Improving pain‐related communication in children with autism spectrum disorder and intellectual disability

Abstract The communication of pain in individuals with co‐morbid Autism Spectrum Disorder and intellectual disability (ASD‐ID) is largely unexplored. The communication deficits associated with ASD‐ID can result in nonverbal behavior such as self‐injurious behavior, aggression, irritability, and reduced activity as a means to communicate that pain is present. The objective of this study was to determine whether a behavioral‐based educational intervention could increase the pain‐related communication of children with ASD‐ID who experience pain frequently. Specifically, the study aimed to determine if children with ASD‐ID can label the location of their pain or quantify pain severity and request pain relief. The sample included three children with ASD‐ID who experienced pain frequently. The intervention utilized educational materials and behavioral reinforcements and the intervention was conducted using a series of case studies. Pain was assessed daily by caregivers using the Non‐Communicating Children's Pain Checklist—Postoperative (NCCPC‐PV) and the ability of the individual to identify and express pain was recorded using the Wong Baker FACES Pain (WBFPS) Scale. Challenging behavior was recorded based on frequency count. The results indicated that all participants displayed the ability to independently respond to a question about how they were feeling by vocalizing the location of pain or indicating their level of pain on the WBFPS and requesting pain relief. The results suggest a role for behavioral‐based educational interventions to promote communication of pain in people with ASD‐ID.


| INTRODUC TI ON
The International Association for the Study of Pain (IASP) defines pain as "an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage", 1 pg. 1976). Chronic pain is generally considered to be pain that is present for at least 3 months. IASP recognizes that pain may be present even in those unable to reliably verbally communicate the presence of pain. 1 People with co-morbid Autism Spectrum Disorder (ASD) and intellectual disability (ID) may fall into this group. Indeed, this population is at increased risk of experiencing pain due to the presence of challenging behavior, | 23 FITZPATRICK eT Al.
including self-injurious behavior and increased risk of accidental injury. 2 Individuals with ASD and ID (ASD-ID) may also present with co-morbid health problems that may be associated with pain, 3 including neurological, musculoskeletal, and gastrointestinal problems.
A small number of studies to date have looked at the prevalence and impact of pain in individuals with ID. Findings from caregiver reports have found wide-ranging estimates of daily pain, ranging from 15%-50% of people with ID. [4][5][6] Pain impacts on the daily functioning of individuals with ID by inhibiting their ability to participate fully in day service activities, 6 and negatively impacts on sleep, 6 emotional well-being, and quality of life. 7 A recent study on the prevalence of pain (based on parental report), in a representative sample of children in the United States found that children with ASD showed a higher incidence (15.6%) compared with children without ASD (8.2%). However, the prevalence of pain was highest among children with ASD and developmental co-morbidities (19.9%). The authors opined that underlying sensory sensitivities, comorbidity of conditions associated with pain such as cerebral palsy and gastrointestinal conditions, as well as more frequent medical procedures, could account for the elevated prevalence of pain. 8 Self-report is considered the "gold standard" in the assessment of pain; however, self-report measures are not always accessible or feasible for use with individuals with ID and ASD. 9 Furthermore, individuals with ASD-ID are often reported to express their pain in "atypical" ways such as through self-injurious behavior (eg, headbanging and biting), aggression and changes in behavior such as irritability, low mood, reduced activity, appetite change, changes in sleep, or crying. 6,10,11 Consequently, pain is frequently unidentified and ineffectively managed among individuals with communication impairments. 7,12 To date, only two studies have examined psychological interventions focused on pain in individuals with a diagnosis of ID, both using cognitive behavioral therapy to teach self-management strategies for chronic pain. 13,14 These studies included individuals who were high functioning, with good verbal communication skills, and did not involve people with ASD and/or moderate ID. There is an ongoing need for research on interventions for people with more severe cognitive impairments and other complex presentations. 9,15 The ways in which children in general learn about the concept of pain is relatively under-researched. The limited literature in this area suggests that age-appropriate language should be used and that tasks such as drawing and vignettes should be used to allow children to communicate their concept of pain, such as asking children about how the person may be feeling and what the person may need to feel better or different. 16 It is also known that children learn about how to express pain and how to manage pain through their observations and interactions with caregivers. 17 Children can, in fact, influence the attention they receive from others depending on the strength of their display of distress and that of their facial cues. 18 From their work on the concept of pain in children, Pate et al identified four themes labeled as follows: (i) "my pain related knowledge," (ii) "pain in the world around me," (iii) "pain in me," and (iv) "communicating my concept of pain." They concluded that the first three themes (ie, pain-related knowledge, pain in the world around me, and pain in me) are inputs that combine in varying proportions to produce the output which is how a child communicates their individual concept of pain.
There is therefore a reasonable conceptual basis to assume that labeling pain in oneself (pain in me) and labeling pain in others and identifying how the person is feeling (pain in the world around me) or identifying what someone would need to feel better (pain-related knowledge) are prerequisites in order for a child to reliably communicate their own pain and their pain management needs.
The current study aimed to use a behavioral-based educational intervention to teach children with ASD and ID to communicate about their pain, by communicating the severity of pain, reporting the location of the pain, and requesting pain relief, using a series of case studies, in which the initial baseline was staggered and ongoing probes were taken to monitor for the emergence of the skill. Data were also gathered on challenging behavior and pain (as measured by the Non-Communicating Children's Pain Checklist) to (i) identify if pain was present for each participant, (ii) utilize this information in order to facilitate skills teaching when pain was present (participants learning the skill of communicating pain), and (iii) determine if there was any reduction in challenging behavior in response to the ability to communicate pain.

| Inclusion criteria
The study was open to children ranging in age from 5-18 years who (i) had a confirmed diagnosis of ASD (made by an interdisciplinary service), (ii) presented with co-morbid ID (as confirmed by formal psychometric and functional assessment), (iii) presented with a delay in language function, (iv) never vocalized or reported pain, and (v) experienced daily or frequent pain based on The Non-Communicating Children's Pain Checklist-Postoperative Version (NCCPC-PV; 19 where a proxy-reported score >6 indicates the presence of pain.

| Participants
Three male children with a diagnosis of ASD and ID were recruited from a school for children with special educational needs. Participant 1 was aged 11 years and 4 months with a diagnosis of ASD, moderate level of ID, co-morbid neurological conditions and limb amputation which resulted in him having a prosthetic limb. The Vineland Adaptive Behavior Scales 20 assessment indicated that the verbal behavior of this participant was at 2 years and 7 months while expressive language was at 24 months. He expressed about 5-10 words verbally and could echo several complete words. Staff reported that they believed he experienced pain frequently from the prosthetic limb and from constipation. He had a history of self-injurious behavior (ie, hand biting) and challenging behavior (ie, crying and refusal) and these behaviors were reported to occur when he experienced pain. A score of 22 was reported on the initial assessment of pain using NCCPC-PV. The participant suffered from constipation at times and staff estimated that pain may be present once a week. Participant 3 was unable to verbally communicate when pain was present and this would result in noncooperation, irritability, guarding part of body that would hurt and he would seek comfort through physical closeness.
A baseline score of 18 was reported on the NCCPC-PV.

| Ethical considerations
Ethical approval was obtained from the academic institution hosting the research, which was conducted in compliance with the Declaration of Helsinki. Consent and assent were obtained before taking any baseline data.

| Experimental design
A series of case studies were undertaken, in which the initial baseline was staggered across three participants and ongoing probes were used to determine whether an increase in the communication of pain (ie, communicating the severity of pain or reporting the location of pain, and requesting pain relief) in participants with ASD-ID could be achieved using principles of applied behavior analysis.
Each of the four components of the intervention was introduced individually and sequentially: (1) label body parts on iPad, (2) label body parts on self, (3) label pain and score severity using the Wong Baker FACES Pain Scale, and (4) request pain relief. Following this, in-situ training was implemented. The intervention was introduced in a staggered fashion across participants following achievement of each component for each participant.

| Challenging behavior
Frequency of challenging behavior was collected by the therapist and teacher using an individualized data collection sheet to record the occurrence of challenging behavior. A "+" was marked for each instance of challenging behavior in order to gather a tally of the occurrence of challenging behavior across the day. A "−" was marked if no instances of challenging behavior occurred for the entire day. All three participants displayed different topographies of challenging behavior. Participant 1 had a history of self-injurious behavior (hand biting) and crying when pain was present. Hand biting was defined as any instance where the child's teeth made contact with their hand with enough force to leave a visual mark on the skin and/or make hand bleed. Crying was defined as any instance in which the child engaged in loud vocalizations accompanied with tears when pain was present. Participant 2 had a history of aggressive behaviors (hitting others and throwing objects) when pain was present. Hitting was defined as any instance where the child would use their dominant hand with force and hit another person on their upper body (including arms, back, chest, and face). Throwing objects was defined as any instance where the child would pick up an object (chair, book, etc) that was in close proximity to them and throw it with force. Participant 3 has a history of noncompliance and irritable behaviors (crying) when pain was present. Noncompliance was defined as any instance where the child physically and/or verbally refused to follow an instruction or complete a task. The definition of crying was identical to that detailed for Participant 1.  19 is a modified version of the Non-Communicating Children's Pain Checklist-Revised. 23 The NCCPC-PV is a pain measurement tool specifically designed for children with cognitive impairment and it has been shown to be valid and reliable for measuring pain in intellectual disability. 19,24 This study used the postoperative version of the measure as it was not possible to observe sleeping and eating across the duration of the study. This study used the English version to assess the level of pain at baseline and throughout the intervention. The English version is composed of 27 items divided into 6 subscales (vocal, social, facial, activity, body and limbs, and physiological). The measure is completed by parents/caregivers. Scores are obtained and calculated by adding the 27 items to obtain the total score. A total score of 6-10 indicates a child has mild pain, and a total score of 11 or greater indicates a child has moderate to severe pain. 19 The NCCPC-PV was not used to make any medical decisions for the three participants.

| Pain
(b) The Wong Baker FACES Pain Scale (WBFPS, 25 is a measurement tool used to rate the severity of pain in children. The assessment tool contains a series of six round cartoon faces rating from 0 (no hurt) to 10 (hurts worst) beginning with a face that contains a smile representing "no hurt" to the sixth picture with a frown and tears coming from the eyes representing the "worst hurt".

| Baseline and probes conditions
During baseline, data were gathered on the communication of pain.
All sessions consisted of a four-hour period during the participant's typical school day. Probes were taken on two consecutive days following components (1) label body parts on iPad, (2) label body parts on self, and (3) label pain and score severity using the Wong Baker During baseline and across all probes, the teacher was working with the participant in their typical classroom environment, therefore, was aware if the participant engaged in any behavioral indicators of pain. The NCCPC-PV was completed by the teacher at the end of each school day, and a score was calculated to quantify the level of pain behaviors. The NCCPC-PV was used as a guide to observe pain-related behaviors of all three participants within the baseline conditions.
Frequency data was also collected on the occurrence of challenging behavior across a school day. At baseline and throughout the intervention phase, staff asked participants "How are you feeling?" when they believed the participant was experiencing pain based on the presence of challenging behavior or based on the NCCPC score.

| Intervention phase
During the intervention phase, four target skills (labeling body parts on iPad, labeling body parts on self, identifying pain, and identifying pain relief) were identified as prerequisites to communicating pain and requesting (rather than just identifying) pain relief. Participants were assessed to verify if these skills were present and in cases where they were absent, the skill was taught to mastery. The mastery criteria for each component was 100% correct once or 90% correct on two consecutive trails. Training on the four components took place for 1 hour per session and was consistent for all three participants. Prior to commencing teaching sessions, each participant was presented with a choice board containing three preferred items and offered to select one item, which they would receive as a reinforcer during the teaching session. Each participant had individualized choices, token boards, and tokens.

| Component one
The first component of the intervention consisted of labeling body parts (head, arms, legs, back, and stomach) using a representation of a body on an iPad using the "Learn Body Parts" app. [Details on how to deliver training to teach this component can be obtained from the corresponding author on request].

| Component two
Component two was introduced, whereby the participant had to point and label the requested body part on themselves. The participant was presented with the instruction "show me ____" (eg, head), participants were required to respond by pointing to the body part on themselves.

| Component three
This component consisted of the presentation of five painful scenarios (ie, falling off a bike, playing football, hit head against a door, falling off a swing, and burning hand with hot water). Each scenario contained three pictures on flashcards outlining a story in which the character experienced pain, each scenario began with an image of a boy or girl doing an activity but without evidence of injury or pain.
The participant was presented with one scenario at a time and then the WBFPS. During the teaching session, each participant was presented with all five scenarios (see Table 1).
The therapist presented the scenario in conjunction with five questions (How is s/he feeling (at the start of the sequence)? What is s/he doing? What happened? Where does it hurt? How is s/he feeling (at the end of the sequence)?). Data was taken on three trials which related to answering "how is s/he feeling?" (at the start of the sequence), "where does it hurt?" and "how is s/he feeling?" (at the end of the sequence). When the researcher gave the instruction "how is the person feeling?" pointing at picture 1, the participant was then presented with the WBFPS to which they had to point to the appropriate face. The purpose of labeling "not hurt" (happy) at the start of the sequence was  Table 1 for pain relief options presented for each scenario.
If the participant selected the correct picture of pain relief, the therapist provided praise (eg, "yes, he needed a band aid") and a token was given to the participant for him to place on the token board.
However, if the participant picked the incorrect picture (ie, selected an ice pack for a cut on the arm instead of selecting a band aid, an error correction procedure was carried out by the therapist. The therapist stated "he needs a band aid" while pointing to the item. The therapist would then give the instruction "what does he need?" to which the participant had to pick the correct picture and place it next to the four-picture sequence. Praise and tokens were withheld for any incorrect responses. Teaching continued until mastery was achieved.

| In-situ training
Individuals with ASD often have difficulty with skill generalization.
To address this, in-situ training was implemented for 5 days. The

Band aid on knee or cream for hand
Note: For pain relief options, the correct option is denoted in italics. Within the current study, the visual sequences were deemed to display a moderate level of pain. Therefore, a correct response was deemed to be a score of 6 or above on the Wong Baker FACES Pain Scale (WBFPS).

| Component 1 and 2
All three participants correctly labeled and vocalized the five targeted body parts on the iPad (Component 1) and on themselves (Component 2). The participants had already acquired the skill of labeling their body parts, as they achieved 100% correct responding first time (see Table 2).

| Component 3
All participants successfully labeled the location of pain in all five scenarios and quantified the severity of pain using the Faces Scale following 2 days of training.

| Component 4
All participants successfully labeled the location of pain on all five scenarios, quantified the severity of pain using the Faces Scale, and requested appropriate pain relief following 2 days of training.

| Baseline
During the baseline phase, communication of pain was not observed for any of the three participants in a reliable or readily identifiable manner.

| Probes 1, 2 and 3
For Participant 1, no communication of pain was observed during

| Baseline
During the baseline phase, pain was believed to be present for all three participants, as indicated by staff scores on the NCCPC-PV.

| In-situ training
During in-situ training, Participant 1 continued to display the presence of pain with scores on the NCCPC-PV above 6 daily. However, in contrast to baseline and probe conditions, challenging behavior was reduced, occurring only on two of the 5 days. In addition, the number of incidents of challenging behavior was lower than baseline (during which challenging behavior occurred on all 3 days).

| Treatment fidelity
The researcher used a fidelity checklist to ensure that the intervention was implemented consistently across all three participants.
The fidelity checklist included a 19-item checklist of each step of the intervention. The researcher delivered the intervention while a teacher/teacher's assistant observed the intervention and completed the fidelity checklist. Fidelity checks were taken for over 30% of sessions and was 100% for all training sessions.

| Inter-observer agreement
Inter-observer agreement was assessed in all components of the study by a second observer. Both the first and second observer took data independently for all targeted responses. For participant 1, the inter-observer agreement was 92.5% with a range of 80%-100%, while for participant 2, inter-observer agreement was 97.5% (range 90%-100%), and for participant 3, inter-observer agreement was 95% (range 80%-100%).

| DISCUSS ION
This study aimed to increase communication of pain among three individuals with autism spectrum disorder and intellectual disability using a behaviorally-based educational intervention by ensuring the presence of prerequisite skills (labeling and identifying) and using in-situ training. All three participants learned how to report when pain was present, by reporting the severity of pain using the Wong

| Limitations and future research
Although all three participants increased their communication of Further prevalence studies are also needed to determine the extent and nature of pain in people with ASD-ID and to elucidate the most common behavioral indicators of pain so that functionally equivalent behaviors can be taught.

| CON CLUS ION
This study showed that in-situ training was required in addition to the presence of the prerequisite skills of being able to identify and