Stroke-Adapted Sickness Impact Profile (SA-SIP30)

Evidence Reviewed as of before: 19-08-2008
Author(s): Lisa Zeltzer, MSc OT
Editor(s): Nicol Korner-Bitensky, PhD OT; Elissa Sitcoff, BA BSc

Purpose

The Stroke-Adapted Sickness Impact Profile (SA-SIP30 – van Straten, de Haan, Limburg, Schuling, Bossuyt, & van den Bos, 1997) was developed from the original 136-item Sickness Impact Profile (SIP-136), and assesses quality of life in patients who have sustained a stroke. The scale was developed specifically for use in stroke outcome research in order to overcome the major problem observed with the SIP-136, its length (Finch, Brooks, Stratford, & Mayo, 2002).

In-Depth Review

Purpose of the measure

The Stroke-Adapted Sickness Impact Profile (SA-SIP30 – van Straten, de Haan, Limburg, Schuling, Bossuyt, & van den Bos, 1997) was developed from the original 136-item Sickness Impact Profile (SIP-136), and assesses quality of life in patients who have sustained a stroke. The scale was developed specifically for use in stroke outcome research in order to overcome the major problem observed with the SIP-136, its length (Finch, Brooks, Stratford, & Mayo, 2002).

Available versions

The SA-SIP30 was adapted from the original SIP-136 first published in 1976 by Bergner, Bobbitt, Pollard, Martin, and Gilson and later revised in 1981 by Bergner, Bobbit, Carter and Gilson.

Features of the measure

Items:

van Straten et al. (1997) followed a three-stage process to eliminate items and subscales that were least relevant to stroke survivors (i.e. those applying to fewer than 10% of patients) as well as those with the lowest levels of reliability from the original SIP (van Straten et al. 1997; Golomb, Vickrey, & Hays, 2001).

A criticism of the SA-SIP30 is that no attempt has been made to enhance the scale with items or domains of potential importance to stroke. Thus, the SA-SIP30 does not assess pain, recreation, energy, general health perceptions, overall quality of life or stroke symptoms (Golomb, Vickrey, & Hays, 2001).

The SA-SIP30 contains 30 items. Each item takes the form of a statement describing changes in behavior that reflect the impact of illness on some aspect of daily life. Patients are asked to mark items most descriptive of themselves on a given day. All responses are “yes” or “no”. Scale items are weighted to reflect the relative importance of the item to health status and are the same as the weights used in the SIP-136. In addition to maintaining much of the original subscale structure of the SIP-136, these weights help facilitate comparisons with studies using the original SIP-136.

Scoring:

The scoring of items, subscales, dimensions and total score is the same as for the original SIP. To score the scale, weights are applied to marked items, summed for each subscale and expressed as a percentage for each subscale ranging from 0 to 100%. Higher scores indicate less desirable health outcomes (van Straten et al., 1997; van Straten, de Haan, Limburg, & van den Bos, 2000; Finch et al., 2002; Cup, Scholte op Reimer, Thijssen, & van Kuyk-Minis, 2003). Regression weights have also been provided to allow for a calculation of estimated SIP-136 scores from SA-SIP30 scores.

Cut-off scores representative of poor health have been defined as the following: patients with scores > 33 are known to be impaired in activities of daily living, unable to live independently, experience difficulties in self care, mobility and in performing their main activity. Similar profiles have been observed for Physical dimension scores > 40, but no cut-off values could be defined using the Psychosocial dimension (van Straten et al., 2000).

Subscales:

There are 8 subscales:

  • Body Care and Movement (5 items)
  • Social Interaction (5 items)
  • Mobility (3 items)
  • Communication (3 items)
  • Emotional Behavior (4 items)
  • Household Management (4 items)
  • Alertness Behavior (3 items)
  • Ambulation (3 items)

Subscales can be combined to form 2 dimensions:

  • Physical: includes the subscales Body care and movement, Ambulation, Household management and Mobility (15 items)
  • Psychosocial: includes the subscales Alertness behavior, Communication, Social interaction and Emotional behavior (15 items)

Equipment:

No special equipment is required to administer the SA-SIP30.

Training:

The scale is intended for self-administration or by interview (Buck, Jacoby, Massey, & Ford, 2000). No special training is necessary, however a user’s manual and trainer’s manual are available for the original SIP (McDowell & Newell, 1996). There is not yet any evidence that the SA-SIP30 can be administered by proxy, however, the original SIP-136 can be used in this fashion (Sneeuw, Aaronson, de Haan, & Limburg, 1997).

Time:

The average scale completion time has not been reported, however, the SA-SIP30 is known to be a shorter scale than the original SIP, which takes 30 minutes on average to administer.

Alternative forms of the SA-SIP30

None.

Client suitability

Can be used with:

Patients with stroke.

Should not be used in:

  • The SA-SIP30 should be administered with caution to patients who have experienced a severe stroke. van Straten et al. (1997) noted that the SA-SIP30 might be less effective for patients with severe stroke because in developing the SA-SIP30, higher item weights were mostly associated with items that were removed, and these had been descriptive of more severe health status. Evidence of this came from the observation that agreement between scores obtained with the original SIP-136 and the SA-SIP30 were lower among more severely ill patients with stroke than among healthier patients (van Straten et al., 1997). However, it is important to note that in a subsequent study by van de Port et al. (2004), this trend was only observed on the Physical dimension of the SA-SIP30 and even then, the trend was less notable than on the SIP-68 (a short version of the original SIP-136).
  • The SA-SIP30 should be administered with caution to patients who have a major physical disability. van Straten et al. (2000) found that the total scores of the SA-SIP30 were largely explained by the Physical dimension of the scale (66% for the subscales of the Physical dimension versus 25% for the subscales of the Psychosocial dimension). This might result in any patient with a serious physical disability being automatically detected by the scale as having poor health-related quality of life.
  • Patients who require a proxy to complete. Although the original SIP has been validated for proxy use, proxy use has not been examined using the SA-SIP30. For patients who have had a stroke and who require a proxy, the Stroke Impact Scale is known to be a reliable and valid measure of quality of life (Duncan, Lai, Tyler, Perera, Reker, & Studenski, 2002).
  • Patients with aphasia. The SA-SIP30 has not been validated for use in patients with aphasia. A French questionnaire, the SIP-65, has been validated to assess quality of life in patients with aphasia, however this scale is not available in English (Benaim et al., 2003). The Stroke and Aphasia Quality of Life Scale-39 (SAQOL-39) is another measure that assesses quality of life and was developed specifically for use in patients with aphasia. This scale has been found to be an acceptable, reliable, and valid measure in patients with long-term aphasia (Hilari, Byng, Lamping, & Smith, 2003).

In what languages is the measure available?

English (van Straten et al., 1997)

Summary

What does the tool measure? Health-related quality of life
What types of clients can the tool be used for? The SA-SIP30 was developed for use in patients with stroke.
Is this a screening or assessment tool? Assessment
Time to administer The average scale completion time has not been reported, however, the SA-SIP30 is known to be a shorter scale than the original SIP, which takes 30 minutes on average to administer.
Versions The SA-SIP30 was adapted from the original SIP-136
Other Languages No translations of the SA-SIP30 have been conducted to date.
Measurement Properties
Reliability

Internal consistency:

Out of 2 studies that examined the internal consistency internal of the SA-SIP30, both studies reported excellent internal consistency

Test-retest:

No studies have examined the test-retest reliability of the SA-SIP30.

Inter-rater:

No studies have examined the inter-rater reliability of the SA-SIP30.

Validity

Content:

Items least relevant to patients with stroke were eliminated. Items with a skewed response pattern or those relevant to < 10% of patients were dropped. Linear regression was used to assess the relevance of remaining items. Item selection for each subscale was completed when items in the model explained 80% of the variance in score of the original total subscale. Least relevant subscales were excluded using a stepwise linear regression with forward inclusion. When adding another subscale to the model did increase the percentage of variance more than 1%, the process was stopped. Unreliable items were excluded, as long as at least 3 items remained in each subscale.

Construct:
Convergent:
Excellent correlations were found between the SA-SIP30 and the SIP-136 total score and subscales; the SIP-68 (shortened version of the SIP-136); and the global functional health score on the Rankin Scale. Adequate correlations with the disability score on the Barthel Index; total Rankin Scale; EuroQol; and the Frenchay Activities Index.

Discriminant:
Poor correlation between the SA-SIP30 and the Canadian Occupational Performance Measure.

Known groups:
The SA-SIP30 was able to distinguish clients with lacunar infarctions from those with cortical or subcortical lesions. One study reported that when using appropriate SA-SIP30 cut-off scores, the SA-SIP30 could classify patients as dependent in their activities of daily living; patients able to live independently; and patients having poor health-related quality of life.

Floor/Ceiling Effects None.
Does the tool detect change in patients?

One study examined found that the SA-SIP30 had only a moderate ability to detect change in patients from 6 months to 1 year post-stroke.

Acceptability The SA-SIP30 is shorter and simpler than the original SIP-136. The original SIP has been tested for use with proxy respondents, however the SA-SIP30 has not yet been tested for use by proxy respondent. The SA-SIP30 should not be administered to patients with aphasia, and should be used with caution in patients with a major physical disability or who have suffered a severe stroke.
Feasibility This shorter, simpler version of the SIP should represent less administrative burden and can be more easily included in both research and clinical setting. The scale is intended for self-administration or by interview. No special training is necessary. A user’s manual and trainer’s manual are available for the original SIP only. The SA-SIP30 is fairly simple to score and is based on weights that are applied to marked items, which are then summed for each subscale and expressed as a % for each subscale ranging from 0 to 100%. Higher scores indicate less desirable health outcomes.
How to obtain the tool? Click here to find a copy of the SA-SIP30. The SA-SIP30 can also be found in van Straten et al. (1997).

Psychometric Properties

Overview

To date, only a few studies have examined the psychometric properties of the Stroke-Adapted Sickness Impact Profile (SA-SIP30). For this reason, we have included for review all of the publications that we could identify on the scale. The SA-SIP30 was originally validated by its authors (van Straten et al., 1997; van Straten et al., 2000) and was later evaluated by van der Port et al. (2004).

Reliability

Internal consistency:
van Straten et al. (1997) developed and examined the reliability of the SA-SIP30 in 319 patients post-stroke. The total SA-SIP30 demonstrated excellent internal consistency (alpha = 0.85), as did the Psychosocial (alpha = 0.78) and Physical dimensions (alpha = 0.82). All subscales had adequate internal consistency with the exception of the Emotional Behavior (alpha = 0.57), and Ambulation (alpha = 0.54) subscales, which were poor. With the exception of the Communication subscale, the internal consistency of the SIP-136 was found to be slightly higher on all items than the internal consistency of the SA-SIP30.

van de Port, Ketelaar, Schepers, van den Bos, and Lindeman (2004) also examined the internal consistency of the SA-SIP30 in 122 patients with stroke and found excellent reliability for the total score (alpha = 0.82), and moderate reliability for the Physical dimension (alpha = 0.76). However, unlike the results of van Straten et al. (1997), the internal consistency of the Psychosocial dimension was found to be poor (alpha = 0.68).

Inter-rater:
Not reported.

Test-retest:
Not reported.

Validity

Criterion:

None.

Content:

van Straten et al. (1997) eliminated the least relevant items for patients with stroke from the SIP-136 . Items that had a skewed response pattern were dropped, as were items relevant to less than 10% of all patients. Linear regression was used to assess the relevance of the remaining items with a forward selection strategy, using the F statistic with p = 0.5 as the criteria level for selection. The item selection for each subscale was completed when the items in the regression model explained 80% of the variance in score of the original total subscale. The least relevant subscales were excluded by applying a stepwise linear regression with forward inclusion to explain the variation of the original total SIP score with the shortened subscales. When adding another subscale to the model did not result in an increase in the percentage of variance more than 1%, the process was stopped. Finally, unreliable items were excluded, while ensuring that at least three items remained in each subscale.

Construct:

A principal component analysis supported two dimensions (Physical and Psychosocial), which is evidence that the original dimension structure of the SIP-136 was retained with the SA-SIP30 (van Straten et al., 1997). Twenty percent of the SA-SIP30-explained score variance could be attributed to the Physical dimension and 11% to the Psychosocial dimension (van Straten et al., 1997).

Convergent:
van Straten et al. (1997) examined the convergent validity of the scale by comparing the scores of the SA-SIP30 with the scores on the 136-item version in 319 patients post-stroke. The SA-SIP30 total score explained 91% of the variance in SIP-136 scores. Furthermore, 87% of the original Physical dimension scores and 88% of the Psychosocial dimension scores could be explained by the SA-SIP30. For the different subscales, the percentages of explained variance ranged from 69% (Social Interaction) to 84% (Emotional Behavior). The Spearman rank correlation coefficient between the SA-SIP30 and the SIP-136 total scores was excellent (r = 0.96). Subscale correlations were also excellent, ranging from r = 0.75 (Emotional Behavior) to r = 0.90 (Body Care and Movement).

Also in this study by van Straten et al., the SA-SIP30 was correlated with the Barthel Index and the Rankin Scale. As expected, SA-SIP30 correlated moderately with the disability score on the Barthel Index (r = 0.50) and had an excellent correlation with the global functional health score on the Rankin Scale (r = 0.68), further demonstrating the convergent validity of the SA-SIP30.

van de Port, Ketelaar, Schepers, van den Bos, and Lindeman (2004) examined the convergent validity of the SA-SIP30 in 122 patients with stroke. The correlation between the SA-SIP30 and total SIP-68 (a shortened version of the SIP-136) scores was excellent (r = 0.98). Similar associations were reported for the Physical (r = 0.89) and Psychosocial (r = 0.84) dimension scores.

Cup et al. (2003) found that the SA-SIP30 correlated adequately with the Barthel Index (r = -0.517), the Rankin Scale (r = 0.468), the EuroQol (r = -0.483), and the Frenchay Activities Index (r = -0.426). The correlations among the SA-SIP30 and the EuroQol, Barthel Index, and Frenchay Activities Index are negative because a high score on the SA-SIP30 indicates poor health outcomes, whereas a high score on these other scales indicates positive health outcomes. The results of this study demonstrate the convergent validity of the SA-SIP30 with other frequently used standardized functional measures in stroke.

van Straten et al. (2000) conducted a linear regression analysis and found that common measures of physical disability were closely associated with SA-SIP30 scores. The Barthel Index accounted for 36% of the variance in total SA-SIP30 scores, the Rankin scale accounted for 53%, and the Euroqol index score accounted for 44%. The results of this study also confirm the convergent validity of the SA-SIP30 with other frequently used standardized functional measures in stroke.

Discriminant.
Cup et al. (2003) examined the discriminant validity of the Canadian Occupational Performance Measure in 26 patients with stroke. As predicted, the correlation between the scores on the Canadian Occupational Performance Measure and the SA-SIP30 was poor (r = 0.102). This was to be expected because the Canadian Occupational Performance Measure was developed to examine issues specific to the individual, whereas the SA-SIP30 is focused on a societal perspective of independence.

Known groups:
van Straten et al. (1997) found that the SA-SIP30 was unable to distinguish between clients with supratentorial and infratentorial strokes, as has been possible with the SIP-136 (de Haan, Limburg, & van der Meulen, 1995). However, the SA-SIP30 was able to distinguish clients with lacunar infarctions from those with cortical or subcortical lesions. Further, clients with lacunar infarcts reported better functional health than those with cortical or subcortial lesions on the Psychosocial dimension of the scale, the total SA-SIP30 score, and on all subscales with the exception of Emotional Behavior and Mobility.

van Straten et al. (2000) identified the cut-off scores for poor health outcomes by examining the area under the ROC curves (AUC). When using a cut-off SA-SIP30 score > 28, the percentage of patients correctly classified as dependent in their activities of daily living on the SA-SIP30 as assessed using the Barthel Index was adequate, 77% (AUC = 0.84). When using a cut-off SA-SIP30 score > 40 for the Physical dimension alone, the percentage of patients correctly classified as dependent in their activities of daily living was excellent, 84% (AUC = 0.90). When using a cut-off SA-SIP30 score > 25, the percentage of patients correctly classified as unable to live independently by the SA-SIP30 as measured by the Rankin Scale was adequate for the total score was excellent, 80% (AUC = 0.90). When using a cut-off of > 36 for the Physical dimension alone, the percentage of patients correctly classified was excellent, 83% (AUC = 0.90). When using a cut-off of > 33, the percentage of patients correctly classified as having poor health-related quality of life as assessed by the EuroQol was adequate, 80% (AUC = 0.80) for the total score. When using a cut-off > 40 for the Physical dimension alone, the percentage of patients correctly classified was also adequate, 79% (AUC = 0.86).

Responsiveness

van de Port et al. (2004) found that the SA-SIP30 demonstrated moderate responsiveness in a longitudinal study. Effect sizes from 6 months to 1 year post-stroke were 0.60 for the total SA-SIP30 scores, and 0.56 and 0.65 for the Physical and Psychosocial dimensions, respectively.

References

  • Benaim, C., Pelissier, J., Petiot, S., Bareil, M., Ferrat, E., Royer, E., Milhau, D., Herisson, C. (2003). A French questionnaire to assess quality of life of the aphasic patient: The SIP-65. [French]. Ann Readapt Med Phys, 46(1), 2-11.

  • Bergner, M., Bobbitt, R. A., Pollard, W. E., Martin, D. P., Gilson, B. S. (1976). The sickness impact profile: Validation of a health status measure. Med Care, 14(1), 57-67. 

  • Bergner, M., Bobbit, R. A., Carter, W. B., Gilson, B. S. (1981). The Sickness Impact Profile: development and final revision of health status measure. Med Care, 19, 787-805.

  • Buck, D., Jacoby, A., Massey, A., Ford, G. (2000). Evaluation of measures used to assess quality of life after stroke. Stroke, 31, 2004-2010.

  • Coons, S. J., Rao, S., Keininger, D. L., Hays, R. D. (2000). A comparative review of generic quality-of-life instruments. Pharmacoeconomics, 17, 13-35.

  • Cup, E. H. C., Scholte op Reimer, W. J. M., Thijssen, M. C., E., van Kuyk-Minis, M. A. H. (2003). Reliability and validity of the Canadian Occupational Performance Measure in stroke patients. Clinical Rehabilitaton, 17(4), 402-409.

  • de Haan, R. J., Limburg, M., van der Meulen, J. H. P. (1995). Quality of life after stroke. Stroke, 26, 402-408.

  • Duncan, P. W., Lai, S. M., Tyler, D., Perera, S., Reker, D. M., Studenski, S. (2002). Evaluation of proxy responses to the Stroke Impact Scale. Stroke, 33, 2593-2599.

  • Finch, E., Brooks, D., Stratford, P. W., Mayo, N. E. (2002). Physical Rehabilitations Outcome Measures. A Guide to Enhanced Clinical Decision-Making (second ed.), Canadian Physiotherapy Association, Toronto.

  • Golomb, B. A., Vickrey, B. G., Hays, R. D. (2001). A review of health-related quality-of-life measures in stroke. Pharmacoeconomics, 19(2), 155-185.

  • Hilari, K., Byng, S., Lamping, D. L., Smith, S. C. (2003). Stroke and Aphasia Quality of Life Scale-39 (SAQOL-39): Evaluation of acceptability, reliability, and validity. Stroke, 34, 1944-1950.

  • Lurie, J. (2000). A review of generic health status measures in patients with low back pain. Spine, 25, 3125-3129.

  • McDowell, I., Newell, C. (1996). Measuring Health. A Guide to Rating Scales and Questionnaires (2nd ed.), New York: Oxford University Press.

  • Sneeuw, K. C. A., Aaronson, N. K., de Haan, R. J., Limburg, M. (1997). Assessing quality of life after stroke. The value and limitations of proxy ratings. Stroke, 28, 1541-1549.

  • van Straten, A., de Haan, R. J., Limburg, M., Schuling, J., Bossuyt, P. M., van den Bos, G. A. M. (1997). A Stroke-Adapted 30-Item Version of the Sickness Impact Profile to Assess Quality of Life (SA-SIP30). Stroke, 28, 2155-2161.

  • van Straten, A., de Haan, R. J., Limburg, M., van den Bos, G. A. M. (2000). Clinical Meaning of the Stroke-Adapted Sickness Impact Profile-30 and the Sickness Impact Profile-136. Stroke, 31, 2610-2615.

  • van de Port, I. G. L., Ketelaar, M., Schepers, V. P. M., van den Bos, G. A. M., Lindeman, E. (2004). Monitoring the functional health status of stroke patients: the value of the Stroke-Adapted Sickness Impact Profile-30. Disability and Rehabilitation, 26(11), 635-640.

See The Measure

How to obtain a copy of the SA-SIP30?

The measure is provided in van Straten et al. (1997). Please click to view a copy of the SASIP-30.

Table of contents
Survey