LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0

J Perinat Neonat Nurs

Vol. 17, No. 4, pp. 304–312


c 2003 Lippincott Williams & Wilkins, Inc.

Management of Pain From

Heel Stick in Neonates
An Analysis of Research Conducted
in Thailand
Tassanee Prasopkittikun, PhD, RN;
Fongcum Tilokskulchai, PhD, RN

The heel stick procedure is the most common painful procedure performed in preterm and full-
term neonates. Various nonpharmacologic interventions have been used for pain relief. However,
the magnitude of the effect of different interventions has received little attention. In this study, 4
eligible studies conducted in Thailand, focusing on the effects of interventions on pain responses
to heel stick procedure in neonates, were obtained for analysis. Swaddling in full-term newborns
was found to have the largest mean effect size (dmn = 0.79). However, the moderate-to-large effect
sizes (dmn = 0.5–0.75) of positioning in preterm newborns tended to exist throughout the poststick
period while the effect sizes of other interventions decreased over time. The effect sizes of these
interventions for physiological responses varied. Key words: heel stick, neonates, pain, research
analysis

BACKGROUND infants.3 Although the heel stick procedure is

A number of research studies demonstrated
that newborn infants have anatomic and func-
tional capabilities for pain response by 20
weeks’ gestational age.1 However, the pain as-
sociated with diagnostic and therapeutic pro-
cedures in preterm and full-term infants has
been largely ignored because of 3 major rea-
sons: (1) the limited means of valid assess-
ment, (2) the fear of using opiates to prevent
pain, and (3) the disregard of possible nega-
tive sequelae to repeated painful events.2
The heel stick procedure is the most
common procedure performed in newborn
From the Department of Pediatric Nursing, Faculty
of Nursing (Siriraj), Mahidol University, Bangkok,
Thailand.
Corresponding author: Tassanee Prasopkittikun, PhD,
RN, Faculty of Nursing (Siriraj), Mahidol Univer-
sity, Bangkok 10700, Thailand (e-mail: tassanee mu@
yahoo.com).
Submitted for publication: June 3, 2003
Accepted for publication: August 25, 2003
304
short in duration, it can affect behavioral and
physiological responses, such as facial expres-
sion, heart rate, and oxygen saturation, of the
infants.4–8 Untreated pain has adverse physi-
ological consequences including blood pres-
sure and glucose alterations.9,10 Short-term
effects may include feeding problems, parent-
infant interaction dysfunction, and interrup-
tion of sleep-wake cycles. Possible long-term
effects of repeated heel sticks include im-
pairments of neurobehavioral development,
learning, and memory.8,9
Nonpharmacological pain management
strategies, such as swaddling,11,12 position-
ing,13 holding and rocking,14 nonnutritive
sucking,14,15 breast-feeding,16 and oral glu-
cose/sucrose administration,17,18 have been
used to reduce pain from heel stick in
preterm and full-term neonates. For example,
Fearon and others12 examined responses
of 15 preterm infants to swaddling after a
heel lance. They found that infants 31 weeks
postconceptional age or older exhibited
protracted behavioral disturbance after the
blood was drawn. However, this behavioral
LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0
Management of Pain From Heel Stick in Neonates
display was significantly reduced by the use of
swaddling. Corff and her colleagues13 found
that preterm neonates who were arranged in
a facilitated tucking (a side-lying or supine
position with flexed arms and legs close
to the infant’s trunk) demonstrated a lower
mean heart rate, a shorter mean crying time,
a shorter mean sleep disruption time, and
fewer sleep-state changes after the heel stick
compared to those who were the controls.
Holding and rocking also was found to be
an effective method of reducing pain-elicited
distress. Campos14 examined the effects of 2
comfort interventions, holding and rocking
and pacifiers, compared with routine care
administered to reduce stress of pain from
heel stick among 60 neonates. Even though
pacifiers had the strongest and most consis-
tent comforting effects, maternal holding and
rocking (in a rocking chair) also proved to be
an effective nonpharmacological intervention
for reducing crying and lowering heart rate
levels.
In Thailand, various methods for relieving
pain from heel sticks in preterm and full-term
neonates have been studied. Kacome5 used
an experimental design in which each baby
served as his or her own control to study the
effects of swaddling on pain response to heel
stick in 18 full-term newborns. Video record-
ing was performed for 3 minutes as baseline
data. Then the infant was swaddled. The pro-
cedure used for swaddling was similar to the
one that was studied by Campos.11 A strip of
cotton cloth was first wrapped upward from
the infant’s feet to waist. Next, a large cot-
ton blanket was pulled snugly over the in-
fant’s extended right arm and tucked under
the left side of the body. Finally, the opposite
edge was pulled over the left arm and tucked
under the body. Then a heel stick was imme-
diately performed and video recording con-
tinued for 4 minutes poststick. The infant re-
mained being swaddled until the end of the
video recording. The Neonatal Facial Coding
System (NFCS)19 was used to assess behav-
ioral and physiological responses. The results
revealed that swaddling significantly reduced
pain scores and the heart rate response to
305
pain of newborn infants at 1, 2, 3, and 4
minutes after heel stick. Using the same ex-
perimental design, Tantapong6 studied the ef-
fects of swaddling on pain responses related
to heel stick in 30 preterm newborns from
32 to 35 weeks of gestational age. The in-
fants were swaddled and then video record-
ing was performed for 3 minutes as baseline
data. Only 1 large piece of cotton blanket
(80 × 60 cm) was used for swaddling in this
study. The method of swaddling with a large
cotton blanket was similar to what was done
in the study by Kacome.5 After a heel stick
was done, video recording continued for 5
minutes. The physiological responses (heart
rate and oxygen saturation) were measured
and the Modified Premature Infant Pain Pro-
file (PIPP)20 was used to assess pain response.
Swaddling significantly reduced the heart rate
response to pain and pain scores of preterm
infants after heel stick. Comeoo7 examined
the effect of positioning on pain responses
from heel stick in 30 preterm infants. A blan-
ket roll was used for containment around an
infant’s body. A side-lying position with legs
and arms flexed close to the infant’s trunk
was arranged. The videotape was recorded
throughout the process, that is, 5 minutes be-
fore the heel stick for baseline data and 5
minutes after the heel stick. The physiolog-
ical responses (heart rate and oxygen satu-
ration) were measured and PIPP20 was used
to assess pain response. The results demon-
strated the significant lower pain scores in
the positioning group than in the nonposition-
ing group. Ngamvittayapong8 investigated the
effects of maternal hold and touch on 30
full-term neonates’ pain responses from heel
stick. A mother cradled the infant in her arms
and used her hand to support infant’s bot-
tom and legs throughout the process, that is,
3 minutes prior heel stick, and 4 minutes post-
stick. The physiological responses (heart rate
and oxygen saturation) were measured and
the NFCS19 was used to assess behavioral re-
sponses to pain. The results revealed that ma-
ternal hold and touch significantly affected
heart rate, oxygen saturation, and pain scores
of infants in the experimental group.
LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0
306 JOURNAL OF PERINATAL AND NEONATAL NURSING/OCTOBER–DECEMBER 2003
Empirical findings in Thailand revealed that
nonpharmacological pain management such
as swaddling, positioning, and maternal hold-
ing and touch on pain responses to heel
stick in neonates is effective. However, the
extent to which each intervention was rela-
tively effective on pain management has never
been evaluated. The purpose of this article
is to carry out a systematic analysis of stud-
ies conducted in Thailand on nonpharmaco-
logical methods of reducing pain responses
to heel stick in neonates in order to exam-
ine the magnitude of effectiveness of specific
interventions.
METHODOLOGY
Selection of studies
All researches conducted in Thailand and
published in any form were considered for in-
clusion. The focus of the current study was on
the studies pertaining to the effects of nursing
interventions on pain response to heel stick
in neonates. Key descriptors including pain,
heel stick, heel prick, heel lance, preterm in-
fant, and newborn were used for computer
search of the literature. Thai Nursing Research
Database 1988–199921 was used as a primary
indexing service for searching. The studies
conducted after 1999 were recruited through
computer and manual searching from Mahidol
University Library.
Requirements for inclusion
A study had to meet 3 criteria to be in-
cluded. The first requirement was inclusion
of a sample of either full-term or preterm
neonates who had to have a heel stick proce-
dure. Second, the study must have used a pain
measure that had established reliability and
validity. Finally, all studies were required to
have reported statistics including means and
standard deviations to calculate effect sizes
through comparison with a control group.
Only 5 studies were obtained and 4 of
them met the criteria for the analysis. One
study was excluded because of inadequate
reporting of baseline data. One author was
contacted and asked to provide necessary
statistics (standard deviations) for effect size
calculation.
Quality assessment of the individual studies
was undertaken independently by the inves-
tigators using the Tool for Research Quality
Assessment (F. Tilokskulchai et al, unpub-
lished data, 2002). This tool was originally
developed by Ariyasinsomboon.22 The origi-
nal one contained 30 items assessing quality
of research methodology and was used for
assessing research in the field of education.
Content validity and interrater reliability
were reported. To make this tool suitable for
research in the field of nursing, Tilokskulchai
and her colleagues (unpublished data, 2002)
modified and deleted some items. The mod-
ified tool contains 28 items and the score
of each item ranges from 1 to 4. The overall
scores of the quality are categorized into
“poor,” “fair,” “good,” and “excellent” levels.
Content validity of the modified tool was per-
formed by an expert in educational research.
Interrater reliability using percent agreement
over 80% was achieved. The quality for all of
the studies was rated as “good.”
Statistical analyses
Each individual study aimed at assessing the
effectiveness of an intervention on pain re-
sponses to heel stick in neonates. The ef-
fect size was calculated using the standard-
ized mean difference as defined by Glass and
colleagues:23,24
Me − Mc
d=
Sc
where Me represents the mean for the exper-
imental group, Mc the mean for the control
group, and Sc the standard deviation for the
control group.
For the magnitude of effect size, Cohen’s
guidelines25 were used for interpretation, that
is, an effect size of 0.2 is considered small, 0.5
is moderate, and 0.8 and above is large.
LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0

Management of Pain From Heel Stick in Neonates 307

Table 1. Characteristics of the participants weeks (SD = 0.5) for preterm infants. The
across studies (N = 4) mean for average birth weight of participants
was 3128.5 g (SD = 126.6) for full-term and
1739.2 g (SD = 103.0) for preterm infants.
Characteristics Mean SD
Interventions implemented for pain preven-
tion or management included maternal hold-
Average gestational
ing and touching, swaddling, and positioning
age (wk)
Full-term newborn 39.15 0.21 the newborns (see Table 2). Characteristics of
Preterm newborn 33.59 0.45 the studies are summarized in Table 3. The
Average birthweight (g) number of subjects in either control or ex-
Full-term newborn 3128.50 126.57 perimental group ranged from 18 to 30. The
Preterm newborn 1739.17 103.00 NFCS19 was used to assess pain in the studies
Average age (d) 4.29 1.23 of full-term neonates, and PIPP20 in the studies
of preterm neonates. Interobserver reliability
for each tool was evaluated in each study and
ranged from. 91 to 1.00.5–8
RESULTS Across the 4 studies, the observation was
divided into 3 phases, including the base-
Four studies published during 1997–2002 line prestick (3–5 minutes prior to the start-
met the criteria for meta-analysis and were ing of the heel stick), the heel stick (about
included.5–8 All were master’s thesis in- 1 minute), and the poststick (4–5 minutes fol-
volving an experimental study with a self- lowing the heel stick). Behavioral responses
controlled design, that is, participants served (pain scores) were videotaped and physiolog-
as their own controls. One hundred and ical responses (heart rate and oxygen satu-
eight neonates (62 males and 46 females) ration) were obtained from pulse oximeters.
participated in the 4 studies. As seen in The measurement of oxygen saturation was
Table 1, the mean for average gestational not included in 1 study. During the base-
age of participants across the 4 studies was line prestick phase, no significant differences
39.2 weeks (SD = 0.2) for full-term and 33.6 in any of the observed variables were found

Table 2. Characteristics of the study interventions (N = 4)

Prestick Poststick
Type of Description of phase phase Starting point of
Study intervention intervention (min) (min) intervention

A Hold & Touch Cradling a baby in mother’s 3 4 The beginning of

arms prestick phase
B Swaddle Using 2 strips of cloth to 3 4 The end of
wrap a baby prestick phase
C Swaddle Using a large blanket to 3 5 The beginning of
wrap a baby prestick phase
D Position Using a blanket roll for 5 5 The beginning of
containment, side-lying prestick phase
positioning with flexion
legs & arms
LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0

308 JOURNAL OF PERINATAL AND NEONATAL NURSING/OCTOBER–DECEMBER 2003

Table 3. Characteristics of the studies (N = 4)

Pain measurement
Study Intervention Subject n∗ Male:Female† tool Reliability

A Hold & touch Full-term 30 14:16 NFCS‡ 0.92

B Swaddle Full-term 18 12:6 NFCS‡ 1.00
C Swaddle Preterm 30 13:17 PIPP§ 0.91
D Position Preterm 30 23:7 PIPP§ 0.94

∗ Number of subjects in either control or experimental group.

† Number of male and female subjects in each group.
‡ Neonatal Facial Coding System.
§ Premature Infant Pain Profile.

between control and experimental trials
across studies. Each of the studies reported a
significantly greater effect of the study inter-
vention on behavioral responses, that is, pain
scores of the experimental group were less
than that of the control group over time. How-
ever, the effect of study intervention on phys-
iologic responses varied across studies.
In the current analysis, the effect size for
pain scores at each observation following
the heel stick was computed. Analysis in-
dicated that the magnitude of the interven-
tion effect for pain relief ranged from 0.56
to 0.98, 0.58 to 0.96, 0.29 to 0.83, and 0.5
to 0.75 when maternal holding and touching,
swaddling in full-term, swaddling in preterm,
and positioning were employed, respectively
(Fig 1). Then a mean effect size across all ob-
servations following the heel stick was com-
puted. As seen in Table 4, maternal hold-
ing and touching, swaddling in full-term,
swaddling in preterm, and positioning pro-
vided more pain relief (dmn = 0.73, 95% CI =
0.41–1.04; dmn = 0.79, 95% CI = 0.53–1.05;

Fig 1. Effect sizes for pain scores by nursing interventions.

LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0

Management of Pain From Heel Stick in Neonates 309

Table 4. Effect sizes for pain scores, heart rate, and oxygen saturation by nursing interventions

Post Oxygen
stick Pain scores Heart rate saturation
period
Intervention Subject (min) dmn 95% CI dmn 95% CI dmn 95% CI

Hold & touch Full-term 4 0.73 0.41–1.04 0.48 0.18–0.77 0.39 0.17–0.60
Swaddle Full-term 4 0.79 0.53–1.05 0.64 0.46–0.81 ... ...
Swaddle Preterm 5 0.53 0.27–0.80 0.23 0.08–0.38 0.13 0.06–0.21
Position Preterm 5 0.64 0.51–0.77 0.12 0.02–0.22 0.14 0.01–0.27

Note: dmn indicates mean effect size; CI, confidence interval.

dmn = 0.53, 95% CI = 0.27–0.80; dmn = 0.64,
95% CI = 0.51–0.77; respectively) than their
controls.
Similar analyses were done to calculate the
effect sizes of the study interventions for heart
rate and oxygen saturation. The magnitude of
the intervention effect for decreasing heart
rate after the heel stick ranged from 0.30 to
0.52 (dmn = 0.48, 95% CI = 0.18–0.77), 0.55
to 0.79 (dmn = 0.64, 95% CI = 0.46–0.81),
0.04 to 0.34 (dmn = 0.23, 95% CI = 0.08–
0.38), and 0.01 to 0.24 (dmn = 0.12, 95%
CI = 0.02–0.22) when maternal holding and
touching, swaddling in full-term, swaddling in
preterm, and positioning were employed, re-
spectively. These findings are summarized in
Fig 2 and Table 4.
The magnitude of the intervention ef-
fect for increasing oxygen saturation after
the heel stick ranged from 0.23 to 0.56
(dmn = 0.39, 95% CI = 0.17–0.60), 0.08 to
0.22 (dmn = 0.13, 95% CI = 0.06–0.21), and
0.01 to 0.29 (dmn = 0.14, 95% CI = 0.01–
0.27) when maternal holding and touching,

Fig 2. Effect sizes for heart rate by nursing interventions.

LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0
310 JOURNAL OF PERINATAL AND NEONATAL NURSING/OCTOBER–DECEMBER 2003
Fig 3. Effect sizes for oxygen saturation by nursing interventions.
swaddling in preterm, and positioning were
employed, respectively. See Fig 3 and Table 4.
DISCUSSION
It is noteworthy that pain measurement
used in the studies had different indicators,
thus limiting generalizability. The NFCS used
in 2 out of the 4 studies assesses only facial
expression while the PIPP used in the other
2 studies assesses facial expression, physi-
ological (heart rate and oxygen saturation)
changes, and contextual (gestational age and
behavioral state) factors. As a result, pain
scores in some studies did not refer to exclu-
sively motoric behavioral responses. Regard-
less of which indicators were used in measur-
ing the neonates’ pain, the mean effect sizes
ranging from moderate (0.53) to large (0.79)
suggest that all study interventions were ef-
fective nonpharmacological methods in at-
tenuating neonates’ pain scores. Moreover,
it is interesting that the effect size of the
pain scores for positioning increased during
the first 3 minutes poststick and remained
somewhat stable after that. The findings sug-
gest that the moderate-to-large effect of po-
sitioning on pain relief continued to exist
throughout the poststick period. In contrast,
the effect of swaddling (in either full-term or
preterm neonates) and maternal holding and
touching tended to decrease over time. Con-
gruent with a previous study,13 positioning ap-
pears to be the most appropriate nonpharma-
cological intervention especially for preterm
neonates.
The mean effect sizes of holding and touch-
ing and swaddling in full-term infants for phys-
iological responses were greater than that
of the others. However, the changes in ef-
fect sizes for physiological responses, within
each study intervention, were mixed through-
out the poststick periods. The inconsistent
patterns of the effect sizes for physiological
responses within each intervention do not
allow us to make a conclusion. A possible
reason may be that physiological changes vary
as a result of many alterations within the
LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0
Management of Pain From Heel Stick in Neonates
infant’s body. For example, the infant’s oxy-
genation will reduce in response to pain but
may increase after vigorous crying.26 It can
be suggested that physiological responses are
just one source of information for pain assess-
ment in neonates while behavioral responses,
especially facial expression, may be better
consistent indicators of pain in infants.27,28
However, additional research is needed to de-
termine which indicator is more reliable.
CONCLUSION
The preliminary findings support that
nonpharmacological nurisng interventions
(swaddling, maternal holding and touching,
and positioning) are beneficial to some de-
gree in neonates who experience heel stick
and may vary from infant to infant. Therefore,
effective nonpharmacological pain manage-
ment in neonates requires that nurses be
willing to try a number of interventions to
gain best results. Neonatal nurses should
develop their knowledge and skills in assess-
ing pain and providing nonpharmacological
interventions that work best for them and the
individual infant.
NURSING IMPLICATIONS
The alleviation of pain is a basic need and
human right regardless of age. Thus, to en-
courage nurses to use nonpharmacological
pain management in newborns is essential.
The current findings suggest that the non-
pharmacological interventions used in these
studies are effective methods for pain relief
when a newborn experiences a heel stick. In
addition to their effectiveness, these interven-
tions are simple to apply, and have minimal ad-
verse effect. Parent involvement in the inter-
ventions is also possible by supporting them
REFERENCES
1. Anand KJS, Phill D, Hickey PR. Pain and its effects
in the human neonate and fetus. N Engl J Med.
1987;31:1321–1329.
311
to hold, touch, swaddle, and position their
newborn undergoing a heel stick. Facilitating
parents in applying the effective pain manage-
ment interventions can improve their ability
to comfort their baby and lessen potential feel-
ings of helplessness in soothing and caring for
their newborn.13
IMPLICATIONS FOR FUTURE RESEARCH
Replication of studies using the study in-
terventions in both full-term and preterm
neonates is recommended so that a larger
number of studies will be included in a future
meta-analysis study resulting in a more conclu-
sive effect size of each pain management in-
tervention for Thai neonates. In addition, the
observation during poststick period should be
taken until the pain indicators return to base-
line so that the extent to which the interven-
tion is relatively effective in reducing pain du-
ration can be evaluated.
LIMITATIONS
The current study was a preliminary anal-
ysis to examine nonpharmacological meth-
ods for reducing pain during heel stick in
neonates. A major limitation of the current
study was the very limited number of individ-
ual studies that were recruited in the analy-
sis. In addition, the samples differed in that in
some studies the subjects were term infants,
while in others the subjects were preterm
infants. The effect size of each specific in-
tervention was analyzed from each individ-
ual study. Thus, the synthesis and integration
of results from multiple individual studies for
each specific intervention could not be fully
performed. Moreover, the small sample size of
4 studies also limited statistical analysis. The
issue of which factors/variables predict the ef-
fect sizes could not be explored.
2. Johnston CC, Stevens BJ, Yang F, Horton L. Differ-
ential response to pain by very premature neonates.
Pain. 1995;61:471–479.
LWW/JPNN AS225-06 October 9, 2003 19:9 Char Count= 0
312 JOURNAL OF PERINATAL AND NEONATAL NURSING/OCTOBER–DECEMBER 2003
3. Barker DP, Rutter N. Exposure to invasive procedure
in Neonatal Intensive Care Unit admissions. Arch Dis
Child. 1995;72(1):F47-F48.
4. Stevens BJ, Johnston CC. Physiological response of
premature infants to a pain stimulus. Nurs Res.
1994;43(4):226–231.
5. Kacome P. Effects of Swaddling on Pain Response
to Heel Prick in Full Term Infants [master’s thesis].
Thailand: Chiang Mai University; 1997.
6. Tantapong E. The Effects of Swaddling on Pain
Response Related to Heel Stick in Prematrue In-
fants [master’s thesis]. Thailand: Mahidol University;
2000.
7. Comeoo P. Effects of Positioning on Pain Response
From Heel Prick in Premature Infants [master’s the-
sis]. Thailand: Mahidol University; 2001.
8. Ngamvittayapong A. Effects of Maternal Hold and
Touch on Full Term Infants’ Response Towards Pain
From Heel Stick [master’s thesis]. Thailand: Mahidol
University; 2002.
9. Stevens BJ. Pain in infants. In: McCaffery M, Pasero
CC, eds. Pain Clinical Manual. St. Louis: Mosby;
1999.
10. Chiswick ML. Assessment of pain in neonates.
Lancet. 2000;355(9197):6–7.
11. Campos RG. Soothing pain-elicited distress in in-
fants with swaddling and pacifiers. Child Dev.
1989;60:781–792.
12. Fearon I, Kisilevsky BS, Hains SN, et al. Swaddling
after heel lance: age-specific effects on behavioral
recovery in preterm infants. J Dev Behav Pediatr.
1997;18(4):222–232.
13. Corff KE, Seideman R, Venkataraman P, Lutes L, Yates
B. Facilitated tucking: a nonpharmacologic com-
fort measure for pain in preterm neonates. JOGNN.
1995;24(2):143–147.
14. Campos RG. Rocking and pacifiers: two comforting
interventions for heelstick pain. Res Nurs Health.
1994;17:321–331.
15. Field T, Goldson E. Pacifying effects of nonnutritive
sucking on term and preterm neonates during heel-
stick procedures. Pediatrics. 1984;74(6):1012–1015.
16. Gray L, Miller LW, Philipp BL, Blass EM. Breastfeed-
ing is analgesic in healthy newborns. Pediatrics.
2002;109(4):590–593.
17. Johnston CC, Stremler RL, Stevens BJ, Horton LJ.
Effectiveness of oral sucrose and simulated rock-
ing on pain response in preterm neonates. Pain.
1997;72(1/2):193–199.
18. Stevens B, Johnston C, Franck L, Petryshen P, Jack
A, Foster G. The efficacy of developmentally sensi-
tive interventions and sucrose for relieving procedu-
ral pain in very low birth weight neonates. Nurs Res.
1999;48(1):35–42.
19. Grunau RVE, Craig KD. Pain expression in neonates:
facial action and cry. Pain. 1987;28:395–410.
20. Stevens BJ, Johnston CC, Petryshen P, Taddio A.
The premature infant pain profile: development
and initial validation. Clin J Pain. 1996;12:13–
22.
21. Tilokskulchai F, Sitthimongkol Y, Phuphaibul R,
Sirapo-ngam Y, Sirorojsakul S. Thai Nursing Re-
search Database 1988–1999 [book on CD-ROM].
Thailand: Mahidol University; 2002.
22. Ariyasinsomboon W. Synthesis of Research in the
Field of Educational Psychology: A Meta-Analysis
[dissertation]. Thailand: Chulalongkorn University;
2001.
23. Glass GV. Primary, secondary and meta-analysis of re-
search. Educ Res. 1976;5:3–8.
24. Glass GV, McGaw B, Smith ML. Meta-Analysis in So-
cial Research. Beverly Hills, Calif: Sage; 1981.
25. Cohen HM. Statistical Power Analysis for the Behav-
ioral Sciences. Hillsdage, NJ: Erlbaum; 1988.
26. Rawling D, Milles P, Engel R. The effect of tran-
scutaneous PO2 in term infants. Am J Dis Child.
1980;134:676–678.
27. Johnston CC, Strada ME. Acute pain response
in infants: a multidimensional description. Pain.
1986;24:373–382.
28. Craig KD, Whitfield MF, Grunau RVE, Linton J,
Hadjistavropoulos HD. Pain in the preterm
neonate: behavioral and physiological indices.
Pain. 1993;52:287–299.