I S S U E S A N D IN N O V A T I O N S IN N U R S I N G P R A C T I C E

Effects of massage on pain and anxiety during labour: a randomized

controlled trial in Taiwan
Mei-Yueh Chang MSc RN
Lecturer, National Tainan Institute of Nursing Tainan, Taiwan

Shing-Yaw Wang MD MPH

Associate Professor, Department of Psychiatry, Kaohsiung Medical University, Kaohsiung, Taiwan

and Chung-Hey Chen PhD RN

Professor, College of Nursing, Kaohsiung Medical University, Kaohsiung, Taiwan

Submitted for publication 25 April 2001

Accepted for publication 7 January 2002

Correspondence: C H A N G M .-Y ., W A N G S .-Y . & C H E N C .-H . ( 2 0 0 2 ) Journal of Advanced Nursing

Chung-Hey Chen, 38(1), 68–73
College of Nursing, Effects of massage on pain and anxiety during labour: a randomized controlled trial
Kaohsiung Medical University,
in Taiwan
no. 100, Shih Chuan 1st Road,
Aims. To investigate the effects of massage on pain reaction and anxiety during
Kaohsiung,
Taiwan. labour.
E-mail: m735007@kmu.edu.tw Background. Labour pain is a challenging issue for nurses designing intervention
protocols. Massage is an ancient technique that has been widely employed during
labour, however, relatively little study has been undertaken examining the effects of
massage on women in labour.
Methods. A randomized controlled study was conducted between September 1999
and January 2000. Sixty primiparous women expected to have a normal childbirth
at a regional hospital in southern Taiwan were randomly assigned to either the
experimental (n ¼ 30) or the control (n ¼ 30) group. The experimental group
received massage intervention whereas the control group did not. The nurse-rated
present behavioural intensity (PBI) was used as a measure of labour pain. Anxiety
was measured with the visual analogue scale for anxiety (VASA). The intensity of
pain and anxiety between the two groups was compared in the latent phase (cervix
dilated 3–4 cm), active phase (5–7 cm) and transitional phase (8–10 cm).
Results. In both groups, there was a relatively steady increase in pain intensity and
anxiety level as labour progressed. A t-test demonstrated that the experimental
group had significantly lower pain reactions in the latent, active and transitional
phases. Anxiety levels were only significantly different between the two groups in
the latent phase. Twenty-six of the 30 (87%) experimental group subjects reported
that massage was helpful, providing pain relief and psychological support during
labour.
Conclusions. Findings suggest that massage is a cost-effective nursing intervention
that can decrease pain and anxiety during labour, and partners’ participation in
massage can positively influence the quality of women’s birth experiences.

Keywords: massage, labour pain, anxiety, partner participation, child birth

68 Ó 2002 Blackwell Science Ltd

Issues and innovations in nursing practice
Introduction
Childbirth is a stressful experience, with pain, fatigue, fear
and negative moods reaching high levels as labour pro-
gresses (Sommer 1980, Melzack et al. 1984, Melzack &
Schaffelberg 1987, Leventhal et al. 1989, Field et al. 1997).
As labour pain is acute and increases quickly, and because
considerable emotions are involved, pain relief poses a
major problem (Harrison et al. 1986). Various analgesics
can be used, but side-effects on women in labour and other
multidimensional phenomena indicate that analgesia alone
may not manage pain adequately (McCaffery & Beebe
1989, Mobily et al. 1994). Pain can prompt the sufferer to
seek comforting contact and to prevent intrusive touch.
Appropriate contact modalities can help to relieve pain and
thereby help to maintain a sense of body boundary intact-
ness (Richardson 1984). In Chinese culture, ‘reen’ (Mandarin,
meaning tolerance or patience) is a traditional virtue. A
Chinese saying often used with women in Taiwan who are
undergoing labour is ‘If you wish to be the best person, you
must suffer the bitterest of the bitter’, with the result that
pain relief is not commonly requested. In addition to
analgesics, self-comforting behaviours and nurse-initiated1
comfort measures are ways in which labour pain can be
managed.
Massage is an ancient technique that has been widely
employed during labour. It is thought to work either by
blocking pain impulses to the brain by increasing A-fibre
transmission, or by stimulating the local release of endor-
phins (McCaffery & Beebe 1989). Pressure from massage
preempts the processing of painful stimuli because pressure
fibres are longer and more myelinated, and therefore relay
signals to the brain more quickly than pain fibres (Melzack &
Wall 1965). McCaffery and Beebe (1989) summarized the
potential benefits of massage as decreasing the intensity of
pain, relieving muscle spasm, increasing physical activity,
distracting from pain, promoting general relaxation, decreas-
ing anxiety, and in some instances the character of the pain
may simply change to a sensation that is more tolerable, for
example, a change from sharp to dull pain. In addition to
this, massage can strengthen the nurse–patient relationship
and conveys caring through socially acceptable physical
contact. Massage is an ideal way to involve family and
friends who would otherwise feel helpless or who want to do
something for a woman during labour (McCaffery & Beebe
1989). It helps to convey caring, sympathy, encouragement,
1
In Taiwan, staff in labour rooms are nurses. They may, or may not,
hold a midwife license.
Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 38(1), 68–73
Effects of massage during labour
acceptance, support, participation, competence and comfort
in nursing/midwifery practice, and as it is believed to reduce
pain, constitutes an important intervention (Weaver 1900,
Birch 1986, Fraser & Kerr 1993, Malkin 1994). Unfortu-
nately, there have been few controlled studies that have
actually measured psychosocial parameters during labour.
Field et al. (1997) investigated the use of massage of the
head, shoulder, back and feet by the partner at approxi-
mately 3–5 cm cervical dilation found that mothers receiving
massage for 20 minutes reported a decrease in anxiety, pain
and agitated activity, as well as a more positive affect and
shorter labour. Furthermore, the presence of a supportive
individual may relieve symptoms of anxiety and decrease
pain levels (Henneborn & Cogan 1975, Herr & Mobily
1992).
There appears to be renewed interest in the use of touch
and massage in nursing and midwifery, although little study
has been undertaken examining the effects of massage on
women in labour. The purpose of this study was to undertake
an empirical evaluation of the effects of massage on pain and
anxiety during labour, and was the first such investigation in
Taiwan.
The study
Methods
Sample
Participants for this study were primiparous women giving
birth in a regional hospital in southern Taiwan between
September 1999 and January 2000. The following criteria
were used for the study: (1) subjects were between 37 and
42 weeks gestation, (2) normal pregnancies and childbirths to
date, (3) the partner was to be present during labour, and (4)
estimated cervical dilatation of no more than 4 cm. Women
eligible for inclusion were sequentially recruited into the
study and randomly allocated to groups by assistants using
four balls, two of which were marked E for the experimental
group, and the other two marked C for the control group.
The next allocation was concealed from the person entering
women into the trial.
Following initial assessment and allocation, 83 women
were recruited. Of the 83 subjects, 22 (massage group ¼ 12,
control group ¼ 10) underwent caesarean section for pro-
longed labour or foetal heart beat deceleration, and one
(control group) declined to participate during the data
collection period. The sample size was predetermined in a
pilot study using power analysis based on a medium size
effect, an alpha of 0Æ05 and power ¼ 0Æ80. Ultimately, a total
of 60 subjects were included in the analysis (Figure 1).
69
M.-Y. Chang et al.
Figure 1 Flow diagram of subject progress through the phases of
randomized trial.
Procedure
After allocation, recruited subjects in the two groups were
told that pain and anxiety were to be compared through the
three phases of labour. Subjects were approached by the
researcher who explained the purpose of the study and
described the data collection procedures. Informed consent
was obtained. After the recruitment preliminaries had been
carried out, subjects were asked to complete a form request-
ing basic demographic information. In accordance with
hospital policies for human experimentation, ethical approval
was obtained for the study. Permission to conduct the study
and access the subjects was obtained from the director of
OBS-GYN and the head nurse of the unit.
Couples in the experimental group were given a detailed
description of the massage protocol. Following this, the
primary researcher gave massage during uterine contractions
in each phase and taught the method to the partner. In phase
1 (latent phase, with the cervix dilated 3–4 cm) women were
encouraged to close their eyes when uterine contractions
began and to take two deep breaths in order to concentrate
on the massage. They received directional, reasonably firm
and rhythmic massage lasting 30 minutes and comprising
abdominal effleurage, sacral pressure and shoulder and back
kneading. Subjects were encouraged to select their favourite
type and site for massage, depending on what felt most useful
and comfortable at the time. The same 30-minute massage
was repeated in phase 2 (active phase, 5–7 cm cervical
dilation) and in phase 3 (transitional phase, 8–10 cm cervical
dilation). After the 30-minute massage at each stage, pain and
anxiety states were evaluated to assess the immediate effects
of the massage. The newly taught partners repeated the
massage at each phase of labour after the 30-minute massage
by the researcher.
Subjects in the control group were not aware that they had
not received massage but nevertheless received standard
70 Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 38(1), 68–73
nursing care and 30 minutes of the researcher’s attendance
and casual conversation during each phase. Control subjects
were told that their participation was complete after the
researcher had attended all three phases of labour.
Measures
A patient data sheet was used for recording demographic and
clinical data from medical records. The nurse-rated present
behavioural intensity (PBI) (Bonnel & Boureau 1985) scale
was used as a measure of present manifestations of pain and
was translated into Chinese. The PBI is a five category
behavioural observation scale used to assess present beha-
vioural manifestations of pain: 0 – normal respiration, 1 – the
frequency or amplitude of respiration changes during con-
tractions, 2 – gasping reactions that cease during contraction
relaxation, 3 – gasping that persists between contractions,
4 – signs of agitation. Concurrent validity was supported by
Bonnel and Boureau (1985) with the self-reported present
pain intensity (PPI) scale, a measure of the overall pain
intensity on a scale of 0–5: 0 represents no pain, 1 mild, 2
discomforting, 3 distressing, 4 horrible, and 5 excruciating
pain. In the present study, the inter-rater reliability of PBI
between the researcher, a clinical faculty member and a nurse
(n ¼ 3) was 100%. For three phases, concurrent validity
correlations between the PBI and PPI were 0Æ45, 0Æ50 and
0Æ44.
The visual analogue scale for anxiety (VASA) consists of a
10-cm horizontal line with the descriptors ‘no anxiety’ at the
left and ‘worst possible anxiety’ at the right. Subjects were
asked to indicate how anxious they were feeling ‘right now’
by marking the appropriate place on the line. Higher values
indicated increased levels of anxiety. The VASA is a reliable,
valid and sensitive self-reported measure for the study of
subjective patient experiences including pain, nausea, fatigue
and dyspnoea (Gift 1989).
Thirty minutes after childbirth, all subjects were asked to
indicate subjective assessments of satisfaction with the child-
birth experience, support levels from their partner (‘How
effectively did your partner help you deal with labour pain?’)
and assistance levels from nurses (‘How effectively did nurses
help you manage your labour pain?’) on 5-point scales. The
partners were also asked to give their subjective assessments
of satisfaction with the childbirth attendance experience, and
the level of support they had given to their wife (‘How
effectively did you help your wife deal with labour pain?’) on
5-point scales. Women in the experimental group were asked
to evaluate the helpfulness of massage on a 5-point scale, and
to answer an open-ended question about the advantages of
massage. Nine experts examined the content validities of all
instruments.
Issues and innovations in nursing practice Effects of massage during labour

Descriptive statistics including the mean (SD), minimum
and maximum were calculated. Two-sample t-tests were used
to test differences between the massage and control groups,
and the mean differences between the control and massage
groups for the PBI and VASA were produced with 95%
confidence intervals.
Results
Sixty primiparous women in labour were included in the
study. The experimental group consisted of 30 women.
Subjects ranged between 21 and 35 years of age with a mean
age of 28 years, and had an average weight of 65Æ22 kg,
mean gestational age of 277Æ23 days, mean infant birth
weight of 3196 g, and mean labour duration of 10Æ96 hours.
The control group consisted of 30 women between 20 and
39 years of age, with a mean age of 27Æ9 years, average
maternal weight of 67Æ07 kg, mean gestational age of
276Æ53 days, mean infant birth weight of 3106 g, and mean
labour duration of 9Æ61 hours. Demographic and obstetric
data for the experimental and control groups were compared.
No significant differences (P > 0Æ05) were found between
the two groups in terms of age, maternal weight, gestational
age, infant birth weight and mean duration of labour
(Table 1).
Pain and anxiety intensity scores reached a peak during the
third phase of labour (Table 2). Mean scores, SD, mean
differences and assessments of pain and anxiety intensity
scores during the three phases for the two groups are shown
in Table 2. PBI and VASA levels showed a relatively steady
increase in both groups as labour progressed. Trend analysis
showed that these were significant, linear trends. Significantly
lower PBI scores in the massage group (0Æ73, 1Æ73 and 2Æ17)
were observed than among the controls (1Æ30, 2Æ17 and 2Æ87)
at phase 1, 2 and 3 of labour, respectively. Differences
between the VASA means were also compared, with only
phase 1 showing a significant difference (massage group
37Æ20, control group 53Æ47).
A sense of satisfaction from massage was noted by
participants, as was the value of massage to the couple’s
experience of labour. From Table 2, two significant differ-
ences were found in the assessments by women of ‘How
effectively did your partner help you deal with labour pain?’
(d ¼ 0Æ57, 95% CI of d ¼ 0Æ09–1Æ04, P ¼ 0Æ019), and partners’
assessments of ‘How effectively did you help your wife deal
with labour pain?’ (d ¼ 0Æ70, 95% CI of d ¼ 0Æ30–1Æ10,
P ¼ 0Æ001). The experimental group returned more positive
evaluations, and 87% of the experimental group reported
that massage was of more than moderate helpfulness during
labour. On the other hand, subject satisfaction with the
childbirth experience, their assessment of nursing assistance
levels and partner satisfaction with the childbirth attendance
experience were not significantly different. More than 90%
(55 of 60) of participants rated the nursing assistance as ‘very
helpful’ and ‘extremely helpful’ in both groups. In total, 73%
(44 of 60) of women and 82% (49 of 60) of partners felt
‘very’ and ‘extremely’ satisfied with the childbirth experience.
Discussion
Although massage is believed to be a simple, effective and
safe method of support and relief for women during child-
birth, few scientifically rigorous assessments exist to substan-
tiate this belief. Chen and Chang (2000) commented that in
the high-tech medical environment the use of touch therapy
such as effleurage, stroking, sacral pressure and patting or
handholding could help women to remain a sense of body
boundary intactness and facilitate their ability to cope with
labour. This study sheds some light on the effects of massage
on women in childbirth. Data from different phases (latent,
active and transition) all suggested that massage reduced pain
behaviour observations during the three phases and also
helped to reduce anxiety levels during the latent phase.
Behavioural observational measurement demonstrated less
agitation following massage at all three phases. Field et al.
(1997) found significant effects only from massage at 3–5 cm
cervical dilation. This is the first controlled study to provide
evidence that massage provides psychosocial support for
women during all three phases of labour. Appropriate touch
at appropriate times may help the woman to feel in control of
her body and maintain a sense of body boundary integrity. In
keeping with earlier work (Chen et al. 2001), comforting
touch was found to be significant and important. This finding
differed from that in a study in Hong Kong (Holroyd et al.

Table 1 Comparison between intervention

Characteristics
and control groups of demographic and
obstetric characteristics. Figures quoted are Age of mother (year)
mean (SD), minimum to maximum Maternal weight (kg)
Gestation age (day)
Newborn weight (g)
Duration of labour (hour)
Experimental group (n ¼ 30)
28 (3Æ74), 21–35
65Æ22 (6Æ83), 54–80
277Æ23 (8Æ21), 263–289
3196Æ67 (356Æ15), 2660–3998
10Æ96 (4Æ81), 4Æ45–28Æ08
Control group (n ¼ 30)
27Æ9 (3Æ85), 20–39
67Æ07 (8Æ19), 54–85
276Æ53 (8Æ54), 259–292
3106 (79Æ77), 2340–4120
9Æ61 (4Æ24), 4Æ40–20Æ97

Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 38(1), 68–73 71

M.-Y. Chang et al.

Table 2 Mean (SD), estimated difference (95% CI) and P-values for pain and anxiety intensity scores during the three phases of labour in two
groups

Massage (n ¼ 30) Control (n ¼ 30) Massage–control (95% CI) P*

PBI
Phase 1 0Æ73 (0Æ52) 1Æ30 (0Æ53) ÿ0Æ57 (ÿ 0Æ84, ÿ0Æ29) 0Æ000
Phase 2 1Æ73 (0Æ45) 2Æ17 (0Æ59) ÿ0Æ43 (ÿ 0Æ71, ÿ0Æ16) 0Æ002
Phase 3 2Æ17 (0Æ53) 2Æ87 (0Æ78) ÿ0Æ70 (ÿ 1Æ04, ÿ0Æ36) 0Æ000

VASA
Phase 1 37Æ20 (20Æ30) 53Æ47 (22Æ18) ÿ16Æ27 (ÿ 27Æ25, ÿ5Æ28) 0Æ040
Phase 2 64Æ93 (24Æ07) 73Æ87 (22Æ64) ÿ8Æ93 (ÿ 21Æ01, 3Æ14) 0Æ144
Phase 3 80Æ67 (19Æ11) 85Æ17 (18Æ29) ÿ4Æ50 (ÿ 14Æ17, 5Æ17) 0Æ355

Subjective assessment by women

Childbirth experience 4Æ17 (1Æ05) 3Æ70 (1Æ32) 0Æ47 (ÿ 0Æ15, 1Æ08) 0Æ135
Partner’s support level 4Æ30 (0Æ79) 3Æ73 (1Æ01) 0Æ57 (0Æ09, 1Æ04) 0Æ019
Nurse’s assistance level 4Æ53 (0Æ68) 4Æ53 (0Æ63) 0Æ00 (ÿ0Æ34, 0Æ34) 1Æ000

Subjective assessment by partners

Attendance experience 4Æ53 (0Æ73) 4Æ20 (0Æ81) 0Æ33 (ÿ0Æ64, 0Æ73) 0Æ098
Effectiveness of assistance 3Æ73 (0Æ83) 3Æ03 (0Æ72) 0Æ70 (0Æ30, 1Æ10) 0Æ001

*Two sample t-test.

PBI ¼ present behavioural intensity; VASA ¼ visual analogue scale for anxiety.

1997), which reported that Chinese people use distance to
regulate their privacy and level of intimacy in encounters. The
results of our study suggest that, although cultural, ethnic,
and religious differences exist between Eastern and Western
societies, the perception of nursing behaviours by Taiwanese
women is similar to that of Western women.
The Gate Control Theory proposes a gating mechanism at
the spinal cord and may explain the effects of massage seen
in this study. The gate is thought to open when increased
activity from small diameter nerve fibres transmit pain
impulses to the spinal cord and brain. Massage or pressure
activity that activates large diameter fibres can close the
gate, thereby inhibiting the transmission of pain (Melzack &
Wall 1965). Another possibility is that massage may
increase the level of endorphins (endogenous opioids)
thereby reducing pain (McCaffery & Beebe 1989). Based
on the subjective assessments of women in this study, the
effects of massage can be categorized as pain relieving
(‘Although pain was still there, massage promoted my
comfort’) and psychologically supportive (‘Massage made
me feel that someone was sharing my suffering’). The tactile
stimulation may also increase vagal activity, which in turn
lowers physiological arousal and stress hormone (cortisol)
levels. Additionally, feelings associated with a caring and
empathic forms of massage may affect higher brain centres,
further influencing perceptions of pain (Melzack & Wall
1965, Field 1998).
Lower levels of self-reported pain and anxiety were
described in Field’s study of massage when women were at
approximately 3–5 cm cervical dilation (Field et al. 1997).
The physical and psychological influences of massage on
couples’ experiences of labour were very positive (Kimber
1998). Of the nonpharmacological methods, massage is an
effective, noninvasive technique for promoting relaxation and
improving communication with women in labour. As mas-
sage requires a supportive personal presence, it is possible
that the favourable effects are, in part, because of continuous
emotional support (McCaffery & Beebe 1989, Anonymous
1998). In order to find an effect over and above the presence
of a supportive person, the control group was to some extent
placebo controlled, in that the person doing and teaching
massage was present and made conversation with control
couples but did not actively intervene.
Conclusion
Caregivers’ physical touch influenced the woman’s reaction
to pain, made her feel safer and calmer, and improved her
well-being during labour. Consequently, massage has the
potential to improve the relationship between nurses/mid-
wives and women in labour, as well as between the couple.
Touch and massage can convey concern, security, closeness
and encouragement, and at the same time serve as a
psychosocial intervention.

72 Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 38(1), 68–73

Issues and innovations in nursing practice
Acknowledgements
We wish to thank the women participating in the trial, and
the delivery suite nursing staff of the Foundation Medical
Center, Chi Mei.
References
Anonymous (1998) Midwifery management of pain in labour.
Journal of Nurse-Midwifery 43, 77–82.
Birch E.R. (1986) The experience of touch received during labour.
Journal of Nurse-Midwifery 31, 270–276.
Bonnel A.M. & Boureau F. (1985) Labour pain assessment: validity
of a behavioral index. Pain 22, 81–90.
Chen C.H. & Chang M.Y. (2000) The application of touch therapy
during labour: an old method newly applied. Journal of Nursing
47, 28–32.
Chen C.H., Wang S.Y. & Chang M.Y. (2001) Women’s perceptions
of helpful and unhelpful nursing behaviors during labour: a study
in Taiwan. Birth 28, 180–185.
Field T.M. (1998) Massage therapy effects. American Psychologist
53, 1270–1281.
Field T., Hernandez-Reif M., Taylor S., Quintino O. & Burman I.
(1997) Labour pain is reduced by massage therapy. Journal of
Psychosomatic Obstetrics and Gynecology 18, 286–291.
Fraser J. & Kerr J.R. (1993) Psychophysiological effects of back
massage on elderly institutionalized patients. Journal of Advanced
Nursing 18, 238–245.
Gift A.G. (1989) Visual analogue scales: measurement of subjective
phenomena. Nursing Research 38, 286–288.
Harrison R.F., Woods T., Shore M., Mathews G. & Unwin A. (1986)
Pain relief in labour using transcutaneous electrical nerve stimula-
tion (TENS). A TENS/TENS placebo controlled study in two parity
groups. British Journal of Obstetrics & Gynaecology 93, 739–746.
Henneborn W.J. & Cogan R. (1975) The effect of husband
participation on reported pain and probability of medication
Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 38(1), 68–73
Effects of massage during labour
during labour and birth. Journal of Psychosomatic Research 19,
215–222.
Herr K.A. & Mobily P.R. (1992) Interventions related to pain.
Nursing Clinics of North America 27, 347–369.
Holroyd E., Yin-King L., Yin-King L., Pui-Yuk L.W., Kwok-Hong
F.Y. & Shuk-Lin B.L. (1997) Hong Kong Chinese women’s
perception of support from midwives during labour. Midwifery
13, 66–72.
Kimber L. (1998) Effective techniques for massage in labour.
Practising Midwife 1, 36–39.
Leventhal E.A., Leventhal H., Shacham S. & Easterling D.V. (1989)
Active coping reduces reports of pain from childbirth. Journal of
Consulting and Clinical Psychology 57, 365–371.
Malkin K. (1994) Use of massage in clinical practice. British Journal
of Nursing 3, 292–294.
McCaffery M. & Beebe A. (1989) Pain: Clinical Manual for Nursing
Practice. C. V. Mosby, St Louis.
Melzack R., Kinch R., Dobkin P., Lebrun M. & Taenzer P. (1984)
Severity of labour pain: influence of physical as well as psy-
chologic variables. Canadian Medical Association Journal 130,
579–584.
Melzack R. & Schaffelberg D. (1987) Low-back pain during labour.
American Journal of Obstetrics and Gynecology 156, 901–905.
Melzack R. & Wall P.D. (1965) Pain mechanisms: a new theory.
Science 150, 971–979.
Mobily P.R., Herr K.A. & Nicholson A.C. (1994) Validation of
cutaneous stimulation interventions for pain management. Inter-
national Journal of Nursing Studies 31, 533–544.
Richardson P. (1984) The body boundary experience of women in
labour: a framework for care. Maternal-Child Nursing Journal 13,
91–101.
Sommer P. (1980) Can touch make a difference in nursing care?
Texas Nursing 54, 30.
Weaver D.F. (1900) Nurses’ views on meaning of touch on obstet-
rical nursing practice. Journal of Obstetric Gynecologic Neonatal
Nursing 19, 157–161.
73