Increased heart rate as a signal of acute distress in non-communicating persons with intellectual disability. A real-world study.

Intellectual disability (ID) affects approximately 1% of the population. Some patients with severe or profound ID are essentially non-communicating and therefore risk experiencing pain and distress without being able to notify their caregivers, which is a major health issue. This study aimed to see if heart rate (HR) monitoring could reveal whether non-communicating persons with ID experience acute pain or distress in their daily lives. We monitored HR in 14 non-communicating participants with ID in their daily environment to see if specic situations were associated with increased HR. We dened increased HR as being > 1 standard deviation above the daily mean and lasting > 5 seconds. In 11 out of 14 participants, increased HR indicated pain or distress in situations that were not previously suspected to be stressful, e.g. passive stretching of spastic limbs or being transported in patient lifts. Increased HR suggesting joy was detected in three participants (during car rides, movies). In some situations that were previously suspected to be stressful, absence of HR increase suggested absence of pain or distress. We conclude that HR monitoring may identify acute pain and distress in non-communicating persons with ID, allowing for improved health care for this patient group.


Introduction
Intellectual disability (ID) is a neurological condition that affects approximately 1% of the population. 1,2 ID is highly heterogeneous both phenotypically and in terms of causality. It may be an acquired condition, e.g. through cerebral hypoxia, ischemia, or infection early in life, [3][4][5][6] or it may have a genetic cause, most often a de novo mutation. 7 The diagnosis implies IQ below 70, reduced adaptive skills (manifesting as di culties in managing everyday life, school, work, interpersonal relations, etc.), and occurrence of the condition before the age of 18. [8][9][10] In the majority of cases, the condition is evident early in life. Approximately 5% of persons with ID have severe ID (IQ: [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] or profound ID (IQ < 20). 11,12 These patients are entirely dependent on their caregivers, but intellectual shortcomings limit their ability to communicate with them. Communication may be further limited if the ID is compounded by autism or by cerebral palsy that affects control of muscles involved in speech, gesticulation, and grimacing. Some of these individuals may be essentially non-communicating.
Persons with ID are prone to painful conditions, e.g. low impact fractures 13 or various abdominal conditions. 14,15 A series of studies has made it clear that persons with ID are at least as sensitive to pain as the general population [16][17][18][19][20] but signs of pain may be lacking or overlooked by caregivers. [21][22][23] Structured interviews, questionnaires, and checklists aimed at caregivers have been developed to identify behavioral signs of pain in persons with ID. [24][25][26] However, even caregivers who are experienced observers may fail to identify painful conditions in non-communicating persons with ID. 17,27,28 Thus, there is a need to explore additional measures of pain detection that are applicable to a real-world situation.

Participant characteristics
The group of participants was heterogeneous with respect to the cause of their ID, but they all had communication di culties that rendered them essentially non-communicating; some also had severe cerebral palsy or autism (Table 1). Some participants used a few words or phrases of uncertain meaning; others had no spoken language at all. Most participants had some form of behavior that was interpreted by caregivers as conveying information about preferences and dislikes, pain and pleasure. This behavior ranged from smiles and grimacing with an emotional value that seemed intelligible to caregivers, to more personal or idiosyncratic signs or movements whose interpretation was less obvious. In their response to the Inventory of Potential Communicative Acts, 24 caregivers mostly agreed among themselves on how to interpret the various behaviors of the participants (data not shown).
Cognitive assessment con rmed that the participants had severe or profound ID, except one woman with severe cerebral palsy whose level of comprehension was equivalent to moderate ID (IQ [35][36][37][38][39][40]. Several participants had restricted mobility, used a wheelchair, and were dependent on being lifted from bed to chair or from one chair to another, often with the use of an electric patient lift (Table 1). Three participants were fully ambulant. Most participants had comorbidities such as cerebral palsy, autism, epilepsy, diabetes, osteoporosis, or skeletal malformations (scoliosis, hip dysplasia). Some participants received medication that would be expected to affect HR, either by targeting the sinoatrial node (candesartan, amlodipine), or by blunting emotional responses (e.g. benzodiazepines) or pain perception (paracetamol, antiepileptic drugs). An increase in HR was detected in speci c situations in 13 out of 14 participants (Table 2). This was the case even in those participants whose medication could be expected to affect HR. The three fully ambulant participants were all monitored during seated activities with few postural changes and little muscle work.
Increased HR that was interpreted to re ect pain or discomfort in speci c situations was detected in 11 out of 14 participants. However, only in six out of the 11 participants (55%) was the HR increase fully reproducible in the sense that HR increased each time a presumed stressful situation arose. In ve participants for whom there were available data on HR in speci c situations, reproducibility varied from 44 to 93%. In one participant, no HR increase could be identi ed in any of the situations that were analyzed, and for one participant data on reproducibility were not available. Speci c and novel ndings during HR monitoring HR increased in several situations that were not previously thought to entail pain or distress by caregivers. In three participants, extension of a spastic arm during dressing or physiotherapy caused increased HR (Table 2). In one of these cases (M48 ; Tables 2 and 3), this realization led to treatment with botulinum toxin of the elbow exors, which allowed the participant to extend his arm voluntarily when being dressed.
Four participants had increased HR at the sight of an electric patient lift used for their transportation, e.g. from bed to a chair. (Tables 2 and 3). In the lift, the patients hangs from a sling that may swing rather freely. A fth participant (F31), who had ts of tremor and sweating in various situations, had two of these during transportation in a patient lift. It was discovered that she had fallen out of the lift and onto the oor some months previously. Two participants, who were not ambulant (F36 and M37), had increased HR at the sight of an apparatus used for helping them stand upright. One of these (F36) was assumed to react with joy at the sight of the equipment, but it was discovered that the apparatus was wrongly adjusted so that it likely had caused discomfort or pain. Table 2 Heart rate (HR) observations. Typical mean HR in beats per minute (bpm) and 1 standard deviation (SD) for each participant. The mean HR, SD and HR increase (> mean HR + 1SD) were calculated daily. Typical situations that were associated with increased HR are given together with number of replications/total number of observed situations. Eight participants were interpreted as experiencing pain or distress during episodes with increased HR; three participants (F56, M27, and M18) were considered to experience joy during some of the episodes with increased HR. One participant had increased HR when being looked at by children, e.g. in shopping malls, a nding that was interpreted to re ect anxiety. One participant had increased HR during generalized epileptic seizures, which was interpreted to re ect the seizure rather than an emotional response.
Three participants had HR increases that were interpreted to re ect joy. In one, the HR increase that occurred during physiotherapy was accompanied by laughter; in another the HR increase occurred during car travels or when his grandmother came to visit. A third participant had increased HR when he watched children's movies.
Change of care practice following HR monitoring For eight participants, HR monitoring led to a change in caregivers' practice (Table 3), including reassurance strategies during transportation in patient lifts, changes in the execution of physiotherapy and daily stretching exercises, and changes in a daily nail ling procedure for a participant who was prone to self-harm by scratching herself. For three participants, HR monitoring did not support the caregivers' impression that physiotherapy was painful, allowing for more extensive physiotherapeutic procedures. For all but two participants, the use of HR monitoring led to some change in caregivers' understanding of the participant. Increased HR may be caused by both adverse and pleasant stimuli. [29][30][31][32][33] This was illustrated in the present study by participants whose HR increased in response to situations that were likely to be painful and by participants whose HR increased during activities they probably enjoyed. Therefore, a HR increase has to be interpreted in light of its context. Such interpretation is indispensable and at the same time a potential source of error. Erroneous interpretation was illustrated in the present study by the participant who displayed increased HR when mounted in an apparatus used to helping her stand upright: her HR increase was interpreted by caregivers as re ecting joy until it was found that the apparatus was wrongly adjusted and probably caused discomfort or pain.
At present there is no gold standard for identifying pain in non-communicating persons with ID. In an experimental study in which non-communicating persons with ID received pressure-induced pain, Benromano et al. found facial expression, monitored by a video camera and analyzed retrospectively, to be a more reliable indicator of pain than HR. 19 In the present study, however, only some of the participants had intelligible facial expression. Further, the use of video monitoring would have been di cult in our study, which took place in the daily lives of the participants in their communal residences or day care centers. At present, therefore, we believe that HR monitoring is a promising mode of identifying at least some acutely painful or distressing situations that non-communicating persons with ID experience.
In the present study we used increased HR as an indicator of acute pain or distress. [29][30][31] Data are scarce on HR during prolonged or chronic pain, but some clinical studies suggest that prolonged pain is not accompanied by increased HR. 34,35 These ndings point to the need to nd parameters other than HR to identify long lasting pain in non-communicating persons with ID.

Limitations
In the present study we used a de nition of increased HR (HR that persisted > 5 seconds at 1 SD above the daily mean) that allowed for individual and day-to-day variations. This de nition is arbitrary, however, and we may have failed to detect some cases of increased HR due to the de nition being too strict. Further, although we found that HR increased in situations that could be expected to be stressful, we do not know our rates of false positive or false negative results. In some participants, HR did not increase consistently in situations that were interpreted as being stressful. This nding may re ect the occurrence of false negative results due to our de nition of increased HR.
Autonomic responses to distress are modi ed by predictability: unpredicted distress causes stronger autonomic responses. 36 We assume that non-communicating persons with ID are prone to experiencing distress as unpredictable because of their limited understanding of the necessity for painful procedures, such as physiotherapy for spasticity or pinpricks for blood glucose monitoring. We further assume that their reduced ability to avoid distress or to notify their caregivers about it makes stressful situations even more distressing than would be expected in the general population. Therefore, it is possible in the present study that the increases in HR that we detected in presumed distressing situations re ected some combination of pain and the fear that the pain provoked. We are at present not able to distinguish between the two.
HR is expected to increase in response to postural changes or muscle work, which could interfere with the interpretation of increased HR as a sign of distress. In the present study, however, such factors were probably not an important source of error, because HR monitoring predominantly took place when participants were seated or lying down, and because the occurrence of postural changes and muscle work was included in the contextual interpretation of HR variations. For instance, the increase in HR that was seen in several participants in transportation situations occurred at the mere sight of a patient lift, indicating an anticipatory reaction rather than a response to postural change.
We did not monitor HR after the introduction of changes to care practice, although doing so could have shed light on the validity of HR monitoring as a means of identifying acute pain or distress. A group of 14 adults with intellectual disability who lived permanently in communal residences was recruited in the following manner ( Fig. 1): The investigators presented the study to the local administrations of two districts of Oslo city, Norway, which contacted the leaders of the districts' communal residences for persons with ID. Professional caregivers at the communal residences approached the parents or wardens of potential participants with written material describing the study. Informed, written consent was obtained from parents or legally authorized representative of participants. All parents or legal representatives that were invited agreed to the participation. None of the participants withdrew from the study. Prior to HR monitoring, full-time caregivers working in the communal residences and with good knowledge of the participants were asked to complete the Inventory of potential communicative acts questionnaire (IPCA) about how the participant conveyed being happy, sad, bored, amused, frightened, in pain, angry, or tired. 24 The cognitive function of all participants was individually assessed by a licensed neuropsychologist (KS).
Adults with profound or severe ID are estimated to have a mental age below six years of age, 8  The participants engaged in varied daily activities, including alternating between staying in their communal residences and visiting day care centers or their family home. Therefore, HR monitoring had to be done as was practically feasible. HR monitoring took place during weekdays. Most participants had the chest strap put in place during dressing in the morning, while some participants had it put on in the afternoon when they returned from day care center. The participants did not sleep with the HR monitor on.
In some participants, the chest strap monitors sometimes lost contact with the skin, so that the electric signals from the heart were lost. This was especially the case in participants with scoliosis or spastic cerebral palsy (Table 1).

De nition of HR increase
An increase in HR was de ned as HR that persisted > 5 seconds at 1 standard deviation (SD) above the mean HR, which was calculated after completion of each HR monitoring session. This de nition allowed for individual and day-to-day variations in HR.

Registration of behaviors, activities, and situations
The study period was 2 months. Participants were accompanied 1:1 by professional caregivers or students of social education. Students of social education (∼20-50 years old) were completing a three-year full-time study at Oslo Metropolitan University. Caregivers or students registered behaviors (e.g. certain movements, grimacing, screams, laughter), activities that the participant took part in on a regular basis (e.g. physiotherapy sessions, meals, naps, swimming), and various situations that arose (e.g. visit from a relative, choking on food during a meal). The registered behaviors, activities, and situations were aligned with HR data retrospectively.
Therefore, increased HR did not cause caregivers to intervene in real time. Because the study period was limited (2 months), HR was not monitored after care practice was changed. Declarations