Measuring sustained interaction in adults with deafblindness and multiple disabilities: Development of an observational coding system

Abstract

Sustained interaction with other people is of crucial importance for persons with deafblindness and multiple disabilities. This article introduces a relatively time-efficient observational coding system to measure sustained interaction directly from video recordings using only two observation categories: turns and interaction breakdowns. Sustained interaction is defined as a sequence of at least three turns between two interaction partners. An interaction breakdown means that a sequence is not being continued, but is broken off by one of the partners. This article also presents the results of a first study applying the coding system to videos of five adults with deafblindness and multiple disabilities and their caregivers. The system seems well applicable and provides supporting evidence regarding inter-observer reliability. The system can stimulate further research and intervention to increase sustained interaction.

Keywords

Adults with deafblindness and multiple disabilities interaction breakdown interaction quality measuring sustained interaction observational coding system

The length of interaction sequences is an important quality of interactions. Interaction has been defined as the process of two individuals mutually influencing each other’s behaviour (Janssen, Riksen-Walraven, & Van Dijk, 2003b). ‘Sustained interaction’ is defined here as an interaction consisting of at least three reciprocal turns (Levinson, 1983; Linell, 2009): an interaction partner takes an initiative, the other reacts, and the first reacts again. Research results suggest that a prolonged lack of experience of reciprocal attunement may cause later problems in elaborating communicative skills and in establishing positive affect (Reddy, 2008; Riksen-Walraven & van Aken, 1997; Tronick, 1989). There is much evidence about young children without disabilities and some evidence about persons with disabilities demonstrating that interactions focused on affect attunement and sensitive responsiveness by the caregiver are crucial for generating a feeling of recognition and sustained flow in early conversations (Trevarthen & Aitken, 2001).

It is important to stimulate sustained interaction and reduce interaction breakdown between caregivers (including teachers) and persons with deafblindness and multiple disabilities. Caregivers experience challenges in following the client’s lead, waiting long enough for a response, and scaffolding interaction routines with a reciprocal and sustained interaction flow, but training them to be sensitive and to respond to the subtle, infrequent, and often atypical signals reduces stress in daily life (Bloeming-Wolbrink et al., 2012; Chen, Klein, & Haney, 2007; Vervloed, Van Dijk, Knoors, & Van Dijk, 2006) and creates opportunities for the students to initiate, sustain, and terminate interactions (Bruce, 2002). Training of caregivers is successful, resulting in more harmonious interactions, smooth-turn-taking, and affective involvement (Janssen, Riksen-Walraven, & Van Dijk, 2003a), not only during interaction but also during sharing experiences and sharing meaning (Martens, Janssen, Ruijssenaars, Huisman, & Riksen-Walraven, in press), although it remains complex to coordinate the flow of interactions and at the same time focus on meanings and intentions and tactilely sharing emotions (Hart, 2006).

In an earlier study, we demonstrated that it is possible to increase sustained interaction and reduce breakdowns in interactions between caregivers and children with congenital deafblindness (Janssen, Riksen-Walraven, Van Dijk, Huisman, & Ruijssenaars, 2012). We identified sustained interactions using a five-category observational coding system that was not designed for that purpose and turned out to be rather time-consuming. First, we coded the occurrence of the five separate categories (initiative, confirmation, answer, intensity, and acting independently by the caregiver). In a second step, we identified sustained interaction sequences and breakdowns based on these records. Although sustained interactions could be reliably assessed, the system seems unnecessarily complicated for this purpose. A more simple and time-efficient observational coding system for sustained interactions could be very conducive to the implementation and evaluation of interventions to foster sustained interactions. In this article, we introduce such a coding system as well as the results of a first empirical study applying the system to five adult clients with deafblindness and multiple disabilities and their caregivers.

The present coding system (see the section ‘Method’ for details) allows observing sustained interaction directly from video using only two behavioural categories: turns and interaction breakdowns. Sustained interaction requires a sequence of at least three turns: a turn by the client, a turn by the caregiver, and another turn by the client, or vice versa. An interaction breakdown means that a sequence of at least three turns is broken off by one of the partners.

The first question of this study was whether the observational coding system could be reliably applied to videos of the five clients and their caregivers. Second, for intervention purposes, we examined which forms of client and caregiver behaviour led to breakdown of sustained interactions. And third, because we also wanted to know if certain daily interaction situations were more beneficial for sustained interaction than others, we compared sustained interaction in different interaction situations per client. We looked particularly at the length of sustained interactions, the longest sequence, the number of turns in a sequence, and the number of interaction breakdowns.

Method

Participants and setting

This study followed the tenets of the World Medical Association Declaration of Helsinki – Ethical Principles for Medical Research Involving Human Subjects – approved by the Institutional Review Board of Talant, the Netherlands. Informed consent was obtained from the clients’ parents and the participating caregivers. Talant is a regular setting for people with intellectual disabilities, with one group home for clients with deafblindness and multiple disabilities.

Five adults with deafblindness and multiple disabilities participated in the study. Fictitious names are used. Table 1 provides an overview of the relevant characteristics of the clients, such as age, score on social competency using the Sociale Redzaamheidsschaal voor Zwakzinnigen (SRZ) Scale (Krayer, Kema, & De Bildt, 2004), visual disability, hearing disability, and communicative and physical functioning. The SRZ score ranges from 3 (lowest) to 9 (highest). For each of the five clients, we made a random selection of video clips of client–caregiver interactions, with an attempt to include more than two caregivers per client. This yielded a total of 11 caregivers for this study.

Table 1.

Clients’ characteristics.

Client age	SRZ score for social competence	Visual disability	Hearing disability	Communicative functioning	Physical functioning
Arne 30 years	3	Severe visual impairment 0.06 (TAC)	Moderate hearing impairment 40 dB	Uses: vocalising, some objects, pointing, directing hand; understands: verbalisations, name calling, some signs, some objects	Epilepsy, paralysis of right side
Daan 51 years	3	Blind with light perception	Moderate hearing impairment 60 dB	Uses: vocalising, directing hand, pushing away; understands: gesturing, name calling, some objects	Motor and balance problems
Erik 61 years	3	Visual impairment 0.23 (TAC)	Severe hearing impairment 60–75 dB	Uses: some objects, vocalising, making eye contact, some signs; understands: short sentences, some objects, name calling	Not able to walk
Peter 40 years	3	Visual impairment 0.30 (TAC)	Severe hearing impairment 70 dB	Uses: reaching, smiling, vocalising; understands: objects, gesturing, intonation	Epilepsy, not able to walk
Maaike 48 years	3	Totally blind	Moderate hearing impairment 40 dB	Uses: reaching hand, turning head away, withdrawing hand; understands: routines, verbalisations, hugging	Motor problems, not able to walk outside

SRZ: Sociale Redzaamheidsschaal voor Zwakzinnigen, TAC: Teller Acuity Cards procedure.

A social competence SRZ score of 3 is appropriate for a developmental age of between 0 and 2 years. Moderate hearing loss: 41–60 dB, severe hearing loss: 61–80 dB, and profound hearing loss: ≥81 dB.

General procedure

We used systematic observation to quantify predefined target categories for sustained interaction (see observation categories) and demonstrated inter-observer reliability (Bakeman & Gottman, 1997). A total of 25 video segments were randomly selected from existing videos recorded for earlier research purposes, controlled for two criteria: for each client (a) two or more caregivers are involved and (b) the interaction situations differ in chronological time, setting, caregiver, and type and rate of activity during a client’s week.

Observation situations

Per client, five interaction situations were observed. Table 2 presents an overview of the variety in observation situations.

Table 2.

Overview of the observed activities, caregivers, settings for each client, and behavioural forms which preceded the breakdowns in each interaction situation.

Client	Activity	C.g.	Setting	Breakdowns	C.g. Beh.	Cl. Beh.
Arne	1. Drinking	1	DAC	2	TSE; HOC
	2. Sensory stimulation	2	DAC	4	TSE
	3. Eating	3	GH	17	TSE; AI
	4. Dressing	4	GH	0
	5. Eating	4	GH	14	AI
Daan	1. Massage	3	GH	0
	2. Undressing	5	GH	3	TSE
	3. Undressing	6	GH	6	AI
	4. Drinking	5	GH	6	HOC
	5. Drinking	7	DAC	0
Erik	1. Laying table	8	GH	5	AI
	2. Eating	8	GH	6	TSE; AI
	3. Washing up	9	GH	1		RI
	4. Eating	10	GH	8	AI	RI
	5. Eating	4	GH	9	AI	RI
Peter	1. Wheelchair	2	DAC	1	TSE
	2. Breakfast	9	GH	5		ES
	3. Dressing	9	GH	0
	4. Movements	1	DAC	4	TSE; AI
	5. Play	5	DAC	5	HOC; AI
Maaike	1. Eating	4	DAC	12		RI
	2. Drinking	3	DAC	2	AI
	3. Sitting couch	4	GH	13	AI	RI
	4. Massage	8	GH	2		RI
	5. Drinking	11	GH	12		RI

GH: Grouphome; DAC: day activity centre; C.g.: caregiver; C.g. Beh.: caregiver behaviours; AI: acting independently; HOC: helping other clients; TSE: talking with someone else; CL. Beh.: client behaviours; RI: regulating intensity; ES: epileptic seizure.

Observation categories

Two main observation categories were used: turns and interaction breakdowns. The turns category was specified in turns caregiver and turns client. Interaction was sustained when a three-turn sequence was performed. Turn caregiver was described in concrete behaviours such as touching the client to make contact, handling the client an object, imitating a movement or a sound, and joining in with what the client is doing. Turn client was described, for example, as taking the caregiver’s arm, laying a hand on the caregiver’s arm, joining in with what the caregiver is doing, and pushing the caregiver’s hand away.

Interactions could be broken by the caregiver or client. A breakdown preceded by caregiver behaviours or client behaviours after a sequence of at least three turns was scored as an interaction breakdown. Different forms of caregiver behaviour could precede a breakdown, such as the caregiver acting independently (AI) of the client, for example, by looking at a watch, stirring food, talking with somebody else (TSE) in the room, or helping other client (HOC). Breakdowns by the client were mainly caused by regulation intensity (RI), which means that the client breaks the interaction to regulate certain tension or arousal during the interaction and/or to process information. Different ‘processing signals’ can be observed, such as, a client withdraws his hand when the caregiver is pushing harder during a hand massage, or the client tilts his head backwards and flaps his hands when the caregiver demonstrates too many initiatives in a row. An epileptic seizure (ES) can also cause a breakdown; characteristics were described for two clients (Arne and Peter).

The coding system was designed to be mutually exclusive, for the categories turns and interaction breakdowns. Clear definitions and clear descriptions in terms of concrete behaviours were given separately for each client. Inter-observer agreement was only calculated for the observation categories turns caregiver, turns client, and interaction breakdowns, not for the variety in behaviours that precede the breakdowns. The described examples are from this study. This observation method was generated from previous work (Janssen et al., 2003a, 2012).

Observation procedure

Two observers watched 25 video segments in random order, the first observer being a master’s student in Special Education (second author [N.C.R.B-G]), and the second observer a certified primary education teacher. The length of all observations was exactly 4 min, because this was the maximum length of some clips.

The observers scored occurrences of behaviours that fit the target categories turns and interaction breakdowns and recorded 24 of them in 0-s intervals on an observation form. Intervals of 10 s were chosen so that infrequent and complex individual behaviours that occurred during client or caregiver turns would not be missed. Every time an interaction breakdown occurred, they noted the preceding client or caregiver behaviour. After every observation session, the observers counted the sequences of sustained interaction, intervals of sustained interaction, intervals for the longest sequence, turns in a sequence, and number of interaction breakdowns. The observers’ notes about the interaction breakdowns were coded afterwards by the researcher (first author [M.J.J.]; e.g., TSE and ES).

Inter-rater agreement

Prior to formal data gathering, the researcher trained the observers to reach 80% inter-observer agreement using three sessions for each client. These sessions were not used in the actual study. The first observer coded all observation sessions, and the second observer coded 40% of these sessions randomly selected per case to determine inter-rater agreement. The inter-rater agreement was computed with the formula agreements divided by agreements plus disagreements multiplied by 100 (Barlow, Nock, & Hersen, 2009).

Average inter-observer agreement percentages ranged from 84% to 94% for turns caregiver, from 82% to 87% for turns client, from 93% to 100% for interaction breakdowns, from 95% to 99% for sustained interaction, from 97% to 100% for longest sequence, from 95% to 100% for mean number of turns in a sequence, and from 99% to 100% for mean number of breakdowns.

Results

Tables 2 and 3 present detailed information about the variables and results per client. Table 3 shows the occurrence (duration or frequencies) of the categories (sustained interaction, longest sequence, turns in a sequence, and interaction breakdowns) per 4-min observation episode per interaction situation and the means for each client. Table 2 gives an overview of behaviours of clients and caregivers which preceded interaction breakdowns in the different interaction situations for each client.

Table 3.

Occurrences (frequencies or duration) for sustained interaction, longest sequence, turns in a sequence, and interaction breakdowns in five different interaction situations and means for each client.

Client	Situation 1	Situation 2	Situation 3	Situation 4	Situation 5	Mean
Sustained interaction
Arne	220	230	150	240	110	190
Daan	240	220	200	210	240	222
Erik	220	200	230	190	180	204
Peter	230	190	240	210	180	210
Maaike	180	220	150	220	140	182
Longest sequence
Arne	190	180	60	240	80	150
Daan	230	140	150	170	240	186
Erik	130	50	230	70	110	118
Peter	150	140	240	82	120	146
Maaike	90	220	60	210	90	134
Turns in a sequence
Arne	39	19	8	29	10	21
Daan	40	30	26	23	91	42
Erik	16	17	39	15	12	20
Peter	28	26	78	11	14	31
Maaike	6	47	7	14	8	16
Interaction breakdowns
Arne	2	4	17	0	14	7
Daan	0	3	6	6	0	3
Erik	5	6	1	8	9	6
Peter	1	5	0	4	5	3
Maaike	12	2	13	2	12	8

Results per client

The next paragraphs summarise sustained interaction patterns for the individual clients.

Arne

Arne’s lowest duration of sustained interaction, lowest duration of the longest sequence, and highest number of interaction breakdowns occurred in situations 3 and 5 (eating). His highest duration of sustained interaction was in situation 4 (dressing). Caregiver behaviours preceded most of the interaction breakdowns, most frequently TSE. Very few interaction breakdowns were observed in the other three situations (drinking, sensory stimulation, and dressing).

Daan

Sustained interaction was high in all five interaction situations. Few breakdowns occurred in general, with the majority of breakdowns in situations 3 (undressing) and 4 (drinking). Caregiver behaviours preceded mainly these breakdowns. The highest number of turns in a sequence was in situation 5 (drinking).

Erik

Sustained interaction was quite high in all situations; most of the breakdowns occurred in situations 4 and 5 (eating). The duration of the longest sequence was highest in situation 3 (washing up); only one interaction breakdown was observed here. The main behaviours preceding breakdowns were regulation of intensity by Erik and the caregivers acting without paying attention to him.

Peter

Duration of sustained interaction was highest in situations 1 (wheelchair) and 3 (dressing) and lowest in situations 2 (breakfast) and 5 (play). Situation 3 had the longest sequence, the highest number of turns in a sequence, and no interaction breakdowns. The interaction breakdowns during breakfast were preceded by ESs. In the other situations, the main preceding behaviours were caregivers’ AI, TSE, and HOC.

Maaike

Her best sustained interactions were in situations 2 (drinking) and 4 (hand massage). She had the most breakdowns in three different situations (eating, drinking, and sitting on the couch). The main behavioural forms for breakdowns were regulation of intensity in four situations and the caregivers AI in two situations.

Breakdown in different situations

Table 2 shows the activities that led to breakdowns. The greatest number of breakdowns for all clients occurred in eating situations (6). The activities in which no or fewest breakdowns occurred differed by client: dressing (2) for Arne and Peter, massage (2) for Daan and Maaike, drinking (2) for Daan and Maaike, washing up (1) for Erik, and wheelchair for Peter (1).

Caregiver and client behaviours related to breakdown

For all five clients, the following caregiver behavioural forms preceded breakdowns in the greatest number of situations: AI (11), TSE (7), and HOC (3).

The client behaviour that directly preceded most of the breakdowns for two of the clients (Erik and Maaike) was ‘regulation of intensity’ (7). One of the other clients (Peter) experienced ESs prior to the breakdowns in one situation (1).

Discussion

This study showed the usefulness of a relatively time-efficient observational coding system to assess sustained interaction based on two behavioural categories (turns and interaction breakdowns). First, our results demonstrate that the new coding system can be reliably applied to video recordings of caregiver–client interactions. This makes it possible to obtain a relatively quick picture of the occurrence and length of sustained interactions and of breakdowns in interactions. Second, our observations with the new system in a small number of adults with deafblindness and multiple disabilities suggest (a) that there are considerable differences between clients as to the caregivers behaviours that precede (and therefore might cause) an interaction breakdown and (b) that the number of interaction breakdowns and the behaviours preceding a breakdown may vary in different interaction situations. Although strong general conclusions cannot be drawn on the basis of the relatively small sample of clients we used in this first study, the results provide valuable hints for further application and development of this observational coding system that can be used as a framework for designing and evaluating interventions to increase sustained interaction in this specific population.

In this study, we focused on differences between clients and differences between situations. It is a limitation that we could not draw any conclusions about differences between caregivers. This would have been possible if different caregivers had been observed while interacting with the same client in the same situations. Table 2 shows that this was not the case in the present study. We recommend to examine this question in further studies, because it may yield information that is relevant for caregiver supervision and training.

Another limitation is that we only assessed inter-rater reliability for the occurrence and number of interaction breakdowns and not for the preceding behaviours. It is recommended to do that in future studies.

The average inter-rater agreement scores differed across the different categories, with a lower percentage for the turns category compared to the other categories. This is understandable because a distinction was made between caregiver turns and client turns. It is more difficult to observe who is performing a turn than to observe whether there is an interaction breakdown.

The target categories on sustained interaction were derived from turns and interaction breakdowns. This method of analysis was well applicable, and the average inter-rater agreement percentages were high on these categories.

Comparing sustained interactions in different situations yielded interesting trends. The observed situations with the greatest number of interaction breakdowns were eating situations. Interaction situations which seemed to be more beneficial for sustained interaction differed among clients. The longest interactions were observed during drinking, dressing, massage, and washing up.

As a starting point for future intervention research, it could be recommended for caregivers to be aware that certain activities are more beneficial for sustained interaction then others and that these situations can differ for each client. Eating situations are not preferable for increasing sustained interaction. It is recommended that caregivers be trained to be more focused on client interactions, also when others are in the room.

Special attention must be given to the clients’ individual regulation signals. When client behaviours precede an interaction breakdown, it does not always mean that the client is the first initiator of the breakdown. The clients may be breaking down the interaction because they just want to regulate the tension or process certain information. Caregivers should be aware of these signals and adjust their behaviours according to those of the client to be better capable of preventing such client breakdowns.

Footnotes

Acknowledgements

We would like to acknowledge the contributions of Andrea Groenendijk-Treep.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Bakeman

Gottman

J. M.

(1997). An introduction to sequential analysis (2nd ed.). Cambridge, UK: Cambridge University Press.

Barlow

D. H.

Nock

M. K.

Hersen

(2009). Single case experimental designs: Strategies for studying behavior change. Boston, MA: Pearson.

Bloeming-Wolbrink

K. A.

Janssen

M. J.

de Weerth

Ruijssenaars

A. J. J. M.

Sweep

F. C. G. J.

Riksen-Walraven

J. M.

(2012). Stress in adults with congenital deafblindness and an intellectual disability: Information from their cortisol curves. British Journal of Visual Impairment, 30, 149–159.

Bruce

S. M.

(2002). Impact of a communication intervention model on teachers’ practice with children who are congenitally deaf-blind. Journal of Visual Impairment and Blindness, 96, 154–168.

Chen

Klein

M. D.

Haney

(2007). Promoting interactions with infants who have complex multiple disabilities. Development and field-testing of the PLAI curriculum. Infants & Young Children, 20, 149–162.

Hart

(2006). Using imitation with congenitally deafblind adults: Establishing meaningful communication partnerships [Special issue: Imitation and socio-emotional processes: Implications for communicative development and intervention]. Infant and Child Development, 15, 263–274.

Janssen

M. J.

Riksen-Walraven

J. M.

Van Dijk

J. P. M.

(2003a). Contact: Effects of an intervention program to foster harmonious interaction between deaf-blind children and their educators. Journal of Visual Impairment and Blindness, 97, 215–229.

Janssen

M. J.

Riksen-Walraven

J. M.

Van Dijk

J. P. M.

(2003b). Towards a diagnostic intervention model for fostering harmonious interactions between deaf-blind children and their educators. Journal of Visual Impairment and Blindness, 97, 197–214.

Janssen

M. J.

Riksen-Walraven

J. M.

Van Dijk

J. P. M.

Huisman

Ruijssenaars

A. J. J. M.

(2012). Enhancing sustained interaction between children with congenital deaf-blindness and their educators. Journal of Visual Impairment and Blindness, 106, 77–183.

10.

Krayer

D. W.

Kema

G. N.

De Bildt

A. A.

(2004). Sociale Redzaamheidsschaal voor Zwakzinnigen (SRZ) [Social Competency Scale]. Lisse, The Netherlands: Swets Test Publishers.

11.

Levinson

S. C.

(1983). Pragmatics. Cambridge, UK: Cambridge University Press.

12.

Linell

(2009). Rethinking language, mind, and world dialogically: Interactional and contextual theories of human sense-making. Charlotte, NC: Information Age Publishing.

13.

Martens

M. A. W.

Janssen

M. J.

Ruijssenaars

A. J. J. M.

Huisman

Riksen-Walraven

J. M.

(in press). Effects of an intervention on affective involvement and expression of emotions in an adult with congenital deafblindness. A single-subject case study. Communication Disorders Quarterly.

14.

Reddy

(2008). How infants know minds. London, England: Harvard University Press.

15.

Riksen-Walraven

J. M.

van Aken

M. A.

(1997). Effects of two mother-infant intervention programs upon children’s development at 7, 10, and 12 years. In Koops

Hoeksma

J. B.

van den Boom

D. C.

(Eds.), Development of interaction and attachment: Traditional and non-traditional approaches (pp. 79–93). Amsterdam, The Netherlands: North-Holland Publishing Company.

16.

Trevarthen

Aitken

(2001). Infant intersubjectivity: Research, theory, and clinical applications. Journal of Child Psychology and Psychiatry, 42, 3–48.

17.

Tronick

(1989). Emotions and emotional communication in infants. American Psychologist, 44, 112–119.

18.

Vervloed

M. P. J.

Van Dijk

R. J. M.

Knoors

Van Dijk

J. P. M.

(2006). Interaction between the teacher and the deaf-blind child. American Annals of the Deaf, 151, 336–344.