Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2008 Jan 18.
Published in final edited form as: Pain Med. 2007;8(7):585–600. doi: 10.1111/j.1526-4637.2007.00316.x

Evaluation of the Iowa Pain Thermometer and Other Selected Pain Intensity Scales in Younger and Older Adult Cohorts Using Controlled Clinical Pain

A Preliminary Study

Keela Herr *, Kevin F Spratt †,, Linda Garand §, Li Li
PMCID: PMC2211278  NIHMSID: NIHMS36917  PMID: 17883743

Abstract

Objective

To evaluate the sensitivity and utility of the Iowa Pain Thermometer (IPT) and other selected pain intensity scales in younger and older adults using a controlled clinical pain condition.

Design

A quasi-experimental study with 61 younger (age 21-55 years) and 36 older (age 65-87 years) adults experiencing arthritic pain at two rheumatology clinics. Before and after joint injection, patients reported current pain intensity with the following scales: IPT, Numeric Rating Scale (NRS), Verbal Numeric Rating Scale (VNS), Faces Pain Scale (FPS), and Visual Analog Scale (VAS).

Results

The IPT demonstrated the lowest failure rate of all pain intensity scales evaluated. Other scale failure rates were relatively low except for the VNS and the VAS. No significant difference was noted in scale failure by age, gender or education level, but cognitive impairment was significantly related to failure on the VAS and the NRS. All five pain scales were sensitive in detecting changes in pain intensity pre and post joint injection. All correlations between the scales were strong and significant; however, the intercorrelations for the older cohort were weaker. The scale most preferred in both cohorts of patients was the IPT, followed by the FPS.

Conclusions

Based on sensitivity to change, lower failure rates, higher preference evaluations, and little appreciable affects associated with cognitive impairment, the IPT was judged to be the best choice for assessing pain intensity for both age cohorts and warrants further study.

Keywords: Pain Intensity, Assessment, Measurement, Pain Thermometer, Numeric Rating Scale, Verbal Numeric Rating Scale, Faces Pain Scale, Visual Analog Scale, Aged, Arthritis

Introduction

Approximately 25-80% of community-dwelling or institutionalized older adults report some degree of pain that may interfere with function and quality of life [1-10]. Yet, researchers have found that pain is often poorly assessed and insufficiently managed in older adults [1,2,4,11-13]. Inadequate assessment of pain and pain relief has been identified as an important barrier to optimal pain management in younger and older adults with either acute or chronic pain [11,14,15]. Both the evaluation of treatments in controlled studies and the implementation and evaluation of interventions in the clinical care of older adults require an accurate assessment of perceived pain.

Subjective pain intensity is probably the component most often measured in both clinical work and in treatment-outcome research and various scales that are quick and easy to understand have been developed to quantify pain intensity [16]. The ability to use various scales and preference for scales to document pain intensity are important factors for consideration when selecting scales for research or clinical use. In recent years, a growing number of studies have begun to address the psychometric properties of a variety of pain intensity assessment scales in older populations, including visual, numeric, verbal, and facial scales. Among them, the Numeric Rating Scale (NRS), the Verbal Descriptor Scale (VDS, called Verbal Rating Scale [VRS] by some researchers), the Verbal Numeric Rating Scale (VNS), the Faces Pain Scale (FPS), and the Visual Analog Scale (VAS) are appropriate for use with older adults [17] and having been evaluated in Caucasian and African American older adults with experimental or clinical pain in various settings using correlational approaches [17-22]. These studies show that the NRS, VDS, and/or FPS were preferred scales for older adults, while the VAS was least preferred with higher error rates, including those with mild to moderate cognitive impairment [17-22]. More-over, the Pain Thermometer (PT), a modified vertical VDS alongside a graphic thermometer, has also been validated as a measure for pain in older adults [23-26].

Although the most important aspect of a pain measure’s validity is often its sensitivity to change, or the ability to detect changes in pain over time or due to treatment [27], few studies have used a sound methodological approach to establish this aspect of scale validity in pain assessment in older adults.

Jensen et al. conducted a study to compare the relative sensitivity of three measures of pain outcome and a composite made up of all three measures for detecting analgesic effects in a sample of older adults with knee surgery (age 64.85 ± 9.97 years) [27]. In this study, two measures of pain intensity (VAS and 4-point VRS) were administered at baseline and post medication 24 hours following surgery. They found that the 4-point VRS showed less sensitivity than the VAS in detecting change in pain from treatment. In another study, Gagliese and Katz [28] presented age comparisons of the sensitivity of three pain scales (McGill Pain Questionnaire; Present Pain Intensity; and VAS) for the assessment of change in postoperative pain completed by younger (N = 95, mean age = 56.4 ± 5.8 years) and older (N = 105; mean age = 66.8 ± 2.7 years) men following radical prostatectomy. They found that all three scales have comparable sensitivity within an age group; however, the different results between the scales for the effect of age suggest that the VAS is not sufficiently sensitive to detect age differences. However, these studies have a limited age range of older adults, did not include cognitively impaired older adults, and did not evaluate a PT or facial scale, which are recommended scales for pain assessment in older adults [29].

As cognitive and motor changes that often accompany aging can interfere with a person’s ability to properly use a given pain intensity rating scale [30], it is important to have empiric validation of scales used with this population. Although studies have demonstrated that mild and even moderate cognitive impairment did not necessarily inhibit older adult’s ability to assess their pain intensity [5,13,20-22,31-35], few studies have examined the relationship between cognitive status, psychomotor abilities, educational level, age, ethnicity, and gender as factors that could impact successful completion of the pain intensity scales.

Findings from our previous psychometric scale evaluation using experimental thermal stimuli to identify suitable scale selection for older adults revealed that the most preferred and valid self-report was obtained from the VDS [20]; however, a PT was not included and there were concerns with limited descriptor choices and difficulty selecting between word choices on the VDS. Additionally, limited choice options on scales impacts scale sensitivity in detecting change [21,27], of particular interest in the conduct of research to evaluate intervention effectiveness. To address these concerns, we adapted the VDS validated in our earlier study to the Iowa Pain Thermometer (IPT) by adding choices of severity between verbal descriptor words and including a thermometer to facilitate understanding of the tool by those with diminished cognitive capacity. The purpose of this study was to compare the sensitivity and utility of the new IPT with four other pain scales: NRS, VNS, FPS, and VAS, using a naturally occurring pain condition and a controlled treatment with rheumatology patients. Results were compared between younger and older patients and between patients with normal and impaired cognitive function. The specific objectives of this study were to assess:

  1. Failure types and rates for each scale, and relationships of failure with cognitive status, psychomotor abilities, as well as subject’s education level, ethnicity and gender;

  2. The sensitivity of each pain rating scale to detect changes in perceived pain intensity, including effect of age;

  3. The concurrent validity of subjects’ pain intensity ratings on the selected pain scales; and

  4. Scale preference and subject factors associated with preference.

Methods

Subjects

A quasi-experimental design with repeated assessments was used to evaluate the psychometric properties for the selected pain intensity scales with younger and older adults experiencing chronic joint pain in a clinical setting. The pain-reducing stimulus was a standard joint injection that included a steroid plus lidocaine. With few exceptions, the injection provided immediate relief of joint pain. Therefore, the standard joint injection provided a clinical situation where subjects reported different levels of pain intensity before and after the joint injection.

Younger (aged 21-55 years) and older (aged 65 years or older) adults were recruited from two rheumatology clinics from a large university medical center and Veteran’s administration medical center. Eligibility criteria included: diagnosis of rheumatoid arthritis, osteoarthritis or other pain-related diagnosis; ability to speak English; and receiving a corticosteroid joint injection with lidocaine during the rheumatology clinics visit. Subjects were excluded if they were: 1) unable or unwilling to give informed consent; 2) verbalized an absence of pre-injection pain; or 3) were unable to complete the protocol due to severe cognitive impairment.

Measures

Cognitive Function

Cognitive function was determined by using 1): the Cognitive Capacity Screening Examination (CCSE) scored 0-30 (30 being highest cognitive state and scores less than 20 indicating cognitive impairment) [36,37]; and 2) the clock drawing procedure scored 1-6 (1 being perfect, 2 mild visuospatial impairment, 3-6 more severe cognitive impairment) [38,39]. Because abstract thinking is a key cognitive process potentially related to difficulty understanding scale use, particularly with the VAS [40], the CCSE was chosen as the cognitive screen as it includes five items related to abstract thinking. Studies have suggested that current cognitive screening procedures may require a certain level of education for accurate measurement of cognitive impairment. The clock drawing procedure is another method of evaluating cognitive impairment that is relatively free from cultural and educational bias and has been used successfully to identify cognitive and visuospatial impairment in older subjects [38,39].

Psychomotor Function

Psychomotor impairment may interfere with the ability to mark a line or circle the desired number on a given scale as intended. To evaluate psychomotor ability, line bisection was used. Line bisection also measures some higher neuropsychological functions that might be important for VAS completion [41]. Subjects were asked to mark on top of a line printed on a VAS at a point that is halfway between the two endpoints. Accuracy of mark placement was determined by measuring the number of mm deviation between the midpoint-printed mark and the subject’s mark.

Pain Intensity

Pain intensity was measured with the following five scales: IPT (Figure 1); NRS; VNS; FPS; and VAS. All scales used in the study were enlarged to 14-point bold typeface and printed on buffcolored paper to accommodate changes in vision most often occurring in older adults [17]. To facilitate comparison between scales, endpoints of all five scales were worded the same with anchors “No Pain” and “The Most Intense Pain Imaginable.” By using the same anchors on each pain rating scale, the subjects were able to focus on rating their pain intensity rather than choosing preferred descriptive anchors [42,43]. Moreover, the IPT, the NRS, and the VAS were all presented in vertical formats, based on suggestions in literature that this format may be more easily understood by older adults [17,19].

Figure 1.

Figure 1

Iowa Pain Thermometer (printed with permission from Keela Herr,The University of Iowa).

IPT

The IPT is a modified VDS consisting of seven pain descriptors describing different levels of pain intensity (no pain, slight pain, mild pain, moderate pain, severe pain, very severe pain, and the most intense pain imaginable) plus response options between words (scores from 0 to 12) and aligns with a PT to assist with the conceptualization of pain identified from preliminary research [19,20]. Other preliminary testing of the IPT has been reported with Caucasian and minority older adults including African American and Hispanics; and found to be reliable and valid and the preferred scale by many subjects [21,22,44]. However, the scale was not evaluated for sensitivity to change.

NRS

Available in a variety of scale ranges and anchors, the NRS is more commonly used in clinical practice and has been found to be a reliable and valid pain intensity measure [18,20]. The NRS is also recommended as a core outcome measure in clinical trials of chronic pain treatments because of more patient preference and less missing and incomplete data [45]. The NRS in this study uses numbers 0 through 20 to measure pain intensity. The patient is instructed to select the number that best reflects current intensity of pain. This scale is similar to the 21-point Box Scale supported for use with elders but uses 5-point intervals rather than consecutive numbers [46]. The NRS is preferred by many older adults; however, a substantial portion of older adults (with and without cognitive impairment) have difficulty with the scale [18,20,34,47].

VNS

The VNS requires subjects to verbally rate pain intensity from 0 to 10. Due to its ease in use by those with psychomotor or visual impairment and common use in clinical practice, the VNS was included in the pain scales evaluated. Although the VNS is the only measurement option that requires a verbal response, this verbal option is comparable to the written numeric format [48].

FPS

The FPS developed by Bieri et al. consists of seven line-drawn faces (0-6) presented in a horizontal format, ranging from a neutral face to a grimacing face [49]. Participants are instructed to point to the face that best represents the intensity of their current pain. Though originally developed for use in children, the FPS was selected over other options for several reasons: the facial depictions appeared less childlike, the absence of tears avoids potential cultural bias about pain expression, and the use of a neutral face to represent no pain instead of a happy face [20,50]. Preliminary validation testing of the FPS in the older adults was satisfactory in previous studies [50,51]. The FPS does not require reading ability or a specific language, thereby facilitating transcultural pain studies and is preferred by many older adults, most preferred by African Americans older adults [21,22,25]. However, it requires abstract thinking and has been difficult for some older adults with cognitive impairment to use and tends to correlate less strongly with other pain intensity scales [20,32,34,46].

VAS

The VAS is a vertical or horizontal 100 mm line anchored by the verbal descriptors described above. Subjects make a mark on the line that represents their current pain intensity, and pain intensity level is determined by measuring the distance between “No Pain” and the mark the subject made on the scale. There is considerable and growing evidence that the VAS is not appropriate for older adults. However, the VAS is highly sensitive to change in pain intensity and because it is a continuous variable, it is well suited for inferential statistics. Thus, it was included in this study for comparison.

Procedures

Oversight and Injection Procedures

The Institutional Internal Human Subjects Review Board approved the study. A standardized joint injection containing a combination of corticosteroid (either betamethasone sodium phosphate combined with betamethasone acetate, triamcinolone acetonide, triamcinolone hexacelonide, or methylprednisolone acetate) and lidocaine hydrochloride was selected as the intervention in order to assure a change in pain intensity. Subjects were identified through daily consultation with the nurse clinician coordinating the rheumatology clinics and review of medical records. Subjects were approached if the physician determined the patient would receive a corticosteroid injection. The research assistant (RA) informed the subject of study procedures and informed consent was obtained. For patients unable to provide informed consent, proxy consent was obtained from the legally authorized representative.

Screening and Data Collection

Following the joint injection, the RA screened for cognitive and psychomotor impairments, as well as gathered demographic information including age, gender, race, ethnicity, and educational level. After the screening, subjects were asked to rate the intensity of their current pain on the selected pain intensity scales again, approximately 15-30 minutes after their joint injection. Then subjects were asked to identify the scale that best helped them describe the severity of pain experienced. It was also believed that prior experience or familiarity with a pain scale may impact the ability to use the scale successfully, so subjects were asked if they had any prior familiarity or exposure to each of the pain scales.

The RA assigned to the study used a standardized script of instructions when introducing the scales to subjects. Following initial scale exposure, the RA provided subjects with structured responses to questions that included repetition of instructions and difficulties with understanding and utilizing the various scales were recorded. Subjects were asked to rate the intensity of their current pain on each of the scales to serve as baseline pain report (pre-injection).

Order of Presentation

Five instruments can be ordered in 5! (5 factorial), or 120 ways. Rather than generating all possible orders and assigning a unique order to all patients and thereby allowing all error associated with context effects to become part of the overall error in the model, a select number of orders were identified. In this way, this order effect and interactions between order and other factors in the design could be explicitly studied to evaluate the nature of potential context effects.

In total, 10 orders were identified under the rule that each instrument should appear at each possible position the same number of times. Under this condition, the 10 selected orders were 12,345, 23,451, 34,512, 45,123, and 51,234, and these same patterns in reverse, 54,321, 15,432, 21,543, 32,154, and 43,215. The booklets were created for each order so that all 10 orders would be presented an equal number of times. Booklets were constructed so that each subject would have the same randomized pattern of instruments before and after joint injection.

Scoring and Error Handling

All survey data were double entered into excel spreadsheets and any discrepancies between entry were identified and resolved by comparing the differences to the original survey.

Five possible failure types were coded: 1) Blank, meaning the subject did not answer the questions; 2) Between, meaning that the subject marked a location on the scale that was between admissible responses; 3) Out of Range, meaning that the subject’s response was lower or higher than the possible values on the scale (e.g., Question: “On a scale of 1-10 how bad is your pain?” Answer: 11); 4) Two Responses, meaning that the subjects indicated two answers; and 5) Multimark, which was specific to the VAS scale and meant that the subject’s mark on the scale crossed the visual analog line more than once. This could occur because the line drawn was a circle or a checkmark and crossed the line more than once or because the respondent drew several lines meant to reflect a single location, but to provide emphasis by making the line thicker or bolder.

The rules for coding a response in the face of these errors were as follows: if the response was: 1) Blank, the coder left this answer missing; 2) Between, the coder entered the number halfway between the two coded values; 3) Out of Range, the closest admissible score was coded (e.g., if on a 1-10 scale the subject indicated 11, then the value was coded as 10); 4) Two Responses, the coder entered the average of the two marked values (e.g., if 4 and 5 were marked, then 4.5); and 5) Multimark, if multiple marks were made in a small area, the midpoint value was coded and if the mark was a circle or a checkmark if knowable, the point where the mark first crossed the line was coded and if not knowable, the midpoint between the multiple line crossing was coded.

Once all values were coded, the missing data remaining because the subject left the item blank was imputed as the average estimated score for the missing value based on a series of regressions on that missing value based on subjects nonmissing responses to the other instruments.

Data Analysis

Scale Failure

Scale failure rates were determined by the frequency with which subjects were unable to correctly use the pain scale (e.g., items left blank, responses falling between two numbers, words or facial expressions, selecting more than one response or options outside the scale range) [19,20].

Preliminary analyses summarize count data regarding failure types by scale showing failure rates for each instrument at each time of assessment to evaluate failure rates at the subject level. Comparison of overall failure rates across scales was performed by using repeated measures Poisson regression (i.e., a member in the family of generalized estimating equations) to estimate the relative risk (RR) of failure responses relative to scale, age group, gender, education, and cognitive functioning based on the CCSE and Clock scores. RRs for each variable were initially evaluated, then changes or adjusted RR for failure responses across the five instruments were estimated controlling for all demographic variables demonstrated to reflect risk for failed response.

Context Effects

The order in which the five instruments were presented to patients was varied to evaluate what has been called context effects, which would be demonstrated if randomly equivalent groups were demonstrated to respond differently to on a scale based on the order in which the scales were presented. Randomly equivalent groups were established by preparing the response booklets based on instrument order. Order effects were evaluated by modeling reported pain within a general linear model (GLM) where scale (5 within subjects levels: the IPT, NRS, VNS, FPS, and VAS scales); age group (2 between subjects levels: young [21-55 years], and older [65-95 years of age]); time (2 within subjects levels: pre and post treatment), and order (10 between subjects levels) were crossed. Context effects were inferred if a significant context main effect or interactions between order and other predictors were observed. If context effects were not observed, then the order effect would be dropped from subsequent models used to evaluate scale sensitivity.

Scale Equating

Each scale was linearly transformed to a 0-100 scale to allow direct comparisons between scales. For example, the NRS (0-20) was converted to 0-100 by multiplying each score by 5 and the IPT (0-12) was converted by dividing each score by 12 and then multiplying by 100. The advantage of rescaling each scale to range from 0 to 100 by linear transformations is that it allows for a more obvious comparison of reported pain intensity without changing the correlations between the scales. In a previous study [20], conversion of all scales to range from 0 to 100 was performed by using equal percentile equating to put all scales onto the 0-100 VAS metric. However, the analysis of those results was in all respects consistent with the results obtained when the scales were all converted to range for 0-100 via linear transformation. In addition, given the known problems with the VAS scale with older patients, it was decided not to equate all instruments to the VAS scale.

Scale Sensitivity

Perhaps the most fundamental aspect of validity for a tool purported to assess pain intensity is that it demonstrate sensitivity to change in pain magnitude. Scale sensitivity was evaluated using the GLM by assessing pre- to post-injection pain reports change as identified as a scale × time interaction. Further, differential scale sensitivity across age groups was assessed by evaluating the scale × time × age group 3-way interaction. As this was an a priori hypothesis, analyses of follow-up scale × time interactions for each age group were planned regardless of the statistical significance of the overall 3-way interaction. These differences were evaluated by assessing least squares means differences in scores from pre- to posttreatment for each scale under the null hypothesis of no change, which, if the pre- and post treatment scale scores were connected by a line segment, would result in a line with a slope of 0. If context effects were detected in preliminary analyses, order and any interactions between order and the other factors in the experiment will be included in the model so that scale sensitivity will be adjusted for these context effects.

Scale Validity

The design of this study allows the assessment of concurrent validity. Concurrent validity can be assessed within the framework of this study only to the extent that differences in scale responsiveness are identified between the two cohorts. Thus, differential sensitivity across the age cohorts could be considered evidence against the concurrent validity of a pain intensity scale.

Scale Preference

Scale preference was based on a single question asking subjects to indicate which scale they preferred. Valid responses included any of the five scales or subjects could indicate no preference. Differences in preferences across scales and relative to subject demographics will be evaluated using chi-squared tests of independence.

All analyses were performed using the SAS™ System, version 9.1.2 running under the Windows XP professional operating system.

Results

Sample Description

A convenience sample of 97 adults with chronic joint pain ranging in age from 21 to 87 years (mean = 55.8; SD = 15.8) was obtained. Subjects were recruited into cohorts of 21-55 (N = 61) and 65-87 (N = 36) to more clearly differentiate younger from older subjects.

Seven subjects approached either declined to participate or were excluded from participation: all were older adults, two male and five female. Reasons for refusal to participate were unknown for five; however, cognitive impairment was suspected. One subject reporting no pain and one who was unable to consent were excluded from participation. Demographic and screening variable information from the younger and older cohorts is summarized in Table 1.

Table 1.

Demographic and screening variables of younger and older cohorts

Younger (N = 61) Older (N = 36) Test (df) P <
Age (years), mean ± SD (range) 46.0 ± 10.7 (21-55) 72.5 ± 5.6 (65-87) F1,95 = 185.7 0.0001
Gender, N (%)
 Male 15 (24.6) 15 (41.7) χ2(1) = 3.09 0.078
 Female 46 (75.4) 21 (58.3)
Education level, N (%)
 <High school 5 (8.2) 14 (38.9) χ2(2) = 16.4 0.0003
 High school 18 (29.5) 12 (33.3)
 >High school 38 (62.3) 10 (27.8)
CCSE, mean ± SD 26.5 ± 3.3 23.6 ± 4.7 F1,95 = 12.6 0.0006
CCSE <20, N (%) 2 (3.3) 8 (22.2) χ2(1) = 8.8 0.003
Psychomotor, N (%)
 0 deviation 54 (88.5) 28 (77.8) χ2(1) = 3.31 0.19
 1-2 mm deviation 7 (11.5) 8 (22.2)
Clock, mean ± SD 1.5 ± 0.7 2.2 ± 1.1 F1,95 = 16.1 0.0001
Osteoarthritis diagnoses, N (%) 5 (8.2) 10 (27.8) χ2(1) = 6.6 0.01

CCSE = Cognitive Capacity Screening Examination.

Significant differences were noted between the younger and older subjects on education level, cognitive function (CCSE, Clock), and diagnosis. Older subjects had significantly lower education levels, increased cognitive impairment, and greater percentage of osteoarthritis diagnosis versus other diagnoses of rheumatoid arthritis, fibromyalgia, bursitis, tendonitis, compared with the younger subjects in this study (P < 0.01). As 100% of the older adult sample was Caucasian and only three younger subjects were non-Caucasians, it was not possible to examine effect of race or ethnicity on study outcomes.

Scale Failure

Failure Types

The error type patterns across the instruments, as summarized in Table 2, were quite interesting. First, at least in this setting, nonresponse or leaving a scale unanswered has a low likelihood. Second, the notion of marking two answers was also not prevalent for any one scale, with the exception of the VAS scale, which was coded separately as multimark to indicate that a subject’s mark crossed the VAS scale more than once, which demonstrated very high prevalence. Third, marking between choices was high for the NRS scale, and relatively common on the VNS and FPS scales, but less so on the IPT scale, which allowed for subjects to legitimately mark a point between two labeled values. Fourth, the VNS scale, where subjects verbally indicated their response was most prone to out of range values, indicating that subject’s verbal responses were often not consistent with the standard verbal choices provided.

Table 2.

Summary of failure types by scale and time of assessment

IPT
NRS
VNS
FPS
VAS
Total
Error Type Pre Pst Pre Pst Pre Pst Pre Pst Pre Pst Pre Pst
None 93 97 91 90 78 86 90 92 84 80 435 445
Blank 0 0 1 0 0 1 1 0 0 0 2 1
Between 2 0 3 7* 4 3 4 4 0 0 13 14
Out of range 1 0 0 0 13* 7* 0 0 1 1 15 8
Two responses 1 0 2 0 2 0 2 1 1 0 8 1
Multimark 0 0 0 0 0 0 0 0 11* 16* 11 16
Total errors 5 0 6 7 19 11 7 5 13 17 50 40
*

Indicates large error rates within a specific instrument.

IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale. Pre = pre-injection; Pst = post-injection.

At the subject level, 61% (59/97) of the patients had no failures for pre-injection pain evaluations, with 29% making one tool error, 9% making two tool errors, and 1% making three tool errors. For post-injection responses, these percentages were 65%, 29%, 6%, and 0%, respectively. Across time, 47.4% (46/97) of the patients made no errors on any instrument on any occasion, 21.6% made at least one error on both the pre- and post-injection evaluations, 17.5% made errors on the pre-injection evaluation only, and 13.4% made errors on the post-injection evaluation only.

Overall Failure

Overall failure rates for each scale for the entire sample and broken down by age group and time of assessment are summarized in Table 3. Initial inspection indicated that failure rates were quite difference across the five scales, being very low across age groups and pre- and post-injection for the IPT, NRS, and FPS scales and substantially higher for the VAS and VNS scales.

Table 3.

Scale failure rates at pre- and post-injection by age group % (N) and for the entire sample

Pre-injection
Post-injection
Scale Younger (N = 61) Older (N = 36) Younger (N = 61) Older (N = 36) Total (N = 194)
IPT 1.7 (1) 8.3 (3) 0.0 0.0 2.1 (4)
NRS 3.3 (2) 11.1 (4) 3.3 (2) 13.9 (5) 6.7 (13)
VNS 19.7 (12) 19.4 (7) 13.1 (8) 8.3 (3) 15.5 (30)
FPS 6.6 (4) 8.3 (3) 4.9 (3) 5.6 (2) 6.2 (12)
VAS 13.1 (8) 13.9 (5) 18.0 (11) 16.7 (6) 15.5 (30)

IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale.

Further analysis of error rates were limited to the pre-injection evaluation for two reasons: 1) in the clinical setting pre- and postassessments of pain in such short proximity are not common; and 2) as shown in Table 3, error rates at post injection for IPT were nonexistent. Failure rates across scales, age groups, and subject demographics, including gender, education, and coded impairments, were evaluated using repeated measures Poisson regression. Results summarized in Table 4 indicated the RR of failure for each scale relative to the IPT scale was 3.25, 1.50, 4.75, and 1.75 times higher for the VAS, NRS, VNS, and FPS scales, respectively. Further, RR of scale failure was 3.14 times greater for those rated as impaired based on a CCSE score of less than 20, and 2.45 times greater for those rated as impaired based on a Clock score of 3 or more. RRs for failure were not significantly higher by age group, gender, or level of education attained.

Table 4.

Summary of unadjusted and adjusted relative risks (RR) for response failure associated with scale and patient variables

Unadjusted
Adjusted*
Variable RR 95% CI RR 95% CI
Scale
 IPT 1.00 1.00
 NRS 1.50 0.48-4.65 1.91 0.59-6.18
 VNS 4.75 1.78-12.66 2.65 0.58-12.10
 FPS 1.75 0.51-5.98 1.52 0.31-7.38
 VAS 3.25 1.21-8.66 4.14 1.40-12.18
CCSE
 Not impaired 1.00 1.00
 Impaired 3.14 1.91-5.15 2.38 1.13-4.96
Clock
 Not impaired 1.00 1.00
 Impaired 2.45 1.43-4.21 1.44 0.71-2.94
Psychomotor
 Not impaired 1.00
 Impaired 0.92 0.45-1.90
Age group
 Young (21-55) 1.00
 Old (65-87) 1.38 0.78-2.42
Gender
 Male 1.00
 Female 1.12 0.57-2.17
Education
 >High school 1.00
 High school 0.60 0.29-1.24
 <High school 1.68 0.90-3.13
*

Adjusted RR models included variables that demonstrated significant unadjusted RR for response failure.

IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale; CCSE = Cognitive Capacity Screening Examination.

When adjusting for impairment levels (based on CCSE or Clock scores), RRs for scale failure were still 4.14 time more likely on the VAS scale in comparison with the IPT scale. However, although RRs were still higher for all other scales compared with the IPT scale, these risks adjusted for impairment levels were no longer significant when type I error was held at 0.05. Interestingly, risk of scale failure remained significant at 2.38 for cognitive impairment based on CCSE scores, but was not considered reliably different for impairment based on Clock scores.

Figure 2 summarizes the percentage of cognitively impaired subjects having difficulty completing selected pain scales correctly. Consistent with the overall findings, scale failure rates were highest for the VAS score, for both patients with CCSE scores less than 20 and Clock scores greater than 2. Otherwise, scale failure rates were slightly higher for those classified as impaired by the CCSE criterion in comparison with the Clock criterion. Overall, scale failure rates were lower on the second administration, with two exceptions: 1) for the VNS scale there was an increase from one to two errors (pre- to post-injection) for patients with CCSE scores less than 20; and 2) for the VAS scale there was an increase from six to eight errors for patients with Clock scores greater than 2.

Figure 2.

Figure 2

scale failure rates by cognitively impaired subjects (%). IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale; CCSE = Cognitive Capacity Screening Examination.

Older adults had less familiarity than younger subjects with the scales used in this study and the only scale for which a substantial number of subjects had familiarity was the VNS. Familiarity with the scales by the younger cohort was as follows: VAS (6.6%), NRS (3.3%), IPT (6.6%), VNS (47.5%), and FPS (11.5%). For the older cohort, percent of familiar with each scale was as follows: VAS (2.8%), NRS (0%), IPT (2.8%), VNS (36.1%), and FPS (0%). Due to the limited number of subjects with prior familiarity with most tools, it was not possible to conduct an analysis of the impact of this factor.

Context Effects

Context effects for the instrument performance was evaluated by considering patient’s responses on each scale within a 2bAgeGroup × 10bOrder × 2wTime × 5wScale anova.

As context was considered a problem and sample size was fixed, power was increased for observing significant order effects, by increasing tolerance for type I error from 0.05 to 0.10.

Results indicated significant order by scale (F36,308 = 1.42, P < 0.062) and order by scale by time interactions (F36,308 = 1.39, P < 0.076). Given the a priori type I error rate of 0.10 for order effects, these results were considered sufficient to suggest that some context effects were present. The order by scale by time interaction is illustrated in Figure 3. The interaction did not demonstrate striking differences in reported pain intensity across scales or from pre-injection to post-injection. Reported pain intensities dropped from pre-injection to post-injection for all scales regardless of order of presentation and the relative ordering of pain intensity reports for each scale were generally quite similar regardless of order. The most striking results were that order 12,345 had significantly higher scores at T1 compared with the other scales and that the change in pre- to post-injection scores were least for order 45,123.

Figure 3.

Figure 3

Summary of order × scale × time interaction for patient-reported pain intensity. VAS = Visual Analog Scale; NRS = Numeric Rating Scale; IPT = Iowa Pain Thermometer; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale.

Scale Validity (Validity in Terms of: Scale Sensitivity and Concurrent Validity, Based on the Lack of Differences in Sensitivity Across Cohorts)

Scale Sensitivity

Given the presence of some order effects, scale sensitivity was assessed within the same 2bAge-Group × 10bOrder × 2wTime × 5wScale anova initially used to check for context effects. In this way, any context effects were controlled when evaluating scale sensitivity.

Results demonstrated significant main effects for scale (F4,308 = 21.61, P < 0.0001), time (F1,77 = 162.74, P < 0.0001), and a significant scale by time interaction (F4,308 = 2.86, P < 0.024). As expected, examination of mean pain scores pre and post injection indicated a strong decrease in reported pain at post-injection for all five instruments. Tukey’s studentized Range tests for highest significant differences between pairs of scores were used to evaluate the scale main effect and revealed that, overall, reported pain intensities were significantly higher for the VNS scale (47.2) and were significantly lower than all others for the FPS (39.4) with the VAS, NRS, and IPT scales (43.9, 43.7, and 42.1, respectively) not significantly different from each other.

The significant scale × time interaction indicated that the magnitude of change from pre- to post-injection was differential across the pain scales. Table 5 summarizes the pre- and post-injection pain ratings for all five scales for the entire sample and broken down by age groups. Follow-up tests indicated significant decreased in reported pain for all five scales in both age groups and for the entire sample, with all statistical tests significant at P < 0.0001. Thus, all scales were sensitive in detecting change in pain report within this experimental study. Figure 4 summarizes the results of the simple effects tests performed to investigate the nature of the scale by time interaction and reveals that the interaction reflected statistically significant but not clinically important differences in slopes. In this case, the differential drop in pain intensity was observed only for the VAS compared with the FPS scales. However, the VAS scales magnitude drop was not significantly different compared with the NRS, IPT, and VNS scales and, similarly, the FPS scale magnitude drop was not significantly different compared with the NRS, IPT, and VNS scales.

Table 5.

Means and standard errors for pre- and post-injection pain intensity scores by scale for all subjects, and the older and younger subject cohorts

All Subjects (N = 97)
Older Subjects (N-36)
Younger Subjects (N = 61)
Pre-injection
Post-injection
Pre-injection
Post-injection
Pre-injection
Post-injection
Scale Mean Se Mean Se Mean Se Mean Se Mean Se Mean Se
IPT 58.44 0.82 25.90 0.82 59.79 1.31 26.81 1.31 57.09 1.00 25.00 1.00
NRS 60.49 0.82 27.42 0.82 62.78 1.31 30.31 1.31 58.19 1.00 24.52 1.00
VNS 64.88 0.82 30.17 0.82 68.13 1.31 32.29 1.31 61.63 1.00 28.06 1.00
FPS 54.91 0.82 24.08 0.82 57.36 1.31 24.72 1.31 52.47 1.00 23.43 1.00
VAS 62.28 0.82 26.31 0.82 65.43 1.31 28.23 1.31 59.14 1.00 24.39 1.00

Means are adjusted for all effects in model, which include: age group, instrument order, scale, time of assessment, and all interactions among these factors.

Se = standard error; IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale.

Figure 4.

Figure 4

Slopes for pre-injection to post-injection pain intensity scores by scale for older and younger cohorts and the total sample. VAS = Visual Analog Scale; NRS = Numeric Rating Scale; IPT = Iowa Pain Thermometer; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale.

As demonstrated in Table 5 and Figure 4, the pattern of change from pre- to post-injection was similar across the two age cohorts, which was consistent with the lack of a significant age group × time × scale interaction, F36,308 = 0.63, P < 0.64.

Concurrent Validity

Pearson correlations between subjects’ pain intensity ratings when the five pain scales were administered consecutively both before and after receiving a steroid injection with lidocaine are summarized in Table 6. All scales correlated strongly at pre- and post-injection for both the younger and the older adult cohorts providing evidence of construct validity; however, the correlations for the older cohort were lower than for the younger sample although still acceptable. Weakest correlations at pre-injection were noted between the FPS and both VNS and IPT. Weakest correlations at post-injection were noted between FPS and both NRS and VAS. The FPS tended to correlate less strongly with other scales.

Table 6.

Pearson correlations between scales for the younger and older cohorts at pre- and post-injection

Time Scale Cohort IPT NRS VNS FPS VAS
Pre-injection IPT Younger 0.92 0.92 0.88 0.90
Older 0.79 0.70 0.68 0.72
NRS Younger 0.92 0.87 0.95
Older 0.79 0.71 0.91
VNS Younger 0.88 0.89
Older 0.67 0.77
FPS Younger 0.86
Older 0.78
VAS Younger
Older
Post-injection IPT Younger 0.95 0.92 0.92 0.93
Older 0.81 0.82 0.84 0.80
NRS Younger 0.90 0.92 0.96
Older 0.96 0.79 0.95
VNS Younger 0.92 0.91
Older 0.80 0.92
FPS Younger 0.91
Older 0.75
VAS Younger
Older

All correlations were significantly greater than 0 at P < 0.05.

IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale.

Scale Preference

The relationships between scale preference and selected subject demographics are summarized in Table 7. Preference rates were significantly different across the five instruments (χ52 = 24.67, P < 0.0002). Overall, 10 of the 97 or 10.3% of the subjects expressed no preference. However, as chisquared tests may be unreliable when cell counts are less than 5, and because some cell counts were less than 5 in the age group, gender, education, CCSE, and Clock and psychomotor cross classifications, little emphasis should be placed on the statistical significance level for these analyses. Nevertheless, note that for 12 of the 14 lines, the IPT scale demonstrated the highest patient preferences. On the two occasions where IPT was not the highest preference, FPS had the highest preference. The IPT scale did not rate highly only for the subgroup of 30 patients whose education ended with a high school diploma.

Table 7.

Summary of relationships between scale preference and selected subject factors

Scale Preference
Variable N None N = 10 IPT N = 31 NRS N = 16 VNS N = 15 FPS N = 20 VAS N = 5 χdf2 df P <
Scale 97 10.31 31.96* 16.49 15.46 20.62 5.15 24.67 5 0.0002
Age group (years) 4.04 5 0.55
 Older 36 16.67 27.78 13.89 13.89 19.44 8.33
 Younger 61 6.56 34.43 18.03 16.39 21.31 3.28
Gender 6.79 5 0.24
 Female 67 8.96 32.84 17.91 10.45 22.39 7.46
 Male 30 13.33 30.00 13.33 26.67 16.67 0.00
Education 21.97 10 0.02
 <High school 19 26.32 31.58 5.26 5.26 26.32 5.26
 High school 30 13.33 13.33 23.33 13.33 26.67 10.00
 >High school 48 2.08 43.75 16.67 20.83 14.58 2.08
CCSE 3.54 5 0.62
 Impaired 10 10.00 20.00 10.00 10.00 40.00 10.00
 Not impaired 87 10.34 33.33 17.24 16.09 18.39 4.60
Clock 2.88 5 0.72
 Impaired 16 18.75 31.25 6.25 12.50 25.00 6.25
 Not impaired 81 8.64 32.10 18.52 16.05 19.75 4.94
Psychomotor 3.42 5 0.64
 Impaired 15 0.00 26.67 26.67 20.00 20.00 6.67
 Not impaired 82 12.20 32.93 14.63 14.63 20.73 4.88
*

Percentages in bold indicate the highest rate for that row of data.

IPT = Iowa Pain Thermometer; NRS = Numeric Rating Scale; VNS = Verbal Numeric Rating Scale; FPS = Faces Pain Scale; VAS = Visual Analog Scale; CCSE = Cognitive Capacity Screening Examination.

When we were over halfway through data collection and it appeared that the IPT was a preferred scale, we considered that we had added two components to the original VDS: additional options between words and the thermometer. So the remaining 39 subjects were asked as to the importance of the thermometer along the word descriptors. Of these subjects, 86.7% (N = 13) of the older adults and 91.7% (N = 22) of the younger subjects indicated that the thermometer was not important to them in using this scale. To verify that subjects were using the in-between options on the IPT, the frequency of responses to each option was examined. Subjects selected all the intermediate options designated by bubbles.

Discussion

This preliminary report provides psychometric evaluation data on a new pain intensity scale designed to improve sensitivity and utility primarily for research purposes. The study also provides follow-up evaluation of four pain intensity scales from an earlier study that evaluated psychometrics and utility with older adults using an experimental pain stimulus [20]. Given that the same factors shaping responses to experimental pain stimuli may also contribute to the experience of clinical pain [52], it is important to examine the sensitivity and utility of the selected scales in a sample with clinical pain. The method was chosen to assure changes in pain intensity and to observe the impact of age and cognitive impairment on report of pain intensity in selected scales. The findings from this study provide preliminary support for further testing of the IPT and additional support for tools evaluated in prior research.

Scale Failure

Selection of a scale that the highest percentage of subjects can complete would provide a sound basis for scale selection if tool validity has been established. In this pilot study, the IPT demonstrated the lowest failure rates suggesting ease of use in older persons including those with cognitive impairment. Of the scales examined in this study, the IPT, NRS, and FPS are clear winners when considering the RR for failure and user ability to provide a self-report successfully for both research and clinical contexts where obtaining self-report from the most older adults possible is advantageous.

As expected, failure rates on the VAS were high (13.1-18.0%) in both younger and older cohort, which is consistent with error rates reported by other studies [20,53,54]. Although the VAS is highly sensitive to detect change in pain intensity, it would not be recommended for use in research or clinical settings with the older adult population because of its high failure rate. Also of interest, regarding familiarity with the selected scales, was that subjects only reported considerable familiarity with the VNS. This is not particularly surprising given that the verbal 0-10 scale is the most common scale used in clinical practice. But it was notable that scale failure rates in older subjects were relatively high for the VNS (8.3-19.4%) and higher than that reported in the literature [46,48]. This scale may be difficult because it is verbal without a written cue, thus requiring abstract conceptualization of pain. These findings reinforce the importance of administering an NRS in written format rather than verbal with older adults. The VNS also solicited higher pain scores, and although the reason for this is uncertain, it may have to do with expectation demand or desire to influence the rater that might have influence in decisions about pain treatment.

Evidence of psychomotor impairment was limited in this sample and did not significantly affect ability to use any of the pain scales. This study demonstrated that cognitive impairment was significantly related to failure on the VAS and NRS, using the CCSE. The Clock drawing cut-off of 3 or greater reflects moderate visuospatial impairment. Using this approach, only the VAS would be inappropriate to try with a cognitively impaired sample. Mild cognitive impairment did not seem to impede the ability to self-report pain and attempts should be made to solicit self-report if possible [29]. Behavioral observation and surrogate reporting may be necessary for more severe cognitive impairment in which self-report is not possible [55].

In this study, most of the five scales’ failure rates decreased from pre- to post-injection. Subjects may have been more familiar with the scales at that point and also experiencing less pain that could impact attention and focus on the task. This may also impact increased correlations between scales found at post-injection.

Scale Validity

All the scales examined in this study, including the IPT, were sensitive in reflecting changes in pain intensity, providing evidence of scale criterion validity. Although a goal of this study was to investigate the possibility of differential validity in these scales across age cohorts, which could be construed as evidence for concurrent validity, the universal sensitivity of all scales to decreased pain intensity after injection suggests that these scales were all valid for the purpose of estimating pain intensity, regardless of the respondents’ age. Though the correlations between scales for the older cohort were lower than for the younger sample, the strength of the correlations is consistent with earlier findings, supporting validity of the selected scales’ measurement of a single construct, pain intensity [20-22,34,35]. The FPS tended to correlate less strongly with other scales, especially with the VAS, similar as reported in previous studies [20-22] and may be related to the restricted range or number of categories in this scale.

Scale Preference

Scale preference reported in this study is similar to that of other researchers with VDSs, including the IPT adaptation of a VDS, preferred by many older adults [18-21]. The FPS was the second most preferred scale in both age groups, although it is not often offered as an option for younger adults. These preferences were similar in those with cognitive impairment. The high preference for the FPS is of interest and supports earlier work in African Americans [21,22]. It is still important to note that although these scales are most preferred, there is still individual variability and, at least in clinical settings, alternative options should be available for those that may not be able to use a selected institutional scale.

Limitations

Findings from this study are limited by the small sample size in general and especially in the older adult cohort and the use of a convenience sample, which contributed to a limited racially diverse and cognitively impaired sample and lack of power to detect differences. However, a recent report of follow-up testing of the IPT in a sample with African American and Hispanic elders provides additional psychometric evaluation in racially diverse and cognitively impaired elders. Moreover, the FPS was recently revised by Hicks et al. to six faces as FPS-R to make it suitable for use with 0-10 scale [56]. It has been validated in Spanish older adults [25] and older minority adults in the United States [44], but not in Caucasian elders. Further clinical controlled studies exploring the sensitivity of IPT, FPS-R, and other pain intensity scales including racially diverse older adults are needed.

Conclusions

In this study, we adapted the VDS to the IPT by adding choices of severity and including a thermometer. It is noteworthy that of the selected pain intensity scales evaluated, the IPT rises to the top as a good choice for older (and younger) adults, including those with mild to moderate cognitive impairment. The failure rate of IPT was very low and the revised scale provides increased options thus increasing its sensitivity, particularly important for evaluating research intervention outcomes. Given that a small subsample indicated that the thermometer may not be a necessary component to the scale, future study is needed to verify the significance of the thermometer versus the additional option choices provided in this revised VDS.

Clinicians should consider scales that have established psychometric properties for older adults and assure that scales address individual needs (e.g., sensory, cognitive, education). Overall, all pain intensity scales with the exception of the VAS are good options for younger or older adults to report pain intensity, but the IPT appears to be a better choice for older adults. Those with lower education levels and cognitive impairment would be best served with the IPT or FPS, because of low failure rates and preference by this population as well.

Acknowledgment

This study was supported by National Institutes of Health National Institute of Nursing Research grant (PI: Herr; R29 NR-03528).

References

  • 1.Bernabei R, Gambassi G, Lapane K, et al. Management of pain in elderly patients with cancer. SAGE Study Group. Systematic Assessment of Geriatric Drug Use via Epidemiology. JAMA. 1998;279:1877–82. doi: 10.1001/jama.279.23.1877. [DOI] [PubMed] [Google Scholar]
  • 2.Desbiens NA, Mueller-Rizner N, Connors AF, Jr, Hamel MB, Wenger NS. Pain in the oldest-old during hospitalization and up to one year later. HELP Investigators. Hospitalized Elderly Longitudinal Project. J Am Geriatr Soc. 1997;45:1167–72. doi: 10.1111/j.1532-5415.1997.tb03765.x. [DOI] [PubMed] [Google Scholar]
  • 3.Elliott AM, Smith BH, Penny KI, Smith WC, Chambers WA. The epidemiology of chronic pain in the community. Lancet. 1999;354:1248–52. doi: 10.1016/s0140-6736(99)03057-3. [DOI] [PubMed] [Google Scholar]
  • 4.Ferrell BA, Ferrell BR, Osterweil D. Pain in the nursing home. J Am Geriatr Soc. 1990;38:409–14. doi: 10.1111/j.1532-5415.1990.tb03538.x. [DOI] [PubMed] [Google Scholar]
  • 5.Ferrell BA, Ferrell BR, Rivera L. Pain in cognitively impaired nursing home patients. J Pain Symptom Manage. 1995;10:591–8. doi: 10.1016/0885-3924(95)00121-2. [DOI] [PubMed] [Google Scholar]
  • 6.Herrick C, Steger-May K, Sinacore DR, Brown M, Schechtman KB, Binder EF. Persistent pain in frail older adults after hip fracture repair. J Am Geriatr Soc. 2004;52:2062–8. doi: 10.1111/j.1532-5415.2004.52566.x. [DOI] [PubMed] [Google Scholar]
  • 7.Jacobs JM, Hammerman-Rozenberg R, Cohen A, Stessman J. Chronic back pain among the elderly. Prevalence, associations, and predictors. Spine. 2006;31:E203–7. doi: 10.1097/01.brs.0000206367.57918.3c. [DOI] [PubMed] [Google Scholar]
  • 8.Thomas E, Peat G, Harris L, Wilkie R, Croft PR. The prevalence of pain and pain interference in a general population of older adults: Cross-sectional findings from the North Staffordshire Osteoarthritis Project (NorStOP) Pain. 2004;110:361–8. doi: 10.1016/j.pain.2004.04.017. [DOI] [PubMed] [Google Scholar]
  • 9.Weiner D, Peterson B, Ladd K, McConnell E, Keefe F. Pain in nursing home residents. An exploration of prevalence, staff perspectives, and practical aspects of measurement. Clin J Pain. 1999;15:92–101. doi: 10.1097/00002508-199906000-00005. [DOI] [PubMed] [Google Scholar]
  • 10.Won AB, Lapane KL, Vallow S, Schein J, Morris JN, Lipsitz LA. Persistent nonmalignant pain and analgesic prescribing patterns in elderly nursing home residents. J Am Geriatr Soc. 2004;52:867–74. doi: 10.1111/j.1532-5415.2004.52251.x. [DOI] [PubMed] [Google Scholar]
  • 11.Herr K, Titler MG, Schilling ML, et al. Evidence-based assessment of acute pain in older adults: Current nursing practices and perceived barriers. Clin J Pain. 2004;20:331–40. doi: 10.1097/00002508-200409000-00008. [DOI] [PubMed] [Google Scholar]
  • 12.Titler MG, Herr K, Schilling ML, et al. Acute pain treatment for older adults hospitalized with hip fracture: Current nursing practices and perceived barriers. Appl Nurs Res. 2003;16:211–27. doi: 10.1016/s0897-1897(03)00051-x. [DOI] [PubMed] [Google Scholar]
  • 13.Feldt KS, Ryden MB, Miles S. Treatment of pain in cognitively impaired compared with cognitively intact older patients with hip-fracture. J Am Geriatr Soc. 1998;46:1079–85. doi: 10.1111/j.1532-5415.1998.tb06644.x. [DOI] [PubMed] [Google Scholar]
  • 14.Gagliese L, Melzack R. Chronic pain in elderly people. Pain. 1997;70:3–14. doi: 10.1016/s0304-3959(96)03266-6. [DOI] [PubMed] [Google Scholar]
  • 15.Barkin RL, Barkin SJ, Barkin DS. Perception, assessment, treatment, and management of pain in the elderly. Clin Geriatr Med. 2005;21:465–90. v. doi: 10.1016/j.cger.2005.02.006. [DOI] [PubMed] [Google Scholar]
  • 16.Jensen MP. The validity and reliability of pain measures in adults with cancer. J Pain. 2003;4:2–21. doi: 10.1054/jpai.2003.1. [DOI] [PubMed] [Google Scholar]
  • 17.Herr KA, Garand L. Assessment and measurement of pain in older adults. Clin Geriatr Med. 2001;17:457–78. vi. doi: 10.1016/s0749-0690(05)70080-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Gagliese L, Weizblit N, Ellis W, Chan VW. The measurement of postoperative pain: A comparison of intensity scales in younger and older surgical patients. Pain. 2005;117:412–20. doi: 10.1016/j.pain.2005.07.004. [DOI] [PubMed] [Google Scholar]
  • 19.Herr KA, Mobily PR. Comparison of selected pain assessment tools for use with the elderly. Appl Nurs Res. 1993;6:39–46. doi: 10.1016/s0897-1897(05)80041-2. [DOI] [PubMed] [Google Scholar]
  • 20.Herr KA, Spratt K, Mobily PR, Richardson G. Pain intensity assessment in older adults: Use of experimental pain to compare psychometric properties and usability of selected pain scales with younger adults. Clin J Pain. 2004;20:207–19. doi: 10.1097/00002508-200407000-00002. [DOI] [PubMed] [Google Scholar]
  • 21.Taylor LJ, Harris J, Epps CD, Herr K. Psychometric evaluation of selected pain intensity scales for use with cognitively impaired and cognitively intact older adults. Rehabil Nurs. 2005;30:55–61. doi: 10.1002/j.2048-7940.2005.tb00360.x. [DOI] [PubMed] [Google Scholar]
  • 22.Taylor LJ, Herr K. Pain intensity assessment: A comparison of selected pain intensity scales for use in cognitively intact and cognitively impaired African American older adults. Pain Manag Nurs. 2003;4:87–95. doi: 10.1016/s1524-9042(02)54210-7. [DOI] [PubMed] [Google Scholar]
  • 23.Weiner DK, Peterson BL, Logue P, Keefe FJ. Predictors of pain self-report in nursing home residents. Aging (Milano) 1998;10:411–20. doi: 10.1007/BF03339888. [DOI] [PubMed] [Google Scholar]
  • 24.Weiner D, Peterson B, Keefe F. Evaluating persistent pain in long term care residents: What role for pain maps? Pain. 1998;76:249–57. doi: 10.1016/s0304-3959(98)00059-1. [DOI] [PubMed] [Google Scholar]
  • 25.Miro J, Huguet A, Nieto R, Paredes S, Baos J. Evaluation of reliability, validity, and preference for a pain intensity scale for use with the elderly. J Pain. 2005;6:727–35. doi: 10.1016/j.jpain.2005.06.005. [DOI] [PubMed] [Google Scholar]
  • 26.Benesh LR, Szigeti E, Ferraro FR, Gullicks JN. Tools for assessing chronic pain in rural elderly women. Home Healthc Nurse. 1997;15:207–11. doi: 10.1097/00004045-199703000-00009. [DOI] [PubMed] [Google Scholar]
  • 27.Jensen MP, Chen C, Brugger AM. Postsurgical pain outcome assessment. Pain. 2002;99:101–9. doi: 10.1016/s0304-3959(02)00063-5. [DOI] [PubMed] [Google Scholar]
  • 28.Gagliese L, Katz J. Age differences in postoperative pain are scale dependent: A comparison of measures of pain intensity and quality in younger and older surgical patients. Pain. 2003;103:11–20. doi: 10.1016/s0304-3959(02)00327-5. [DOI] [PubMed] [Google Scholar]
  • 29.AGS Panel on Persistent Pain in Older Persons The management of persistent pain in older persons. J Am Geriatr Soc. 2002;50(suppl 6):S205–24. doi: 10.1046/j.1532-5415.50.6s.1.x. [DOI] [PubMed] [Google Scholar]
  • 30.Herr KA, Mobily PR. Complexities of pain assessment in the elderly. Clinical considerations. J Gerontol Nurs. 1991;17:12–9. doi: 10.3928/0098-9134-19910401-04. [DOI] [PubMed] [Google Scholar]
  • 31.Manz BD, Mosier R, Nusser-Gerlach MA, Bergstrom N, Agrawal S. Pain assessment in the cognitively impaired and unimpaired elderly. Pain Manag Nurs. 2000;1:106–15. doi: 10.1053/jpmn.2000.19332. [DOI] [PubMed] [Google Scholar]
  • 32.Krulewitch H, London MR, Skakel VJ, Lundstedt GJ, Thomason H, Brummel-Smith K. Assessment of pain in cognitively impaired older adults: A comparison of pain assessment tools and their use by nonprofessional caregivers. J Am Geriatr Soc. 2000;48:1607–11. doi: 10.1111/j.1532-5415.2000.tb03871.x. [DOI] [PubMed] [Google Scholar]
  • 33.Fisher SE, Burgio LD, Thorn BE, Hardin JM. Obtaining self-report data from cognitively impaired elders: Methodological issues and clinical implications for nursing home pain assessment. Gerontologist. 2006;46:81–8. doi: 10.1093/geront/46.1.81. [DOI] [PubMed] [Google Scholar]
  • 34.Closs SJ, Barr B, Briggs M, Cash K, Seers K. A comparison of five pain assessment scales for nursing home residents with varying degrees of cognitive impairment. J Pain Symptom Manage. 2004;27:196–205. doi: 10.1016/j.jpainsymman.2003.12.010. [DOI] [PubMed] [Google Scholar]
  • 35.Pautex S, Herrmann F, Le Lous P, Fabjan M, Michel JP, Gold G. Feasibility and reliability of four pain self-assessment scales and correlation with an observational rating scale in hospitalized elderly demented patients. J Gerontol A Biol Sci Med Sci. 2005;60:524–9. doi: 10.1093/gerona/60.4.524. [DOI] [PubMed] [Google Scholar]
  • 36.McDougall GJ. A review of screening instruments for assessing cognition and mental status in older adults. Nurse Pract. 1990;15:18–28. [PMC free article] [PubMed] [Google Scholar]
  • 37.Jacobs JW, Bernhard MR, Delgado A, Strain JJ. Screening for organic mental syndromes in the medically ill. Ann Intern Med. 1977;86:40–6. doi: 10.7326/0003-4819-86-1-40. [DOI] [PubMed] [Google Scholar]
  • 38.Shulman KI, Shedletsky R, Silver IL. The challenge of time: Clock-drawing and cognitive function in the elderly. Int J Geriatr Psychiatry. 1986;1:135–40. [Google Scholar]
  • 39.Wolf-Klein GP, Silverstone FA, Levy AP, Brod MS. Screening for Alzheimer’s disease by clock drawing. J Am Geriatr Soc. 1989;37:730–4. doi: 10.1111/j.1532-5415.1989.tb02234.x. [DOI] [PubMed] [Google Scholar]
  • 40.Cohen D, Wu S. Language and cognition during aging. In: Eisendorfer C, editor. Annual Review of Gerontology and Geriatrics. Vol. 1. Springer Publishing; New York: 1980. pp. 71–96. [Google Scholar]
  • 41.Pierce CA, Jewell G, Mennemeier M. Are psychophysical functions derived from line bisection reliable? J Int Neuropsychol Soc. 2003;9:72–8. doi: 10.1017/s1355617703910083. [DOI] [PubMed] [Google Scholar]
  • 42.Price DD, Harkins SW. Psychophysical approaches to pain measurement and assessment. In: Turk D, Melzack R, editors. Handbook of Pain Assessment. The Gilford Press; New York: 1992. pp. 111–34. [Google Scholar]
  • 43.Kremer E, Atkinson JH, Ignelzi RJ. Measurement of pain: Patient preference does not confound pain measurement. Pain. 1981;10:241–8. doi: 10.1016/0304-3959(81)90199-8. [DOI] [PubMed] [Google Scholar]
  • 44.Ware LJ, Epps CD, Herr K, Packard A. Evaluation of the Revised Faces Pain Scale, Verbal Descriptor Scale, Numeric Rating Scale, and Iowa Pain Thermometer in older minority adults. Pain Manag Nurs. 2006;7:117–25. doi: 10.1016/j.pmn.2006.06.005. [DOI] [PubMed] [Google Scholar]
  • 45.Dworkin RH, Turk DC, Farrar JT, et al. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain. 2005;113:9–19. doi: 10.1016/j.pain.2004.09.012. [DOI] [PubMed] [Google Scholar]
  • 46.Chibnall JT, Tait RC. Pain assessment in cognitively impaired and unimpaired older adults: A comparison of four scales. Pain. 2001;92:173–86. doi: 10.1016/s0304-3959(00)00485-1. [DOI] [PubMed] [Google Scholar]
  • 47.Kaasalainen S, Crook J. An exploration of seniors’ ability to report pain. Clin Nurs Res. 2004;13:199–215. doi: 10.1177/1054773804265692. [DOI] [PubMed] [Google Scholar]
  • 48.Paice JA, Cohen FL. Validity of a verbally administered numeric rating scale to measure cancer pain intensity. Cancer Nurs. 1997;20:88–93. doi: 10.1097/00002820-199704000-00002. [DOI] [PubMed] [Google Scholar]
  • 49.Bieri D, Reeve RA, Champion GD, Addicoat L, Ziegler JB. 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. doi: 10.1016/0304-3959(90)90018-9. [DOI] [PubMed] [Google Scholar]
  • 50.Herr KA, Mobily PR, Kohout FJ, Wagenaar D. Evaluation of the Faces Pain Scale for use with the elderly. Clin J Pain. 1998;14:29–38. doi: 10.1097/00002508-199803000-00005. [DOI] [PubMed] [Google Scholar]
  • 51.Stuppy DJ. The Faces Pain Scale: Reliability and validity with mature adults. Appl Nurs Res. 1998;11:84–9. doi: 10.1016/s0897-1897(98)80229-2. [DOI] [PubMed] [Google Scholar]
  • 52.Edwards RR, Sarlani E, Wesselmann U, Fillingim RB. Quantitative assessment of experimental pain perception: Multiple domains of clinical relevance. Pain. 2005;114:315–9. doi: 10.1016/j.pain.2005.01.007. [DOI] [PubMed] [Google Scholar]
  • 53.Briggs M, Closs JS. A descriptive study of the use of visual analogue scales and verbal rating scales for the assessment of postoperative pain in orthopedic patients. J Pain Symptom Manage. 1999;18:438–46. doi: 10.1016/s0885-3924(99)00092-5. [DOI] [PubMed] [Google Scholar]
  • 54.Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: A comparison of six methods. Pain. 1986;27:117–26. doi: 10.1016/0304-3959(86)90228-9. [DOI] [PubMed] [Google Scholar]
  • 55.Herr K, Coyne PJ, Key T, et al. Pain assessment in the nonverbal patient: Position statement with clinical practice recommendations. Pain Manag Nurs. 2006;7:44–52. doi: 10.1016/j.pmn.2006.02.003. [DOI] [PubMed] [Google Scholar]
  • 56.Hicks CL, von Baeyer CL, Spafford PA, van Korlaar I, Goodenough B. The Faces Pain Scale-Revised: Toward a common metric in pediatric pain measurement. Pain. 2001;93:173–83. doi: 10.1016/S0304-3959(01)00314-1. [DOI] [PubMed] [Google Scholar]

RESOURCES