Mother-Infant interaction in the health centre, Weight -Cas electronic scales, Height - collected in triplicate and averaged. Social demographics and feeding information - questions were also collected at time of selection.

Mother's Perception of Infant's Temperament (defined as the individual's emotional reactivity and behavioural style in interacting with the environment), Mother-Infant Interactions, Height, Weight.

Not fully randomised in design.Sample size. Study length. Lack of similarity between participants. Difference in length and quality of daily massage.  Limited collection of physiological parameters. All babies lived with their families.

t-test and chi-square test



Prospective Block - Randomised Controled design

62 Mothers scoring 13+ on the EPDS at 4 weeks post partum. Control Group - 34 (56 dropped out) scoring less than 9 on the EPDS at 4 weeks post partum. Control groups were split equally between 2 interventions. Mother's attending 4 or more sessions were included.

6 Sessions of either IAIM Infant Massage or a support group

Edinburgh Post Natal Depression Scale, Spielberg State Anxiety Inventory, Infant Characteristics Questionairre, Video interactions using the Global Ratings for Mother-Infant interactions. Social economic questions.

At 4 weeks and 9 to 12 weeks Questionairres and Videos in Clinic. Repeated at 19 weeks in clinic. Social ecomomic profile questions asked at the beginning.

Depression and Anxiety State Maternal Sensitivity in Interaction, Infant Performance in Interaction, Overall Interaction, Fussiness/Difficulty Scale.

The group were well educated with no socio economic stress. Participants were predominantly white and married. No treatment for depression group was not included for ethical reasons. Initial fall in EPDS scores may be attributed to anticipated support (Appleby, et al. 1997).

ANOVA with post hoc Bonferroni Tests.



Qualitative Single Pretest-Posttest Design

4 (3 dropped out) Mothers of healthy 1-3 months old infants responding to a poster advert.

One off training session in non specific massage and parental sensitivity. Applied 5 x weekly for 15-30 mins per day. Journal completion. Contacted at midpoint to discuss progress and concerns.

Parenting Stress Index - Short Form

Pre and post questionairres. Narrative analysis was done on the journals.

Total Stress Score. Parent's ratings of Infant's temperament, parental competancy, attachment, social support and role habit disturbance. Sensitivity to infant cues and other themes (eg. Time of day).

Very small sample size. No control group. Not randomised in design. All participants were married, proffessional and had existing social support networks. There were no high levels of risk for stress or relationship difficulty. Recruitment via poster.

Visual comparison of Pre and Post stress scores.



Pre and Post Intervention Evaluation

94 (52 dropped out) Parents who had completed IAIM baby massage courses. 60 (140 dropped out) Parents who had not taken part in the massage courses

IAIM 5 week programme

Programme Evaluation Questionairre, Parenting Sense of Competence Questionairre. The self esteem Scale. Focus group Interviews.

Post interaction questionairre by mail, requiring mailing back and Focus Group Interview.

Items about the baby, Parent's Competancy Scale, Items about the parents, breastfeeding, reading, chatting. Impact of Social Interaction and Parenting.

Post intervention data only. Most of the data was from the focus group and just used one trainer. Short length of time so unknown longer term effects. Sensitivity of tools was limited. Non intervention group was not well matched to intervention group so comparisons are limited.

Qualitative data entered into SPSS and analysed using standard non parametric non statistical tests. Tapes were transcribed and analysed using Open Coding and Thematic Analysis.



Intervention research - Non-equivalent Control

40 First time mums, three months post natal who attended infant massage classes. Full term infants.

10-15 minutes massage per day over four weeks. Control group held baby for 15 mins per day

The Profile of Mood States and the Postpartum Bonding Questionnaire. Pre and post intervention.


Tension-anxiety, depression, anger-hostility, vigour, fatigue and confusion.

No author comment but we could consider limitations regarding number, shortness of the interval between outcome measures (no follow up). Unknown care to control group.




Critical Realist research design (mixed method - quantative and qualatative)

39 mothers  attending a six week infant massage programme in a group (six did not complete).  72% were white-British and 28% were other ethnic minorities.

A six week massage course (group) led by the International Association of Infant Massage (IAIM) or another infant massage facilitator

Working model of the child interview (WMCI), the Edinbrugh Post-Natal Depression scale (EPDS) and the care index.

Interviews (unknown environment). Videos. Pre and post intervention

Parental response classified into one of three broad attachment classifications: balanced, disengaged and distorted. EPDS scores for levels of depression and Care Index (CI)measuring three aspects of maternal behaviour (sensitivity, covert and overt hostility and unresponsiveness). Four aspects of infant behaviour (cooperativeness, compulsive compliance, difficultness and passivity).

Diverse range of infant practices observed between parents and infants. They had a high proporiton of white British which may impact on the contextual generalisation. Possible bias in recruitment e.g 1/4 of participants had degrees, meaning they may have been more motivated to attend having heard of the offer of high quality infant massage programmes.   High level of low risk women meaning that no significant change was likely.                                                                                                                     

Quantative data was entered directly into SPSS as well as demographic details, WMCI classifications and EPDS and CI scores. Appropriate statistical t-test were performed to compare the means for quantative data collected before and after the programme. Qualatative data was also entered into NVIVO whcih is a qualatative data analaysis computer-software packaged and analysed thematically.



Randomised Controled

25 (9 dropped out) Mothers scoring >13 on the EPDS at 4 weeks postpartum, with healthy babies. 12 took part in a massage group and 13 took part in a support group

5 weekly sessions (1 hour for massage group and 1 hour for support group).

Edingburgh Postnatal Depression Scale, Video's coded using Global Ratings for mother-infant Interactions at 2 months by Fiori-Cowley and Murray

First and Last Sessions mother's given EPDS and videoed playing with their child for 5 mins.

Depression status. Maternal contribution to interaction. Infant contribution to interaction. Quality of interaction.

Very small sample size. High drop out rate. Differing lengths of sessions. Short period of intervention.

data was analysed by non parametric methods due to sample size and Mann-Whitney U-test (two tailed) or Fisher's prbability test as appropriate



Quasi-experimental design.  Controled

117 (3 excluded) Breastfeeding singleton Mothers of healthy babies living in a specific city and with a minimum of highschool level of education.

Baby massage taught at 5-7 days by IAIM instructor, CD given and technique checked at 15 and 38 days). 15 minute massage given daily for 48 days

Implementation of Mother-Infant Identifier Poll.                              Maternal Attachment Inventory

Test between 24-48 hours after birth in hospital and on the last day of the study at home.

Maternal Affectionate Attachment

MAI conducted in different environments. All mothers well educated.

Analysis used SPSS. Percentiles and means. Chi-square test. One-way analysis of varience. Paired sample t-test. Independent paired sample t-test




Appendix 4
Bias and limitations cited by the authors

  • Intervention and Design Limitations:
  • Short period of intervention, 
  • Post intervention measures delayed, 
  • Not fully randomised in design, 
  • Only one trainer used, 
  • Ethical reasons prevented ideal control group, 
  • Small sample size, 
  • Post intervention data only, 
  • Short interval between outcome measures, 
  • Conducted in differing environments.
  • Participants were usually highly motivated and educated with low social need.

Participant Bias:

  • Motivation necessary to complete post intervention measures.
  • All babies living with both birth parents, groups were well educated with no economic stress, good social support, predominantly white and married, recruitment control group poorly matched to intervention group, 
  • Diverse range of infant practices observed between parent and infant (parenting style), predominantly low risk women, 
  • High dropout rates, 

Limitation in the use of Measures: 

  • Requirement of participants to mail back responses, 
  • Limited collection of physiological parameters, 
  • Sensitivity of tools,

Differences in the Analysis of Measures:

  • Wilcoxon rank sum test,
  • Student t-test,
  • Cortisol levels determined by enzyme linked immunosorbent assay,
  • Percentages increase,
  • t-test,
  • Chi square test,
  • ANOVA with post hoc Bonferroni tests,
  • Visual comparison of pre and post test scores,
  • SPSS
  • Analysis using standard non parametric non statistical tests,
  • Video tape transcription analysed using open coding and thematic analysis, 
  • Qualitative data entered into NVIVO and analysed thematically,
  • Mann-Whitney U test (Two tailed),
  • Fisher’s Probability test
  • Mean,
  • Paired sample t-test,
  • One way analysis of variance,