ARTHRITIS HELPLESSNESS INDEX (AHI)/RHEUMATOLOGY ATTITUDES INDEX (RAI)
The Arthritis Helplessness Index (AHI) and its variants were designed to assess patients' perceptions of helplessness in coping with arthritis as delineated by learned helplessness theory. Helplessness is considered a psychological state in which individuals expect their efforts will be ineffective and become more passive and more likely to be depressed. Learned helplessness theory postulates that this helplessness results from experiencing unpredictable and uncontrollable aversive events. Helplessness has also been postulated to mediate relationships between disease or treatment and health outcomes.
The Rheumatology Attitudes Index (RAI) is conceptually identical to the AHI. Item wording and response format were modified slightly to reduce respondent confusion when the instrument was used with individuals with other rheumatic conditions such as fibromyalgia or bursitis.
Four variants of arthritis helplessness measures are available: the original 15-item Arthritis Helplessness Index (1), a 5-item AHI Helplessness Subscale (2), the 15-item Rheumatology Attitudes Index (RAI) (3), and a 5-item RAI helplessness subscale (4).
The AHI/RAI consists of 2 types of items: items measuring patients' perceptions of their abilities (“I can reduce my pain by staying calm and relaxed”), and their inabilities (“No matter what I do or how hard I try, I just can't get relief from my pain”), to control their arthritis.
Further investigation revealed 2 distinct factors on both the AHI and RAI, internality (“Managing my arthritis is largely my responsibility,” 7 items) and helplessness (“Arthritis is controlling my life,” 5 items).
(AHI) Perry M. Nicassio, PhD, Daley Hall, Room 104, CSPP SD AIU,10455 Pomerado Road, San Diego, CA 92131. E-mail: firstname.lastname@example.org. (RAI) Leigh F. Callahan, PhD, Thurston Arthritis Research Center, University of North Carolina, 3310 Thurston Building CB#7280, Chapel Hill, NC 27599-7280. E-mail: email@example.com.
Four versions are available, the original 15 item AHI and its 5-item helplessness subscale, and the 15 item RAI and its 5-item helplessness subscale. RAI items are identical to AHI items except the word arthritis was replaced by the word condition. This change affected 12 of the original 15 items. The 5-item subscales are considered conceptually cleaner because each consists of a single factor. They are also easier and faster to complete. A Spanish language version of the 5-item RAI-helplessness subscale has been evaluated for cross-cultural equivalence (5). RAI has been translated/adapted to Swedish.
Number of items in scale.
AHI/RAI has 15 items. AHI/RAI Helplessness subscales have 5 items.
The AHI has a 5-item helplessness subscale and 7-item internality subscale. The RAI has a 5-item helplessness subscale. A factor analysis revealed an internality subscale on the RAI as well, but no other information is provided.
All psychometric work on the AHI and RAI was done on individuals with physician-confirmed rheumatoid arthritis.
RAI has been widely used, including with patients with osteoarthritis, fibromyalgia, systemic lupus erythematosis, and scleroderma.
WHO ICF Components.
Self-administered written self-report questionnaire. Easy to administer.
No training required.
Time to administer/complete.
The 5-item helplessness scales estimated to take less than a minute. The 15-item scales may take up to 3 minutes.
Items and scoring available from the literature (see references.) Copy available at the Arthritis Care & Research Web site at http://www.interscience.wiley.com/jpages/0004-3591:1/suppmat/index.html.
Original AHI has a 4-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly agree); AHI Helplessness subscale has a 6-point Likert scale (1 = strongly disagree, 2 = moderately disagree, 3 = disagree, 4 = agree, 5 = moderately agree, 6 = strongly agree); RAI has a 4-point Likert scale with 5 response options (1 = strongly disagree, 2 = disagree, 2.5 = do not agree or disagree, 3 = agree, 4 = strongly agree); RAI Helplessness Subscale has a 5-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = do not agree or disagree, 4 = agree, 5 = strongly agree).
For all versions, higher scores indicate greater helplessness. Original AHI 15–60, AHI Helplessness subscale 5–30, RAI 15–60, RAI Helplessness Subscale 5–30.
Interpretation of scores.
For the AHI Helplessness subscale, empirically derived cut-points, which demonstrated statistically significant differences among groups on psychological, behavioral, and symptom severity measures have been published. The “low helplessness” group achieved better scores while “high helplessness” group scored worse on psychological, behavioral, and symptom severity scales. Cut-points are Low helplessness (< 11), Normal (11–19, combines low normal, normal and high normal), High helplessness (≥ 20). In other versions there is no guidance in interpretation of scores; since the RAI items are identical except for the substitution of “condition” for “arthritis” in 4 items, it is likely to have similar cut-points but this has not been tested.
Method of scoring.
All versions can be scored by hand. In the AHI/RAI reverse the 9 items indicating perceived control (items 2, 3, 5, 6, 8, 9, 11, 13, 15) and sum all items for total score. In the AHI/RAI helplessness subscale reverse scoring on item 4; sum all items for total score.
Time to score.
Not documented, likely to be very brief, but requires score conversion.
Training to score.
Need instructions or template to reverse the appropriate items before scoring (9 items on the 15-item scales, 1 item on the 5-item scales).
Training to interpret.
None required; these instruments have been used primarily in research. Clinicians who use them have relied on clinical judgment.
No formal norms have been published. Cut-points scores used to determine low, midrange, and high helplessness scores using the AHI helplessness subscale were empirically derived to categorize 20% of the sample as low helplessness, and 20% as high helplessness.
Internal consistency reliability, Cronbach's alpha 0.69 (borderline acceptable for a presumed unidimensional scale). The 12-month test-retest reliability was 0.53.
AHI helplessness subscale.
Internal consistency reliability with Cronbach's alpha was 0.63 (item-to-total correlations ranging from 0.29 to 0.47). Internal consistency alpha via the Spearman-Brown prophecy formula (used to equate subscales to the number of items on the full scale) was 0.84. The 6-month test-retest reliability was 0.64.
AHI internality subscale.
Internal consistency reliability via Cronbach's alpha was 0.75 (item-to-total correlations ranging from 0.38 to 0.58). Internal consistency alpha via the Spearman-Brown prophecy formula (used to equate subscales to the number of items on the full scale) was 0.88 and the 6-month test-retest reliability was 0.59.
Internal consistency reliability via Cronbach's alpha was 0.68. Two items had weak correlations with the parallel item on the AHI (I have considerable ability to control my pain”, and “If I do all the right things, I can successfully manage my condition.”)
RAI helplessness subscale.
Internal consistency reliability via Cronbach's alpha was 0.70 Sample A and 0.67 Sample B.
AHI construct validity.
AHI correlates in expected ways with theoretically relevant variables such as health locus of control, self esteem, anxiety, and depression measures. All correlations were significant and remained significant when age- and education level-adjusted. AHI also significantly correlated with measures of functional status (Modified Health Assessment Questionnaire), dissatisfaction with functional status, pain, and general rating of perceived limitations.
AHI helplessness subscale construct validity.
Subscale correlated in expected ways with theoretically relevant measures including chance- and powerful other-health locus of control, depression, non-compliance, information seeking, pain rating, and Arthritis Impact Measurement Scales physical, pain, depression and global health status subscales. The 5-item subscale accounted for more variance in these measures than did the 15-item measure. Changes in the helplessness subscale were more strongly associated with changes in pain and depression than were changes in the scores on the 15-item scale.
AHI internality subscale construct validity.
Scores were associated with internal health locus of control scores. High scores were associated with less pain, depression, and behavioral ineffectiveness. Changes in the full scale were more strongly associated with changes in psychological and disease impact measures than were changes in the internality subscale.
RAI construct/criterion validity.
Correlation of 0.78 between RAI and AHI completed 1–24 hours apart. Validation article (3) presents a variety of correlations of RAI with 4 measures of disease activity, 3 measures of physical performance, and 3 self- report measures of functional status. No relationships between RAI and these measures are hypothesized a priori, so it is difficult to evaluate these as evidence of validity. RAI scores were significantly correlated with physical performance scores, and self-report measures of function; the latter had larger correlations.
RAI helplessness subscale construct/criterion validity.
Correlations of 0.79 between full RAI and RAI helplessness scale. Brief measure had significantly higher correlations to measures of self-reported functional status than did the full RAI.
Sensitivity/responsiveness to change.
An observational study found a 1-point decrease in helplessness on 12- month retesting. Changes in the AHI were significantly correlated with changes in functional status at 12 months.
AHI helplessness subscale.
The 6-month retesting indicated changes over time, and these were more strongly related to changes in pain and depression than were changes in the full AHI.
RAI and RAI Helplessness Subscale.
Comments and Critique
The original 15-item AHI was developed to capture the construct of learned helplessness, the psychological state in which individuals believe their efforts will be ineffective. These perceptions are hypothesized to produce affective, motivational, and behavioral deficits. The original 15-item AHI has been correlated with theoretically relevant variables and has demonstrated construct validity. True criterion validity is not possible because there is no pre-existing gold standard helplessness measure in the literature. The 15-item AHI had modest internal-consistency reliability, however, suggesting it was not a uni-dimensional measure. Factor analysis found 2 factors or subscales, internality and helplessness. The internality subscale has received little attention, but the helplessness subscale has been demonstrated to be more conceptually clear and to account for more variance in variables of interest than does the full AHI. In addition, Stein et al (6) empirically developed a classification schema for the helplessness subscale that has predicted scores on psychological, behavioral, and symptom severity measures even after 2 years.
The RAI is a modification of the AHI, created by replacing the word arthritis with the word condition in 9 items, and creating a fifth response category (“do not agree or disagree”). The RAI appears to behave fairly similarly to the AHI but true psychometric data are thin. There is a 0.78 correlation between the AHI and its variant, the RAI. Factor analysis was also used to identify subscales for the RAI. The same 2 factors emerged, internality and helplessness with items identical to AHI subscales. As with the AHI, the RAI helplessness scale had a reasonably strong correlation with the full 15-item RAI and had stronger correlations to measures of functional status than did the full RAI.
As a measure of the construct of helplessness, the brief 5-item scales appear to be superior to the 15-item scales, because of their speed and ease of use and because of stronger correlations with other health status variables. The 5-item AHI has greater psychometric support than the 5-item RAI, and has demonstrated predictive utility. For these reasons it may be preferable to use the AHI helplessness subscale in situations where having the word arthritis in the item language is not problematic.
DeVellis and Callahan (4) note that reliabilities of these measures are at the low end of the acceptable range, however, with 30% of the score variance due to error. They suggest these measures are adequate for research and general screening purposes but should not be used alone for clinical decision-making.
1. (Original AHI) Nicassio PM, Wallston KA, Callahan LF, Herbert M, Pincus P. The measurement of helplessness in rheumatoid arthritis: the development of the Arthritis Helplessness Index. J Rheumatol 1985;12:462–7.
2. (Original AHI Helplessness) Stein MJ, Wallston KA, Nicassio PM. Factor structure of the Arthritis Helplessness Index. J Rheumatol 1988;15:427–32.
3. (Original RAI) Callahan LF, Brooks RH, Pincus T. Further analysis of learned helplessness in rheumatoid arthritis using a “Rheumatology Attitudes Index.” J Rheumatol 1988;15:418–26.
4. (Original RAI Helplessness) Devellis RF, Callahan LF. A brief measure of helplessness in rheumatoid disease: the helplessness subscale of the Rheumatology Attitudes Index. J Rheumatol 1993;20:866–69.
5. Escalante A, Cardiel MH, del Rincon I, Sudrez-Mendosa AA. Cross cultural equivance of a brief helplessness scale for Spanish speaking rheumatology patients in the United States. Arthritis Care Res 1999;12:341–50.
6. Stein MJ, Wallston KA, Nicassio PM, Castner CM. Correlates of a clinical classification schema for the arthritis helplessness subscale. Arthritis Rheum 1988;31:876–81.