What are the contributions mentioned in the paper "Validity of four pain intensity rating scales" ?

Ferreira-Valente et al. this paperarreira-Belas, Belas, and Belaseira-Vásquez studied the ability of different measures of pain intensity to detect differences in painful stimulation and for detecting sex effects in response to painful stimulation.

(Open Access) Validity of four pain intensity rating scales (2011) | Maria Alexandra Ferreira-Valente

Q: What was the effect size for each pair of temperature differences?

Effect sizes for the comparisons for each pair of temperature differences ranged from medium (0.17) to large (0.59), consistent with the differences between temperatures.

Validity of four pain intensity rating scales

Maria Alexandra Ferreira-Valente

a,b,c,

⇑

, José Luís Pais-Ribeiro

a,c

, Mark P. Jensen

Faculdade de Psicologia e Ciências da Educação da Universidade do Porto, Porto, Portugal

Portuguese Foundation for Science and Technology, Lisbon, Portugal

Unidade de Investigação em Psicologia e Saúde (Psychology and Health Unit), Lisbon, Portugal

Department of Rehabilitation Medicine, University of Washington School of Medicine, Seattle, WA, USA

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

article info

Article history:

Received 22 February 2011

Received in revised form 25 June 2011

Accepted 11 July 2011

Keywords:

Pain assessment

Validity

Numerical Rating Scale

Visual Analogue Scale

Faces Pain Scale

Verbal Rating Scale

abstract

The Visual Analogue Scale (VAS), Numerical Rating Scale (NRS), Verbal Rating Scale (VRS), and the Faces

Pain Scale-Revised (FPS-R) are among the most commonly used measures of pain intensity in clinical and

research settings. Although evidence supports their validity as measures of pain intensity, few studies

have compared them with respect to the critical validity criteria of responsivity, and no experiment

has directly compared all 4 measures in the same study. The current study compared the relative validity

of VAS, NRS, VRS, and FPS-R for detecting differences in painful stimulus intensity and differences

between men and women in response to experimentally induced pain. One hundred twenty-seven sub-

jects underwent four 20-second cold pressor trials with temperature order counterbalanced across 1°C,

3°C, 5°C, and 7°C and rated pain intensity using all 4 scales. Results showed statistically signiﬁcant dif-

ferences in pain intensity between temperatures for each scale, with lower temperatures resulting in

higher pain intensity. The order of responsivity was as follows: NRS, VAS, VRS, and FPS-R. However, there

were relatively small differences in the responsivity between scales. A statistically signiﬁcant sex main

effect was also found for the NRS, VRS, and FPS-R. The ﬁndings are consistent with previous studies sup-

porting the validity of each scale. The most support emerged for the NRS as being both (1) most respon-

sive and (2) able to detect sex differences in pain intensity. The results also provide support for the

validity of the scales for use in Portuguese samples.

1. Introduction

The Visual Analogue Scale (VAS), Numerical Rating Scale (NRS),

Verbal Rating Scale (VRS), and Faces Pain Scale-Revised (FPS-R) are

among the most common measures of pain intensity used by

clinicians and researchers. Evidence supports the reliability and

validity of each of these measures across many populations

[10,26–28]. However, each measure has strengths and weaknesses.

For example, research indicates that VASs have more ratio scale

qualities than other pain intensity scales for groups of patients

(but not necessarily for individuals) [13,42,44,45], although some

authors note that VASs scales do not always have linear qualities

and are not always normally distributed [37,55]. Pain scales with

more response levels (eg, the VAS or 0–10 NRS relative to the 6-

point FPS-R or 4-point VRS) have the potential to be more sensitive

[6,9,55], although more response categories do not necessarily

translate to more responsivity [7,19,29]. Furthermore, research

ﬁndings suggest that no single measure is consistently more

responsive than any of the other measures [6,9–11,28,30,36],

although the responsivity of the VAS, NRS, VRS, and FPS-R has

yet to be directly compared in the same study. Also, the validity

of these scales has never been examined in a sample of individuals

from Portugal. Evaluations of common pain measures in samples

from different countries and cultures can help establish the

cross-cultural generalizability of validity ﬁndings.

Perhaps the most important validity criterion for a pain measure

is its ability to detect changes in pain with pain treatment or proce-

dures known to produce pain. One method for doing this would be to

use an experimental design in which the amount of stimulation is

highly controlled [14]. The cold-pressor test is an experimental

method for inducing pain that is thought to reﬂect many (but not

all) ofthecritical components of clinical pain [25], and its advantages

are discussed in the literature [14,22,40,53]. Moreover, an increase

in pain intensity as water temperature decreases is well docu-

mented, with small variations in water temperature resulting in sig-

niﬁcant differences in pain intensity [22,53,54].

A number of studies have examined the inﬂuence of sex on pain

perceptions and pain response to experimental pain [14,40], with

doi:10.1016/j.pain.2011.07.005

⇑

Corresponding author. Address: Rua 25 de Abril, n° 5, Idanha – Belas, Belas

2605-119, Portugal. Tel.: +351 969082988.

E-mail address: mafvalente@gmail.com (M.A. Ferreira-Valente).

PAIN

152 (2011) 2399–2404

www.elsevier.com/locate/pain

normally menstruating women usually being more sensitive to

painful stimuli than men [34,36,51,54]. Laboratory pain experi-

ments that include both men and women provide an opportunity

to compare the ability of pain ratings to detect these well-estab-

lished sex differences.

The primary aim of the current study was to compare the relative

validity of VAS, NRS, VRS, and FPS-R for detecting differences in pain-

ful stimulation and for detecting sex effects in response to painful

stimulation. Based on previous research, we hypothesized that all

4 scales would be able to detect a sex effect and differences in pain

resulting from 4 temperatures (1°C, 3°C, 5°C, and 7°C). Given the lar-

ger number of response levels of the VAS and NRS, we also antici-

pated that these would evidence greater responsivity than the VRS

or FPS-R. Finally, this study also sought to evaluate the validity of

the pain intensity rating scales in a Portuguese sample.

2. Methods

2.1. Participants

Participants were 127 volunteer university students. Exclusion

criteria included: (1) being under 18 years of age; (2) reporting a his-

tory of any of the following diagnoses or medical problems: muscu-

loskeletal problems, cancer, heart disease, stroke, epilepsy, diabetes

or Raynaud syndrome; (3) having an open wound, cut, or fracture in

any of the upper limbs; (4) having a cognitive or physical disability

that could prevent participation; or (5) refusal to participate.

Of the 127 subjects who expressed an interest in participating in

the study, 112 were eligible and completed the entire experimental

procedures. Of the 112 completers, 3 subjects were excluded from

the analyses because they were unable to understand how to use

the VAS. Thus, complete data were available for 109 subjects, 56 of

whom were female (51.4%). The ages of the participants ranged from

18 to 40 years old (M = 22.27, SD = 3.92; Female: M = 21.24,

SD = 3.45; Male: M = 23.34, SD = 4.13). Most of the sample had their

permanent residence in an urban area (74.3%), and the remainder

(25.7%) lived in a rural area. Ninety-nine participants were under-

graduate college students (33.0%, 17.4%, 21.1%, 11.0%, and 8.3% in

their ﬁrst, second, third, fourth, and ﬁfth year, respectively). Ten

(9.2%) of the participants were in graduate school.

2.2. Material

The cold-pressor apparatus used consisted of 4 thermal insu-

lated containers with 18.1 L of capacity containing water chilled

to 4 different temperatures. The apparatus was capable of main-

taining water temperatures within ±0.5°C of the desired tempera-

tures throughout the experimental procedures. Each container

had 2 compartments separated by a metal ﬁlter, one of which held

water, ice, and a water pump, and the other (the immersion tank)

contained water alone with an armrest; the participants’ hands

therefore never came in direct contact with the ice. Four water

pumps (JAD, model SP-602, 200 L/h, Guangdong, China) made the

water ﬂow continuously between the 2 compartments of each con-

tainer to prevent warm water pockets from forming near the partic-

ipants’ hands [53]. The temperature of the hand and water was

monitored and controlled by asking the participants to hold a mer-

cury thermometer in the palm of their hands and immersing a ther-

mometer in the water, respectively. One extra thermal insulated

container without divisions contained an armrest and tepid water.

2.3. Measures

Fig. 1 presents all the pain intensity rating scales used in this

study. The VAS [23] consists of a horizontal line 100 mm in length,

with the end points ‘‘No pain’’ and ‘‘Worst imaginable pain’’ placed

at each end of the line. Respondents are asked to make a mark on

the line that best represents the level of pain intensity that they are

experiencing. The NRS is an 11-point scale consisting of integers

from 0 through 10; 0 representing ‘‘No pain’’ and 10 representing

‘‘Worst imaginable pain.’’ Respondents select the single number

that best represents their pain intensity. Although validity studies

for the Portuguese versions of these measures have not been pub-

lished, to our knowledge, both the VAS and the NRS used in this

study have been previously used in research with Portuguese sam-

ples [1,16,17,48], and are recommended for use by the Portuguese

Ministry of Health (Normative Circular n° 9/DGCG of June 14,

2003). The VRS is a 5-point scale consisting of a list of phrases

(no pain, mild pain, moderate pain, intense pain, maximum pain)

that describe increasing levels of pain intensity. Respondents select

the single phrase that best characterizes their pain intensity. The

VRS used in this study is commonly used by Portuguese research-

ers (eg, [12]). The FPS-R [4,21] is a 6-point scale, with 6 different

faces that represent increasing levels of pain intensity. Respon-

dents are asked to select the one expression that best characterizes

his or her pain intensity, from the left-most face (‘‘No pain’’), to the

right-most face (‘‘Very much pain’’). Each illustration corresponds

to a numeric score (0, 2, 4, 6, 8, or 10). Research supports the valid-

ity of each of the pain measures used in this study as measures of

pain intensity [24,27,32,39,44,46,47]. Although the FPS-R was ini-

tially developed for use with children, researchers also use the

measure in samples of individuals with cognitive and communica-

tion impairment. The Portuguese (Portugal) translation of the FPS-

R was performed by Batalha [2] and is available online

(www.painsourcebook.ca).

2.4. Procedure

The cold-pressor procedures closely followed the guidelines for

this task described in the literature, and were adapted to ﬁt the

study aims [40,52,53]. The study had Institutional Review Board

approval. The study procedures were described to all potential par-

ticipants and each was given a written consent form to read and

sign before any measures were administered. After signing the

consent form, participants completed a demographic and medical

history questionnaire in order to identify potential medical condi-

tions that would prevent participation. Participants who met any

of the exclusion criteria were then excluded from participation.

Fig. 1. The Numerical Rating Scale, Visual Analogue Scale, Faces Pain Scale –

Revised, and Verbal Rating Scale; Faces Pain Scale – Revised, copyright Ó2001,

International Association for the Study of Pain, reproduced with permission,

www.painsourcebook.ca.

2400 M.A. Ferreira-Valente et al. / PAIN

152 (2011) 2399–2404

The nondominant hand temperature was measured in all par-

ticipants, followed by hand washing to the wrist. The nondominant

hand was then immersed to the wrist in the container with tepid

water (36°C±1°C) for 2 minutes, in order to reduce preexisting dif-

ferences in hand temperature [53]. Hand temperature was again

measured, and participants were instructed to immerse the hand

to the wrist in the ﬁrst cold water container for 20 seconds. Partic-

ipants were also told they could take the hand from the cold water

at any moment if it felt too uncomfortable to continue. For each of

the 4 study conditions, only 3 participants took their hands out

of the water before the established tolerance time (20 seconds).

After the 20-second cold immersion, when we anticipated that

pain would be at its most intense, we administered paper-and-

pencil versions of the 4 pain measures (VAS, NRS, VRS, and FPS-

R). The measures were presented in a random order using a Latin

Square design. After a 3-minute break, the participants were asked

to immerse his or her hand into the tepid warm water again for

2 minutes, and hand temperature was again assessed. Participants

underwent 4 trials with 4 different water temperatures (1°C, 3°C,

5°C, and 7°C), in counterbalanced order, as proposed by Mitchell

and colleagues [40]. Each participant experienced each water tem-

perature only once, and provided ratings using each of the 4 scales

for each temperature; thus, each participant provided 16 ratings.

Participants were not given any information regarding the water

temperature in each container.

2.4.1. Data analysis

We ﬁrst computed medians, means, and SDs for demographic

and study variables for descriptive purposes. We next computed

Pearson correlations between the VAS, NRS, FPS-R, and VRS, for

descriptive purposes. In order to compare the ability of VAS, NRS,

FPS-R, and VRS to detect differences in pain stimuli resulting from

4 different temperatures, as well as the hypothesized sex main ef-

fect on pain intensity ratings, we then performed 4 mixed-design

repeated-measures analyses of variance (ANOVAs), with the pain

intensity ratings as the dependent variables, and sex and temper-

ature as the independent variables. Prior to these analyses, we

evaluated test assumptions, namely normality and sphericity of

the variance–covariance matrix, by analysing skewness (Sk) and

kurtosis (Ku), with values of Sk and Ku lower than 1 indicating ab-

sence of severe violation of normality assumption and Mauchly

test, respectively [5,18]. If a violation of the assumption of spheric-

ity was found, we planned to use Huynh-Feldt epsilon to set the

degrees of freedom [8,38]. In the event that a signiﬁcant tempera-

ture effect was found, we planned to perform between-tempera-

ture comparisons using post hoc Fisher’s least signiﬁcant

difference tests. Effect sizes were estimated using

, and, along

with P values and F statistic magnitudes, were used to compare

the hypothesized differences in responsivity of the 4 pain mea-

sures, with larger

and F statistics, as well as smaller P values,

indicating greater sensitivity [10,11,24,49]. Finally, we performed

power analyses to determine the sample size required to obtain

signiﬁcant effects for each of the 4 measures, both for the omnibus

ANOVA and for each planned temperature paired comparison. Al-

pha was set at 0.05 and power at 0.95 for these analyses. Statistical

analyses were computed using software PASW Statistics 18 (v. 18,

SPSS Inc. Chicago, IL, USA) and G

⁄

Power (v. 3.1) [15].

3. Results

Table 1 lists the descriptive statistics of the study variables, and

Table 2 presents the correlation coefﬁcients between the pain mea-

sures. As can be seen, the pain scales showed strong to very strong

and statistically signiﬁcant inter-scale correlations (rs ranging

from 0.79 to 0.96) for all 4 water temperatures.

Pain ratings, as measured by VAS, NRS, VRS, and FPS-R, showed

normal distributions for each of the 4 temperatures (Sk <1 and Ku

<1). However, for each measure, we noted a violation of the

assumption of sphericity [VAS: W = 0.70, X2 (5) = 38.28,

p < 0.001; NRS: W = 0.59, X2 (5) = 55.53, p < 0.001; VRS: W = 0.82,

X2 (5) = 20.54, p < 0.01; FPS-R: W = 0.81, X2 (5) = 21.74, p < 0.01].

We therefore used Huynh-Feldt epsilon to determine the degrees

of freedom in the analyses [8,38].

As hypothesized, there were statistically signiﬁcant tempera-

ture main effects for each of the 4 scales used in the study [VAS:

Huynh-Feldt

(2.48, 265.24) = 85.74; p < 0.001,

= 0.45;

= 1; NRS:

Huynh-Feldt

(2.28, 243.64) = 93.49; p < 0.001,

= 0.47;

= 1; VRS:

Huynh-Feldt

(2.74, 287.48) = 76.36; p < 0.001,

= 0.42;

=1;

FPS-R: F

Huynh-Feldt

(2.70, 286.42) = 72.62; p < 0.001,

= 0.32;

= 1]. Moreover, all of the effect sizes associated with the temper-

ature main effects in the omnibus ANOVAs for each measure were

large. Effect sizes for the comparisons for each pair of temperature

differences ranged from medium (0.17) to large (0.59), consistent

with the differences between temperatures. These differences also

follow the same pattern for each scale (Table 3 and Fig. 2).

Statistically signiﬁcant sex main effects emerged for the NRS

[F(1, 107) = 4.40; p < 0.05,

= 0.04;

= 0.55], VRS [F(1, 107) =

8.38; p < 0.01,

= 0.07;

= 0.82], and FPS-R [F(1, 107) = 14.13;

p < 0.001,

= 0.12;

= 0.96], and a nonsigniﬁcant trend emerged

for the main effect of gender on pain ratings on VAS

[F(1, 107) = 3.49; p = 0.07,

= 0.03;

= 0.46], with women report-

ing higher pain ratings in every experimental condition.

Post hoc Fisher’s least signiﬁcant difference paired tempera-

tures comparisons for each pain scale found statistically signiﬁcant

differences between all observations (p < 0.001). As would be ex-

pected, in every case, lower temperatures resulted in higher pain

intensity ratings for all 4 scales.

The NRS and VAS evidenced slightly higher effect sizes (0.47

and 0.44, respectively) and higher F statistics (93.49 and 85.74,

respectively) than VRS (0.42 effect size and F statistic of 76.36)

and FPS-R (0.32 effect size and F statistic of 72.62). Power analyses

based on these effect sizes indicated that the number of partici-

pants needed to be able to detect an overall difference between

temperatures, as tested by an ANOVA, would be 5, 5, 5, and 7 for

the NRS, VAS, VRS, and FPS-R, respectively. Power analyses to

Table 1

Means and SDs of the pain ratings for each temperature condition.

Overall

Mean (SD)

Male

Mean (SD)

Female

Mean (SD)

VAS

7°C 3.65 (2.35) 3.23 (2.29) 4.03 (2.36)

5°C 4.38 (2.23) 3.90 (2.10) 4.82 (2.28)

3°C 5.04 (2.24) 4.67 (2.31) 5.38 (2.14)

1°C 5.83 (2.17) 5.58 (2.24) 6.05 (2.12)

NRS

7°C 3.87 (2.39) 3.43 (2.38) 4.29 (2.33)

5°C 4.64 (2.21) 4.13 (2.17) 5.12 (2.15)

3°C 5.32 (2.15) 4.96 (2.24) 5.66 (2.01)

1°C 6.18 (2.21) 5.83 (2.29) 6.52 (2.10)

FPS-R

7°C 3.33 (2.51) 2.52 (2.30) 4.11 (2.48)

5°C 4.15 (2.32) 3.32 (2.19) 4.95 (2.21)

3°C 4.95 (2.55) 4.08 (2.60) 5.78 (2.23)

1°C 5.76 (2.46) 5.17 (2.46) 6.29 (2.35)

VRS

7°C 2.49 (0.93) 2.26 (0.88) 2.67 (0.93)

5°C 2.79 (0.86) 2.55 (0.82) 3.00 (0.85)

3°C 3.08 (0.83) 2.87 (0.81) 3.30 (0.79)

1°C 3.50 (0.80) 3.34 (0.85) 3.63 (0.71)

VAS, Visual Analogue Scale; NRS, Numerical Rating Scale; FPS-R, Faces Pain Scale-

Revised; VRS, Verbal Rating Scale.

M.A. Ferreira-Valente et al. / PAIN

152 (2011) 2399–2404

2401

compute the number of participants needed to detect differences

between each pair of temperatures for each of the pain scales are

presented in Table 3, and range from 5 to 19. These are consistent

with the distance between temperatures, with larger distances

(and effect sizes) corresponding to fewer participants needed.

4. Discussion

The results of this study provide strong support for the validity

of all 4 scales studied for detecting changes in pain intensity in Por-

tuguese university students. All of the scales were able to detect

differences in pain resulting from 4 different temperatures, with

variations in temperature resulting in statistically signiﬁcant dif-

ferences in pain intensity ratings, and lower temperatures result-

ing in higher pain ratings for each of the 4 scales studied. These

results are consistent with previous studies that support each

scale’s validity [9,10,24,33,44,45], and show that small variations

in water temperature result in signiﬁcant differences in pain inten-

sity ratings [22,40,53,54].

As predicted, we found some differences in the relative respon-

sivity, with NRS being the most responsive, followed by VAS, VRS,

and FPS-R. This is consistent with previous studies demonstrating

the superiority of VAS and NRS responsiveness, due perhaps to the

larger number of response levels that these 2 scales provide

[6,9,41,50,55]. In our study, the NRS has shown to be slightly more

responsive than the VAS, as indicated by its larger effect size and F

statistic value. This is consistent with a group of studies showing a

similar sensitivity between the NRS and VAS or a slight superiority

of the NRS over the VAS [6,9–11]. Nevertheless, the results of

power analyses to determine the number of subjects needed to de-

tect a signiﬁcant effect indicate that the 4 measures have very sim-

ilar levels of responsivity, with very small differences between

scales in the numbers of subjects needed to detect differences. This

ﬁnding is consistent with the strong to very strong associations we

found among the study measures (a ﬁnding also consistent with

previous research, eg, [11,24,33,46]), indicating that all 4 measures

tap into the same overall dimension (ie, pain intensity).

Thus, the results indicate that all else being equal, any of the 4

scales could be used for detecting changes in pain, although the

NRS and VAS might be considered ﬁrst when particularly sensitive

and responsive measures of pain intensity are needed. Based on

other considerations, however, researchers and clinicians may

Table 2

Inter-scale correlation coefﬁcients between the VAS, NRS, FPS-R, and VRS.

7°C5°C3°C1°C

VAS NRS FPS-R VAS NRS FPS-R VAS NRS FPS-R VAS NRS FPS-R

VAS––– ––– ––– –––

NRS 0.96 – – 0.95 – – 0.94 – – 0.95 – –

FPS-R 0.84 0.85 – 0.79 0.81 – 0.80 0.80 – 0.84 0.84 –

VRS 0.80 0.82 0.81 0.80 0.79 0.82 0.81 0.81 0.86 0.80 0.80 0.84

VAS, Visual Analogue Scale; NRS, Numerical Rating Scale; FPS-R, Faces Pain Scale-Revised; VRS, Verbal Rating Scale.

Table 3

Effect sizes (and number of participants required to obtain signiﬁcant effects) for the VAS, NRS, FPS-R, and VRS.

Effect size (n)

VAS NRS FPS-R VRS

5°C3°C1°C5°C3°C1°C5°C3°C1°C5°C3°C1°C

7°C 0.22 (14) 0.47 (6) 0.58 (5) 0.25 (12) 0.49 (6) 0.59 (5) 0.20 (16) 0.44 (7) 0.55 (5) 0.17 (18) 0.42 (7) 0.58 (5)

5°C – 0.20 (16) 0.46 (7) – 0.22 (14) 0.48 (6) 0.20 (16) 0.45 (7) – 0.17 (19) 0.29 (10)

3°C – – 0.30 (10) – – 0.36 (8) – 0.20 (15) – – 0.29 (10)

VAS, Visual Analogue Scale; NRS, Numerical Rating Scale; FPS-R, Faces Pain Scale-Revised; VRS, Verbal Rating Scale.

Fig. 2. Average pain intensity ratings across temperatures. Error bars represent SD.

2402 M.A. Ferreira-Valente et al. / PAIN

152 (2011) 2399–2404

elect to choose the NRS over the VAS in many settings. First,

although the NRS has not consistently been shown to have ratio

properties [43,45], its scores can provide data for parametric anal-

ysis [6,13,25,55]. Also, the NRS has been shown to be at least as

sensitive as the VAS, whether a 0–10 NRS or a 0–100 NRS is used

[6,9–11,24]. Third, the NRS is preferred over the VAS by patients

and clinicians for its relative simplicity and ease of administration

and scoring even when administered verbally [3,6,10,11,

13,20,24,55]. Fourth, the VAS tends to have higher failure rates

than the NRS or VRS, probably because both the NRS and the VRS

are very easy to understand and complete by patients

[3,6,10,11,13,20,24]. This latter point was also supported in the

current study, given our ﬁnding that the only participants excluded

from the sample for not being able to understand the measures

were excluded because they were unable to understand the VAS.

The VRS is a categorical measure that might not have ratio prop-

erties [43]. As a result, VRSs do not necessarily have equal intervals

between levels, which limits the conclusions that can be drawn

about the magnitude of differences over time or between patient

groups [25,26]. Also, when differences in responsivity are found,

VRSs, as well as the FPS-R, tend to be less sensitive than VAS and

NRS, consistent with the results from our study. This lower level

of responsivity may be related to the lower number of response

categories of these measures. However, we did ﬁnd that both of

these scales were able to detect changes in pain associated with

differences in water temperature, indicating that they are valid

and could be used when responsivity is not a critical issue

[6,9,24,25,31]. Also, statisticians note that it is possible to draw va-

lid conclusions using parametric analysis with ordinal data, such as

data from VRSs, especially if number of categories in the scale is 5

or more [5,18,38]. Thus, the VRS can be considered a viable choice

in settings with patients or research subjects who might be less

able to use the NRS (eg, very young individuals or individuals with

signiﬁcant cognitive impairment). Likewise, our ﬁndings provide

support for the validity of the FPS-R in adults, supporting its use

in clinical and research settings. Although the FPS-R has been

developed for use with children and proven useful with people

with cognitive and communication disabilities, there may be situ-

ations or settings in which it might be useful for other populations.

For example, the FPS-R might be useful for samples that include

both adults without cognitive impairment and children (or the el-

derly) in the same study, and a measure is needed that all partici-

pants can complete. The FPS-R might also be considered for use in

cross-cultural studies with adults, where researchers cannot be

certain of the meaning equivalence of the verbal endpoint descrip-

tors. In this situation, a measure based on facial expressions might

show a greater cultural equivalence.

Regarding sex effect on pain intensity ratings, our results are in

line with previous research showing signiﬁcant sex effects on

intensity ratings following painful stimulation [34,35,51,54], with

women reporting higher pain intensity ratings across tempera-

tures. The sex effects were statistically signiﬁcant for 3 of the scales

(NRS, VRS, FPS-R), and showed a nonsigniﬁcant trend for the VAS. It

is interesting to note that other researchers have not found sex

main effects when they used the VAS to test for these effects

[40,45]. These ﬁndings suggest the possibility that for experiments

speciﬁcally designed to test for or explain sex effects in pain inten-

sity, it might be best to avoid using the VAS.

One signiﬁcant limitation of the study is that it was performed

with healthy young participants. Although it has been argued by

some that the cold-pressor test mimics the effects of chronic pain

conditions [40,53] due to its unpleasantness and to the fact that

the painful stimuli is conducted by the C ﬁbers, which are impli-

cated in chronic pain, experimental pain is different from clinical

pain. Clinical pain, for example, is less predictable and controllable.

Also, participants in experimental pain studies can be assured that

the pain is not associated with any tissue damage, whereas pa-

tients with clinical pain cannot always be so sure of this. For these

reasons, clinical pain has an emotional signiﬁcance and quality-of-

life implications that may inﬂuence pain perception [14]. There-

fore, the study ﬁndings do not necessarily generalize to patients

with clinical pain conditions. It would be useful to examine the rel-

ative responsivity of the 4 pain measures in response to treatments

or procedures known to impact clinical pain conditions to help

determine their generalizability.

Nevertheless, the ﬁndings provide support for the validity and

sensitivity of all 4 pain scales studied, with the exception that it

might be best to avoid the use of VAS in studies seeking to examine

sex effects. The ﬁndings also suggest that the NRS may be (very)

slightly more sensitive than the other measures in our sample of

individuals from Portugal; a ﬁnding consistent with some other

studies that have compared the NRS to other pain measures, and

supporting their cross-cultural reliability. Research is needed to

compare these measures in clinical settings to conﬁrm the general-

izability of the current ﬁndings to clinical populations.

Conﬂict of interest statement

None of the authors have any conﬂicts of interest with respect

to this study.

Acknowledgements

The authors gratefully acknowledge José Elísio Pereira for his

contribution in the construction of the apparatus, and to Rita Ferre-

ira and Vera Melo for their assistance in data collection. M. Alexan-

dra Ferreira-Valente received PhD grant SFRH/BD/40956/2007 in

the past year from the Portuguese Foundation for Science and

Technology. José L. Pais Ribeiro received a sabbatical grant from

FCT (SFRH/BSAB/982/2010) between January and April 2010. Mark

P. Jensen received research support, consulting fees, or honoraria in

the past year from Analgesic Research, Consultants in Behavioral

Research, Endo, Fralex, Medtronic, Merck, Pﬁzer, Smith & Nephew,

US Department of Education, US Department of Veterans Affairs,

and the US National Institutes of Health.

References

[1] Arantes S, Ferreira C, Lobo S, Moutinho R, Correia J, Carvalho J, Marcos A. Low

back pain: the reality of our pain treatment unit. Dor 2007;15:22–7.

[2] Batalha L. Instructions for administering the Faces Pain Scale-Revised in

languages other than English. Available from: <http://www.

painsourcebook.ca> [accessed 20.06.11].

[3] Bergh I, Sjöström B, Odén A, Steen B. An application of pain rating scales in

geriatric patients. Aging (Milano) 2000;12:380–7.

[4] Bieri D, Reeve R, Champion GD, Addicoat L, Ziegler J. The Faces Pain Scale for

the self-assessment of the severity of pain experienced by children:

development, initial validation and preliminary investigation for ratio scale

properties. Pain 1990;41:139–50.

[5] Bollen KA. Structural equation with latest variables. New York: Wiley; 1989.

[6] Bolton JE, Wilkinson RC. Responsiveness of pain scales: a comparison of three

pain intensity measures in chiropractic patients. J Manipulative Physiol Ther

1998;21:1–7.

[7] Bone M, Critchley P, Buggy DJ. Gabapentin in postamputation phantom limb

pain: a randomized, double-bind, placebo-controlled, cross-over study. Pain

2002;27:481–6.

[8] Box GEP. Some theorems on quadratic form applied to the study of analysis of

variance problems: II. Effects of inequality of variance and the correlation

between errors in the two-way classiﬁcation. Ann Math Stat 1954;25:484–98.

[9] Breivik EK, Bjornsson GA, Skovlund E. A comparison of pain rating scales by

sampling from clinical trial data. Clin J Pain 2000;16:22–8.

[10] Bryce TN, Budh CN, Cardenas DD, Dijkers M, Felix ER, Finnerup NB, Kennedy P,

Lundeberg T, Richards JS, Rintala DH, Siddall P, Widerstrom-Noga E. Pain after

spinal cord injury: an evidence-based review for clinical practice and research.

Report of the National Institute on Disability and Rehabilitation Research

Spinal Cord Injury Measures meeting. J Spinal Cord Med 2007;30:421–40.

[11] Chanques G, Viel E, Constantin JM, Jung B, de Lattre S, Carr J, Cissé M, Lefrant JY,

Jaber S. The measurement of pain in intensive care unit: comparison of 5 self-

report intensity scales. Pain 2010;151:711–21.

M.A. Ferreira-Valente et al. / PAIN

152 (2011) 2399–2404

2403

Validity of four pain intensity rating scales

Figures

Citations

A retrospective cohort study: 10-year trend of disease-modifying antirheumatic drugs and biological agents use in patients with rheumatoid arthritis at Veteran Affairs Medical Centers.

The validity of pain intensity measures: what do the NRS, VAS, VRS, and FPS-R measure?

Psychometric Properties of the Numerical Rating Scale to Assess Self-Reported Pain Intensity in Children and Adolescents: A Systematic Review.

Chondroitin for osteoarthritis

Exogenously Applied Muscle Metabolites Synergistically Evoke Sensations of Muscle Fatigue and Pain in Human Subjects

References

G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences

Measurement of pain.

The validation of visual analogue scales as ratio scale measures for chronic and experimental pain

Measurement of pain.

Pain: a review of three commonly used pain rating scales.

Related Papers (5)

Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale

Core outcome domains for chronic pain clinical trials: IMMPACT recommendations.

The Pain Catastrophizing Scale: Development and validation.

The Hospital Anxiety and Depression Scale.

Statistical Power Analysis for the Behavioral Sciences

Frequently Asked Questions (10)

Q1. What are the contributions mentioned in the paper "Validity of four pain intensity rating scales" ?

Q2. What are the common measures of pain intensity used by clinicians and researchers?

Q3. What is the important criterion for a pain measure?

Q4. What can be done to establish the validity of the VAS?

Q5. How many subjects were eligible for the study?

Q6. What is the common measure of pain intensity used by clinicians and researchers?

Q7. What software was used to compute the temperature main effects?

Q8. What was the effect size for each pair of temperature differences?

Q9. How long did the participants have to immerse their hands in cold water?

Q10. How many participants took their hands out of the water before the established tolerance time?