Chronic pain is defined as pain that is unpleasant and long lasting with sensory, emotional, cognitive, and social components [1,2]. In children and young people with juvenile idiopathic arthritis (JIA), pain is also unpredictable, which hugely contributes to the burden of living with this long-term disease [3-5]. Children and young people conceptualize their JIA in terms of pain experienced [6], suggesting that for some, this is the most salient feature of their illness. However, health care professionals in pediatric rheumatology sometimes neglect to assess pain because of a lack of time and tools available to do so. In addition, professionals may not perceive pain as a priority for their patients because the focus of consultations may be on disease activity and function rather than pain [7,8]. Pain conversations are difficult to have with patients. The use of assessment tools, which remotely and efficiently collect rich pain data for these patients, could help to overcome this clinical problem.

In the general field of pediatric pain, multidimensional assessment using electronic mobile health (mHealth) apps is growing widely and quickly [9]. These tools offer several advantages to researchers, clinicians, and patients, compared with unidimensional, paper-based versions. They enable accurate documentation of complex pain data, reporting patterns and pain fluctuations through time- and date-stamped reports [10-12]; demonstrate better adherence and engagement; avoid recall bias through timely monitoring [13]; and allow for wireless data transport, which can be useful for remote management of symptoms [14].

Remote monitoring of pain for children and young people with JIA is valuable; however, there are some specific administrative challenges for this patient group that need to be further explored before these tools can be effectively implemented. Given that the nature of pain qualities can change daily (and within days) for children with JIA [15], pain assessment should be not only multidimensional but also regular and frequent. However, in diary studies of pediatric pain, there is no consensus on the frequency of reporting. Some studies have adhered to once-a-day reporting, and others have implemented 3-times-a-day sampling strategies [16,17]. Studies also often only collect data for short periods (eg, 1 or 2 weeks) [18]. To our knowledge, no research to date has explored patient preferences in the timing and frequency of ecological momentary assessments (EMAs). EMA is a technique used to assess an individual’s current experiences, as they occur in real time and real-world settings [19]. In most chronic pain studies using EMA, the time sampling strategies are based on researcher decisions, which have little rationale, and are limited in regard to what is acceptable and feasible for patients [20].

Decisions regarding the frequency of pain assessment administration rarely appear to be evidence based. This is a particular concern for health care professionals managing children and young people with JIA because they fear that more intense, regular pain assessment may lead to over-reporting or overexaggeration of pain or pain-related problems [7]. This phenomenon is called measurement reactivity, which is defined as a change in the variable being measured, because of the nature of the measurement method [21,22]. To our knowledge, this effect has not been explored in electronic multidimensional pediatric pain assessment, but now that real-time data collection techniques are becoming commonplace in pediatrics [18], it is important to ensure that these assessment methods are not detrimental to well-being. To this end, changes in the degree of interference caused by pain may provide a useful indicator of the impact of frequent pain assessment. Therapists involved in the care of those with JIA use pain interference rather than measures, such as simple pain intensity scales, to evaluate the outcome of their interventions [7,23]. This approach fits with a broader concept of pain as a motivational state rather than simply a somatic experience [24].

This study aims to investigate patient preferences, feasibility, and influence of several time sampling strategies in remote multidimensional pain reporting. Feasibility studies are used to determine whether an intervention or method is appropriate for real-world use in particular patient groups [25]. In the context of this research study, feasibility referred to the practicability of different self-report schedules, and an N-of-1 trial design allowed for a comparison of these within the same individuals [26]. We aimed to explore which pain reporting patterns were nonburdensome for children and young people with a complex long-term disease and why from both patients’ and parents’ perspective (as little qualitative research into patients’ reasons for disengaging exists [20]). We also studied the effects of different pain reporting intensities to investigate whether there was any evidence of measurement reactivity in response to using these tools more or less frequently.

This study was a randomized N-of-1 cross-over trial design, which explored the use of 4 different time sampling strategies: once-a-day, twice-a-day, once-a-week, and as-and-when children and young people had pain. These time sampling strategies were chosen based on earlier pilot work in which children and young people with JIA completed My Pain Tracker (MPT; an mHealth multidimensional pain assessment tool, discussed further in the section Materials and Measures) daily for 1 week and discussed how often they thought would be feasible to complete the tool. This study has been reported in accordance with the Consolidated Standards of Reporting Trials extension for reporting N-of-1 trials (Multimedia Appendix 1) [27].

Children and young people and their parents were recruited from a UK prospective inception cohort study of childhood-onset inflammatory arthritis (the Childhood Arthritis Prospective Study [CAPS]). The CAPS study collects data longitudinally from individuals who were diagnosed with inflammatory arthritis (in at least one joint) present for at least 2 weeks under the age of 16 years and who attend 1 of 5 UK pediatric rheumatology centers [28]. Exclusion criteria are septic arthritis, hemarthrosis, and arthritis caused by malignancy/trauma or connective tissue disorders. CAPS began recruitment in 2001 and collects up to 10 years of data following initial presentation to pediatric rheumatology. New children continue to be recruited currently. Written informed consent was provided by proxies for all participants. Assent was also provided by children where appropriate. The study obtained ethical approval as an amendment to CAPS from the UK National Health Service (NHS): Health Research Authority (REC 02/8/104, IRAS 184042).

Eligible children and young people and their parents were identified and contacted by the lead researcher (RRL) to invite them to a participant recruitment event in August 2017 in Manchester, the United Kingdom. Suitability for inclusion in this substudy was based on age (between 5 and 16 years) and English speaking in addition to the CAPS inclusion criteria [28]. Invitation letters and information sheets were sent. Interested children and young people and their parents were encouraged to contact the study team to register and to confirm their attendance at the recruitment event where a presentation about MPTs development was given, consent/assent was taken, and instructional guidance packs were provided. Children and young people and their parents who were interested but unable to attend the event were enrolled in their own homes. These participants were visited by the lead researcher (RRL), who gave a brief demonstration of how MPT worked before enrollment (completion of consent/assent forms and provision of instructional guidance pack for the study).

Children and young people were randomized to 1 of 4 groups (group A, B, C, or D), as they presented to the study using a list of random numbers generated before recruitment by the lead researcher (RRL). Each group followed a different pain reporting schedule, which ran for 8 consecutive weeks, divided into 2 blocks of 4 weeks. Each pain reporting schedule included 4 different time sampling strategies (once-a-day, twice-a-day, once-a-week, and as-and-when the pain was experienced), which were randomized to repeat once within each of the 2 blocks. The randomized scheduling was created using block randomization and finalized before children and young people random allocation (see Figure 1). Children and young people were instructed not to complete MPT on the final day of each week (washout period) before cross over into another schedule to prevent carryover effects into the next schedule. As participants were not switching between interventions (or on any active treatments as part of the study design), only switching between administrative patterns, the anticipation of carryover and lingering effects was minimal. However, this 1-day washout time frame was chosen to balance the risk of any small carryover effects while ensuring that participants did not lose the momentum of using MPT at home. Information about reporting schedules and schedule changes was provided to participants in an instructional guidance pack, which also detailed who to contact if they had problems with MPT or the iPad. All randomization processes were created using Stata version 14.0.

Data collection occurred in participants’ own homes throughout the North West of England.

A total of 373 eligible CAPS participants were contacted to take part in the study, of which 20 potential participants registered. Of 373 participants, 8 children and their parents were able to attend a recruitment event, and a further 6 wished to be enrolled in the study at their homes. Moreover, 6 children and young people and their parents did not respond when contacted again directly by the researcher to organize a time to visit them.

Furthermore, 14 children and young people and their parents took part in the study. Children’s and young people’s demographics and disease characteristics are presented in Table 1. In addition, 2 participants experienced technical difficulties during the study, which meant that MPT data did not save to iPads. This highlighted issues with the technical feasibility of some of the MPT software and the tool itself (but this was not linked to the feasibility of the time sampling strategies tested). Therefore, MPT data from these 2 participants were excluded from the analysis of adherence. Their data were still included in other statistical analyses, as pain interference questionnaires were returned, and in the qualitative analyses, as semistructured interviews were still conducted. All data were collected between August 2017 and January 2018. The full age range for the participants included in the study was 7 to 16 years.

Overall adherence (n=12) for each time sampling strategy demonstrated that adherence to once-a-week reporting was highest (15/24, 63% possible reports) followed by once-a-day (85/168, 50.6% total possible reports) and twice-a-day (127/336, 37.8% possible reports) reporting. As-and-when reporting ranged from 0 to 7 reports during the 2 weeks this strategy occurred for participants.

There were no systematic differences in pain interference between children and young people (n=14) in any of the different pain time sampling strategies using MPT (Figure 3; P=.77).

There was no correlation between week of participation and pain interference score (r=−0.04; P=.68). All time sampling strategies generated high test-retest reliability (all ICC over 0.6; see Table 2). When autocorrelation at the patient level was accounted for in a multilevel regression model, there were no significant differences in pain interference between time sampling strategies, compared with once-a-day strategy (compared with once-a-day strategy: once-a-week, P=.98; twice-a-day, P=.59; and as-and-when pain is experienced, P=.56).

A total of 2 major themes were deducted from the manifest content of interview data: theme 1: perceived advantages/disadvantages of each time sampling strategy and theme 2: perceived changes in pain experiences during the study (subjective measurement reactivity). Children and young people and their parent preferences for pain reporting strategies are presented in theme 1, and perceptions of subjective measurement reactivity are presented in theme 2, alongside narratives of findings (subthemes). Supporting interview quotations for each theme are presented in Multimedia Appendix 2.

An equal number of children and young people preferred once-a-day and as-and-when reporting (6/14, 43%). About 14% (2/14) of children and young people preferred twice-a-day reporting, with only one (1/14, 7%) participant reporting preference for once-a-week reporting. Within parents, 36% (5/13) reported that they most preferred once-a-day reporting, and slightly fewer preferred as-and-when strategy (4/13, 29%). Moreover, 14% (2/13) of parents preferred twice-a-day reporting, with one of these parents suggesting that an even more intensive time sampling strategy would have been appropriate (3 times a day). One parent (1/14, 7%) had no opinion on frequency of administration. Another parent preferred not to speak to the researcher about their preferences for reporting to encourage their child to be independently involved with the study by providing feedback about their use of the app on their own. This parent still provided consent and was present for the young person’s enrollment into the study.

The majority of children and young people (n=11) and their parents (n=11) did not perceive their pain to have been influenced by assessment frequency. For those who did report subjective reactivity (children and young people=3 and parents=2), it became apparent that there were 2 kinds of subjective measurement reactivity being referred to: cognitive/emotional and actual physical changes in the level of pain.

Although mHealth tools are increasingly being used to collect pain data in pediatric chronic pain [9], studies have failed to explore the impact of different time sampling strategies with patients and families [20]. It is important to explore this to develop administrative strategies, which balance reporting burden with the highest quality data collection techniques. Findings from this N-of-1 trial of different time sampling strategies suggest that statistically, there is no objective measurement reactivity with different pain reporting frequencies (in terms of pain interference). Qualitative findings suggest that children and young people and their parents have a preference for once-a-day and as-and-when reporting, but the disadvantages of as-and-when reporting (problematic flexibility and difficulties in remembering to report) far outweigh those cited for once-a-day. For some children and young people, there are perceived changes in emotion and fatigue in response to more intense pain reporting (subjective measurement reactivity). Children and young people demonstrate better adherence to less frequent time sampling strategies (once-a-week and once-a-day) in the short term. However, when reporting more flexibly (as-and-when pain is experienced strategy), some children and young people do not report pain at all, and for others, it can be problematic that good pain days are not captured.

Qualitative findings suggest that daily pain reporting is most feasible and preferred for children and young people with JIA, and quantitative data support that frequency of reporting has no impact on pain experience. Once-a-day strategy captured rich pain data, which children and young people valued and often looked back on in their pain report history (documented in MPT). This schedule enabled children and young people to explore patterns in pain (which would be useful for pain discussions with parents and/or health care professionals), and administration was perceived to be easily incorporated into routine. Although an equal proportion of children and young people preferred as-and-when reporting, there were more disadvantages cited for this time sampling strategy overall. For some, the flexibility of reporting was useful, but for others, this aspect of administration was burdensome because there were challenges associated with deciding when and how to report pain and which pains were significant. There were also difficulties associated with reporting pain experienced during school.

In addition to children’s and young people’s cited disadvantages, there are several methodological issues with event-based reporting, such as as-and-when pain is experienced (participant-initiated reports in response to pain occurrences [19]). In the interpretation of reports, it is impossible to know whether children did not have any pain or whether they did have pain but did not report it for whatever reason. This challenge would make it difficult to compare pain over time within individuals, which would be unfeasible in a clinical/home setting where there is remote data collection involved. Another problem concerns the data that are reported using as-and-when schedules. It is important to understand how and when children and young people define a painful event as occurring, and this would inevitably differ between participants [19]. As highlighted in the findings, some chose to only report unusually bad pain in as-and-when reporting. This schedule is problematic because the open interpretation of when it is necessary to report means that pain, which may be of interest to clinicians and researchers, is not captured.

In this study, patients’ preference was for daily pain reporting, whereas data entries were complete during once-a-week time sampling strategy, which may be an indication of the latter being a less demanding task. A strength of our work is that both subjective perspectives and objective indicators of completion were collected; however, the relative merits of each are important to consider. Lower adherence with a more frequent time sampling strategy still provides a more detailed picture (capturing daily variations) compared with complete but less intensive schedules. For example, a richer dataset is collected when a patient misses 3 of 7 daily reports (less adherent) compared with 1/1 weekly report (more adherent). From both research and health care professionals’ perspectives, a daily time sampling strategy with an incomplete dataset may appear more challenging to analyze, but it provides better contextual information about the pain experience overall. This advantage was also identified in the participant accounts.

In the adult literature, studies using electronic pain diaries have found no evidence of measurement reactivity (objective or subjective) in participants’ data responses [35,41-43]. These papers, however, focused only on reactivity in terms of effects on pain intensity, rather than exploring the relationship between measurement artifacts and other pain-related variables (such as pain interference). Pain-related variables may offer richer contextual information, which is why we chose to explore it in this study. Measurement of pain interference is now recommended as a key outcome in clinical trials [44] and refers to measurement of the extent to which daily activities are interfered with or limited by pain. The assessment of interference is important, especially given that pain intensity does not necessarily correspond with lived experience [45].

It is difficult to disentangle changes in actual experiences (eg, natural fluctuations in pain interference) and whether these would occur regardless of any potential reactive effect, which is an inherent risk of bias in studies such as this [46]. There are ongoing discussions about how to better control for this bias, but to partly address this issue in this study design, the number of measurements was balanced, and the timing of measurement was randomized within individuals and between groups. Furthermore, participants completed measurements of reactivity separate to the intervention, which was being explored (MPT), as they were asked to complete an additional measure (PROMIS pain interference scale), which was administered at weekly time points.

In terms of the implications for clinical pain assessment for those with JIA, this study provides some reassurance for health care professionals that fears about pain focusing or pain overreporting may not be justified [7]. The children and young people and their parents in this study reported few differences in their pain experiences during more intense time sampling strategies. Our findings also have implications for pediatric pain research and clinical assessment of pain. Many pediatric pain studies use scales with time reference periods, which encompass substantial periods, such as 2 weeks or a month’s worth of pain [47]. This study showed that once-a-week reporting was one of the least appropriate time sampling strategies from patients’ and parents’ perspectives because children and young people struggle to condense large time frames of information into 1 singular report. This task is inevitably more difficult in even wider time reference periods.

The substantial data collection period of this study should be considered a strength, as usually studies drawing on momentary assessment methodology in children and young people with chronic pain have narrow data collection periods (usually between 1, 2, or 3 weeks) [11,17,48]. A further strength of the study was the considerable sample size we recruited for the N-of-1 cross-over trials, which can be a burdensome study design for those involved. Very few N-of-1 trials have been conducted with children and where conducted in adults, it is not unusual to combine trial data from samples of less than 10 participants [26,49]. In N-of-1 trials, the number of data points per participant is considered to be more important than the total sample size [50].

A limitation of data collection is that participants were asked to verbally comment at the end of the study on whether they noticed changes in pain levels in response to the intensity of pain reporting. These subjective perceptions about measurement reactivity were thus based on their memory rather than objective assessment.

A limitation of this specific study design is that we did not explore the impact of random sampling protocols in children and young people (where randomly programmed iPad alerts would have prompted individuals when to report pain). This is the third main sampling type in real-time data collection techniques (in addition to time driven [based on preset schedules, such as once-a-day] and event-triggered [such as as-and-when schedule]). Exploring the feasibility of random scheduling would have allowed for a more complete picture of children’s experiences of using a wider range of momentary assessment techniques. A limit of the study findings is that N-of-1 trial findings are applicable to the patients the trial are conducted with and not for identifying population-level conclusions [51]. The findings of this intensive trial design possibly reflect the attitudes and impact of pain assessment frequency in some children and young people with JIA (we included participants of different ages and of different subtypes), but other patients may have different pain reporting needs that need to be investigated.

Future research should aim to explore the experience of perceived subjective measurement reactivity and the quality of these perceptions/changes in more detail. Several children and young people in our study reported some changes in mood and fatigue in their perceived response to more intense reporting schedules, which aligns with concerns that attention to pain leads to overexaggeration of symptoms expressed by health care professional’s perspectives [7]. Given these concerns, it would be interesting to explore health care professionals’ perspectives on these approaches. Further research should identify those patients for whom more frequent assessment of pain might be problematic and particularly emotionally or cognitively demanding. From this, we should develop phenotypes of these individuals to ensure appropriate pain data collection with minimal harm in pediatric pain studies and clinical care using these complex, multidimensional pain assessment tools.

In conclusion, our study highlights that daily reporting of pain using mHealth multidimensional assessments is most feasible in terms of patient preference and adherence in long-term data collection with children and young people with JIA. There was no evidence to support that any timing schedule had an objective impact on pain interference, although there were some perceived changes in mood and fatigue in more intense reporting schedules for some participants. These findings are important for the development of administrative guidelines for remote pain monitoring tools, which accommodate momentary assessment techniques. Overall, our findings support the use of mHealth multidimensional pain assessment tools regularly and frequently to better capture pain patterns in children and young people with JIA.