- Email: [email protected]
Functional grouping in residential homes for people with intellectual disabilities
Research in Developmental Disabilities 24 (2003) 170–182
Functional grouping in residential homes for people with intellectual disabilities Jim Mansell*, Julie Beadle-Brown, Susan Macdonald, Bev Ashman Tizard Centre, University of Kent at Canterbury, Beverley Farm, Canterbury, Kent CT2 7LZ, UK Received 5 July 2002; received in revised form 31 October 2002; accepted 19 November 2002
Abstract The effects of functional grouping of people with intellectual disabilities on care practices in small residential homes in the community were investigated. A group comparison and a matched-pairs comparison were carried out in settings where less than or more than 75% residents were non-verbal, non-ambulant, had severe challenging behaviour, severe social impairment or were verbal and ambulant. Further analysis, focused on those with challenging behaviour was carried out using ordinal regression. In the group-comparison study, no significant differences were found for three of the five groups. Residents who were non-ambulant were rated as receiving care with less interpersonal warmth in grouped settings; residents with severe challenging behaviour were rated as receiving less good care practices in four respects (interpersonal warmth, assistance from staff, level of speech and staff teamwork) in grouped settings. The matched-pairs comparison found significant differences only for people with challenging behaviour, where grouped settings achieved less good results in terms of interpersonal warmth and staff teamwork. Higher adaptive behaviour and mixed settings were predictive of better care practices on 13 of 14 items of the Active Support Measure (ASM), with some setting variables also predictive for some items. Care practices only appear to vary for people with challenging behaviour, where grouped settings appear to offer less good results in some respects. # 2003 Elsevier Science Ltd. All rights reserved. Keywords: resident characteristics; grouping; residential homes
*
Corresponding author. Tel.: þ44-1227-823076; fax: þ44-1227-763674. E-mail address: [email protected] (J. Mansell). 0891-4222/03/$ – see front matter # 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0891-4222(03)00027-1
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
171
1. Introduction In a paper entitled ‘‘The less you’ve got the less you get: functional grouping, a cause for concern’’ Raynes (1980) commented that grouping people with intellectual disabilities who share similar characteristics in residential settings had negative effects on their care. She showed that the most severely disabled residents in her study received more institutional care practices, less stimulating speech from staff, less community involvement and a poorer physical environment. She argued that these differences were not all functionally related to the nature of resident disabilities, and concluded that ‘‘. . . for severely mentally retarded individuals, homogenous functionally based groupings appear to generate an environment in which direct-care staff consistently provide care which is characteristically unstimulating, undifferentiated, depersonalized, and rigid’’. Instead, Raynes argued that research should be carried out on environments in which people with different needs and characteristics lived together. She acknowledged that there might be disadvantages of such arrangements but suggested that this should be empirically investigated. Raynes’ data were of course derived from institutional settings; in the succeeding 20 years these have been substantially replaced by small-scale housing, with staff support, based in the community (Mansell & Ericsson, 1996). Predominantly, these services provide much richer environments, with much higher staffing levels, and espouse much more individualised care practices. Nevertheless, they typically group people together (albeit in much smaller groups than in institutions). Sometimes functional grouping in community settings reflects an explicit attempt to provide specially tailored support and care for individuals with similar needs. Here the evidence for relative effectiveness is mixed. For example, Hatton, Emerson, Robertson, and Henderson (1995) found that specialised group homes for people with intellectual disabilities who also had sensory impairments achieved better results than ‘‘ordinary’’ homes in which people with comparable needs lived with other people with intellectual disabilities. Emerson, Beasley, Offord, and Mansell (1992) found that homes which grouped people who had very severe challenging behaviour achieved no improvement in client outcomes over institutional care and Mansell (1995) found that specialised placements in ‘‘mixed’’ homes achieved better results than grouped settings. However, functional grouping also occurs in community-based services without special labels. This finding may reflect the belief of case managers and providers that homes which group people together on the basis of some shared characteristic will provide better quality care. These data may also reflect the common practice of leaving the most disabled residents in institutions until last, so that the only available residents for new services in the community are all very disabled. Since there is evidence of cost reduction over time in community-based services, without commensurate gains in client adaptive behaviour (Cambridge,
172
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
Hayes, Knapp, Gould, & Fenyo, 1994), it may also reflect the interaction of individual case management and downward pressure on residential costs leading to decisions to place people with less complex needs in less expensive services which can only serve such people. This approach leaves the most disabled in care settings which are de facto functionally grouped. There are good reasons for expecting differences between grouped and mixed settings in the way individuals with particular needs are treated, but it is not clear in which direction these differences might lie. For example, if none of the residents in a home can speak, one might expect staff to give up speaking to residents because of lack of reinforcement and the competition of more reinforcing interaction with staff. If everyone in a home has challenging behaviour, one might expect staff to adopt a more distant and cautious approach to interaction because of the risk that resident reactions might be aversive. On the other hand, one might expect staff working with residents who all share the same functional characteristics to develop greater expertise in respect of this. It is not clear whether the potential benefits of greater expertise among staff in grouped settings might outweigh the possible disadvantages. The aim of this study was to find out whether people with particular needs get treated differently by staff if they live in grouped rather than mixed settings. The study used data on resident characteristics, and on the nature of staff support and care provided from another study (Mansell, Ashman, Macdonald, & Beadle-Brown, 2002). Three analyses—a group comparison, a matched-pairs comparison and an ordinal regression—were carried out using the same set of data.
2. Method 2.1. Participants Participants were 303 people with intellectual disabilities living in 68 small homes provided across England by a national charity. Sixty percent of the participants were men and 95% were white. Average age was 39 years (range 16–78). These participants were selected from a larger study of 495 people, 95% of the total population of people with intellectual disabilities served by this charity, because they had assessments of their characteristics and also had ratings made by observers of how staff provided support to them. Adaptive behaviour of participants was rated using the American Association on Mental Retardation Adaptive Behavior Scale Part 1 (ABS) (Nihira, Leland, & Lambert, 1993). The ABS measures independence and adaptive behaviour and has a possible maximum total score of 322. It is divided into 10 behaviour domains: physical development, language development, domestic activity, numbers and time, economic activity, independent functioning, vocational activity, self-direction, responsibility and socialisation. Normative data are provided to calculate percentile ranks from raw scores. Mean total score in this study was 130 (range 12–298).
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
173
The BPI rates 29 specified problem behaviours in groups relating to selfinjurious behaviour, aggression and stereotyped behaviour, with the opportunity to specify additional problems. Problems were rated for frequency (never occurred, less than monthly, monthly, weekly, daily, hourly or more than hourly). Following McGill, Hughes, Teer, and Rye (2001), frequency of problem behaviours was summed to give a total score (maximum 222) indicating the overall burden represented by each person’s challenging behaviour. Average total frequency score was 13 (range 0–96). Each behaviour was also rated for whether it presented a behaviour management problem for staff, distinguishing between those causing severe management problems (staff had to intervene, upset other residents, marked effect on social atmosphere or unacceptable in public), lesser problems, those not causing a problem and potential problems (where the problem was controlled in the present environment but was likely to reoccur as a severe problem if the environment changed). Thirty-one percent of residents were rated as presenting a lesser management problem, 12% as presenting a potentially severe problem and 34% a severe problem. The SIS comprised seven items relating specifically to social impairment from the Abnormal Behaviour section of the Handicaps, Behaviour and Skills Schedule (Wing & Gould, 1978). These items related to whether the person made and used eye contact with other people, spontaneous shows of affection, their response to age peers, social play, willingness to join in leisure activities and overall quality of social interaction. These items were summed to give a total score, expressed as a percentage, with scores from 0 (profound social impairment) to 100 (no social impairment). Average score was 50 (range 0–100). These assessment schedules were sent to the manager of each home, who was asked to ensure that they were completed by a member of staff who knew the individual resident well. Questions and clarification about the information required were dealt with by the fourth author, who also followed up data collection to obtain the fullest information possible and dealt with queries arising during data processing. The reliability and validity of the ABS, BPI and the HBS (from which the SIS was drawn) have been studied and reported as acceptable by their authors. Reliability was measured in the study of which this was part by asking the same member of staff to repeat the rating a few weeks after initial data collection (i.e., the measure was of pre-test/post-test reliability). Pairs of ratings were made for 4% of the population studied (19 residents for the ABS and 21 for the BPI and the SIS). Pearson product-moment correlation coefficients were calculated for pairs of total scores on each measure and were .96 for the ABS (p < :001), .75 for the frequency score of the BPI (p < :001), .79 for the BPI severity score (p < :001), and .85 for the SIS (p < :001). 2.2. Settings Residents lived in 68 homes providing on average 6.5 places (range 2–14). The staff ratio averaged 0.65 (range 0.3–3.1). The length of service of staff averaged
174
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
44 months (range 5–104) and staff turnover (defined as the number of staff leaving the service in the previous year divided by the mean number of staff in post, expressed as a percentage) averaged 47% (range 0–227%). Staff teams had widely varying compositions and training: the percentage of senior staff (Manager, Deputy Manager or Senior Team Member) ranged from 31 to 100%, with an average of 64%. The percentage of the charity’s own management development programme completed by each Service Manager averaged 69% (range 0–100%). The percentage of the Manager and Deputy Managers in each service who had attended ‘‘active support’’ training (concerned with care practices designed to promote and facilitate resident engagement in meaningful activity) averaged 70% (range 0–100%). 2.3. Measurement Each home was visited in order to observe the way staff provided support to residents. Observations were made over a 3–4-hr period around a meal time because this seemed likely to provide many opportunities to see staff providing support. For each resident, the nature and quality of staff support was rated for the whole session using a 15-item rating scale, the Active Support Measure (ASM) (Mansell & Elliott, 1996). Each item is scored on a scale of 0 (poor, inconsistent performance) to 3 (good, consistent performance). The items are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
Age-appropriateness of activities and materials ‘‘Real’’ rather than pretend or very simple activities Choice of activities Demands presented carefully Tasks appropriately analysed to facilitate client involvement Sufficient staff contact for clients Graded assistance to ensure client success Speech matches developmental level of client Interpersonal warmth Differential reinforcement of adaptive behaviour Staff notice and respond to client communication Staff manage serious challenging behaviour well Staff work as a coordinated team to support clients Teaching embedded in everyday activities Specific written individual programmes in routine use
For 38 residents, a second rater made independent assessments. Inter-rater reliability was assessed using Cohen’s kappa (Cohen, 1960). For all but two items, kappa values exceeded 0.6 (ranging from 0.64 to 0.96) and were therefore judged acceptable (Bakeman & Gottman, 1986). For ‘‘Differential reinforcement of adaptive behaviour’’ kappa was 0.55. For ‘‘Staff manage serious challenging behaviour well’’ too little challenging behaviour was observed to assess inter-rater reliability and this variable has been excluded from the analysis.
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
175
2.4. Analysis 2.4.1. Group comparison The analysis was carried out for residents with any of five characteristics of interest: those who were non-verbal (ABS item 42 ¼ 0), non-ambulant (ABS item 28a ¼ 0), had severe behaviour management problems (any item on BPI scored as severe or potentially severe management problem), had severe social impairment (SIS score less than 50%), or who were verbal (ABS item 42 > 0) and ambulant (ABS item 28a ¼ 1). For each service, the percentage of people with each of these characteristics was calculated. Examination of the distribution of the data suggested that such differences as might be found were between homes where more than threequarters of residents shared the particular characteristic of concern (‘‘grouped’’) and homes where a lower proportion than this lived (‘‘mixed’’). The number of people in each group yielded by this procedure is given in Table 1. These sets of two groups were then used in a series of Mann–Whitney nonparametric tests of differences in ranks using SPSS (SPSS Inc., 1998) to examine differences between the groups in: (i) Participant characteristics: ABS, BPI or SIS scores (ii) Settings: staff ratio, length of service and turnover (iii) The pattern of support received from staff using each of the ASM categories Since this involved 21 comparisons for each of the five characteristics, only results significant at the .01 level are reported; all tests were two-tailed. 2.4.2. Matched-pairs comparison Each participant with one of the characteristics of interest living in a grouped setting (in which more than three-quarters of residents shared the same characteristic) was matched with another resident with the same characteristic living in a mixed setting. Residents were matched on their total ABS Part 1 score, on their BPI Aggression and Self-Injury subscale scores and on their score on the SIS. This procedure yielded 49 matched pairs of non-verbal residents, 36 pairs of non-ambulant residents, 35 with severe challenging behaviour, 64 with severe social impairment and 36 who were verbal and ambulant. The matched pairs were compared using Wilcoxon’s signed ranks test. As before, only results significant at the .01 level are reported. Table 1 Number of residents in group comparison
Grouped settings Mixed settings
Non-verbal
Non-ambulant
Challenging behaviour
Socially impaired
Verbal/ambulant
98 50
36 60
36 105
69 86
58 60
176
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
2.4.3. Ordinal regression In order to further investigate the relationship between client characteristics, setting characteristics, grouping and staff practices for people with severe challenging behaviour, a series of ordinal regression analyses were carried out using MINITAB (MINITAB Ltd., 1998). The dependent variables were total ASM score and each of the ASM items. In each case the independent variables were client characteristics (ABS, BPI); setting characteristics (number of clients, staff:client ratio, seniority of staff group, management development, active support training, length of service and staff turnover) and whether or not people lived in a grouped setting (more than 75% of people in the home with challenging behaviour). The independent variables were entered into the model and then, in order to emulate a backward stepwise regression procedure, the variable with the least significant coefficient was removed and the regression re-run. This procedure was repeated until only significant coefficients remained.
3. Results Results of the group comparisons are shown in Table 2. In terms of care practices, there were no significant differences between grouped and mixed settings for residents who were non-verbal, had severe social impairment or were verbal and ambulant. Residents who were non-ambulant were rated as receiving care with less interpersonal warmth in grouped settings. These settings had significantly more staff per resident. Residents with severe challenging behaviour were rated as receiving less good care practices in grouped settings in four respects (interpersonal warmth, assistance from staff, level of speech and staff teamwork). No significant differences between participant or setting characteristics for this group comparison were noted. Table 3 presents the results of the matched-pairs comparison. No significant differences between grouped and mixed settings were found in terms of participant or setting characteristics. For four of the five groups (non-verbal, nonambulant, socially impaired and verbal and ambulant) there were no differences in care practices between grouped and mixed settings. For people with severe challenging behaviour, those living in homes where more than three-quarters of residents had severe challenging behaviour received care from staff rated as significantly lower in interpersonal warmth and the extent to which staff worked as a coordinated team. In an attempt to explore whether there were differences in some aspect of participant characteristics that might help explain the differences between the results of the group comparison and the matched-pairs comparison a Mann– Whitney test was carried out. This test was used to compare the individuals in mixed settings in the group comparison who were matched with people in grouped settings in the matched-pairs comparison, with those who were not included in the matched-pairs comparison. No significant differences in participant or setting characteristics or in care practices were found.
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
177
Table 2 Advantages of grouped and mixed settings (group comparison) Non-verbal
Non-ambulant Challenging behaviour
Socially impaired
Z
Z
Z
p
p
Z
p
Verbal/ ambulant p
Z
p
Participant characteristics ABS 1.336 .182 1.885 .059 0.409 .683 1.908 .056 3.795 .000 BPI 0.613 .540 2.066 .039 1.580 .114 3.268 .001 1.146 .252 SIS 1.829 .067 0.573 .566 0.193 .847 0.348 .728 3.091 .002 Setting characteristics Staff ratio Length of service Turnover Care practices Total ASM score Age-appropriateness ‘‘Real’’ activities Choice of activities Demands presented carefully Tasks appropriately analysed Sufficient staff contact Graded assistance Speech matches development Interpersonal warmth Differential reinforcement Staff notice and respond Staff work as a team Teaching in everyday activities Individual programmes
0.061 .951 3.394 .001 1.830 .067 2.813 .005 1.309 .190 1.196 .232 0.386 .699 0.033 .974 1.698 .089 4.078 .000 0.262 .794 1.141 .254 1.830 .067 2.626 .009 2.089 .037 0.150 0.922 1.527 0.321 0.227
.881 .356 .127 .748 .821
0.337 0.216 0.659 0.464 1.111
.736 .829 .510 .643 .267
2.224 2.234 1.558 2.560 2.529
.026 .025 .119 .010 .011
1.198 1.645 1.541 0.240 0.389
.231 .100 .123 .810 .697
0.849 0.934 0.079 1.434 2.248
.396 .350 .937 .152 .025
0.645 .519 1.386 .166 2.474 .013 1.697 .090 0.840 .401 1.043 .297 2.041 .041 0.688 .491 2.059 .039 0.231 .817 0.624 .533 0.799 .424 2.706 .007 0.344 .731 1.092 .275 1.399 .162 1.271 .204 2.585 .010 1.245 .213 0.442 .658 0.143 .886 2.608 .009 2.723 .006 0.576 .564 1.651 .099 0.535 .593 0.148 .883 1.429 .153 0.155 .877 1.943 .052 2.064 .039 1.384 .166 0.367 .714 1.732 .083 0.966 .334 0.482 .630 2.185 .029 2.913 .004 1.124 .261 0.581 .561 0.646 .518 1.076 .282 2.117 .034 1.763 .078 0.810 .418 0.412 .680 0.539 .590 1.579 .114 0.766 .444 1.355 .175
Table 4 presents the significant results of the regression analysis on ASM for those with challenging behaviour, after stepwise elimination of the least significant variables to find the model of best fit. The regression was significant for total ASM score and for 13 out of 14 of the individual components. For all but ‘‘interpersonal warmth’’, ABS remained in the model of best fit, with higher ABS scores predictive of higher scores on ASM variables. For all but ‘‘real activities’’ and ‘‘staff notice and respond to client communication’’ being in a mixed setting as opposed to a group setting was predictive of better scores on the ASM variables. Staff ratio also figured in the model for ‘‘sufficient staff contact for clients’’, ‘‘speech matches developmental level of client’’, ‘‘differential reinforcement of adaptive behaviour’’, ‘‘staff notice and respond to client communica-
178
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
Table 3 Advantages of grouped and mixed settings (matched-pairs comparison) Non-verbal
Nonambulant
Z
p
Z
Participant characteristics ABS BPI aggression BPI SIB SIS
1.663 1.584 1.308 1.988
0.096 0.113 0.191 0.047
1.524 2.024 1.724 0.281
Setting characteristics Staff ratio Length of service Turnover
0.224 1.298 0.085
0.823 1.360 0.194 1.029 0.933 0.723
0.787 1.666 2.256 0.816 0.291 0.748 0.798
0.431 0.096 0.024 0.414 0.771 0.454 0.425
0.841 1.523
0.400 0.886 0.128 1.217
0.375 1.741 0.224 2.272
0.276 0.330
0.782 2.449 0.741 1.387
1.104
Care practices Total ASM score Age-appropriateness ‘‘Real’’ activities Choice of activities Demands Tasks analysed Sufficient staff contact Graded assistance Speech matches development Interpersonal warmth Differential reinforcement Staff notice and respond Staff work as a coordinated team Teaching in everyday activities Individual programmes
Challenging behaviour
Social impairment
Verbal/ ambulant
p
Z
p
Z
p
Z
p
0.127 0.043 0.085 0.778
0.074 0.735 0.597 1.022
0.941 0.462 0.550 0.307
0.669 3.565 3.427 1.441
0.504 0.000 0.001 0.150
1.803 0.244 2.622 0.601
0.071 0.807 0.009 0.548
0.010 0.879 0.232 1.982 0.081 1.284
0.379 0.048 0.199
0.267 0.123 0.571 0.490 1.326 0.350 0.577
0.789 0.902 0.568 0.624 0.185 0.726 0.564
0.082 0.325 0.023 1.126
0.745 0.355 0.260 0.226
0.722 0.821
0.014 3.066 0.166 1.604
0.002 0.024 0.109 0.592
0.981 0.577 0.554 1.180
0.564 0.238
0.270 0.243
0.808 0.662
0.508 0.343
0.732 0.713
0.476
0.409
0.682 1.047
0.295 2.699
0.007 1.778
0.075 0.680
0.497
0.651
0.515 0.985
0.325 1.073
0.283 0.988
0.323 0.099
0.921
0.113
0.910 0.302
0.763 1.312
0.190 1.321
0.186 0.220
0.826
0.257 0.655 0.081 0.592 0.224 1.229 2.507
0.174 0.442 0.303 0.246 0.470 1.528 0.797 0.512 0.935 0.554 0.823 0.219 0.012
1.958 1.290 0.869 1.342 1.045 1.752 0.751
0.658 2.588 0.806 1.196 0.127 1.745 0.050 0.197 0.385 0.180 0.296 0.080 0.453
0.358 0.921 0.560 1.485 0.443 0.600 1.266
0.721 0.357 0.576 0.138 0.657 0.549 0.205
tion’’, and ‘‘staff work as a coordinated team to support clients’’. For ‘‘sufficient staff contact for clients’’, ‘‘staff notice and respond to client communication’’, and ‘‘staff work as a coordinated team to support clients’’ other setting variables such as seniority, client number, management training, length of service, turnover and training in active support are also important.
4. Discussion 4.1. Adequacy of the data There are three important possible limitations of these data. First, since the data depend on ratings by independent observers of staff support and care to
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
179
Table 4 Significant results of regression analysis on ASM for people with challenging behaviour ASM variable
G-statistic ( p)
Goodnessof-fit ( p)
Concordant pairs (%)
Variables remaininga
Total ASM Age-appropriateness ‘‘Real’’ activities Choice of activities Demands Tasks analysed Sufficient staff contact Graded assistance Speech matches development Interpersonal warmth Differential reinforcement Staff notice and respond Staff work as a coordinated team
31.66 (.000) 25.06 (.000) 40.09 (.000) 27.259 (.000) 27.17 (.000) 22.612 (.000) 36.30 (.000)
.232 .810 .779 .677 .417 .464 .938
65.5 70.9 73.9 72.5 68.9 66.1 72.1
14.388 (.001) 46.888 (.000)
.258 .611
62.8 75.4
6.998 (.008) 14.343 (.002)
.05 .319
37 66.4
34.269 (.000)
.676
73.6
ABS, grouped ABS, grouped ABS ABS, grouped ABS, grouped ABS, grouped ABS, seniority, AS training, size staff ratio, grouped ABS, grouped ABS, staff ratio (p ¼ .113), grouped grouped ABS, staff ratio (p ¼ .084), grouped (p ¼ .275) ABS, seniority, size, staff ratio
33.163 (.000)
.743
75.2
Teaching in everyday activities
22.238 (.000)
.539
66.9
ABS (p ¼ .593), BPI (p ¼ .294), management development, AS training, length (p ¼ .366), turnover, size, staff ratio, grouped ABS, grouped
a Wald statistic for variables remaining in model of best fit significant at p < :05, unless otherwise indicated.
residents, the possibility of reactivity should be considered. The likely direction of any reaction would be to improve the quality of support and care offered by staff. This possibility cannot be excluded. However, it seems unlikely that reactivity differed between grouped and mixed settings because the possibility of this comparison was not known to staff or researchers when data were collected. The second possible limitation is that the data were all obtained from services provided by the same agency. Therefore, they may not be representative of all community-based residential services for people with intellectual disabilities. In the absence of comparable data from a national sample this possibility cannot be excluded and caution should be used in generalising from these findings to other services. The study of services all provided by one agency does remove the possibility that differences in agency philosophy and orientation might confound the comparison between grouped and mixed settings. The third possible limitation is that the grouped settings were not set up to offer special treatment and it might be argued that the potential benefits of grouping require this to be evident. The charity had no policy on whether to group people together or not, regarding the individual needs and preferences of residents as the
180
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
important determinant (though also recognising the power exercised by case managers and commissioning authorities). These data do not, therefore, comment on the advantages or disadvantages of functionally grouped residences set up to provide special kinds of support or care. They comment on functional grouping as it occurs in ordinary community-based services for people with intellectual disabilities.
5. Conclusions The group comparison showed no difference in patterns of staff support to residents between mixed and grouped settings for three of the characteristics of interest. In a fourth, people who were non-ambulant in grouped settings received care judged as providing less interpersonal warmth. The matched-pairs comparison found no difference for any of these four groups. Given the large number of statistical tests involved the finding in relation to people who are non-ambulant should be treated with caution. The general picture, then, is that for these groups the negative effects of functional grouping found by Raynes are not present in small residential homes in the community. This finding is consistent with the intention to provide more individualised care through the policy of replacement of institutions with small homes in the community. It is worth noting in particular that people who were verbal and ambulant were also treated in the same way by staff whether they lived in settings in which more or less than three-quarters of residents were verbal and ambulant. The possible concern that living with many more disabled people might impair the quality of care experienced by more able individuals is not borne out by these data. For people with severe challenging behaviour, all three studies show some detrimental effects of living in a setting where three-quarters or more of the residents have severe challenging behaviour. The differences found in the comparison studies are not large or numerous but the results of the regression do suggest a relatively consistent picture. With the most conservative interpretation, they suggest that any greater expertise staff do develop in grouped settings is not detectable in care practices and they counsel caution in making simplistic assumptions about the advantages or disadvantages of functional grouping. It is not possible from these data to determine why the effects of grouping should only be evident with people who have severe challenging behaviour. The most common forms of challenging behaviour are aggression and self-injury (Kiernan & Qureshi, 1993; Mansell et al., 2002). Staff are the most frequent victims of aggression in residential care, and nearly half the people showing aggression usually require physical intervention (Emerson et al., 2000). Aggression often results in negative emotional consequences for staff (Emerson & Hatton, 2000). Many staff also report negative emotional reactions to selfinjurious behaviour (Hastings, 1995). It may therefore be that the additive effect of grouping people with serious challenging behaviour together is so great that it overwhelms staff attempts to provide individualised care.
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
181
Further research is needed to replicate these findings among services provided by other organisations, to understand in more detail how and why care practices differ between grouped and mixed settings and to find out how staff practices interact with client characteristics and develop over time.
Acknowledgments The authors wish to thank the people with intellectual disabilities and their staff who provided the information used in this study; and the charity which provided access to its services and funded part of the study. Thanks are also due to Peter McGill for comments on the draft manuscript. This study was also partly funded by the National Health Service Executive.
References Bakeman, R., & Gottman, J. M. (1986). Observing interaction: An introduction to sequential analysis. Cambridge: Cambridge University Press. Cambridge, P., Hayes, L., Knapp, M., Gould, E., & Fenyo, A. (1994). Care in the community: Five years on. Aldershot: Arena, Ashgate Publishing Ltd. Cohen, J. A. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37–46. Emerson, E., Beasley, F., Offord, G., & Mansell, J. (1992). An evaluation of hospital-based specialized staffed housing for people with seriously challenging behaviours. Journal of Intellectual Disability Research, (36), 291–307. Emerson, E., & Hatton, C. (2000). Violence against social care workers supporting people with learning difficulties: A Review. Unpublished manuscript. Emerson, E., Robertson, J., Gregory, N., Hatton, C., Kessissoglou, S., Hallam, A. et al. (2000). Treatment and management of challenging behaviours in residential settings. Journal of Applied Research in Intellectual Disabilities, 13(4), 197–215. Hastings, R. P. (1995). Understanding factors that influence staff responses to challenging behaviours: An exploratory interview study. Mental Handicap Research, 8(4), 296–320. Hatton, C., Emerson, E., Robertson, J., & Henderson, D. (1995). The quality and costs of residential services for adults with multiple disabilities: A comparative evaluation. Research in Developmental Disabilities, 16(6), 439–460. Kiernan, C., & Qureshi, H. (1993). Challenging behaviour. In C. Kiernan (Ed.), Research to practice?: Implications of research on the challenging behaviour of people with learning disability (p. 367). Clevedon: British Institute of Learning Disabilities Publications. Mansell, J. (1995). Staffing and staff performance in services for people with severe or profound learning disability and serious challenging behaviour. Journal of Intellectual Disability Research, 39, 3–14. Mansell, J., Ashman, B., Macdonald, S., & Beadle-Brown, J. (2002). Residential care in the community for adults with intellectual disabilities: Needs, characteristics and services. Journal of Intellectual Disability Research, 46(8), 625–633. Mansell, J., & Elliott, T. E. (1996). Active support measure. Tizard Centre. Mansell, J., & Ericsson, K. (Eds.). (1996). Deinstitutionalization and community living: Intellectual disability services in Britain, Scandinavia and the USA. London: Chapman and Hall. McGill, P., Hughes, D., Teer, K., & Rye, L. (2001). Variability in staff reports of the frequency of challenging behavior. Research in Developmental Disabilities, 22(3), 221–231.
182
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
MINITAB Ltd. (1998). Minitab (Version 12.1). Coventry: MINITAB Ltd. Nihira, K., Leland, H., & Lambert, N. (1993). AAMR adaptive behavior scale—residential and community (2nd ed.). Austin, TX: Pro-Ed. Raynes, N. (1980). The less you’ve got, the less you get: Functional grouping, a cause for concern. Mental Retardation, 18, 217–220. SPSS Inc. (1998). SPSS for Windows (Statistical Package for the Social Sciences) (Version 9.0.0). Chicago: SPSS Inc. Wing, L., & Gould, J. (1978). Systematic recording of behaviors and skills of retarded and psychotic children. Journal of Autism and Childhood Schizophrenia, 8(1), 79–97.
Functional grouping in residential homes for people with intellectual disabilities Jim Mansell*, Julie Beadle-Brown, Susan Macdonald, Bev Ashman Tizard Centre, University of Kent at Canterbury, Beverley Farm, Canterbury, Kent CT2 7LZ, UK Received 5 July 2002; received in revised form 31 October 2002; accepted 19 November 2002
Abstract The effects of functional grouping of people with intellectual disabilities on care practices in small residential homes in the community were investigated. A group comparison and a matched-pairs comparison were carried out in settings where less than or more than 75% residents were non-verbal, non-ambulant, had severe challenging behaviour, severe social impairment or were verbal and ambulant. Further analysis, focused on those with challenging behaviour was carried out using ordinal regression. In the group-comparison study, no significant differences were found for three of the five groups. Residents who were non-ambulant were rated as receiving care with less interpersonal warmth in grouped settings; residents with severe challenging behaviour were rated as receiving less good care practices in four respects (interpersonal warmth, assistance from staff, level of speech and staff teamwork) in grouped settings. The matched-pairs comparison found significant differences only for people with challenging behaviour, where grouped settings achieved less good results in terms of interpersonal warmth and staff teamwork. Higher adaptive behaviour and mixed settings were predictive of better care practices on 13 of 14 items of the Active Support Measure (ASM), with some setting variables also predictive for some items. Care practices only appear to vary for people with challenging behaviour, where grouped settings appear to offer less good results in some respects. # 2003 Elsevier Science Ltd. All rights reserved. Keywords: resident characteristics; grouping; residential homes
*
Corresponding author. Tel.: þ44-1227-823076; fax: þ44-1227-763674. E-mail address: [email protected] (J. Mansell). 0891-4222/03/$ – see front matter # 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0891-4222(03)00027-1
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
171
1. Introduction In a paper entitled ‘‘The less you’ve got the less you get: functional grouping, a cause for concern’’ Raynes (1980) commented that grouping people with intellectual disabilities who share similar characteristics in residential settings had negative effects on their care. She showed that the most severely disabled residents in her study received more institutional care practices, less stimulating speech from staff, less community involvement and a poorer physical environment. She argued that these differences were not all functionally related to the nature of resident disabilities, and concluded that ‘‘. . . for severely mentally retarded individuals, homogenous functionally based groupings appear to generate an environment in which direct-care staff consistently provide care which is characteristically unstimulating, undifferentiated, depersonalized, and rigid’’. Instead, Raynes argued that research should be carried out on environments in which people with different needs and characteristics lived together. She acknowledged that there might be disadvantages of such arrangements but suggested that this should be empirically investigated. Raynes’ data were of course derived from institutional settings; in the succeeding 20 years these have been substantially replaced by small-scale housing, with staff support, based in the community (Mansell & Ericsson, 1996). Predominantly, these services provide much richer environments, with much higher staffing levels, and espouse much more individualised care practices. Nevertheless, they typically group people together (albeit in much smaller groups than in institutions). Sometimes functional grouping in community settings reflects an explicit attempt to provide specially tailored support and care for individuals with similar needs. Here the evidence for relative effectiveness is mixed. For example, Hatton, Emerson, Robertson, and Henderson (1995) found that specialised group homes for people with intellectual disabilities who also had sensory impairments achieved better results than ‘‘ordinary’’ homes in which people with comparable needs lived with other people with intellectual disabilities. Emerson, Beasley, Offord, and Mansell (1992) found that homes which grouped people who had very severe challenging behaviour achieved no improvement in client outcomes over institutional care and Mansell (1995) found that specialised placements in ‘‘mixed’’ homes achieved better results than grouped settings. However, functional grouping also occurs in community-based services without special labels. This finding may reflect the belief of case managers and providers that homes which group people together on the basis of some shared characteristic will provide better quality care. These data may also reflect the common practice of leaving the most disabled residents in institutions until last, so that the only available residents for new services in the community are all very disabled. Since there is evidence of cost reduction over time in community-based services, without commensurate gains in client adaptive behaviour (Cambridge,
172
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
Hayes, Knapp, Gould, & Fenyo, 1994), it may also reflect the interaction of individual case management and downward pressure on residential costs leading to decisions to place people with less complex needs in less expensive services which can only serve such people. This approach leaves the most disabled in care settings which are de facto functionally grouped. There are good reasons for expecting differences between grouped and mixed settings in the way individuals with particular needs are treated, but it is not clear in which direction these differences might lie. For example, if none of the residents in a home can speak, one might expect staff to give up speaking to residents because of lack of reinforcement and the competition of more reinforcing interaction with staff. If everyone in a home has challenging behaviour, one might expect staff to adopt a more distant and cautious approach to interaction because of the risk that resident reactions might be aversive. On the other hand, one might expect staff working with residents who all share the same functional characteristics to develop greater expertise in respect of this. It is not clear whether the potential benefits of greater expertise among staff in grouped settings might outweigh the possible disadvantages. The aim of this study was to find out whether people with particular needs get treated differently by staff if they live in grouped rather than mixed settings. The study used data on resident characteristics, and on the nature of staff support and care provided from another study (Mansell, Ashman, Macdonald, & Beadle-Brown, 2002). Three analyses—a group comparison, a matched-pairs comparison and an ordinal regression—were carried out using the same set of data.
2. Method 2.1. Participants Participants were 303 people with intellectual disabilities living in 68 small homes provided across England by a national charity. Sixty percent of the participants were men and 95% were white. Average age was 39 years (range 16–78). These participants were selected from a larger study of 495 people, 95% of the total population of people with intellectual disabilities served by this charity, because they had assessments of their characteristics and also had ratings made by observers of how staff provided support to them. Adaptive behaviour of participants was rated using the American Association on Mental Retardation Adaptive Behavior Scale Part 1 (ABS) (Nihira, Leland, & Lambert, 1993). The ABS measures independence and adaptive behaviour and has a possible maximum total score of 322. It is divided into 10 behaviour domains: physical development, language development, domestic activity, numbers and time, economic activity, independent functioning, vocational activity, self-direction, responsibility and socialisation. Normative data are provided to calculate percentile ranks from raw scores. Mean total score in this study was 130 (range 12–298).
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
173
The BPI rates 29 specified problem behaviours in groups relating to selfinjurious behaviour, aggression and stereotyped behaviour, with the opportunity to specify additional problems. Problems were rated for frequency (never occurred, less than monthly, monthly, weekly, daily, hourly or more than hourly). Following McGill, Hughes, Teer, and Rye (2001), frequency of problem behaviours was summed to give a total score (maximum 222) indicating the overall burden represented by each person’s challenging behaviour. Average total frequency score was 13 (range 0–96). Each behaviour was also rated for whether it presented a behaviour management problem for staff, distinguishing between those causing severe management problems (staff had to intervene, upset other residents, marked effect on social atmosphere or unacceptable in public), lesser problems, those not causing a problem and potential problems (where the problem was controlled in the present environment but was likely to reoccur as a severe problem if the environment changed). Thirty-one percent of residents were rated as presenting a lesser management problem, 12% as presenting a potentially severe problem and 34% a severe problem. The SIS comprised seven items relating specifically to social impairment from the Abnormal Behaviour section of the Handicaps, Behaviour and Skills Schedule (Wing & Gould, 1978). These items related to whether the person made and used eye contact with other people, spontaneous shows of affection, their response to age peers, social play, willingness to join in leisure activities and overall quality of social interaction. These items were summed to give a total score, expressed as a percentage, with scores from 0 (profound social impairment) to 100 (no social impairment). Average score was 50 (range 0–100). These assessment schedules were sent to the manager of each home, who was asked to ensure that they were completed by a member of staff who knew the individual resident well. Questions and clarification about the information required were dealt with by the fourth author, who also followed up data collection to obtain the fullest information possible and dealt with queries arising during data processing. The reliability and validity of the ABS, BPI and the HBS (from which the SIS was drawn) have been studied and reported as acceptable by their authors. Reliability was measured in the study of which this was part by asking the same member of staff to repeat the rating a few weeks after initial data collection (i.e., the measure was of pre-test/post-test reliability). Pairs of ratings were made for 4% of the population studied (19 residents for the ABS and 21 for the BPI and the SIS). Pearson product-moment correlation coefficients were calculated for pairs of total scores on each measure and were .96 for the ABS (p < :001), .75 for the frequency score of the BPI (p < :001), .79 for the BPI severity score (p < :001), and .85 for the SIS (p < :001). 2.2. Settings Residents lived in 68 homes providing on average 6.5 places (range 2–14). The staff ratio averaged 0.65 (range 0.3–3.1). The length of service of staff averaged
174
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
44 months (range 5–104) and staff turnover (defined as the number of staff leaving the service in the previous year divided by the mean number of staff in post, expressed as a percentage) averaged 47% (range 0–227%). Staff teams had widely varying compositions and training: the percentage of senior staff (Manager, Deputy Manager or Senior Team Member) ranged from 31 to 100%, with an average of 64%. The percentage of the charity’s own management development programme completed by each Service Manager averaged 69% (range 0–100%). The percentage of the Manager and Deputy Managers in each service who had attended ‘‘active support’’ training (concerned with care practices designed to promote and facilitate resident engagement in meaningful activity) averaged 70% (range 0–100%). 2.3. Measurement Each home was visited in order to observe the way staff provided support to residents. Observations were made over a 3–4-hr period around a meal time because this seemed likely to provide many opportunities to see staff providing support. For each resident, the nature and quality of staff support was rated for the whole session using a 15-item rating scale, the Active Support Measure (ASM) (Mansell & Elliott, 1996). Each item is scored on a scale of 0 (poor, inconsistent performance) to 3 (good, consistent performance). The items are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
Age-appropriateness of activities and materials ‘‘Real’’ rather than pretend or very simple activities Choice of activities Demands presented carefully Tasks appropriately analysed to facilitate client involvement Sufficient staff contact for clients Graded assistance to ensure client success Speech matches developmental level of client Interpersonal warmth Differential reinforcement of adaptive behaviour Staff notice and respond to client communication Staff manage serious challenging behaviour well Staff work as a coordinated team to support clients Teaching embedded in everyday activities Specific written individual programmes in routine use
For 38 residents, a second rater made independent assessments. Inter-rater reliability was assessed using Cohen’s kappa (Cohen, 1960). For all but two items, kappa values exceeded 0.6 (ranging from 0.64 to 0.96) and were therefore judged acceptable (Bakeman & Gottman, 1986). For ‘‘Differential reinforcement of adaptive behaviour’’ kappa was 0.55. For ‘‘Staff manage serious challenging behaviour well’’ too little challenging behaviour was observed to assess inter-rater reliability and this variable has been excluded from the analysis.
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
175
2.4. Analysis 2.4.1. Group comparison The analysis was carried out for residents with any of five characteristics of interest: those who were non-verbal (ABS item 42 ¼ 0), non-ambulant (ABS item 28a ¼ 0), had severe behaviour management problems (any item on BPI scored as severe or potentially severe management problem), had severe social impairment (SIS score less than 50%), or who were verbal (ABS item 42 > 0) and ambulant (ABS item 28a ¼ 1). For each service, the percentage of people with each of these characteristics was calculated. Examination of the distribution of the data suggested that such differences as might be found were between homes where more than threequarters of residents shared the particular characteristic of concern (‘‘grouped’’) and homes where a lower proportion than this lived (‘‘mixed’’). The number of people in each group yielded by this procedure is given in Table 1. These sets of two groups were then used in a series of Mann–Whitney nonparametric tests of differences in ranks using SPSS (SPSS Inc., 1998) to examine differences between the groups in: (i) Participant characteristics: ABS, BPI or SIS scores (ii) Settings: staff ratio, length of service and turnover (iii) The pattern of support received from staff using each of the ASM categories Since this involved 21 comparisons for each of the five characteristics, only results significant at the .01 level are reported; all tests were two-tailed. 2.4.2. Matched-pairs comparison Each participant with one of the characteristics of interest living in a grouped setting (in which more than three-quarters of residents shared the same characteristic) was matched with another resident with the same characteristic living in a mixed setting. Residents were matched on their total ABS Part 1 score, on their BPI Aggression and Self-Injury subscale scores and on their score on the SIS. This procedure yielded 49 matched pairs of non-verbal residents, 36 pairs of non-ambulant residents, 35 with severe challenging behaviour, 64 with severe social impairment and 36 who were verbal and ambulant. The matched pairs were compared using Wilcoxon’s signed ranks test. As before, only results significant at the .01 level are reported. Table 1 Number of residents in group comparison
Grouped settings Mixed settings
Non-verbal
Non-ambulant
Challenging behaviour
Socially impaired
Verbal/ambulant
98 50
36 60
36 105
69 86
58 60
176
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
2.4.3. Ordinal regression In order to further investigate the relationship between client characteristics, setting characteristics, grouping and staff practices for people with severe challenging behaviour, a series of ordinal regression analyses were carried out using MINITAB (MINITAB Ltd., 1998). The dependent variables were total ASM score and each of the ASM items. In each case the independent variables were client characteristics (ABS, BPI); setting characteristics (number of clients, staff:client ratio, seniority of staff group, management development, active support training, length of service and staff turnover) and whether or not people lived in a grouped setting (more than 75% of people in the home with challenging behaviour). The independent variables were entered into the model and then, in order to emulate a backward stepwise regression procedure, the variable with the least significant coefficient was removed and the regression re-run. This procedure was repeated until only significant coefficients remained.
3. Results Results of the group comparisons are shown in Table 2. In terms of care practices, there were no significant differences between grouped and mixed settings for residents who were non-verbal, had severe social impairment or were verbal and ambulant. Residents who were non-ambulant were rated as receiving care with less interpersonal warmth in grouped settings. These settings had significantly more staff per resident. Residents with severe challenging behaviour were rated as receiving less good care practices in grouped settings in four respects (interpersonal warmth, assistance from staff, level of speech and staff teamwork). No significant differences between participant or setting characteristics for this group comparison were noted. Table 3 presents the results of the matched-pairs comparison. No significant differences between grouped and mixed settings were found in terms of participant or setting characteristics. For four of the five groups (non-verbal, nonambulant, socially impaired and verbal and ambulant) there were no differences in care practices between grouped and mixed settings. For people with severe challenging behaviour, those living in homes where more than three-quarters of residents had severe challenging behaviour received care from staff rated as significantly lower in interpersonal warmth and the extent to which staff worked as a coordinated team. In an attempt to explore whether there were differences in some aspect of participant characteristics that might help explain the differences between the results of the group comparison and the matched-pairs comparison a Mann– Whitney test was carried out. This test was used to compare the individuals in mixed settings in the group comparison who were matched with people in grouped settings in the matched-pairs comparison, with those who were not included in the matched-pairs comparison. No significant differences in participant or setting characteristics or in care practices were found.
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
177
Table 2 Advantages of grouped and mixed settings (group comparison) Non-verbal
Non-ambulant Challenging behaviour
Socially impaired
Z
Z
Z
p
p
Z
p
Verbal/ ambulant p
Z
p
Participant characteristics ABS 1.336 .182 1.885 .059 0.409 .683 1.908 .056 3.795 .000 BPI 0.613 .540 2.066 .039 1.580 .114 3.268 .001 1.146 .252 SIS 1.829 .067 0.573 .566 0.193 .847 0.348 .728 3.091 .002 Setting characteristics Staff ratio Length of service Turnover Care practices Total ASM score Age-appropriateness ‘‘Real’’ activities Choice of activities Demands presented carefully Tasks appropriately analysed Sufficient staff contact Graded assistance Speech matches development Interpersonal warmth Differential reinforcement Staff notice and respond Staff work as a team Teaching in everyday activities Individual programmes
0.061 .951 3.394 .001 1.830 .067 2.813 .005 1.309 .190 1.196 .232 0.386 .699 0.033 .974 1.698 .089 4.078 .000 0.262 .794 1.141 .254 1.830 .067 2.626 .009 2.089 .037 0.150 0.922 1.527 0.321 0.227
.881 .356 .127 .748 .821
0.337 0.216 0.659 0.464 1.111
.736 .829 .510 .643 .267
2.224 2.234 1.558 2.560 2.529
.026 .025 .119 .010 .011
1.198 1.645 1.541 0.240 0.389
.231 .100 .123 .810 .697
0.849 0.934 0.079 1.434 2.248
.396 .350 .937 .152 .025
0.645 .519 1.386 .166 2.474 .013 1.697 .090 0.840 .401 1.043 .297 2.041 .041 0.688 .491 2.059 .039 0.231 .817 0.624 .533 0.799 .424 2.706 .007 0.344 .731 1.092 .275 1.399 .162 1.271 .204 2.585 .010 1.245 .213 0.442 .658 0.143 .886 2.608 .009 2.723 .006 0.576 .564 1.651 .099 0.535 .593 0.148 .883 1.429 .153 0.155 .877 1.943 .052 2.064 .039 1.384 .166 0.367 .714 1.732 .083 0.966 .334 0.482 .630 2.185 .029 2.913 .004 1.124 .261 0.581 .561 0.646 .518 1.076 .282 2.117 .034 1.763 .078 0.810 .418 0.412 .680 0.539 .590 1.579 .114 0.766 .444 1.355 .175
Table 4 presents the significant results of the regression analysis on ASM for those with challenging behaviour, after stepwise elimination of the least significant variables to find the model of best fit. The regression was significant for total ASM score and for 13 out of 14 of the individual components. For all but ‘‘interpersonal warmth’’, ABS remained in the model of best fit, with higher ABS scores predictive of higher scores on ASM variables. For all but ‘‘real activities’’ and ‘‘staff notice and respond to client communication’’ being in a mixed setting as opposed to a group setting was predictive of better scores on the ASM variables. Staff ratio also figured in the model for ‘‘sufficient staff contact for clients’’, ‘‘speech matches developmental level of client’’, ‘‘differential reinforcement of adaptive behaviour’’, ‘‘staff notice and respond to client communica-
178
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
Table 3 Advantages of grouped and mixed settings (matched-pairs comparison) Non-verbal
Nonambulant
Z
p
Z
Participant characteristics ABS BPI aggression BPI SIB SIS
1.663 1.584 1.308 1.988
0.096 0.113 0.191 0.047
1.524 2.024 1.724 0.281
Setting characteristics Staff ratio Length of service Turnover
0.224 1.298 0.085
0.823 1.360 0.194 1.029 0.933 0.723
0.787 1.666 2.256 0.816 0.291 0.748 0.798
0.431 0.096 0.024 0.414 0.771 0.454 0.425
0.841 1.523
0.400 0.886 0.128 1.217
0.375 1.741 0.224 2.272
0.276 0.330
0.782 2.449 0.741 1.387
1.104
Care practices Total ASM score Age-appropriateness ‘‘Real’’ activities Choice of activities Demands Tasks analysed Sufficient staff contact Graded assistance Speech matches development Interpersonal warmth Differential reinforcement Staff notice and respond Staff work as a coordinated team Teaching in everyday activities Individual programmes
Challenging behaviour
Social impairment
Verbal/ ambulant
p
Z
p
Z
p
Z
p
0.127 0.043 0.085 0.778
0.074 0.735 0.597 1.022
0.941 0.462 0.550 0.307
0.669 3.565 3.427 1.441
0.504 0.000 0.001 0.150
1.803 0.244 2.622 0.601
0.071 0.807 0.009 0.548
0.010 0.879 0.232 1.982 0.081 1.284
0.379 0.048 0.199
0.267 0.123 0.571 0.490 1.326 0.350 0.577
0.789 0.902 0.568 0.624 0.185 0.726 0.564
0.082 0.325 0.023 1.126
0.745 0.355 0.260 0.226
0.722 0.821
0.014 3.066 0.166 1.604
0.002 0.024 0.109 0.592
0.981 0.577 0.554 1.180
0.564 0.238
0.270 0.243
0.808 0.662
0.508 0.343
0.732 0.713
0.476
0.409
0.682 1.047
0.295 2.699
0.007 1.778
0.075 0.680
0.497
0.651
0.515 0.985
0.325 1.073
0.283 0.988
0.323 0.099
0.921
0.113
0.910 0.302
0.763 1.312
0.190 1.321
0.186 0.220
0.826
0.257 0.655 0.081 0.592 0.224 1.229 2.507
0.174 0.442 0.303 0.246 0.470 1.528 0.797 0.512 0.935 0.554 0.823 0.219 0.012
1.958 1.290 0.869 1.342 1.045 1.752 0.751
0.658 2.588 0.806 1.196 0.127 1.745 0.050 0.197 0.385 0.180 0.296 0.080 0.453
0.358 0.921 0.560 1.485 0.443 0.600 1.266
0.721 0.357 0.576 0.138 0.657 0.549 0.205
tion’’, and ‘‘staff work as a coordinated team to support clients’’. For ‘‘sufficient staff contact for clients’’, ‘‘staff notice and respond to client communication’’, and ‘‘staff work as a coordinated team to support clients’’ other setting variables such as seniority, client number, management training, length of service, turnover and training in active support are also important.
4. Discussion 4.1. Adequacy of the data There are three important possible limitations of these data. First, since the data depend on ratings by independent observers of staff support and care to
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
179
Table 4 Significant results of regression analysis on ASM for people with challenging behaviour ASM variable
G-statistic ( p)
Goodnessof-fit ( p)
Concordant pairs (%)
Variables remaininga
Total ASM Age-appropriateness ‘‘Real’’ activities Choice of activities Demands Tasks analysed Sufficient staff contact Graded assistance Speech matches development Interpersonal warmth Differential reinforcement Staff notice and respond Staff work as a coordinated team
31.66 (.000) 25.06 (.000) 40.09 (.000) 27.259 (.000) 27.17 (.000) 22.612 (.000) 36.30 (.000)
.232 .810 .779 .677 .417 .464 .938
65.5 70.9 73.9 72.5 68.9 66.1 72.1
14.388 (.001) 46.888 (.000)
.258 .611
62.8 75.4
6.998 (.008) 14.343 (.002)
.05 .319
37 66.4
34.269 (.000)
.676
73.6
ABS, grouped ABS, grouped ABS ABS, grouped ABS, grouped ABS, grouped ABS, seniority, AS training, size staff ratio, grouped ABS, grouped ABS, staff ratio (p ¼ .113), grouped grouped ABS, staff ratio (p ¼ .084), grouped (p ¼ .275) ABS, seniority, size, staff ratio
33.163 (.000)
.743
75.2
Teaching in everyday activities
22.238 (.000)
.539
66.9
ABS (p ¼ .593), BPI (p ¼ .294), management development, AS training, length (p ¼ .366), turnover, size, staff ratio, grouped ABS, grouped
a Wald statistic for variables remaining in model of best fit significant at p < :05, unless otherwise indicated.
residents, the possibility of reactivity should be considered. The likely direction of any reaction would be to improve the quality of support and care offered by staff. This possibility cannot be excluded. However, it seems unlikely that reactivity differed between grouped and mixed settings because the possibility of this comparison was not known to staff or researchers when data were collected. The second possible limitation is that the data were all obtained from services provided by the same agency. Therefore, they may not be representative of all community-based residential services for people with intellectual disabilities. In the absence of comparable data from a national sample this possibility cannot be excluded and caution should be used in generalising from these findings to other services. The study of services all provided by one agency does remove the possibility that differences in agency philosophy and orientation might confound the comparison between grouped and mixed settings. The third possible limitation is that the grouped settings were not set up to offer special treatment and it might be argued that the potential benefits of grouping require this to be evident. The charity had no policy on whether to group people together or not, regarding the individual needs and preferences of residents as the
180
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
important determinant (though also recognising the power exercised by case managers and commissioning authorities). These data do not, therefore, comment on the advantages or disadvantages of functionally grouped residences set up to provide special kinds of support or care. They comment on functional grouping as it occurs in ordinary community-based services for people with intellectual disabilities.
5. Conclusions The group comparison showed no difference in patterns of staff support to residents between mixed and grouped settings for three of the characteristics of interest. In a fourth, people who were non-ambulant in grouped settings received care judged as providing less interpersonal warmth. The matched-pairs comparison found no difference for any of these four groups. Given the large number of statistical tests involved the finding in relation to people who are non-ambulant should be treated with caution. The general picture, then, is that for these groups the negative effects of functional grouping found by Raynes are not present in small residential homes in the community. This finding is consistent with the intention to provide more individualised care through the policy of replacement of institutions with small homes in the community. It is worth noting in particular that people who were verbal and ambulant were also treated in the same way by staff whether they lived in settings in which more or less than three-quarters of residents were verbal and ambulant. The possible concern that living with many more disabled people might impair the quality of care experienced by more able individuals is not borne out by these data. For people with severe challenging behaviour, all three studies show some detrimental effects of living in a setting where three-quarters or more of the residents have severe challenging behaviour. The differences found in the comparison studies are not large or numerous but the results of the regression do suggest a relatively consistent picture. With the most conservative interpretation, they suggest that any greater expertise staff do develop in grouped settings is not detectable in care practices and they counsel caution in making simplistic assumptions about the advantages or disadvantages of functional grouping. It is not possible from these data to determine why the effects of grouping should only be evident with people who have severe challenging behaviour. The most common forms of challenging behaviour are aggression and self-injury (Kiernan & Qureshi, 1993; Mansell et al., 2002). Staff are the most frequent victims of aggression in residential care, and nearly half the people showing aggression usually require physical intervention (Emerson et al., 2000). Aggression often results in negative emotional consequences for staff (Emerson & Hatton, 2000). Many staff also report negative emotional reactions to selfinjurious behaviour (Hastings, 1995). It may therefore be that the additive effect of grouping people with serious challenging behaviour together is so great that it overwhelms staff attempts to provide individualised care.
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
181
Further research is needed to replicate these findings among services provided by other organisations, to understand in more detail how and why care practices differ between grouped and mixed settings and to find out how staff practices interact with client characteristics and develop over time.
Acknowledgments The authors wish to thank the people with intellectual disabilities and their staff who provided the information used in this study; and the charity which provided access to its services and funded part of the study. Thanks are also due to Peter McGill for comments on the draft manuscript. This study was also partly funded by the National Health Service Executive.
References Bakeman, R., & Gottman, J. M. (1986). Observing interaction: An introduction to sequential analysis. Cambridge: Cambridge University Press. Cambridge, P., Hayes, L., Knapp, M., Gould, E., & Fenyo, A. (1994). Care in the community: Five years on. Aldershot: Arena, Ashgate Publishing Ltd. Cohen, J. A. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37–46. Emerson, E., Beasley, F., Offord, G., & Mansell, J. (1992). An evaluation of hospital-based specialized staffed housing for people with seriously challenging behaviours. Journal of Intellectual Disability Research, (36), 291–307. Emerson, E., & Hatton, C. (2000). Violence against social care workers supporting people with learning difficulties: A Review. Unpublished manuscript. Emerson, E., Robertson, J., Gregory, N., Hatton, C., Kessissoglou, S., Hallam, A. et al. (2000). Treatment and management of challenging behaviours in residential settings. Journal of Applied Research in Intellectual Disabilities, 13(4), 197–215. Hastings, R. P. (1995). Understanding factors that influence staff responses to challenging behaviours: An exploratory interview study. Mental Handicap Research, 8(4), 296–320. Hatton, C., Emerson, E., Robertson, J., & Henderson, D. (1995). The quality and costs of residential services for adults with multiple disabilities: A comparative evaluation. Research in Developmental Disabilities, 16(6), 439–460. Kiernan, C., & Qureshi, H. (1993). Challenging behaviour. In C. Kiernan (Ed.), Research to practice?: Implications of research on the challenging behaviour of people with learning disability (p. 367). Clevedon: British Institute of Learning Disabilities Publications. Mansell, J. (1995). Staffing and staff performance in services for people with severe or profound learning disability and serious challenging behaviour. Journal of Intellectual Disability Research, 39, 3–14. Mansell, J., Ashman, B., Macdonald, S., & Beadle-Brown, J. (2002). Residential care in the community for adults with intellectual disabilities: Needs, characteristics and services. Journal of Intellectual Disability Research, 46(8), 625–633. Mansell, J., & Elliott, T. E. (1996). Active support measure. Tizard Centre. Mansell, J., & Ericsson, K. (Eds.). (1996). Deinstitutionalization and community living: Intellectual disability services in Britain, Scandinavia and the USA. London: Chapman and Hall. McGill, P., Hughes, D., Teer, K., & Rye, L. (2001). Variability in staff reports of the frequency of challenging behavior. Research in Developmental Disabilities, 22(3), 221–231.
182
J. Mansell et al. / Research in Developmental Disabilities 24 (2003) 170–182
MINITAB Ltd. (1998). Minitab (Version 12.1). Coventry: MINITAB Ltd. Nihira, K., Leland, H., & Lambert, N. (1993). AAMR adaptive behavior scale—residential and community (2nd ed.). Austin, TX: Pro-Ed. Raynes, N. (1980). The less you’ve got, the less you get: Functional grouping, a cause for concern. Mental Retardation, 18, 217–220. SPSS Inc. (1998). SPSS for Windows (Statistical Package for the Social Sciences) (Version 9.0.0). Chicago: SPSS Inc. Wing, L., & Gould, J. (1978). Systematic recording of behaviors and skills of retarded and psychotic children. Journal of Autism and Childhood Schizophrenia, 8(1), 79–97.