Effects of a cognitive-behavioral intervention program on the health of caregivers of people with autism spectrum disorder

The sample comprised 17 parents of ASD-diagnosed people (10 women and 7 men) who had cared for their offspring for approximately 14 years. Participants mean age was 52 years, most of them were married (76%) and with a high educational level (64.7%). With regard to the care recipients, 13 were males and 4 females with a mean percentage of disability of 75%. Caregivers were recruited from members of an association of relatives of people with ASD ( Asociación Valenciana de Padres de Personas con Autismo, APNAV ) in the region of Valencia (Spain). Firstly, a meeting was conducted with caregivers to explain the aim of the research and the criteria for participation, namely, being the mother or father of an individual with a clinical diagnosis of an ASD; living at home with the care recipient; and having been the primary caregiver (that is, the person with the greatest responsibility for the daily care and rearing of the child) for at least two years before the study. In this meeting, parents were told about the content and procedure of the intervention program. After indicating that they wished to participate in the study, selected parents were interviewed to collect demographic data and information about their caregiving status. They all took part in the study voluntarily and completed an informed consent form in accordance with the ethical principles for research involving human subjects (Declaration of Helsinki).

The intervention program consisted of eight 2-hour sessions once a fortnight. The program is fundamentally based on a cognitive-behavioral approach, employing therapeutic strategies that have demonstrated efficacy in stress reduction. Further, the therapy was adapted to the ASD care context, taking into account factors that have been shown to be modulators of health status in family caregivers of people with ASD, in order to increase its effectiveness in this population. The development of the content and the implementation of the program were undertaken by the first author, with the supervision of the second author. The content of the intervention is summarized in Table 1.

Mood state . Mood was evaluated using the abbreviated version of the Profile of Mood States (POMS), developed by Fuentes, Balaguer, Meliá, & García-Merita, 1995 . This version is composed of 29 items configured by adjectives that respond to the question ‘How are you feeling right now?’ Items are grouped in five subscales: tension (e.g., tense, nervous), depression (e.g., sad, unhappy), anger (e.g. angry, bad-tempered), vigor (e.g., lively, energetic), and fatigue (e.g. fatigued, exhausted). A total score was also calculated by summing scores on all subscales, subtracting the vigor scale score and adding a constant of 100. The higher the total score, the worse the mood. Cronbach's alpha of this instrument is higher than .80.

Somatic symptoms . Somatic symptoms were assessed with the revised Spanish version of the Somatic Symptoms Scale (ESS-R) created by Sandín and Chorot (1995) . This instrument is composed of 80 items focused on the frequency of immunological (e.g., influenza, general state of low energy), cardiovascular (e.g., tachycardia, elevated blood pressure), respiratory (e.g., repetitive sneezing, difficult breathing), gastrointestinal (e.g. nausea, constipation), neurosensory (e.g., gait instability, dizziness), muscular (e.g., backache, joint pain), dermatological (e.g., eczema, pimples on the skin), and genito-urinary (e.g., discomfort when urinating, difficulty urinating) complaints. Each scale is composed of ten items ranked on a five-point Likert scale from 0 ( never) to 4 (more than five times in the last 2 years ), with reliability coefficients ranging from .79 to .84. The total number of symptoms is also calculated.

Depressive symptoms . Depressive symptomatology was evaluated by the Beck Depression Inventory ( Beck & Steer, 1993 ). This questionnaire assesses the intensity and severity of emotional, cognitive, and somatic aspects of depressive symptomatology through 21 self-report items ranked on a four-point Likert scale (e.g., crying: I don¿t cry anymore than I used to/I cry more than I used to/I cry over every little things/I feel like crying, but I can¿t; sadness: I do not feel sad/I feel sad much of the time/I am sad all the time/I am so sad or unhappy that I can¿t stand it). It has a reliability coefficient of .83.

Burden . Burden on the caregivers was evaluated using the Burden Interview created by Zarit, Reever, and Bach-Peterson (1980) . This instrument is composed of 22 items ranked on a five-point Likert scale from 0 ( never) to 4 (nearly always ) with a maximum score of 88. The reliability coefficient is .92. The items are related to health, social, and personal lifestyle, and interpersonal relations with the care recipient who has functional and behavioral disabilities (e.g., “Do you feel that your relative currently affects our relationships with other family members or friends in a negative way?” “Do you feel your health has suffered because of your involvement with your relative?” “Do you feel that your social life has suffered because you are caring for your relative?”). For each of these domains, caregivers express their feelings of burden with the higher scores representing a greater burden.

Burden was evaluated before and after the intervention. Somatic symptoms and depression were evaluated before the implementation of the intervention, immediately after the intervention and at 1 month of follow-up. Mood state was evaluated in session 1 (pre-treatment), session 4 (mid-treatment), and session 8 (post-treatment), and again 1 month later during the follow-up session. In the case of mood state, each participant completed the mood questionnaire before and after each of the aforementioned sessions.

ANOVA for repeated measures was performed to explore changes over time in somatic and depressive symptoms (comparing pre-treatment, post-treatment and follow-up scores) and in mood state (comparing pre-treatment, mid-treatment, post-treatment and follow-up scores). Greenhouse-Geisser adjustments to degrees of freedom and Bonferroni adjustments for multiple contrasts were applied as appropriate. For burden, a paired-sample t-test was employed (assessing differences between pre-and post-treatment scores).

All statistical analyses were performed using IBM SPSS Statistics (version 21.0), considering p < .05 to be significant. The descriptive values are expressed as mean and standard deviation (M and SD, respectively).

• a)

  Somatic symptoms . There were significant changes over time in respiratory symptoms, F (1.845, 29.525) = 6.602, p = .005, η2 partial = .292; neurosensory symptoms, F (1.824, 29.188) = 9.787, p = .001, η2 partial = .380; and genito-urinary symptoms, F (1.777, 28.431) = 5.234, p = .014, η2 partial = .246; as well as the total number of symptoms, F (1.435, 22.968) = 6.361, p = .011, η2 partial = .284. For respiratory, neurosensory and total number of symptoms, post-hoc analyses revealed differences between pre-treatment scores and both post-treatment ( p = .029, p =.004, and p = .050, respectively) and follow-up ( p = .013, p =.050, and p = .006, respectively) scores. In all cases, caregivers had fewer symptoms post-treatment and at follow-up than pre-treatment. With regard to genito-urinary symptoms, caregivers had significantly fewer symptoms at follow-up than pre-treatment ( p = .033) (Table 2).

  Table 2

  Scores (mean and standard deviation) in Somatic Symptoms and Burden in Pre- and Post-treatment and One Month Follow-up for Caregivers.

  	Pre-treatment	Post-treatment	Follow-up
  Somatic symptoms
  Immunological	8.76 ± 6.18	8.29 ± 4.95	7.41 ± 5.30
  Respiratory**	9.53 ± 7.70	6.29 ± 4.59	6.65 ± 6.72
  Cardiovascular	10 ± 7.55	6.35 ± 5.55	7.65 ± 6.50
  Gastrointestinal	9.59 ± 7.87	8.76 ± 7.02	8.71 ± 7.43
  Neurosensory***	8.12 ± 8.44	5.24 ± 7.07	6.47 ± 8.14
  Genital-urinary*	7.35 ± 7.21	5.35 ± 4.60	4.71 ± 5.34
  Muscular	13.53 ± 10.01	11.24 ± 8.91	12 ± 9.63
  Dermatological	11.18 ± 8.13	10.06 ± 6.91	10.65 ± 8.89
  Total symptoms*	78.05 ± 53.16	61.58 ± 39.78	64.23 ± 45.91
  Burden*	67.52 ± 12.50	61.58 ± 12.73

  *p < .05, **p < .01, ***p < .001

• b)

  Burden, depression, and negative mood state . In the case of burden, a significant reduction was found after the intervention, t(16) = 2.775, p = .014, d = 0.47.

For depressive symptoms, changes over time were also significant, F(1.349, 21.581) = 5.329, p = .022, η2 partial = .250, caregivers having fewer depressive symptoms at follow-up than pre-treatment ( p =.014) (Figure 1).

Figure 1

Scores (mean and standard deviation) in Depressive Symptomatology in Pre and Post-treatment and One Month Hollow-up.

With regard to mood state, there were significant changes over time in scores for total negative mood, F (2.347, 37.559) = 9.913, p =.001, η2 partial = .383, and all the subscales, namely, for depression, F (2.224, 35.583) = 7.568, p = .001, η2 partial = .321; vigor F (2.484, 39.736) = 4.728, p = .010, η2 partial = .228; anger, F (2.147, 34.347) = 4.833, p =.013, partial = .232); fatigue F (2.528, 0.447) = 3.762, p = .023, η2 partial = .190); η2 partial and tension, F (2.099, 33.590) = 5.845, p = .006, η2 partial = .268). In the case of depression, vigor, and total negative mood, significant differences were observed from pre-treatment to mid-treatment ( p = .010, p = .027, and p = .010 for each subscale respectively) and to follow-up( p = .004, p = .050, and p = .007 respectively). For fatigue and tension, differences were significant comparing pre-treatment and follow-up scores ( p = .050 and p = .020, respectively). We only observed significant differences between post-treatment and follow-up in the case of total negative mood ( p = .011). Post-hoc analyses did not identify any significant differences in anger. On all subscales except for vigor, scores fell progressively over the course of the whole intervention program. No significant interaction was observed between the time of measurement with respect to the session (pre-or post-session) and time of measurement with respect to the treatment (pre-/mid-/post-treatment or follow-up) for any of the mood scores evaluated ( Figure 2)

Figure 2

Effect of Each Evaluated CBT Session on Mood State

*p < .05; **p < .01.