April 24, 2005
The Benefits of Infant Massage: a Critical Review
EVIDENCE BASED PRACTICE
Baby massage classes are increasingly popular and many health visitors are undertaking training to run them. This paper looks at the evidence base for baby massage. A systematic review of the research-based evidence identified four articles that met the inclusion criteria i.e. they were research-based articles looking at the effects of hands-on massage of babies and infants. The findings are synthesised and the research is critically reviewed. Insufficient evidence is found for the continued massage of pre- term and low birthweight infants in neonatal intensive care units. Benefits are claimed for the use of massage in specific medical conditions such as asthma and dermatitis, but the review reporting on these benefits was methodologically flawed. A randomised controlled trial found no significant difference between colicky babies that were massaged and a control group. By far the most compelling evidence relates to the benefits of baby massage on the maternal-child relationship and postnatal depression. This evidence would thoroughly justify the provision of infant massage classes in the community by the health visitor.
Key words: Massage, infancy, childhood, health visitors, evidence- based practice, maternal-child relationship
Community practitioner 2005; 78, 3: 98-102
Infant massage is one area of health visiting that always attracts enthusiastic interest and participation from the public, an interest that seems to be primarily derived from media presentation and word-of-mouth recommendation. Given the current demand for evidence-based practice, it seemed timely to conduct a critical review of the current evidence base supporting the teaching of massage techniques to parents.
An electronic search was carried out on specialist databases (MEDLINE Plus, AMED, British Nursing Index, CINAHL, EMBASE, the Cochrane Database of Systematic Reviews, Psych Info, Clinical Evidence and NeLH) looking for articles that met the following inclusion criteria:
1. They considered the effects of baby massage (on either the parent or on the baby).
2. They were research-based.
3. They were published in peer-reviewed journals. (Many of the articles identified came from parenting journals or from alternative magazines which were not peer-reviewed and therefore had to be excluded).
4. They were published in the last 10 years, a period chosen because earlier research in the field focuses almost entirely on the effect of massage on the premature hospitalised baby.
The author also contacted a number of experts in the field - Suzanne Adamson, Vivette Glover, Sylvie Htu - and their comments are included where relevant.
Seventy-six references were gathered in total. As two of the studies identified were systematic reviews,1-2 the papers included in these reviews were not looked at separately.
Of the 76 papers identified, four met the inclusion criteria:
* A systematic review1 of articles on massage therapy and therapeutic touch in children.
* A systematic review2 of articles on massage of the pre-term or low birthweight infant from the Cochrane Database of Systematic Reviews.
* A recent study3 looking at infant massage in the treatment of colic.
* A randomised controlled trial4 to determine whether massage classes could reduce maternal depression and improve mother-infant interaction.
The findings of the four studies are summarised in Table 1.
As the Cochrane Review2 provided a comprehensive analysis of the benefits of massage for pre-term and low birthweight infants in the hospital setting, a field which is of only marginal interest to health visitors, it was not thought necessary to include subsequent studies of the same subject.
Critical review of the findings
Pre-term/low birthweight babies
The vast majority of research on baby massage relates to the pre- term and low birthweight baby. This may be because touch is considered particularly significant to a premature baby: as Vimala McClure describes,5 'the premature baby's first contact with human touch may bring pain - needles, probes, tubes, rough handling, bright lights'. (p158) It may be because the cost implications of looking after such children make government funding easier to obtain, or it may be because research breeds research.
Both of the literature reviews1-2 looked at the research relating to the pre-term infant. As a Cochrane systematic review, Vickers' article2 is a rigorous analysis which critically assesses randomisation, blinding, outcome assessment, sample size and heterogeneity.
Blinding means keeping group assignment (for example, to treatment or control) secret from the study participants or investigators or both. Blinding is used to protect against the possibility that knowledge of allocation may influence patient response.
Of the seven studies of pre-term babies identified by Ireland and Olson,1 only two (Field et al 1986(6) and Scafidi et al7] meet the high standards of Cochrane. Three are excluded from the Cochrane Review2 altogether:
* Solkoff and Matuszak 1975(8) (no details given as to method of treatment allocation)
Table 1: Summary of findings
* Rausch9 (historical control group)
* Morrow et al10 (did not meet specified selection criteria).
The remaining two articles (Solkoff 1969(11) and White and Labarba12) are included but are nonetheless criticised for their 'unclear blinding' of randomisation and 'inadequate' blinding of the intervention.2 (p34, 36)
Furthermore, the Cochrane Review2 included 21 studies not identified by Ireland and Olson.1 The authors state that they were keen to include as many trials as possible; they may therefore have trawled more widely, tracked more citations, included unpublished studies and corresponded with experts to elicit the maximum number of studies for inclusion. They also state that they included studies on infants who only just met criteria for low birthweights or prematurity.
It is perhaps not surprising that the two studies reach different conclusions on the benefits of baby massage. Ireland and Olson1 decide that 'there is sufficient evidence to recommend the use of massage therapy in preterm neonates'. (p62)
The benefits identified include rapid weight gain, increased levels of activity, more rapid habituation and fewer days hospitalised. However, these claims are poorly justified in the text.
The authors list the limitations of the studies (no baseline measures, absent statistical procedures, small sample size and lack of control for confounding variables) but no attempt is made to discriminate between them. It is therefore very difficult to make any judgement on the validity of the various stated results.
Vickers2 looks much more critically and in much greater detail at the studies. The finding that pre-term babies gained more weight was found to be of low clinical significance with statistically significant heterogeneity between the trials.
The finding that hospital stay decreased in length was strongly influenced by one study with a small sample size.
Statistical heterogeneity is said to occur when the results of individual studies are to some extent incompatible with one another.
The results regarding improvement on the Brazleton scales for habituation, motor maturity, and range of state were also found to be methodologically flawed and of low statistical significance.
Vickers2 concedes that for medically stable infants, infant massage carries a low risk of adverse effects. However, his overall conclusion is that 'there is insufficient evidence of effectiveness to warrant wider use of pre-term infant massage', (p10) a conclusion which is in direct contradiction to the Ireland and Olson study and which, given the rigour of the review, must be seen to occupy a high position in the hierarchy of evidence related to the issue.
The sick child
A second area of research on the benefits of baby massage has been to look at children with specific health conditions. Ireland and Olson1 included in their review seven such studies, which they examined together as a body of work.
They reported consistent findings of lower anxiety, improved mood, and lower stress hormones (salivary cortisol), findings which may explain some of the specific benefits related to particular conditions such as:
* increased co-operation and sleep in emotionally disturbed children13
* improved relaxation in children with post-traumatic stress14
* reduced pain in children with juvenile rheumatoid arthritis15
* reduced distraction to noise, increased classroom attention and improved relating to teachers in children with autism16
* better clinical response in children with atopic dermatitis17
* increase in peak air flow in children with cystic fibrosis.18
It is noteworthy that all these studies were carried out at the same site, the Touch Research Institute at the University of Miami, directed by Tiffany Field, which is dedicated to studying the effects of massage. Although this means that the Institute is partisan, it will also be aware of the particular requirements of researching this area. It is encouraging that, in his critical review, Vickers2 did not find fault with the two studies6-7 produced by this Institute.
Ireland and Olson1 are critical of several findings in the studies relating to sick children, in particular that no reliability or validity data are provided for measuring mood, relaxation, emotional problems or pain. However, the main problem for the re\ader of the paper is the methodological vagueness of the review itself.
Ideally, a systematic review should synthesise the findings of different studies to reach reliable, accurate and consistent conclusions.19 In this case, despite being written by an assistant professor and published in a peer-reviewed journal, the review is of dubious quality and rigour and its conclusions are therefore questionable.
The colicky child
The study by Huhtala et al3 looked at the effect of infant massage on the colicky symptoms of 58 healthy infants of six weeks or less. The randomised controlled trial was meticulously conducted but found no significant difference between those babies that were massaged and the control group which used a crib vibrator.
The effectiveness was measured using a structured 'cry diary' and a parental interview. Both groups showed a similar decrease in colicky crying and the authors concluded that the results reflect more the natural course of early infant crying and colic than a specific effect of the interventions.
Postnatal depression and mother-infant interaction
The effect of baby massage on postnatally depressed mothers and their babies was the subject of one study20 which found that during massage babies appeared less stressed and showed decreased levels of crying and salivary cortisol. After massage their sleep improved and less time was spent in 'active alert states'. Compared to controls they gained more weight and had improved temperament and lower salivary cortisol levels.
More recently, Onozawa et al4 conducted a randomised controlled trial which allocated 34 mothers who scored >12 on the Edinburgh Postnatal Depression Scale (EPDS)21 to an infant massage group and postnatal support group or to a support group only (control group).
Changes in maternal depression and maternal-infant interaction were compared using EPDS scores and a five-minute video recording of face-to-face play. The video recordings were assessed using the global ratings for mother-infant interactions at two months by Fiori- Cowley and Murray (Onozawa et al4) and a random selection of recordings was similarly rated by an experienced independent rater who was blind to the study.
The study found there was a greater improvement in depression scores in the massage group than in the control group and that mother-infant interaction improved significantly in the massage group. In every dimension measured the massage group showed an improvement whereas the control group remained the same. The authors conclude that attending an infant massage class substantially facilitated mother-infant interaction for women with postnatal depression.
The research was well conducted, with no difference in baseline depression scores or other demographic variables between the two groups. The outcome measures were clear and appropriate and the assessment was independently validated. The major problems with the study were the small size and the high dropout rate.
Eight-hundred-and-thirty women were screened initially using the EPDS. Of the 581 who completed the questionnaire 91 scored >12 on the EPDS. Of these, 59 agreed to take part, 34 started the class and nine dropped out.
The main reason for dropping out appeared to be the time of the class. Thus, only 25 women completed all the sessions, 12 in the massage group and 13 in the control group. Of these, two mothers in the massage group and one in the control group could not complete the video recordings because their infants were not settled. The researchers have recently received funding to repeat the study on a larger scale and to follow up mothers and babies for a year.
If the results of the initial study are confirmed, it would be of considerable clinical significance as mothers with postnatal depression are known to have impaired interactions with their infants22-23 and it is now well established that children of mothers who are depressed in their early months have more behavioural and cognitive problems later.24
Implications for practice
In his systematic review Vickers2 expresses his exasperation that 'the nursing literature abounds in unsystematic reviews, nonrandomised and uncontrolled trials and even general discussion papers cited as evidence in support of massage'. It is unfortunate that one of the literature reviews considered in this study1 did not appear to tackle the quality of the trials or their scientific basis in any systematic way, thus diminishing its credibility.
However, the fact that a study is not randomised or is inadequately blinded does not invalidate it completely. In the hierarchy of evidence to support the teaching of baby massage, a systematic review such as Vickers,2 must take a prominent position, but other evidence should also be considered. Such evidence would include well-designed trials without randomisation (including some of the 72 studies excluded by Vickers2 for various reasons), the experience of parents as stated in informal evaluations and questionnaires, and the testimony of respected authorities in the field.
Furthermore, there is no suggestion that baby massage is ineffective, dangerous, expensive, or in any way damaging. As Vickers2 concedes, it is non-invasive, does not require specialist equipment and can be implemented without undue disruption to routine care procedures. If taught to and implemented by the parents, the cost implications are small.
On the basis of this review, the strongest evidence by far supports the use of baby massage to enhance the maternal-child relationship and reduce postnatal depression.
The evidence is compelling, the research basis is sound and the findings are supported by more than one study.
John Bowlby25 found two conditions to enhance the development of bonding: the ability of the caregiver to be sensitive in understanding and responsive to the infant's cues and the amount and nature of the interactions between them. It is not surprising therefore if massage enhances bonding and attachment. Setting up infant massage classes in the community to enhance bonding could definitely be advocated on the basis of the evidence reviewed here and the study by Onozawa et al4 suggests that this will have a significant effect on postnatal depression.
With regard to the pre-term baby, the evidence-base for infant massage is weak, with the possible exception of length of stay outcomes. However, the cost implications of reducing the length of stay are of such significance that a randomised controlled trial of a high standard looking at the use of massage in a UK hospital may be justified.
On the basis of this review it would be difficult to report with any confidence on the benefits of massage in the management of particular medical conditions. The practitioner would be advised to view the conclusions of Ireland and Olson1 with a critical eye and refer back to the original research before making any changes to practice.
The benefits of baby massage on the normal healthy child should not be overlooked. Nearly 50 years ago Harlow26 demonstrated with infant monkeys that given the choice between an artificial mother made of wire offering liquid nourishment and a soft cuddly artificial mother with no nourishment the infant monkeys chose the latter, demonstrating that contact and comfort appear to be more important than nursing comfort. For a resume of the many animal studies demonstrating the power of touch, the interested reader is directed to Elaine Schneider's fascinating article on the subject.27
Conversely, the evidence from orphanages at the beginning of the 20th century, or more recently in Romania, suggests that touch deprivation can result in profound neurological damage or even death in an otherwise healthy baby.28
The benefits of baby massage on the healthy baby is a little researched area, and one which would be particularly relevant to the work of the health visitor with its emphasis on health promotion rather than treatment of illness.
This systematic literature review found that a number of benefits have been claimed for infant massage. Main claims include:
* rapid weight gain in pre-term infants
* fewer days hospitalised for pre-term infants
* specific benefits in specific medical conditions
* improvement in postnatal depression scores
* improvement in mother-infant interaction.
There is no evidence that baby massage is dangerous or damaging in any way. The most compelling research relates to the maternal- child bond and it would thoroughly justify the setting up of classes teaching baby massage techniques to parents.
The benefits of baby massage on the healthy baby is a little researched area, and one which would be particularly relevant to the work of the health visitor with its emphasis on health promotion rather than treatment of illness
On the basis of this review, the strongest evidence by far supports the use of baby massage to enhance the maternal-child relationship and reduce postnatal depression. The evidence is compelling, the research basis is sound and the findings are supported by more than one study
1 Ireland M, Olson M. Massage therapy and therapeutic touch in children: state of the science. Alternative Therapies in Health and Medicine 2000; 6, 5: 54-63.
2 Vickers A, Ohlsson A, Lacy JB, Horsley A. Massage for promoting growth and development of pre-term and/or low birth-weight infants (Cochrane Review). The Cochrane Library, Issue 4, 2002, Oxford: Update software, 2002.
3 Huhtala V, Lehtonen L, Heinonen R, Korvenranta U. Infant massage compared with crib vibrator in the treatment of colicky infants. Pediatrics 2000; 105, 6.
4 Onozawa K, Glover V, Adams D, Modi N, Channi Kumar R. Infant massage improves mother-infant interaction for mothers with postnatal depression. Journal of Affective Disorders 2001; 63, 201 - 207.
5 McClure V. Infant massage: a handbook for loving parents. London: Souvenir Press, 2001.
6 Field T, Schanberg SM, Scafidi FA et a\l. Tactile/kinaesthetic stimulation effects on preterm neonates. Pediatrics 1986; 77, 654- 658.
7 Scafidi FA, Field TM, Schanberg SM et al. Massage stimulates growth in preterm infants: a replication. Infant behaviour and Development 1991; 13, 167-188.
8 Solkoff B, Matuszak D. Tactile stimulation and behavioural development among low birthweight infants. Child Psychiatry and Human Development 1975; 1, 33-37.
9 Rausch PB. Effects of tactile and kinaesthetic stimulation on preterm infants. Journal of Obstetric, Gynaecologic and Neonatal Nursing 1981; 10,34-37.
10 Morrow CJ, Field TM, Scafidi FA et al. Differential effects of massage and heelstick procedures on transcutaneous oxygen tension in preterm neonates. Infant behaviour and Development 1991; 14, 397- 414.
11 Solkoff N, Yaffe S, Weintraub D, Blase B. Effects of handling on the subsequent developments of premature infants. Developmental Psychology 1969; 1, 765-768.
12 White JL, Labarba RC. The effects of tactile and kinaesthetic stimulation on premature infants. Developmental Psychobiology 1976; 9, 569-577.
13 Field TM, Kilmer T, Hernandez-Reif M. Pre-school children's sleep and wake behaviour; effects of massage therapy. Early Child Development and Care 1996; 120, 39-44.
14 Field TM, Seligman S, Scafidi FA. Alleviating posttraumatic stress in children following Hurricane Andrew. Journal of Applied Developmental Psychology 1996; 17, 37-50.
15 Field TM, Hernandez-Reif M, Seligman S. Juvenile rheumatoid arthritis: benefits from massage therapy. Journal of Pediatric Psychology 1997; 5, 607-617.
16 Field TM, Lasko D, Mundy P et al. Brief report: autistic children's attentiveness and responsivity improve after touch therapy. Journal of Autism Developmental Disorder 1997; 27, 333- 338.
17 Schachner L, Field TM, Hernandez-Reif M. Atopic dermatitis symptoms decreased in children following massage therapy. Pediatric Dermatology 1998; 15, 390-395.
18 Hernandez-Reif M, Field TM, Krasnegor J. Children with cystic fibrosis benefit from massage therapy. Journal of Pediatric Psychology 1999; 24, 2: 175-181.
19 Cook D, Mulrow C, Haynes R. Systematic reviews: synthesis of best evidence for clinical decisions. Annals of Internal Medicine 1997; 126, 378.
20 Field TM, Grizzle N, Scafidi FA. Massage therapy for infants of depressed mother. Infant Behavioural Development 1996; 19, 107- 112.
21 Cox JL, Holden JM, Sagovsky R. The detection of postnatal depression: development of the Edinburgh Postnatal Depression Scale. British journal of Psychiatry 1987; 150, 782-786.
22 Murray L. The impact of postnatal depression and infant development. Journal of Child Psychology and Psychiatry 1992; 33, 543-61.
23 Murray L, Cooper PJ. The role of infant and maternal factors in postpartum depression, mother-infant interactions, and infant outcome. In: Murray L, Cooper PJ. Postpartum depression and child development. London: Guilford Press, 1997.
24 Murray L, Fiori-Cowley A, Hooper R, Cooper P. The impact of postnatal depression and associated adversity on early mother- infant interactions and later infant outcome. Child Development 1996; 67, 2512-2526.
25 Bowlby J. Attachment and loss. New York: Basic Books, 1969.
26 Harlow HF. The nature of love. American Psychologist 1958; 13, 673-685.
27 Schneider E. The power of touch: massage for infants. Infants and Young Children 1996; 8, 3: 40-55.
28 Carlson M. Understanding the mother's touch. On the Brain 1997/ 8; 7, 1.
South Hams and West Devon PCT
Copyright TG Scott & Son Ltd. Mar 2005