Background
Social communication deficits are among the core features of autism spectrum disorder (ASD) [
1]. Symptoms and support requirements for ASD vary in form and intensity with communication deficits ranging from mild to severe [
2]. Approximately 30% of people with an ASD diagnosis also have intellectual disability (ID) [
3], a neurological condition that affects 1% of the population [
4,
5]. For low-functioning individuals with ASD or ID, the symptoms often overlap, and clear-cut diagnosis constitute a challenge [
6‐
8]. The diagnosis of ID involves an intelligence quotient (IQ) below 70, reduced adaptive behaviour, and occurrence of the condition before the age of 18 [
1,
8]. Approximately 5% of those with ID have grade severe (IQ: 20–34) or profound (IQ < 20) and need round-the-clock supervision and help [
9,
10]. Although ID can be an acquired condition, for instance due to cerebral hypoxia, ischemia or infection early in life [
11‐
14], genetic conditions, such as a de novo mutation [
15], are a common cause. Severe or profound ID often coincide with cerebral palsy that affects control of muscles involved in speech, gesticulation, and grimacing. These patients with severe ASD and ID may be unable to notify their caregivers when they are In pain [
16‐
20] and are essentially non-communicative.
There is a need to improve health services for non-verbal patients. Despite the rights of people with disabilities being enshrined in the UN Convention from 2006 [
21,
22], age of death remains lower and mortality rates higher for people with intellectual and developmental disabilities [
23,
24]. Patients with comorbid ASD and ID often have complex medical needs and many cannot convey their needs to their caregivers [
25]. They will typically need the equivalent of 5–10 full-time staff per year–- a major challenge both in terms of competence needs and resource use [
26]. Due to the demanding and person-intensive nursing attention required, there is considerable interest in technological solutions to aid communication and participation for this vulnerable group [
27,
28]. Overall, technology can improve independence, participation and quality of life among people with complex needs [
29].
Sensors may be used to monitor physiological parameters and activity patterns so that individuals and their caregivers can learn about their health. As a neurally mediated phenomenon, heart rate (HR) is regarded a noninvasive window into the central nervous system [
30]. Changes in HR are widely used as markers of reactivity to painful events [
30‐
37], and clinically HR is used in neonatal pain assessment and care [
31]. Pain is an inferred latent process associated with several physiological and psychological markers including changes in neural activity as measured by electroencephalogram (EEG), pupil dilation, skin conductance, HR, blood pressure, and respiration, as well as self-reported pain magnitude [
38,
39]. Although pain-induced increases in blood pressure may trigger vagal baroreflexes that counteract sympathetically-mediated tachycardia; sympathetic changes are found to dominate [
40], making HR a reliable non-invasive way to identify probable painful events [
40‐
42]. Although the central nervous system, cardiovascular system, and pain are closely interrelated [
43,
44], HR is not a specific indicator of pain [
30]. States like depression, being tense, angry or frightened may also involve an increase in HR due to a release of stress hormones like cortisol and adrenaline. The clinical application of such reactivity is shown in studies using HR to predict acute stress in ASD [
45,
46]. As a more general reactivity measure HR is promising as a tool, and when context is given, it may be useful in non-verbal patient pain-management.
We have previously shown that HR can be used as a person-specific measure of reactivity for non-verbal patients [
47]. To evaluate HR as a tool to assist pain management for non-verbal patients, other well-established biomarkers of pain, such as heart rate variability (HRV) and cytokine biomarkers in blood, should be utilized. HRV is a non-invasive measure of imbalances in the autonomic nervous system with a higher variance indicating less stress [
48,
49]. A growing body of psychological research has shown HRV to be a stable biomarker of prolonged pain [
50], and research supports an association between HRV and emotional responses [
51,
52]. Recent studies suggest that certain circulating inflammatory cytokines can be utilized as reliable biomarkers for detecting chronic pain in various diseases [
53‐
55]. Blood biomarkers of pain include MCP-1, IL-1RA, IL-8, TGFβ1, and IL-17 [
53‐
55]. Elevated levels of these cytokine biomarkers have been associated with experimental pain and self-rated pain, and their levels have been shown to closely follow the temporal phenomenology of pain episodes [
54]. The combination of HR, HRV and cytokine biomarkers may prove a novel tool for pain assessment.
Despite the potential of this approach, the use of sensors in care for non-verbal patients is not yet widespread. Most studies on sensors for pain management and communication has been conducted among patients with an ability to communicate [
48,
56‐
60]. Non-verbal patients are frequently left out of research studies, even though advancements in medical technology could bring significant benefits to this patient population.
The proposed study is a randomized controlled trial that seeks to reduce incidence of pain for non-verbal individuals with autism and ID through use of HR-sensors. This study aims to 1) assess the effectiveness of HR as a tool to identify potentially painful care procedures; 2) test the impact of HR-informed changes on pain biomarkers; and 3) examine the impact of six weeks of HR monitoring on the quality of communication between patients and caregivers.
Discussion
In this randomized blinded controlled trial, we aim to study HR to improve communication between patient and caregiver. This clinical study of HR-informed intervention represents an important contribution to the research and practice of developmental disabilities and communication difficulties. Studies on patients with limited communicative abilities are lacking (Table
4). Existing studies investigating severely limited communicative abilities are small and mainly descriptive. Larger studies on technologies for patients with communication disability exclude the most severely debilitated patients. Furthermore, considerable heterogeneity of participant level of communication disability makes for less clinically relevant findings. In the present study we strive to correct this.
Table 4
Methodological advances of current trial relative to past sensor-technology intervention trials among patients with communication difficulties
Few controlled trials | Randomized controlled design |
Small sample sized/insufficient power/case studies | A sample of 38 patients allowing us to detect an 0.80 effect size |
Absence of studies using comparator interventions | Controlling for effect of study participation by including a group with delayed intervention |
Exclusion of most vulnerable patients | Inclusion of profoundly disabled patients |
Lack of standardized measures and biomarkers | HR, HRV and biochemical biomarkers in blood |
No adherence data | Routine signing of protocol by secondary participants to measure adherence |
Considerable heterogeneity of participants regarding type and level of communication need | Patients with ID ranging from severe to profound and ranging from ambulant to wheelchair user, loss of eyesight and hearing |
Some limitations of the present study design should be noted.
First, HR responses to events are not a specific measure of pain per se [
30], and thus must be interpreted in the situational context to be meaningful. However, four potentially painful situations are predetermined, and based on our pilot [
47] there is reason to believe a pattern of HR-increase across situations will become evident over a two week period. Furthermore, to bolster the certainty of pain, we complement the HR measures with other extensively studied biomarkers of pain, such as HRV and pain-related cytokines.
Second, the caregiver is an intermediary between the researchers and the patient. This means we have no direct objective way to classify situations as all reports of context are given through the caregivers. However, caretakers working with this patient group in Norway are generally well educated and have extensive experience [
90], indicating high quality of the observations that are made. We recognize that this indirect method of obtaining contextual information may introduce subjectivity and potential bias. However, we believe that testing the technology through the caregivers is vital to achieving our overarching goal to improve communication and reduce pain in the patient’s everyday life. Setting the study in the patient’s everyday life provides high external validity and ensures that the results are applicable to real-world scenarios. Therefore, while we acknowledge the limitations of our study, we believe that our approach is appropriate and necessary to achieve our research objectives.
Third, as the current study targets stressors relevant for each participant, the specific stressors are idiosyncratic. The current design allows for rich individual case data, yet the predefined situations and interventions will provide the benefit of group analysis. This may make the method and design unorthodox and initially more difficult to understand. However, adopting this approach allows for a deeper understanding of the complexity of stressors and interventions as the design permits us to gather exploratory, descriptive, and case-specific data. The current approach provides a unique opportunity to combine the rich individual case data with group trends, leading to a more comprehensive understanding of stressors and interventions.
The ethical implications of this study are complex, particularly regarding the patients’ inability to provide informed consent. The mentioned issue of the caregiver as an intermediary is relevant to the ethical considerations taken when designing the study. An apparent way to bypass the intermediary of the caregiver would be to use video recording in addition to HR. However, HR is generally regarded as less intrusive than video recording and considered more ethically just [
97]. If video recordings of caregivers were to be conducted, it would be necessary to obtain informed consent from every caregiver involved, as well as any non-participating patients present in the facility. This would introduce additional ethical considerations related to the proper treatment of human subjects in research. We believe HR is informative yet acceptable, a non-invasive technology, possibly of great beneficence to the patients. Another ethical concern here is the aspect of delayed intervention. Obvious painful situations will be detected during the registration phase, and it is not ethical to postpone making changes in these routines for two or four weeks. This ethical aspect makes for more challenging methodology. However, as these instantaneous adaptations are registered, as well as having a delayed intervention group, results will still be valid for evaluation of HR as a communicative aid.
In conclusion, this trial is a randomized blinded controlled clinical trial that test the applicability of HR to reduce incidence of pain in the participant’s everyday life. We will evaluate how HR can be used to identify potentially painful care procedures that should be re-evaluated in terms of the approach taken; test the effect of HR-informed changes in potentially painful care procedures on biomarkers of pain; and assess how six weeks of communication through HR affects the quality of communication between patient and caregiver. Regardless of outcome, the current study will advance the field of wearable physiological sensor-use in patient care.
Trial status
Inclusion to the study is starting 27th of February 2023. We aim to enrol 38 participants by 2025. The end of data collection will be the end of 2025. Protocol version 1, February 2023.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.