Primary Image

RehabMeasures Instrument

West Haven-Yale Multidimensional Pain Inventory

Last Updated

Purpose

The WHYMPI assesses chronic pain in individuals and is recommended for use in conjunction with behavioral and psycho-physiological strategies.

Link to Instrument

Instrument Details

Acronym WHYMPI (MPI)

Area of Assessment

Pain

Assessment Type

Patient Reported Outcomes

Administration Mode

Paper & Pencil

Cost

Free

Diagnosis/Conditions

  • Arthritis + Joint Conditions
  • Brain Injury Recovery
  • Multiple Sclerosis
  • Pain Management
  • Spinal Cord Injury
  • Stroke Recovery

Key Descriptions

  • The WHYMPI is based on a cognitive-behavioral theory of pain emphasizing the role of cognitive, emotional, and behavioral factors in the experience of pain and related disability.
  • The MPI is composed of 52 items across twelve subscales and three overall domains.
  • The first domain assesses the pain experience, with the subscales measuring:
    1) Patient’s perceived interference of pain in various areas of functioning
    2) Patients' perceptions of support received by others
    3) The level of pain severity
    4) The patients' perceptions of control
    5) Patients' level of affective distress
  • The second domain assesses the responses of others to the patient's communicated pain, with subscales measuring:
    1) The patients' perceived frequency of punishing responses
    2) The patients' perceived frequency of solicitous Responses
    3) The patients' perceived frequency of distracting responses
  • The third domain assesses the extent to which patients participate in daily activities with subscales measuring:
    1) The performance of household chores
    2) The performance of outdoor work
    3) The performance of activities away from home
    4) The performance of social activities
  • Each item is rated on a 0-6 scale, and the scores for each subscale are calculated by adding the scores for each item in that subscale, divided by the number of items in that subscale to yield a mean score.
  • Higher scores indicate a greater intensity in that subscale. For example, higher subscale scores for pain interference and severity indicate greater levels of pain and interference in life, whereas higher subscale scores for activity indicate greater ability in daily activities.

Number of Items

52

Time to Administer

15-30 minutes

Required Training

No Training

Age Ranges

Adult

18 - 64

years

Elderly Adult

65 +

years

Instrument Reviewers

Allison Peipert, Edeth Engel

Body Part

Upper Extremity
Back
Lower Extremity

ICF Domain

Body Function

Measurement Domain

Sensory

Considerations

Translated into several languages:

  • Spanish 
  • German
  • Dutch
  • Swedish
  • Turkish
  • Chinese

Do you see an error or have a suggestion for this instrument summary? Please e-mail us!

Cancer

back to Populations

Criterion Validity (Predictive/Concurrent)

Cancer: (Cetin et al., 2016; n = 520; mean age= 52.05 (10.06) years; Turkish Sample- translated version)

  • The criterion validity was determined through item analysis of the differences between averages of upper-lower group item scores. The differences between the mean scores of the items of the upper and lower groups were statistically significant.

Construct Validity

Cancer: (Cetin et al., 2016; n = 520; mean age= 52.05 (10.06) years; Turkish Sample- translated version)

  • Convergent/divergent validity of the Turkish WHYMPI was investigated by correlation with the Turkish Beck Depression Inventory (BDI). Results found that section 1 of the WHYMPI had statistically significant positive correlations with the BDI. When perceived pain intensity increased, depression also increased. They also found that section 3 of the WHYMPI had statistically significant negative correlations with the BDI. When patients' performance of common activities increased, depression decreased.

Content Validity

Cancer: (Cetin et al., 2016; n = 520; mean age= 52.05 (10.06) years; Turkish Sample- translated version)

  • Content validity was established through seven experts. The Kendall’s coefficient of concordance was nonsignificant across the sample (Kendall’s W = 0.188, p = 0.064). This indicates that there is no statistically significant difference between between the expert opinions of each item

Chronic Pain

back to Populations

Standard Error of Measurement (SEM)

 Chronic Pain Patients: (Kerns et al, 1985; n = 120; mean age = 50.8 (14.5) years; mean duration to pain = 10.2 years; 36.4% reported lower back pain)

Subscale

SEM

Interference

0.398

Support

0.616

Pain Severity

0.587

Self-Control

0.719

Negative Mood

0.686

Punishing Responses

0.376

Solicitous Responses

0.539

Distracting Responses

0.566

Household Chores

0.486

Outdoor Work

0.499

Activities Away from Home

0.455

Social Activities

0.484

Normative Data

Chronic Pain Patients: (Kerns et al, 1985; n = 120; mean age = 50.8 (14.5) years; mean duration to pain = 10.2 years; 36.4% reported lower back pain)


WHYMPI Normative Data:

 


Domain

Mean (SD)

Interference

3.74 (1.26)

Support

4.31 (1.47)

Pain Severity

3.55 (1.11)

Self-control

3.63 (1.57)

Negative Mood

3.23 (1.32)

Punishing response

0.97 (0.94)

Solicitious response

2.57 (1.15)

Distracting responses

1.72 (1.31)

Household chores

2.71 (1.30)

Outdoor work

1.19 (1.04)

Activities away from home

1.79 (0.83)

Social activities

1.94 (0.95)

Test/Retest Reliability

Chronic Musculoskeletal Pain: (Bergstrom et al, 1998; n = 682; mean age = 41.8 (9.4); duration of pain = 67.9 (70.4) months; Swedish sample)

Test-retest Data:

Scale

Mean (SD)

Strength

Test-retest ICC

Pain Severity

3.73 (1.16)

Adequate

0.75

Interference

3.74 (1.08)

Excellent

0.85

Life Control

3.12 (1.16)

Excellent

0.81

Affective Distress

2.95 (1.49)

Adequate

0.74

Support

4.30 (1.57)

Excellent

0.88

Punishing Responses

1.23 (1.29)

Adequate

0.75

Solicitous Responses

2.95 (1.35)

Adequate

0.73

Distracting Responses

3.14 (1.60)

Excellent

0.76

Household Chores

3.58 (1.39)

Excellent

0.89

Outdoor Work

1.68 (1.46)

Excellent

0.86

Leisure Activities

3.08 (1.18)

Excellent

0.79

General Activity

2.97 (0.89)

Excellent

0.80

Chronic pain (Lousberg et al., 1999; = 733; mean age = 42.5 (9.7) Dutch sample- translated version)

Scale

Test Retest ICC

Strength

Pain Severity

0.71

Adequate

Interference

0.87

Excellent

Life Control

0.69

Adequate

Affective Distress

0.74

Adequate

Support

0.88

Excellent

Punishing Responses

0.79

Excellent

Solicitous Responses

0.78

Excellent

Distracting Responses

0.75

Adequate

Household Chores

0.94

Excellent

Outdoor Work

0.85

Excellent

Activities Away from Home

0.88

Excellent

Social Activities

0.90

Excellent

General Activity

0.87

Excellent

Chronic Pain: (Verra et al., 2012 n=204 for test-retest correlation, mean age= 46.8 (16.7–72.8), Swiss sample)

Test-retest reliability of mean MPI scale scores and Internal consistency of the items at scale level for all subjects (n = 204)

MPI subscales

T0 (m,s)

T1 (m,s)

ICC (95% CI)

CA (95% CI)

Pain severity (100 = most)

76.1 (16.5)

73.9 (16.0)

0.77 (0.70–0.82) Excellent

0.81 (0.77–0.86)

Interference (100 = worst)

74.7 (15.9)

72.5 (16.1)

0.82 (0.77–0.86) Excellent

0.83 (0.76–0.86)

Life control (100 = best)

49.4 (22.1)

50.2 (21.4)

0.57 (0.47–0.66) Adequate

0.76 (0.69–0.81)

Negative mood (100 = worst)

61.0 (19.6)

58.4 (22.1)

0.72 (0.65–0.78) Adequate

0.60 (0.49–0.69)

Support (100 = most)

69.4 (24.8)

69.0 (26.2)

0.85 (0.80–0.88) Excellent

0.82 (0.77–0.86)

Negative responses (100 = most)

25.6 (26.1)

24.3 (24.4)

0.75 (0.69–0.81) Adequate

0.86 (0.82–0.89)

Solicitous responses (100 = best)

58.3 (25.0)

58.6 (25.0)

0.87 (0.83–0.90) Excellent

0.81 (0.76–0.85)

Distracting responses (100 = best)

53.6 (25.3)

55.3 (24.8)

0.77 (0.71–0.82) Excellent

0.69 (0.61–0.76)

General activity (100 = most)

34.5 (13.7)

35.1 (13.4)

0.86 (0.82–0.90) Excellent

0.82 (0.78–0.86)

Internal Consistency

Chronic Pain Patients: (Kerns et al, 1985; n = 60)

WHYMPI Internal Consistency

Domain

Strength

Cronbach’s alpha

Interference

Excellent

0.90

Support

Excellent

0.83

Pain Severity

Adequate

0.72

Self-control

Adequate

0.79

Negative Mood

Adequate

0.73

Punishing response

Excellent

0.84

Solicitous response

Adequate

0.78

Distracting responses

Adequate

0.74

Household chores

Excellent

0.86

Outdoor work

Adequate

0.77

Activities away from home

Adequate

0.70

Social activities

Adequate

0.74

Chronic Musculoskeletal Pain: (Bergstrom et al, 1998; n = 682; mean age = 41.8 (9.4); duration of pain = 67.9 (70.4) months; Swedish sample)

Internal Consistency:

Scale

mean (SD)

Strength

Cronbach's alpha

Pain Severity

3.73 (1.16)

Excellent

0.80

Interference

3.74 (1.08)

Excellent

0.86

Life Control

3.12 (1.16)

Poor

0.66

Affective Distress

2.95 (1.49)

Excellent

0.80

Support

4.30 (1.57)

Excellent

0.81

Punishing Responses

1.23 (1.29)

Excellent

0.86

Solicitous Responses

2.95 (1.35)

Adequate

0.77

Distracting Responses

3.14 (1.60)

Adequate

0.76

Household Chores

3.58 (1.39)

Excellent

0.81

Outdoor Work

1.68 (1.46)

Excellent

0.80

Leisure Activities

3.08 (1.18)

Poor

0.67

General Activity

2.97 (0.89)

Adequate

0.74

Construct Validity

Convergent Validity:

Chronic Pain Patients: (Kerns et al, 1985)

  • Excellent correlation between the affective stress dimension of the WHYMPI and the State-Trait Anxiety Inventory-State form (STAI-S): r = 0.86
  • Excellent correlation between the affective stress dimension of the WHYMPI and the Beck Depression Inventory (BDI): r = 0.76
  • Excellent correlation between affective stress dimension of the WHYMPI and the Depression Adjective Check List (DACL): r = 0.68
  • Excellent correlation between measured support from others by the WHYMPI and the Marital Adjustment Scale (MAS): r = 0.71
  • Adequate correlation between measured support from others of the WHYMPI and the Multidimensional Health Locus of Control (MHLC)-Powerful Others Scale: r = 0.35
  • Adequate correlation between the activity level dimension of the WHYMPI and the McGill Pain Questionnaire (MPQ)- Total Pain Rating Index and Present Pain Intensity: r = 0.47 and 0.44, respectively

Chronic Pain: (Riley et al., 1999; n = 472 mean age = 44.1 (13.6)) Construct validity of each section of the WHYMPI was assessed with principal axis factor analysis procedure.

Section one yielded four factors with an eigenvalue greater than 1, which accounted for 69.6% of total variance:

  • interference scale
  • support scale
  • self-control and negative mood scale (lack of psychological adjustment)
  • pain severity scale

Problematic cross loading across factors were observed for 3 items: item 12 (amount of suffering experienced because of pain), item 13 (pain affects marital and family relationships), and item 19 (pain affects friendships with other than family members).

Section two yielded three factors with an eigenvalue greater than one, which accounted for 62.7% of variance:

  • solicitous response scale
  • punishing response scale
  • distracting response scale

Cross loadings were observed for item 3 (reads to me) and item 6 (talks to me to take my mind off the pain), providing evidence for lack of divergence between distracting and solicitous scale items.

Section 3 yielded three factors with an eigenvalue greater than 1, which accounted for 54% of total variance:

  • household chores scale
  • outdoor work scale
  • activities away from home scale
  • social activities scale

No cross loadings observed for this section.

Chronic Pain in Elderly Adults (76-99 years old): (Jakobsson et al., 2006; n= 175 mean age = 84.9 (5.4); Swedish sample- translated version)

  • The full length instrument showed acceptable validity and reliability for some parts of the instrument but showed no satisfactory factor structure was found for section 1 or section 2 of the instrument. Overall, It had acceptable convergent and discriminant validity, but the factor analysis did not show a good model fit. The brief screening version, however, showed higher reliability and validity.

Low Back Pain: (Bernstein et al., 1995; n = 194, mean age = NR; SD = NR)

Construct validity of the WHYMPI was assessed with factor analysis with commonly utilized measures found relevant to pain including:

  • Pain duration in days
  • Self-efficacy: worry
  • Self-efficacy: confidence
  • Pain disability
  • Self-esteem
  • The somatization scale of the Symptom Checklist 90 revised (SCL-90-R)
  • The GSI index of the SCL-90-R

Results found that the MPI possesses convergent validity because it significantly correlates with other measures of psychosocial adjustment in pain patients.

 

Content Validity

Chronic Pain Patients: (Kerns et al, 1985; from the original validation study)

Three separate sections of the West Haven-Yale Multidimensional Pain Inventory, WHYMPI) were created:

  • The first section addresses 5 important aspects of the pain experience: interference in areas of functioning, support from others, severity, self-control, and negative mood.
  • The second section addresses perceived responses of others to the communicated pain in three types: negative (punishing), solicitous, and distracting responses.
  • The third section of the WHYMPI assesses the patient’s reported activity in 4 areas: household chores, outdoor work, activities away from home, and social activities.

Mixed Populations

back to Populations

Test/Retest Reliability

Inpatient Sample: (Broderick et al, 2004; = 199; mean age approximately 50 (11) years; interval between the two samples examined was significantly different (treatment sample = 19.4 days versus 33.7 days for the mail sample)

 

Test-retest Reliability Between Two Validation Studies:

 

 

 

 

 

 

 

MPI Scales

 

Correlation coefficients

 

 

 

Broderick: means (SDs)

 

Pain severity

Excellent

0.74

Excellent

0.82

43.4 (10.9)

42.3 (11.6)

Interference

Excellent

0.87

Excellent

0.86

43.1 (11.6)

41.6 (11.8)

Life control

Excellent

0.62

Excellent

0.68

49.2 (7.1)

49.0 (7.4)

Affective distress

Adequate

0.53

Excellent

0.69

47.1 (9.5)

46.9 (10.5)

Support

Excellent

0.84

Excellent

0.80

43.2 (11.3)

42.2 (11.3)

Punishing responses

Excellent

0.79

Excellent

0.84

49.9 (9.2)

50.5 (9.3)

Solicitous responses

Excellent

0.83

Excellent

0.89

47.2 (10.2)

46.9 (9.8)

Distracting responses

Excellent

0.80

Excellent

0.62

47.9 (9.6)

47.0 (9.4)

Household chores

Excellent

0.80

Excellent

0.91

57.2 (7.4)

57.1 (7.5)

Outdoor work

Excellent

0.70

Excellent

0.83

50.6 (9.8)

51.3 (9.5)

Activities away from home

Excellent

0.83

Excellent

0.88

55.6 (10.0)

54.9 (9.6)

All correlations significant at p < 0.001

 

 

 

 

 

 

Internal Consistency

Temporomandibular Disorder: (Andreu et al, 2006; n = 114, mean age = 35 (14) years)

Domain level Internal Consistency:

 

 

 

 

Section

Domain

Strength

Chrnobach's alpha

Section 1

Repercussion of pain

Excellent

0.85

 

Social support

Excellent

0.82

 

Pain severity

Adequate

0.75

 

Negative mood

Adequate

0.75

 

Interference in daily activities

--

--

Section 2

Negative responses

Adequate

0.78

 

Support responses

Excellent

0.85

Section 3

Household chores

Excellent

0.80

 

Social and leisure activities

Adequate

0.73

 

Taking care of the car

Excellent

0.88

Criterion Validity (Predictive/Concurrent)

Fibromyalgia: (Turk and Rudy, 1988; n = 48) Three subtypes based on WHYMPI scores were created; these include:

  • Dysfunctional (DYS)
  • Interpersonally Distressed (ID)
  • Adaptive Copers (AC)

6 treatment sessions spaced over 4 weeks resulted in significantly different levels of functioning.  DYS patients experienced reductions in pain, affective distress and perceived interference of pain while ID patients exhibited poor responses to treatment.  AC patients demonstrated reductions in pain, but not affective distress or disability.

Bibliography

Andreu, Y., Galdon, M. J., et al. (2006). "An examination of the psychometric structure of the Multidimensional Pain Inventory in temporomandibular disorder patients: a confirmatory factor analysis." Head Face Med 2: 48. 

Bergstrom, K. G., Jensen, I. B., et al. (1999). "A psychometric evaluation of the Swedish version of the Multidimensional Pain Inventory (MPI-S): a gender differentiated evaluation." Eur J Pain 3(3): 261-273. 

Broderick, J. E., Junghaenel, D. U., et al. (2004). "Stability of patient adaptation classifications on the multidimensional pain inventory." Pain 109(1-2): 94-102. 

Kerns, R. D., Turk, D. C., et al. (1985). "The West Haven-Yale Multidimensional Pain Inventory (WHYMPI)." Pain 23(4): 345-356. 

Lousberg R, Van Breukelen GJ, Groenman NH, Schmidt AJ, Arntz A, Winter FA. (1999), Psychometric properties of the Multidimensional Pain Inventory, Dutch language version (MPI-DLV). Behavior Research and Therapy. 1999;37(2):167-182.

Riley JL, Zawacki TM, Robinson ME, Geisser, ME. Empirical test of the factor structure of the West Haven-Yale Multidimensional Pain Inventory. The Clinical Journal of Pain. 1999;15(1): 24-30.

Turk, D. C., Okifuji, A., et al. (1998). "Differential responses by psychosocial subgroups of fibromyalgia syndrome patients to an interdisciplinary treatment." Arthritis Care Res 11(5): 397-404. 

Verra ML, Angst F, Staal JB, et al. Reliability of the Multidimensional Pain Inventory and stability of the MPI classification system in chronic back pain. BMC Musculoskeletal Disorders. 2012;13:155.