randomised, controlled trial

Method Fifty-seven parents randomised to I0 weeks ofex~erimental Habilitation programmes for intellectual disability are primitive in developing countries (Heron & Myers, 1983). Resources to develop specialist care are scarce in these nations. One compensatory option for this deficit is to facilitate the primary care-giver to take on the role of therapist (McLoughlin, 1992), because parents are the focus of intervention (Myreddi, 1992). Parental attitude influences the development and training of the developmentally disabled child (Beckett-Edwards, 1994) and is a dynamic adaptational process subject to change (Gallimore et al, 1993). Changes in and control therapy were assessed using parental attitude occur with intervention the Parental Attitude Scale towards the (Bruiner & Beck, 1984; Sameroff & Managementof Intellectual DisabilityThe 1990). Interventions with parents are varpreand post-intervention measurements ied (Girimaii, 19931, including a model were done by a single-blinded rater and with an O ~ ~ O r m n i t y raise questions and discuss problems over a period of time (Stecompared. phens & Wyatt, 1969; Cunningham et al, Results The intervention group had a 1993). This randomised-controlled ma1 evalustatistically significant increase in the ates the efficacy of Interactive Group outcome scores and clinical improvement psychoeducation (IGP) in changing attiin the total parental attitude score, tudes towards children with intellectual orientation towards child-rearing, disability.

Method Fifty-seven parents randomised to I0 weeks ofex~erimental Habilitation programmes for intellectual disability are primitive in developing countries (Heron & Myers, 1983). Resources to develop specialist care are scarce in these nations. One compensatory option for this deficit is to facilitate the primary care-giver to take on the role of therapist (McLoughlin, 1992), because parents are the focus of intervention (Myreddi, 1992). Parental attitude influences the development and training of the developmentally disabled child (Beckett-Edwards, 1994) and is a dynamic adaptational process subject to change (Gallimore et al, 1993). Changes in and control therapy were assessed using parental attitude occur with intervention the Parental Attitude Scale towards the (Bruiner & Beck, 1984;Sameroff & Managementof Intellectual DisabilityThe 1990). Interventions with parents are varpre-and post-intervention measurements ied (Girimaii, 19931, including a model were done by a single-blinded rater and with an O~~O r m n i t y raise questions and discuss problems over a period of time (Stecompared.

Results
The intervention group had a 1993). This randomised-controlled ma1 evalustatistically significant increase in the ates the efficacy of Interactive Group outcome scores and clinical improvement psychoeducation (IGP) in changing attiin the total parental attitude score, tudes towards children with intellectual orientation towards child-rearing, disability.
knowledge towards intellectual disability and attitude towards management of intellectual disability, but no change in attitude towards the intellectual disability subscale. Declaration of interest None.

Subjects
Recruits from the facility for intellectually disabled children were screened to see whether they met the trial criteria, including DSM-IV criteria (American Psychiatric Association, 1994) for intellectual disability. Inclusion criteria were: age below 13 years; IQ of c 7 0 , confirmed by the Binet Kamat Scale of Intelligence (Kamat, 1967) or Gesell Developmental Schedule (Gesell, 1940); and conversation skill in English, Hindi, Bengali or Tamil. Exclusion criteria were: presence of any psychiatric morbidity in the parent (mood disorders, anxiety disorder, psychoses, etc.) or the intellectually disabled child (autism, behavioural problem, etc.); physical handicap in the intellectually disabled child; and exposure to any therapy before. The intelligence scale or developmental schedule was administered by a psychologist as pan of the clinical assessment. Presence of physical handicap and psychiatric morbidity was eliminated with clinical examination as well as interviews by an occupational therapist and psychiatrist, respectively. Consecutive, biological parents of intellectually disabled children who satisfied the selection criteria, gave signed consent and enrolled either in the day-care or residential therapy programme formed the study population. Only one parent, either the mother or the father, who agreed after enrolment to stay or accompany the child during the training period, was included in the trial.
Sample size was calculated based on the pilot data at the tenth week from the subscore on the outcome of orientation towards child rearing. The mean (s.d.) difference between the two arms was 10 (eight). By keeping alpha and beta errors at 5% and 20%, respectively, the sample size needed in each arm was 15, with an expected drop-out of 8-10%.

Study design
At Week 1, 57 parents who were suitable for the trial were randomised to one of two conditions: the experimental IGP or the didactic lecture control. Simple randomisation was done on s o h a r e used by a special educator who was involved in neither the assessment procedures nor the therapy procedures. The entire training duration was 12 weeks for both groups, consisting of multiple therapeutic modules focused on the children and their parents. The experimental and control therapy for the parents was only for 10 weeks out of the total 12 weeks of training. Other training components (play-material-making, etc.) for the parents of both groups continued during the last two weeks.
The IGP was in the form of closed group sessions, which were conducted twice a week for 10 weeks by psychologists and special educators. Each one-hour session was on the aspects of child-rearing skills, developmental milestones and delays, common causes of intellectual disability, comorbidities, skill deficit, problem behaviour, behavioural techniques, sexuality and marriage and legal and social support systems for the intellectually disabled in India. Each session alternated between information dissemination, discussion among parents and with therapist and problem-solving tasks.
For the wnaol group, classes, which were only for information dissemination, were conducted on the same topics and for the same period as the didactic lecture but without the discussion. Roll call was taken to encourage attendance in both ~roups.
The children attending the therapy programme in both groups were given training in self-care skills, social skills, pre-vocational skills, control of problem behaviour using special education and behavioural techniques, irrespective of their being in the experimental or control group.
Participating parents completed the questionnaire in the first and tenth week of the therapy programme. The assessor, one of the authors (P.S.S.R.), who was blinded to randomisation and therapies, administered the Parental Amtude Scale towards Management of Intelleaual Disability (PAM-ID; Bhatti et al, 1985).
The PAM-ID is a standardised instrument developed for the Indian population to measure the total parental attitude (PAT) towards intellectual disability in four areas: orientation of parents towards childrearing skills (OCR), knowledge on intellectual disability (KID), attitude towards intellectual disability (AID) and attitude towards management of intellectual disability (AMID). The self-rated assessment scale was 56 items, which are scored on a Likert scale of 1-4. The Binet Kamat Scale of Intelligence is the Indian adaptation of the 1934 version of the Stanford-Binet Scale of Intelligence. Same of the tests, items and materials were amended to suit Indian conditions, such as Indian coins, typically Indian pictorial scenes, vocabulary and Indian concepts. The intelligence scale assessed the child's skills in six areas: memory, language, conceptual thinking, reasoning, numerical reasoning, visuomotor coordination and social intelligence. Gesell Developmental Schedule was used in children below three years of age.

DATA ANALYSIS
Categorical data from the first and renth week were compared using d~ $ test with Yates' wrrection when required. Continuous outcome measures and the diffmna (post-pre) in raw scores benvcen groups were compared using the Mann-Whimey U-test The mean attendance between the groups was compared using Student's t-test. For the outcomes, multiple regression analysis was used to take into account the possible confounding effect of level of intellectual disability, education of parents, profession of parents, age of parents, number of children, residence and socio-economic status. confidence interval was obtained for the difference in means in the experimental and control groups; P<O.OS (twotailed test) was considered significant. Statistical analysis was performed using the SPSSlPCc and Epilnfo software packages.

Subject flow
A total of 804 children underwent diagnostic i n t e~e w s and screening tests at the centre: 138 had average intelligence, 182 had low average intelligence and 484 had various levels of intellectual disability. In this study: 185 children enrolled themselves for a residential or day-care therapy programme; 128 intellccntally disabled children had to be excluded from the study because 93 (73%) had physical or psychiatric cornorbidity; 18 (14%) parents did not have the working knowledge of the language of training; 11 (8%) children did not fulfil the age criteria; and six (5%) children had prior therapy elsewhere.

sample c)llractc*frtkr
Among the 57 intellectually disabled children who entered the study, there were 39 (68.5%) mildly, 12 (21%) moderately, four (7%) severely and two (35%) profoundly disabled children. Thirty-two (56%) of the children had an identifiable c a w , such as Down's syndrome, hypoxic ixhaernic encephalopathy or encephalitis, as opposed to 25 (44%) who had no identifiable cause for the intellecnral disability. Nearly 75% and 72% of the children were male and from a rural area, respectively. Ninety-four  The experimental and control group parents revealed no statistically significant differences in intake socio-demographic variables. The distribution of sociodemographic variable according to treatment and control groups is presented in Table 1.

Completer analysis
There was no statistically significant difference between the experimental and control groups in the first-week data on the subscales of OCR, AID or AMID. However, there was a statistically significant higher score in the control group (Table 2) than the experimental group on the subscales of KID (P=0.0002) and the PAT score (P=O.Ol). Table 2 shows a statistically significant increase in the post-intervention OCR, KID, AMID and PAT scores among the experimental group over the control group. The maximum increase in scores was for PAT and OCR, followed by KID and T d e 3 Dierence in raw score of total parental attitude (PAT) and subxale measures for treatment and control groups  AMID. The difference in the raw PAT score and the sub-scales also showed a statistically significant improvement in the treatment group (Table 3) in the areas of OCR, KID, AMID and PAT when compared with the control group parents.
There was a statistically significant improvement in the difference in the raw scores among the experimental group parent after adjusting for the confounding variables with multiple regression. The difference in the raw PAT score in the experimental group was significantly (t=4.2, P=0.0001) higher (by 14.0 units) than the control group. Also, as the socioeconomic status increased there was a significant (t=2.3, P=0.02) increase in the difference in the raw score (regression coefficient=8.0). The difference in the raw score of OCR in the experimental group was significantly (t=3.2, P=0.002) higher (by 8.0 units) than the control group. As the socio-economic status increased there was a significant (t=2.1, P=0.04) increase in the difference in the raw score (regression coefficient=6.8). The difference in the raw score of KID in the experimental group was significantly (t=S.O, P=0.0001) higher (by 4.7 units) than the control group. The difference in the raw score of AMID in the experimental group was significantly (t=2.0, P=O.OS) higher (by 1.3 units) than the control group, and for AID the difference in raw score in the urban area was significantly (t=1.97, P=O.OS) higher (by 1.6 units) than the rural area. These results were comparable with the findings of bivariate analysis.
The improvement in attitude was statistically and clinically significantly better among the experimental group parents of children with mild intellectual disability: PAT (Z=2.9, P=0.003), OCR (Z=0.4, P=0.01) and KID (Z=4.2, P=0.0001). There was no improvement in the PAT score or any of the sub-scale scores of the experimental group parents with children of moderate intellectual disability, although clinically the parents had shown improvement.

Intention-to-treat analysis
Nine per cent of the parents dropped out from the study by Week 2, which included three parents from the experimental group, and two from the control group. Children dropped out of the study because of father's job transfer (n=l), mother's illness (n=l) and unknown reasons (n=3). Data for parents who did not complete the trial were projected as though they had not improved and the intention-to-treat analysis was conducted. As with the completer analysis, the post-intervention scores in the experimental group showed a statistically significant improvement in the areas of PAT (Z=3.1, P=0.001), OCR (Z=2.7, P=O.OOS), KID (Z=2.3, P=0.01) and AMID (Z=2.1, P=0.03); AID did show some improvement (Z=0.4, P=0.6).

Is IGP of benefit?
At final assessment, among the experimental group parents there was a significant clinical improvement in childrearing skills and knowledge on intellectual disability, with a statistically significant increase in PAT, OCR, KID and AMID subscale scores; clinically, AMID had improved significantly, as evidenced by increased therapy compliance by the experimental group parents. Knowledge, in spite of being statistically significantly low among the experimental group in the first week, showed a significant increase over the control group after intervention. There was also a clinical improvement in AID, reflected by an improved parent-child interaction, although the AID score did not reach statistical significance.
The improvement in attitude was statistically significant among the parents of children with mild intellectual disability when compared with moderate intellectual disability. This could be because of the small subsample size of moderately disabled children resulting in a Type 2 error. A comparison with parents of severely and profoundly disabled children could not be made because of the smaller numbers.
The effect of subculture on the differences in attitude cannot be ruled out because the majority of the study population was from a nual, lowlmiddle socioeconomic background. However, this effect has not been evident in this study because of the equal distribution (Table 1) of these two variables among the experimental and control groups as well as the confounding effect due to these variables having been adjusted by multiple regression.
The statistically significant higher score in the control group than the experimental group on the subscales of KID and PAT score pre-intervention could also have been due to randomisation.

Therapeutic components
These improvements in the experimental group and the lack of decline in the control group could be attributed partly to the increase in knowledge (Narayan, 1993), although it does not explain the lack of improvement in the AU) score. The other explanation for these changes among the experimental group may be the content of the intervention as well as the circumstances in which this occurred. Parents in the control group, during lectum, were given information without a chance for discussion. The intervention group, in addition to the information given, had specific additional factors operating, such as repetition with clarification of the information being conveyed, opportunity to discuss views with the therapist and other group members about the information being conveyed and exercises to improve problemsolving skills. This may have a distinct role in enhancing the assimilation of the acquired knowledge, bringing about changes in attributional style; ruminative style, dysfunctional attitude, expectations and assumptions, thus gaining mastery over the knowledge they have gained and eventually changing their attitude. These cognitive changes need to be tested.
The clinical improvement in the AID, reflected by an improved parent-child interaction, may be explained by the improvement perceived in the children with the multimodal training offered. Post-study interviews with parents suggested that they were unsure as to how to manage the new skills at home. Mothers were particularly concerned about their ability to utilise their time between the regular home-making and training responsibilities. This concern might have contributed partly to the lack of change in the AID score. Subsequently, time management skills were added to the training that the parents were beiig given.