Descriptions of joint pain by American Indians: Comparison of inflammatory and noninflammatory arthritis


  • B. Josea Kramer,

    Corresponding author
    1. Department of Veterans Affairs Greater Los Angeles Healthcare System, Geriatric Research Education Clinical Center, Sepulveda, California, and University of California at Los Angeles School of Medicine
    • Geriatric Research Education Clinical Center (11E), VA Greater Los Angeles Healthcare System, 16111 Plummer Street, Sepulveda, CA 91343
    Search for more papers by this author
  • Judith O. Harker,

    1. Department of Veterans Affairs Greater Los Angeles Healthcare System, Sepulveda, California
    Search for more papers by this author
  • Andrew L. Wong

    1. Olive View-UCLA Medical Center, Sylmar, California, University of California at Los Angeles School of Medicine, and Department of Veteran Affairs Greater Los Angeles Healthcare System, Sepulveda, California
    Search for more papers by this author



To improve assessment of arthritis joint pain in American Indians by describing how symptoms are communicated.


In-depth interviews were conducted with American Indians who experience chronic joint pain (n = 56), to elicit descriptions and self-reported ratings of pain, disability, and beliefs associated with the pain for affected joints (n = 326).


Discrete sets of specific verbal descriptions distinguished inflammatory arthritis (n = 20 terms) from noninflammatory arthritis (n = 22 terms), and indicated levels of pain intensity. An additional set of 14 vague but commonly used verbal descriptors did not distinguish the type of joint disease or pain intensity.


Subtle pain complaints and vague verbal descriptions, such as “ache,” “hurt,” and “discomfort,” may reflect severe pain symptoms, disability, and more serious joint disease in American Indian patients. In addition, certain specific sensory descriptions used by American Indians suggest inflammatory arthritis and may warrant further evaluation.


American Indians are a culturally heterogeneous special population composed of over 300 nations, tribes, rancherias, and bands. Although federal reservation communities are concentrated in 13 states, about 60% of this diverse population reside in urban areas, with the largest concentration living in southern California. The prevalence of rheumatoid arthritis (RA), which may have originated in the New World (1), is greater among American Indians than non-Indians and Alaskan Natives, and this difference is most likely genetic in origin (2). Research that has focused on characterizing the epidemiologic, immunologic, and clinical manifestations in selected reservation communities has documented higher prevalence rates of rheumatoid arthritis (3–6), systemic lupus erythematosus (4, 7), and scleroderma (8), cross-cutting cultural differences between American Indian peoples and occurring in tribes indigenous to the US southwest, northwest, and plains regions. Despite the stereotype that American Indians are stoic and generally not communicative about painful experiences, secondary analysis of the 1992–1993 (Wave 1) Health and Retirement Study (∼hrswww) found that American Indians aged 51–61 years were significantly more likely to report chronic pain than whites, and were also more likely to report pain than other ethnic groups (9). Regression modeling of demographic, social, medical, and health characteristics estimated that the most significant predictor of chronic pain was a diagnosis of arthritis (10). Although one anecdotal report suggests that standardized patient self-report instruments may not yield meaningful results because of cultural differences in disclosing disabilities (11), no studies of cultural assessment of patients with arthritis in this population have been conducted. Moreover, studies of arthritis in urban American Indians have not been published, except as part of more general health surveys.

A number of studies have attempted to derive clinically useful information from subjective, culture-bound communications of arthritis pain, modeled, in part, on the McGill Pain Questionnaire (MPQ) (12), which assesses the multidimensional nature of pain through a forced-choice word list. Although the MPQ distinguishes arthritis from 7 other pain syndromes, (ie., labor pain, chronic back pain, or cancer pain) (13), its usefulness is limited. The MPQ is not able to more specifically characterize arthritis pain by distinguishing between inflammatory and noninflammatory arthritis. With the goal of creating such an instrument, various researchers developed forced-choice word lists of 17–28 items, with terms derived from the MPQ, patient descriptions, and the rheumatology literature (14–17). These studies were generally unable to distinguish inflammatory arthritis from noninflammatory arthritis pain. Nor did these studies suggest how to assess the more general terms used indiscriminately by patients to describe inflammatory and noninflammatory arthritis conditions. Particularly problematic are 10 terms, (i.e., aches, constant, gnawing, hurts, sharp, burning, sore, throbbing, annoying, and stabbing), which typically are used to describe joint disease pain syndromes, appear on all standard word lists, and are commonly volunteered or selected by patients. Development of a standardized word list to distinguish pain caused by either inflammatory or noninflammatory arthritis has, to this point, been unsuccessful.

American Indians have been reported to have a higher prevalence of specific rheumatic conditions, including RA, systemic lupus erythematosus (SLE), and scleroderma, than non-Indians. Little is known about appropriate, culturally-sensitive pain assessment strategies for this population or how standardized assessments may impact diagnosis and treatment. This study is the first to describe how American Indians may communicate their symptoms of chronic joint pain. Findings of systematic categorical descriptions are reported. The implications of these findings for improving arthritis pain assessment and care to American Indian patients are discussed.


Study sample

The study was conducted in Los Angeles and northern Orange counties in southern California. Arthritis/rheumatism had been previously found to be a leading health problem for American Indians living in the study area (18). This region has the highest population density of urban-dwelling American Indians in the US, and no reservations are located in the study area. As many as 200 American Indian and Alaskan Native tribes and villages may be represented in this population according to the Los Angeles City/County Native American Indian Commission. The majority are affiliated with cultural groups from the historic Oklahoma Indian territories (e.g., Choctaw), the Southwest (e.g., Navajo), and the Northern Plains (e.g., Sioux) (19–21). In comparison, relatively few indigenous California peoples live in the study area. This pan-Indian community is defined by interaction, rather than geography, and participation in American Indian community activities is an important marker of ethnic identity (21).

Convenience sampling was used as the most appropriate methodology for a community-based psychosocial study of urban American Indians (22). This strategy has been demonstrated to reflect the general American Indian population in this region (23). Study participants were recruited by local American Indian organizations (i.e., a multipurpose community organization and several churches). These agencies are Indian-operated and serve a pan-tribal clientele from which study participants were recruited. Subject recruitment criteria were 1) recognition as an American Indian by the community, and 2) currently experiencing any type of joint pain. Persons who were identified as potential participants were contacted and, with the subjects' permission, audio-taped interviews were conducted by American Indian interviewers. Bilingual interviews and translations were performed as needed.

A pretested semistructured interview protocol was followed to conduct in-depth, face-to-face interviews that could be completed in 45–60 minutes. The interview included both closed-ended responses to standardized instruments and open-ended questions. Standardized closed-ended questions were used to describe the sample population, likelihood of inflammatory or noninflammatory arthritis, and disease activity. The connective tissue disease questionnaire for autoimmune and rheumatic diseases (24), with other information, was used to classify subjects by arthritis category. American College of Rheumatology core disease activity (25) as measured by patient self-report, was obtained from the disability section of the Health Assessment Questionnaire (HAQ) (26), mannequin-assisted self-reported tender and swollen joint counts (27), and pain visual analog scale (VAS).

Open-ended questions were used to document the verbal descriptors of present pain in each involved joint. No standard word list was offered as a prompt and no limit was set on the number of words that could be volunteered. Followup questions elicited other words that would be used to describe this particular painful sensation to family members or to a doctor. After a complete list of verbal descriptors had been reported, respondents rated pain intensity using a VAS. Because the most appropriate scale for this population was unknown, respondents were given the option to select among 5 formats all shown on a single page: 10-point VAS scales with or without word anchors, 5-point VAS scales with word or both word and color anchors, and a facial expressions scale. The entire process was repeated for up to 5 painful joints, thereby documenting, in descending order, the most painful joints. Other open-ended questions designed to elicit experiential descriptions of joint pain, meanings, and beliefs associated with this type of pain, coping strategies, and medical experiences are reported elsewhere (27).

Quantitative and qualitative analyses

The analysis compared responses of study participants according to 3 disease categories, inflammatory arthritis, noninflammatory arthritis, and localized musculoskeletal pain syndrome. Assignment to a likely disease category was made by a clinical rheumatologist based on participant responses to the connective tissue disease screening questionnaire, the pattern of joint involvement on the mannequin-assisted self-reported tender and swollen joint counts, and self-reported use of prescribed or over-the-counter medications. Using these screening criteria, the diagnoses likely to be identified in the inflammatory category were RA, SLE, sero-negative spondylarthropathy, scleroderma, and Sjögren's syndrome; diagnoses in the noninflammatory category were generalized and localized osteoarthritis; and diagnoses in the localized musculoskeletal pain syndrome category were tendinitis, bursitis, and localized myofascial pain syndromes.

Quantifiable responses from the voice recording, note-based documents, and transcripts were entered into Microsoft Excel database files (Microsoft Corporation; Seattle, WA). These responses included qualitative data that were amenable to coding, scaling, and statistical analyses, such as use of prescription or over-the-counter medications, duration in years of joint pain and self-reported pain ratings. Quantitative data were exported to a SPSS database (Statistical Package for the Social Sciences, 1999, Chicago, IL) for descriptive, parametric, and nonparametric analyses. Qualitative text data were also managed using Microsoft Excel to facilitate sorting and comparisons by disease classification.



Interviews were conducted with 45 subjects identified by American Indian community organizations as currently experiencing joint pain. Respondents ranged in age from 23–90 years (mean 55.4 years), 76% were women, and they described a total of 326 tender and/or swollen joints. The sample represented 29 tribal affiliations, and the majority were affiliated with tribes from the Southwest, Oklahoma historic Indian territories, and Northern Plains. Consistent with other community samples, relatively few indigenous California tribes were represented. All respondents spoke English, and 2 respondents preferred to conduct the interview discussions in both English and Navajo. Although most respondents selected the VAS 10-point scale with word anchors, persons with limited English speaking skills tended to use a 5-point VAS scale anchored with words and colors. No respondents selected the 10-point VAS without word anchors, or the facial expressions scale.

Respondents were classified into 3 disease categories: inflammatory arthritis (n = 12), noninflammatory arthritis (n = 24), and localized musculoskeletal pain syndrome (n = 9) (Table 1). The group with noninflammatory arthritis was slightly older on average (59.4 years) than the inflammatory arthritis group (54.0 years), and the group with localized musculoskeletal pain syndrome tended to be younger (46.7 years). The average duration of pain for the noninflammatory arthritis group was slightly longer (10.4 years, n = 21) than the inflammatory arthritis group (9.6 years, n = 12); and the localized musculoskeletal pain syndrome group had a much shorter average duration of pain (4.3 years, n = 7).

Table 1. Characteristics of sample
CategorySwollen joints* mean ± SDTender joints mean ± SDHAQ disability§ range: 0–3 mean ± SDHighest pain rating reported range: 0–10 mean ± SD
  • *

    P = <0.001 (one-way analysis of variance [ANOVA], F [2, 42] = 11.21)

  • P <0.001 (Tukey's honestly significant difference [HSD]), inflammatory arthritis > non-inflammatory arthritis, localized musculoskeletal pain syndrome

  • P = <0.001 (one-way ANOVA, F = 21.30)

  • §

    P = <0.036 (one-way ANOVA, F [2, 42] = 3.60

  • P = <0.001 (Tukey HSD), inflammatory arthritis > localized musculoskeletal pain syndrome

Inflammatory joint disease (n = 12)10.9 ± 11.717.8 ± 11.90.958 ± 0.7586.67 ± 2.6
Noninflammatory joint disease (n = 24)1.4 ± 2.04.4 ± 4.10.464 ± 0.5115.6 ± 2.5
Musculoskeletal pain (n = 9)0.0 ± 0.01.3 ± 0.90.347 ± 0.5484.7 ± 2.6

In general, the group classified as having inflammatory arthritis had significantly more severe arthritis symptoms: more swollen joints, more tender joints and higher level of disability (HAQ). Although the mean pain rating was highest in the inflammatory arthritis group, there were no statistically-significant differences between groups in pain rating. However, pain was correlated with HAQ scores (r = 0.703, P < 0.001). For all 3 groups, mild to moderate pain was associated with less disability (HAQ scores of <1.0) while severe to worst pain was associated with greater disability and reduced functional independence (HAQ scores of >1.0).

Verbal descriptors

Responses of subjects with inflammatory and noninflammatory arthritis were analyzed in more detail to better assess their experiences of chronic joint pain. Complete descriptions of joint pain, pain intensity, and disability were recorded for 192 joints in subjects with inflammatory arthritis and 72 joints in subjects with noninflammatory arthritis. A total of 56 English-language terms were used, excluding modifiers such as “very.”

Two unique sets of terms emerged that differentiated inflammatory from noninflammatory arthritis. A set of 20 terms and similes exclusively described inflammatory arthritis pain and a separate set of 22 terms and similes exclusively described noninflammatory arthritis pain. In addition, these terms were generally associated with certain specific levels of pain intensity (Table 2). For instance, mild, inflammatory arthritis joint pain was described as “sore,” “tender,” and “puffy,” but when severe, the joint pain was described as “stretched to pieces” and as “inside of knee is growing.” Inflammatory arthritis joint pain descriptors also focused on the sensory qualities of the pain (e.g., “throbbing,” “heat”). In contrast, noninflammatory arthritis was described with a broader range of experience: sensory (e.g., “pulling,” “stabbing”), affective (e.g., “tiring”), evaluative (e.g., “annoying,” “stabbing”) and pain intensity (e.g., “horrible”). These disease-specific terms (Table 2) communicated the pain experience for 81% of the inflammatory and noninflammatory arthritis joints described. Often participants offered several responses to fully communicate the pain experience, such as “puffy,” “mushy,” and “swollen.” In contrast to those with other types of arthritis, persons who experienced inflammatory arthritis joint pain were likely to provide a more detailed, focused description of pain suggestive of inflammation to their doctors than to family and acquaintances (Fisher's exact test, P = 0.01).

Table 2. Disease-specific descriptions of joint pain according to pain intensity: comparing inflammatory and noninflammatory arthritis descriptors
Disease categoryPain intensity
Mild (VAS 1–3)Moderate (VAS 4–6)Severe (VAS 7–9)Worst (VAS 10)
  • *

    Terms that are used for more than one level of pain intensity.

  • VAS = visual analog scale.

Inflammatory arthritis joint diseaseHeat penetrating skinBones got fat insideInside of knee is growing
Like little burnLike pliers squeezingNo control
MushyPretty badSharp, intense
NumbSensitive to touchStretched to pieces
Noninflammatory arthritis joint diseaseFeels hard in there TiredAggravatingCan't moveCuss words
AnnoyingHated (limitation)Horrible
Bothers me/BotheringLike constant toothache*Severe*
Can't work like used toLike beating on drumTerrible
Like got hitTiringUnbearable
Like muscle spasmSevere*
Like toothache*Stabbing*

However, joint pain was also described by 14 common, vague, general terms that did not distinguish either arthritis disease categories or intensities of pain (Table 3). General terms, such as “hurts,” “constant,” “aches,” and “painful,” were used to describe mild/moderate pain as well as severe/worst pain. These general terms were used to describe pain in 51 joints, of which 37% were classified as inflammatory arthritis. Respondents were just as likely to offer a vague descriptor of pain (Table 3) or a vague descriptor plus a disease-specific descriptive term (Table 2) as either a first response or subsequent response.

Table 3. Vague general verbal descriptions of joint pain*
Mild and moderate pain (VAS 1–6)Severe and worst pain (VAS 7–10)
  • *

    General terms that do not effectively distinguish inflammatory from noninflammatory arthritis joint pain or self-report of pain as measured with a visual analog scale (VAS).

  • Descriptors used for both mild/moderate and severe/worst pain.

  • Describes VAS 1–3 only.

  • §

    Describes VAS 4–6 only.

  • Describes VAS 10 only.

Constant (or won't go away)Constant (or won't go away)
Dull painExcruciating
Locks upPain/Painful
Mild painSharp pain
Moderate pain§Uncomfortable
Sharp pain

Although the sample size is small, the relationship between disability and choice of verbal expressions was suggestive that dimensions other than pain might also be communicated. The vague, general terms “pain,” “discomfort,” and “uncomfortable” were reported by people who experienced functional impairment (HAQ >1) while “ache” and “hurt” were reported by people experiencing less disability (HAQ <1).


This study found 2 sets of descriptive terms that allowed discrimination between the experiences of inflammatory and noninflammatory arthritis joint pain, and a third set of terms that conveyed pain but lacked discriminatory power. The sets of specific verbal descriptors also indicated the level of pain intensity. The experience of joint pain was communicated with disease-specific terms and/or with vague descriptors. Thus, the overlap of verbal descriptions suggests practical clinical responses to patients' free-listed descriptions and additional considerations for interpreting a forced choice word list to screen for arthritis.

Many of the descriptive terms elicited from American Indians correspond to other arthritis word lists, and nearly all of the general terms would be found on the MPQ, with the exception of the words “locks up” and “uncomfortable.” Although Papageorgiou and Badley (16) noted that patients with inflammatory arthritis volunteered affective terms not found on the MPQ, the additional terms volunteered by American Indians to describe inflammatory arthritis joint pain focused on sensory terms suggestive of active inflammation. This focus on sensory terms contrasts with other studies, notably by Charter and colleagues (14) who found emotional-affective properties of pain dominated the experience. Some expressions used for non-inflammatory pain were colorful, reflecting other painful experiences, such as “a constant toothache” or “getting hit.” If a forced choice list had been used, rather than free listing, the number of terms may have been reduced. For instance, “sensitive to touch” may have been equated to “tender” and “bones got fat inside” may have been equated to “tight.”

This small, heterogeneous sample is limited in its ability to detect statistically significant differences, and the extent to which its findings can be generalized. Like other semeiological analyses, the number of respondents is small and limited to a single convenience population sample. Descriptive pain word lists generated from respondents are subject to idiosyncratic interpretations of the relatively few subjects in comparison to the number of joints described. In addition, the arthritis pain word lists (Table 2) omitted any terms for which the interview inadvertently failed to document pain intensity rating for the joint being described, including the term “crampy.” Qualifiers such as “very” or “not very” were also omitted, and these adjectives were generally consistent with level of pain intensity. Although inadvertent word prompts from the visual analog scale word anchors were unavoidable when multiple joints were described, these anchor terms were also volunteered prior to exposure to the analog scales. A final limitation is that no independent clinical examination or review of medical records was performed to confirm arthritis diagnoses in the study sample population. Therefore, unexpected findings, such as use of the verbal description “hot” by persons with noninflammatory arthritis could not be evaluated or the disease classification corroborated. Validation of this pilot study would require a rheumatology evaluation for each participant.

Although other studies noted the overlap of some terms across disease categories, none have previously isolated distinct terms by disease category and intensity. These terms are familiar, everyday language and are not specialized words or phrases. Our findings suggest that verbal descriptions used by American Indians to describe types of arthritis joint pain would be useful information for screening purposes. However, the general pain descriptions, which do not distinguish either disease category or pain intensity, should be interpreted with caution. The assumption that the term “ache” indicates a low level of pain intensity (29, 30) was not reliable for American Indians with chronic joint pain. We found that American Indians used the expression “ache” to describe pain of any intensity, including severe pain. Equally vague terms, such as “pain,” “discomfort,” and “uncomfortable” may appear to understate symptoms in persons having significant functional disabilities from chronic joint disease.

A word list that distinguishes inflammatory from noninflammatory arthritis was constructed based on the verbal descriptors commonly used by American Indians. Further research is needed to determine if these particular terms will constitute a standard word list that will be valid and consistent in populations composed by other ethnic groups. Use of vague, general pain terms such as “ache,” “hurt,” and “painful” may indicate either mild or severe pain, and therefore should warrant further evaluation of the quality and intensity of pain. Recognition that moderate to severe pain reports may be understated by American Indians should be part of the cultural assessment evaluation performed by healthcare providers for determining the most appropriate individualized approach to patient care and pain management.