Qualitative development of the ‘Questionnaire on Pain caused by Spasticity (QPS),’ a pediatric patient‐reported outcome for spasticity‐related pain in cerebral palsy

The chronology of activities in the development of the QPS is summarized in Fig. 1 and comprised four principal phases: (1) literature review, (2) concept elicitation (CE) interviews, (3) QPS revision, and (4) field testing with cognitive interviews. The development of the QPS conformed to the United States Food and Drug Administration (FDA) and other good research practice guidelines, which point out the importance of establishing proper content validity for PRO’s and ObsRO’s in the target patient population [25‐27]. This study adhered to ethical principles for human research studies and to good clinical practice guidelines [28]. Appropriate institutional review board approval was obtained before study activities began. Caregivers received financial compensation for their time and travel expenses associated with the study participation.

The decision to start with the QPS development was based on a qualitative publications search in the PubMed database (search terms: spasticity, pain, cerebral palsy, children, and botulinum). The aim was to identify the current peer-reviewed literature on SRP and available pain assessments in children with CP and for botulinum toxin treatment. General information about SRP was reviewed, and findings on important concepts were outlined in four initial draft questionnaire modules.

CE interviews were conducted with the child/adolescent and their parent/caregiver, following a semi-structured interview guide. Initial open-ended questions explored the presence of spasticity and SRP in patient language (e.g., ‘What happens when your (leg/arm) gets tight or stiff?’ ‘Does your (leg/arm) hurt sometimes?’ and ‘What happens when it hurts?’). These questions were followed by probing questions that explored the presence of the latter symptoms in particular situations (e.g., ‘Do your arms ever tighten up when you’re doing things like getting washed and dressed?’). Interviewers also recorded visual clues, such as children’s facial expressions and pointing, and explored whether the Wong-Baker FACES^® scale [29] was understood by the children and acceptable to the adolescents.

Interviews were audio-recorded, transcribed, and coded using Atlas.ti software for content analysis. Inter-rater agreement between coders was assessed on approximately 10 % of the transcript database. To assess saturation of concepts, transcripts were ordered chronologically, grouped into quartiles, and the newly appearing concept codes for each new transcript group were compared to the prior group. Saturation of concept was determined to have been reached when there were no longer new concepts being coded, and thus, all relevant concepts were captured. Based on previous experience conducting qualitative research and the degree of homogeneity in this population, we estimated that saturation of concept would likely be achieved within 14–16 CE interviews.

The study was conducted at four different clinical sites in the USA with three trained and experienced interviewers. Cognitive interviews were subsequently conducted at two of the four sites. Sites were selected in different geographic areas for their large pool of eligible subjects, to avoid the clustering of regional effects. Sites were specialized for treatment of children with CP and were regularly administering therapies such as botulinum toxin, baclofen, or phenol to minimize spasticity. Patient records were initially reviewed by clinical staff to identify suitable families, and primary caregivers were approached if eligible.

Purposive sampling was used to maximize variation in both CE and cognitive interviews, so children in different age brackets and with different gross motor function classification system (GMFCS) levels were recruited. Male and female children (2–11 years) and adolescents (12–17 years) with unilateral or bilateral CP with spasticity and intermittent SRP in either the upper limb or the lower limb (or both) on at least a weekly basis were eligible for participation. Children were categorized into the following age bands for recruitment: 2–4, 5–7, and 8–11-year old. To include the full spectrum of children with CP, recruitment targeted children with upper or lower limb SRP and with GMFCS of I–III (ambulatory, less impaired) and IV–V (non-ambulatory, greater impairment). Subjects had already received intramuscular botulinum toxin treatment of spasticity or were likely candidates for this type of treatment. Subjects were excluded if they had fixed contractures, predominant forms of muscle hypertonia other than spasticity, constant pain, pure, or predominantly dyskinetic CP, or had undergone surgery for pes equinus in the 12 months prior to recruitment. Subjects unable to answer interview questions due to profound cognitive impairment were also excluded.

The CE interview results guided revisions of the initial draft measures to develop the preliminary modules of the QPS. Content validity assessment included the relevance of concepts to subjects and caregivers, the specific language used to describe symptoms and observed pain behaviors, the appropriateness of the aspects of each concept being measured (frequency, severity, duration of pain), and the appropriateness of the recall period.

In order to field test the QPS, four separate waves of cognitive interviews were conducted (approximately four interviews per wave) during which subjects and their caregivers were asked to complete a preliminary module of the QPS. Participants were asked to ‘think aloud’ and describe their thought process for interpreting each question and arriving at an answer. A sample of the questions asked would be ‘Please tell me in your own words what this item is asking you about’ (for caregivers) or ‘What did you remember when you read this question?’ (for children).

The cognitive interview process was used to assess and document the participant’s understanding of the underlying concept presented in the QPS items and to refine language and restructure difficult items for children and caregivers. Other aspects of the preliminary QPS were also assessed, including the appropriateness of the recall period, the fit of the response options, the suitability of the Wong-Baker FACES^® scale [29], and overall clarity of the format and instructions.

An additional 12 child and parent/caregiver pairs were enrolled in four separate waves of the cognitive interviews to assess the preliminary measure.

The main findings of the literature review can be summarized as followed: (1) Pain in children with CP is an under-recognized and emerging topic in this population due to the significant influences on QoL [3, 7‐9]; (2) SRP is one of multiple pain sources and needs to be distinguished from others such as hip dislocation, dystonia, surgeries, or contractures [6, 10]; (3) several general, acute pain assessments, and QoL questionnaires with pain items are available, but none specific for SRP in children and adolescents with CP and following actual guidelines [15, 16, 25‐27]; (4) pain severity, frequency, and location in dependence on different activity situations are important [8, 10, 16]; and (5) a modular approach for an assessment is necessary to differentiate between upper and lower limb SRP and to account for different ages or cognitive abilities of the patients [1, 8]. This supports the development of a parent/caregiver module to capture observed signs of SRP in children who were either too young or too impaired to communicate. Based on the review, four initial draft modules were created: one upper and lower extremity child/adolescent module and one upper and lower extremity parent/caregiver module. These modules corresponded to each other in structure and item content across a general pain item and three different activity situations (at rest, usual activities, and active mobilization) plus an item on location of pain. A 7-day recall period was selected based on considerations that (1) SRP is characterized as intermittent and not constant pain, so sufficient time was needed for the subject to have experienced pain to report about; (2) specific activities such as physical exercises and clinical therapy sessions relate more to a weekly schedule than a daily one; and (3) qualitative evidence with both children and parents indicated reliable memory for the past week (7 days) for activity and pain reporting, but not beyond that.

CE interviews were conducted with 21 children (aged 2–16 years) and caregiver pairs (Table 1 ). Subjects’ cognitive skills varied widely within age groups and GMFCS levels. Findings confirmed that SRP and spasticity itself were relevant to children and caregivers and affected their everyday lives. In total, 40 relevant symptom concepts were identified in the CE interviews. Larger domains such as ‘pain related to spasticity’ included sub-concepts such as ‘pain at rest,’ ‘pain triggered by specific activities,’ ‘pain severity,’ or ‘pain frequency’.

Saturation of concept was achieved by the fourth transcript group. Inter-rater agreement was between 92.3 and 94.9 % for the assignment of specific concept codes.

The most frequently mentioned symptom-related concept was ‘Pain experienced in general,’ with 50 expressions (10.9 % of all symptom expressions) in the child interviews and 36 expressions (7.8 % of all symptom expressions) in the caregiver interviews. This code category mainly captured children saying ‘it hurts,’ and caregivers reporting ‘pain’ and ‘hurt.’ The pain experienced by the children differed depending on the type of activity they engaged in. The next most frequently expressed symptom concept was ‘tightness’ (used to describe spasticity), with 27 expressions in child interviews (5.9 %) and 62 in caregiver interviews (13.5 %).

Generally, the children were not able to describe frequency or duration of their SRP in a reliable way, but they could identify pain intensity/severity using the Wong-Baker FACES^® scale [29].

All caregivers reported having observed SRP behaviors in their children during the last 7 days; yet, only 16 (76.2 %) of the children acknowledged having pain. One child directly stated he would deny pain: ‘Sometimes it hurts so bad, but I’ll never tell you.’ Caregivers most frequently detected pain by ‘observing body movement’ (n = 113, 17.4 % of the caregivers’ expressions for pain detection) followed by other signs such as ‘the child articulating pain,’ ‘changing the position of their body,’ ‘having mood changes,’ and ‘having different facial expressions’ [n = 102 (15.7 %); n = 68 (10.5 %); n = 45 (6.9 %); and n = 39 (6.0 %), respectively]. Most caregivers were able to report on the frequency of pain based on their observations of pain behaviors and their child’s verbalization of pain.

CE interviews confirmed that certain children were able to communicate about their pain, but did not have the cognitive or motor abilities to answer a questionnaire on their own. Accordingly, interviewer-administered modules were needed to be developed to cover a higher percentage of the target population. A total of six QPS modules were next drafted (Table 2): two child/adolescent self-administered modules, two child/adolescent interviewer-administered modules, and two parent/caregiver modules.

Content analysis of the CE interview results determined the selection of 11 symptom concepts (assessed in 11 of 12 items) for the child/adolescent modules and 12 symptom concepts (assessed in 12 of 17 items) for the parent/caregiver modules of the QPS (Table 3). It was also confirmed that a 7-day recall period was appropriate for this age group.

Items in the four child/adolescent modules probe the occurrence of SRP (‘Yes/No’) and pain intensity (Wong-Baker FACES^® scale) in different activity situations (pain in general, at rest, during usual activities, during physical therapy, and while performing a self-defined very difficult task) during the last 7 days. The caregiver modules probe for behaviors caregivers recognize as indicating pain and how frequently these behaviors were observed in the different activity situations on a five-point response scale from 0 = never to 4 = always. An example item of the upper extremity child/adolescent module and the corresponding items in the parent/caregiver module are presented in Fig. 2. Pain location was evaluated by caregivers and children together and included only in the caregiver’s questionnaire.

Cognitive interviews were conducted with 12 child (aged 7–16 years) and caregiver pairs. The demographic and clinical characteristics of subjects and caregivers are detailed in Table 1.