«Yona Lunsky (University of Toronto) and Paul Bramston (University of Southern Queensland) Approximate word count: 4468 words Keywords: Intellectual ...»
An additional measure of stressful life events, a short form of the Inventory of Negative Social Interactions (INSI; Lakey, Tardiff & Drew, 1994) was included in this study as a measure of convergent validity of the Lifestress-Inf. The INSI focuses on stressful interpersonal interactions and is structured identically to the Inventory of Socially Supportive Behaviours (ISSB; Barrera, Sandler, & Ramsey, 1981). The INSI has 40 items where the respondent rates the frequency of specific types of interactions that have occurred in the last month. For individuals with intellectual disabilities, the instrument is presented verbally, rather than in written form, and there are only three response options (not at all, sometimes, and a lot) and items are modified slightly to be better understood. For example, instead of “criticized you” the item is reworded as “said bad things about you”. This scale was modified by Lunsky and Benson (2001) with demonstrated reliability and validity. For the purposes of this study, the 25 items most commonly endorsed by adults with intellectual disability in that study were
6 Lifestress Informant 7
administered. The informant version of this questionnaire was modified similarly to the Lifestress-Inf, substituting “s/he” for “you”.
The Birleson Depressive Short form Self- Rating Scale (BDS-S) was included in this study to further assess convergent validity of the Lifestress-Inf and was administered as a self-report and an informant scale. The BDS-S is an 18-item measure, based on a 37-item questionnaire developed for 7 to 12-year-old children (Birleson, 1981). There are three response categories: “Most of the time,” “Sometimes,” and “Never.” These are scored 0, 1 or 2 with the higher score reflecting depression. This scale was adopted in Benson and Ivins‟ 1992 study involving adults with ID, where scores on it correlated significantly with informant ratings. The informant version of the Birleson consists of 5 items scored on a 3 point Likert scale (Benson & Ivins, 1992).
Procedures Potential participants were contacted by their service providers from 15 agencies who provide residential, therapeutic or vocational services to adults with ID and asked about the project and for permission to be contacted by the researcher, who then telephoned them and explained the study in greater detail and arranged a meeting date. Interviews occurred in settings comfortable for the person with ID, following an oral and written consent procedure. To minimized response sets, prior to administration of each self-report scale, participants were given two to three training items using the same response choices as those used with the scale. These response choices were presented both verbally and visually, with coloured symbols/words. If the participant appeared to answer each item in the same way (e.g., always picking the last option), response choices were presented in varied order. At the end of this interview, participants with ID nominated a caregiver informant. With participants‟ consent, the researcher then contacted the nominated caregivers by telephone, explained the study and asked if they would be willing to complete several questionnaires. The questionnaires were then mailed to the informants, with a postage-paid return envelope. Informants were telephoned twice following the mailing of the questionnaires if they were not returned within a month. All questionnaires reported here were returned within 8 weeks of the initial participant interview. This study received ethics approval from the University of Toronto Department of Psychiatry Institutional Review Board.
Results In order to reveal how many of the 30 potentially stressful life events were endorsed as occurring in participants‟ lives, a frequency analysis was conducted.
Results revealed that 14.76 (SD 4.34) life events occurred according to informants and 12.45 (SD 5.17) events according to participants with ID. This suggests that on average, nearly half the potentially stressful events listed in the scale did actually occur in participants‟ lives. Using the same 30 life events, Bramston et al found the number of events self-reported as occurring by 459 people with mild or moderate intellectual disability was 8.59 (SD 6.14) (Bramston, Fogarty & Cummins, 1999), a
7 Lifestress Informant 8
mean difference of less than one SD from the current study (that is not statistically significant). In the current study, adding the Likert scale ratings (maximum possible score of 60) produced a mean Impact score of 21.21 (SD 7.12) according to informants and 20.78 (SD 8.67) according to self-report.
Mean Frequency and Impact scores were compared according to the age and tested intelligence level of the participants with ID and were found to be unrelated.
However, women were rated as having higher stress Frequency scores than men according to informants [t (69) = 2.27, p.05] and higher stress Impact scores according to both informants [t(69) = 2.96, p.05] and self report [t(69) = 2.21, p.05].
Relationship between responses of informants and people with ID Impact scores on the informant and self-report versions of the scale were modestly correlated (r (70) =.41, p.005) suggesting a total of 16% of the variance is shared. Interestingly, Impact scores of family member informants and self reports were strongly correlated (r (29) =.70, p.001) and Impact scores on the staff and self report versions were only modestly correlated (r (41) =.34, p.05). This difference in strength of correlations was statistically significant (z=2.02, p.05). Mean Frequency scores of informants were significantly higher than mean Frequency scores of self reports (t(69)3.282, p.01), but there were no differences between the groups according to mean Impact scores (t(69)=0.35, p.05). Again, we see a different relationship between scores of family members compared to self and scores of staff compared to self in this regard. Families gave significantly higher scores than self reports for both Frequency (t(28)=4.85, p.001) and Impact of stress (t(28)=3.67, p.005). Staff, in contrast, did not give significantly higher scores for Frequency (t(40)=-.95, p.05) or stress Impact (t(40)=1.63, p.05) than their clients.
The most common and least common items/events reported by self and informant are presented in Table 1. Of the four most frequently reported items, only one was common to both self-report and informants (hearing others argue), and two items were in common amongst the least often reported items (not getting enough privacy and not making own choices).
Reliability Initial reliability was calculated for the present sample, with a recognition for the need to test again in a larger sample. The Lifestress-Inf had a Cronbach‟s alpha of 0.77, which is comparable to the.80 of the Lifestress Inventory (self-report) in this study, suggesting adequate internal reliability for both versions.
Validity In order to estimate the convergent validity of the Lifestress-Inf, the mean Impact score was compared to the mean total score for the Inventory of Negative Social Interactions (INSI). Both the self-report and informant versions were significantly correlated with the INSI (r (69) =.64 p.001 and r (57) =.58 p.001
8 Lifestress Informant 9
respectively) and the Birleson Depression Scales, self-report and informant versions (r (70) =.78 p.001 and r (65) =.31 p.05 respectively).
Discussion This study examined the qualities of the Lifestress-Inf, an informant version of the Lifestress Inventory, a self-report scale designed to measure stress in people with ID. The self-report nature of the Lifestress Inventory has restricted its use to people with higher-level communication skills, whereas the informant version enables
clinicians and researchers to quantify the stress from another important perspective:
the caregiver. In the present study, both scales were related to reports of interpersonal difficulties and depressed mood. However, there were some differences between caregiver and self-report perspectives. Given the modest but far from perfect relationship between the two perspectives, it is important to examine both sides and attempt to understand differences in perspectives when they do exist.
The items on the Lifestress Inventory that were most commonly self-reported in this Canadian study were largely the same as items most frequently reported in an earlier study amongst a large Australian sample of people with mild or moderate levels of intellectual disability (see Bramston, Fogarty and Cummins, 1999) and a recent study with an American sample (Sigan & MacLean, in press). Hence, it would
seem that across different nations people with ID consistently report as stressful:
hearing other people argue, death or serious illness of someone close, constantly being interrupted, and not being quick enough. Apart from the latter event, these common stressors are „other oriented‟ involving interpersonal relationships and networks. This finding can be understood in the light of research literature showing that disruption of the social environment and relationships are known to be highly stressful because attachment to others and the social support gained from such attachment is a fundamental human need (O‟Leary, 1990).
According to the results of this study, on average, participants experienced almost half of the 30 potentially stressful listed items/life events. It is interesting that women were reported to experience stressors more frequently than men, and that their impact was perceived as more severe according to both self-report and informants.
Higher reported stress by women has also been reported in the general research literature (Mazure & Maciejewski, 2003). Women with intellectual disability have also been found to report higher levels of depressed mood than men (Lunsky, 2003), as have women in the general population (McGrath et al., 1990). Higher levels of stress in women may be one explanation for gender differences in the onset of depression in the general population and in those with ID (see Mazure & Mciejewski, 2003 for further discussion on link between stress and depression in women in the general population).
The Lifestress-Inf, as used by 70 informants with respect to a person with disability known well to them, was shown to have adequate reliability and modest correlations with an informant measure of negative social interactions and an informant measure of depression. The Lifestress-Inf also showed modest correlations
9 Lifestress Informant 10
with the equivalent self-report stress scale. However, some informants may be better able to rate stress in those that they care for better than others. Specifically, findings from this study suggest that family member ratings were more highly correlated with self report ratings than were ratings of staff. Only 14 (48%) of these family members were living with the individuals they rated. Whether informants (family or staff) can accurately report stress of individuals who are non-verbal is a separate issue that was not addressed by this study. Clearly, the informant can provide some information, but their perspective cannot replace the perspective of the person with ID and should not be adopted on its own as a proxy measure.
The items/events most often and least often thought by informants in this study to be troublesome for people with ID differed appreciably from the self-report.
This suggests that the orientation or focus of an informant may differ from the selffocus of the individual, a finding replicating earlier research (Bramston & Fogarty, 2000). In this study the informants seemed to focus more on inadequacies or coping deficits in the people with disability, for example, not being able to find or maintain a job or getting „stuck‟ in a difficult situation and not knowing what to do. Interestingly, in this sample such issues were not as highly emphasized in the ID group. It may be that such stressors are a greater concern for the agencies that serve these individuals than the individuals themselves.
In contrast to the deficit orientation of caregivers, people with ID focused on current challenges in their interpersonal relationships as their greatest stress. It may be that caregivers are more aware of skill deficits because that is what they see and that is often the very reason caregiving is needed, while relationship difficulties are less obvious and less well understood. The obvious implication here is that informant and self-report scales should not be considered interchangeable. Self-report has the clear advantage of being able to potentially access feelings and attitudes not available to informants (see also Bramston et al., 2000; Cummins, 2002 for review on proxy measures).
Parallel versions of the Lifestress Inventory, given their relationship to self reported and informant rated depressed mood, have potential clinical utility, particularly when the two perspectives differ. For example, if a clinician relies primarily on the caregiver to report on relevant life events that may have contributed to the current problems and the caregiver fails to report on significant interpersonal issues or losses to the individual with ID, the clinician might not intervene accordingly. Alternatively, the clinician could rely exclusively on the self-report of an individual with ID who is very concerned about how others perceive him. Such an individual might not report that he is experiencing stressors at work or that he is worrying about how to complete work tasks because of fears that his job could be in jeopardy, but a caregiver might be able to offer this perspective. The process of caregivers and individuals with ID talking together about the kinds of things they think are stressful to the person with ID might help them to better understand and support one another. The clinician, having obtained both perspectives through the Lifestress Inventory, could easily facilitate such a discussion.
10 Lifestress Informant 11