This study was conducted at a hospital-affiliated pain management and physical medicine and rehabilitation outpatient clinic in southwestern Ontario. It was staffed by an interdisciplinary team of health care professionals, consisting of a physiatrist, physical therapist, and kinesiologist. Patients who are referred to this outpatient clinic receive a thorough medical evaluation, including assessment of pain, and are then informed of the management plan including pharmacological, injection, and physical therapies. They may also be referred for psychological therapy (eg, group counseling, cognitive behavioral therapy) if needed. All patients are reassessed at timely intervals and treatments are adjusted according to clinical need.

Informed written consent was obtained from all participants, and the study was approved by the locally responsible Research Ethics Boards. A health care provider known to patients identified eligible individuals by screening the patient lists of consecutively scheduled clinic appointments. Individuals were eligible to participate if they were (1) aged 18 years or older, (2) able to speak and read English, and (3) currently experiencing pain of any intensity according to self-report. Individuals were excluded if they had severe cognitive impairment or major comorbid medical or psychiatric illness that could preclude their ability to self-report pain or take part in a verbal interview, according to their health care provider. Individuals were also excluded if they had severe vision or hand dexterity impairments that could prevent independent use of a computer and mouse.

Following consent, each participant completed a Demographic and Health Questionnaire, which collected data on age, sex, computer comfort, weekly computer use, language proficiencies, education level, and date of pain problem onset.

All participants took part in an individual semistructured interview (20-30 minutes) with a trained investigator (author CL). The investigator was experienced in conducting qualitative interviews and used techniques to minimize the power differential between the interviewer and participant (eg, established rapport, engaged in active listening, used relaxed body language) [25]. The investigator also stressed that the research team wished to ensure that Pain-QuILT addressed the needs of adults with chronic pain and thus encouraged participants to freely express opinions about good and bad aspects of the tool. As a first step, participants self-reported their pain using the Pain-QuILT, MPQ, and BPI (described in detail below). These tools were administered in a randomized order for each participant, in order to minimize potential order effects. Investigator observation and participant comments were used to identify any difficulties or confusion with using each tool; these were recorded as field notes. The time required to complete each tool was recorded. Next, a semistructured interview format was used to discuss participant preferences for pain self-report. A 0-10 NRS ranging from “not easy at all” to “very easy” was used to appraise each tool. Qualitative written feedback on the ease of using each tool was also collected. Finally, participants were asked to indicate their preference of methods for self-reporting pain and explain the reason for their choice. All interviews were conducted by the same investigator in a quiet room within the clinic.

Qualitative written data and field notes from the semistructured interview were transcribed verbatim and imported into the qualitative software program, HyperRESEARCH [30]. This software was used to facilitate a simple content analysis of the data [31]. A line-by-line coding analysis was used to identify key concepts from the interview transcripts and field notes. Concepts addressed during the semistructured interviews were used to thematically code and organize participant responses [31]. Participant quotations were selected to illustrate each key interview concept with the aim of representing the balance of opinion among participants.

Quantitative data from the Demographic and Health Questionnaire, MPQ, BPI, and Pain-QuILT were coded, scored, and entered into a Statistical Package for the Social Sciences database [32]. As described by Lalloo and colleagues, the extracted parameters from each Pain-QuILT report were the number of unique painful sites (range 0 to 110) and number of different pain quality descriptors (range 0 to 16) used to express current pain [24]. Additionally, a cumulative mean pain intensity score was calculated across all painful body sites. While this cumulative score provided a convenient indicator of the central tendency of data, it was also sensitive to outliers. Thus, we also extracted the lowest and highest single NRS intensity score to provide an indicator of data dispersion. For example, if a participant reported a 5/10 burning pain in their foot, a 3/10 burning pain in their hand, and a 3/10 stabbing pain in their back, then the number of unique painful sites would be recorded as 3, the number of unique pain quality descriptors would be 2, the cumulative intensity score would be calculated as [(5+3+3)/3]=3.7, the lowest reported NRS score would be 3, and the highest reported NRS score would be 5.

All data were analyzed descriptively to assess measures of central tendency (mean, median) and dispersion [standard deviation, interquartile range]. Data were also evaluated to ensure that they met the assumptions of parametric statistical analysis (ie, the normal distribution). When these assumptions were not met, the non-parametric equivalent test was used. Repeated measure analysis of variance (ANOVA) was used to determine whether there were any differences between Pain-QuILT, MPQ, and BPI in terms of time to complete or ease of use ratings. Pearson correlations were used to examine the association between pain intensity scores on Pain-QuILT and BPI. The a priori criterion for evidence of convergent validity was a moderate correlation of r=.5 between Pain-QuILT and BPI scores for current pain intensity. Using the guidelines from Streiner and Norman pertaining to sample size for correlation coefficients, assuming alpha=.05 and beta=.05, the required sample size was N=50 [33]. The level of significance was set at P<.05 for all tests.

A total of 50 adults completed the study over a 5-month period in 2013. Sample characteristics are summarized in Table 1. Nearly all participants (48/50, 96%) had a computer at home as well as Internet access (45/50, 90%). Of the 50 participants, 84% (42/50) reported being “comfortable” or “very comfortable” with using computers, while 10% (5/50) were “a little comfortable” and 6% (3/50) were “not at all comfortable”. The self-reported frequency of computer use among participants was none (3/50, 6%), once per week (3/50, 6%), twice per week (2/50, 4%), three times per week (4/50, 8%), five times per week (1/50, 2%), and every day (37/50, 74%).

All participants reported the relative ease of using each tool for self-reporting pain. The mean ratings were 5.9 (SD 2.6) for the MPQ, 7.0 (SD 2.6) for the BPI, and 8.3 (SD 2.0) for Pain-QuILT. Overall, there was a significant difference between the tools in terms of perceived ease of use, F _2,96=20.6, P<.001. Pairwise comparisons also indicated significant differences between the MPQ and BPI (P=.009), MPQ and Pain-QuILT (P<.001), as well as BPI and Pain-QuILT (P=.002).

Overall, 46% (23/50) of participants indicated that they had difficulties in completing the MPQ, while 22% (11/50) reported difficulties with the BPI and 16% (8/50) specified difficulties with Pain-QuILT.

The most commonly reported issue with the MPQ was trouble with understanding the qualitative word descriptors (10/23, 43%) due to language barriers (eg, English as second language) or uncommon vocabulary, such as “taut” and “smarting”. Participants (7/23, 30%) also reported that the available pain words “...weren’t very good to describe [pain]” (ie, lack of descriptiveness). Other participants (7/23, 30%) noted that it was difficult to select the right words to express their pain due to ambiguity (“what is difference between cool, cold, freezing?”), the number of available options (“too many choices”), and the presence of more than one relevant word from certain subclasses. Last, participants (2/23, 9%) expressed concern about potentially misrepresenting their pain to their health care providers: “more fear of not describing your pain properly with this test”.

The most commonly reported issues with the BPI were communicating pain location using the body-manikin (2/11, 18%; “hard to pull out meaning”) and choosing intensity ratings to describe pain (2/11, 18%; “hard time with pain numbers”). Other reported difficulties included recalling pain over the last 24 hours (“hard to simplify pain”), reporting pain from multiple sites (“varying intensities of pain from different injuries”), and questionnaire design (“cumbersome to complete, too general”). One participant also indicated a “fear of not explaining properly what is happening”.

The most commonly reported issues with Pain-QuILT were related to the virtual body-map (3/8, 37.5%). Specifically, participants identified a need for orientation labels (left, right) and to make it easier to isolate specific painful body areas (“hard to find specific regions on [the] back versus a ‘paint’ tool, because some pain radiates”). In addition, participants (2/8, 25%) indicated difficulty in choosing pain quality icons due to “too many choices...sometimes it aches, sometimes it burns”, and a dislike of using descriptors because “pain just hurts”. Other participants (3/8, 37.5%) identified a “bug” in the software related to an inability to remove icons that were mistakenly added to the body-map.

The mean time required by participants to complete a single pain report using each tool was 4.2 minutes (SD 1.5) for the MPQ, 4.0 minutes (SD 1.4) for the BPI, and 4.1 minutes (SD 2.2) for Pain-QuILT. There was no significant difference between the tools in terms of time to complete, F _1.4,44.8=0.13, P=.81.

Overall, 16% (8/50) participants chose the MPQ as their preferred method for self-reporting pain, while 14% (7/50) chose the BPI, and 58% (29/50) chose Pain-QuILT. Four of the 50 participants (8%) indicated that they preferred the “other” method of verbally explaining pain to their health care provider. Finally, one participant (1/50, 2%) indicated an equal preference between the BPI and Pain-QuILT.

Reasons for selecting the MPQ included a preference for paper versus electronic pain reporting and greater perceived precision in describing pain, for example, “[it has] words that exactly indicate what is happening to [my] leg—bang on”.

Reasons for choosing the BPI included familiarity, the ability to describe how pain changes over time, and ease of choosing ratings on a set scale, for example, “[it] seems more easy to answer personally. Fits the way that I speak”.

Explanations for choosing Pain-QuILT included greater ease of use, ability to pinpoint different locations and types of pain, preference for computer versus paper-based pain reporting, as well as the visual language to express pain, for example, “[I would] feel more confident being treated by a doctor if they used this tool because [they] would know exactly what you are feeling”.

Our previous work has established the acceptability, usability, and content validity of Pain-QuILT in samples of adults with central post-stroke pain [21], adults in the community with a range of different types of chronic pain [22], as well as adults and adolescents with arthritis pain [34]. Clinical feasibility testing, the focus of the present study, is an important stepping-stone toward widespread implementation of a new assessment tool [35]. Our group has recently demonstrated clinical feasibility of Pain-QuILT in the context of an interdisciplinary pediatric chronic pain clinic among adolescents aged 12-18 years [24]. The present study sought to extend these findings by evaluating clinical feasibility of Pain-QuILT from the perspective of adults attending an outpatient pain clinic for treatment of chronic pain. This study included a comparison of Pain-QuILT with standard methods of pain assessment.

As a tool for self-reporting pain, Pain-QuILT was rated as significantly easier to use than the MPQ and BPI, which are two of the most commonly used pain assessment tools in research and clinical practice. Almost half (46%) of participants reported difficulties in using the MPQ. Most of these difficulties related to understanding the pain descriptors and finding accurate words to express pain from a large number of options. These findings of the present study are consistent with a meta-analysis of 51 studies involving 3624 patients, which found that most MPQ words (75%) are rarely endorsed by patients to describe their pain [36]. Although the BPI was associated with fewer reported difficulties, participants indicated that its design was not conducive to reporting different intensities of pain in different body sites. Numerous studies have demonstrated that chronic pain is rarely confined to a single body region [37-39]. For instance, in a study involving 2445 patients, Carnes and colleagues found that 73% experienced pain across multiple body sites [37]. Among patients with low back pain, only 13% experienced regionally isolated pain. In terms of implications for pain assessment and management, these authors concluded, “self-reported measures of multi-site pain are problematic with pain measures that are site-specific. Pain in other areas may render them less reliable and responsive. Future intervention studies should consider recording other pain sites to identify predictors of response to treatment” (p. 1170) [37]. Overall, Pain-QuILT was associated with the fewest reported difficulties among participants. Most of the identified issues (75%) will be resolved in the next iteration of Pain-QuILT software (eg, adding orientation labels to body map, fixing “bug” related to deleting unwanted icons). Participant concerns related to the changing nature of pain (“sometimes it aches, sometimes it burns”) will be addressed in future longitudinal studies, which will allow patients to use Pain-QuILT as a diary to document symptoms as they occur. A major identified strength of Pain-QuILT was the ability to record multiple sites, types, and intensities of pain.

The average time required to complete each assessment tool was less than 5 minutes. While there was no significant time difference between the tools, it is important to note that patients can enter Pain-QuILT data independently, while the MPQ and BPI are usually administered by a health care provider in the context of a clinic appointment. Moreover, Pain-QuILT data are generated and stored in a digital format, while information from MPQ and BPI must be manually transcribed into a spreadsheet (paper or computer-based) in order to facilitate tracking over time. Thus, Pain-QuILT has the potential to increase efficiency of clinic appointments by (1) empowering patients to self-report pain on their own time (eg, at home and/or in the clinic waiting room), (2) providing health care providers with digital summaries of tracked pain data to evaluate and inform their management plan, and (3) standardizing the assessment of pain outcomes for use across multiple providers.

Given the inherently personal nature of pain, it is important to consider patient preferences regarding the most effective way of expressing symptoms. The majority of participants (58%) indicated positive preference for Pain-QuILT over alternate methods. It is well recognized that patient engagement is a critical factor in the successful management of chronic disease [40]. In particular, effective doctor-patient communication is known to enhance the health outcomes of pain management [41]. The interactive and dynamic format of Pain-QuILT may also help patients forge a stronger emotional connection to the tool as a means for portraying and conveying their pain experience, compared to static questionnaires. Moreover, there is a growing body of literature documenting the rise of “self-tracking” among people living with chronic illness. A recent Pew Research Center report found that 40% of adults with 1 chronic condition and 62% of adults with 2 chronic conditions currently self-track their symptoms [42]. In terms of patient benefits, respondents indicated that self-tracking influenced their overall approach to maintaining health (56%), prompted them to ask their doctor new questions (53%), or influenced a treatment decision (45%) [42]. Thus, by providing a user-friendly method for communicating with health care providers and self-tracking painful symptoms, Pain-QuILT may encourage greater patient involvement in the long-term management of their own disease.

There is a growing number of patient-oriented mobile applications (apps) designed to aid the self-tracking of pain. In 2011, Rosser and Eccleston identified 111 pain management apps, of which 24% included a self-monitoring function [43]. A more recent scoping review, conducted in 2013, identified 224 pain apps, of which 14% allowed users to self-track their symptoms [44]. Unfortunately, both studies identified major limitations in the current field of pain apps, including a lack of formal evaluation and limited involvement of health care professionals and patients in their development. Pain-QuILT has been iteratively evaluated and refined through consultation with patients as well as health care professionals and thus has potential to address these identified gaps in the field, as one component of chronic pain management.

Given that participants were asked to self-report their current pain using three different methods, we expected to observe consistency in reported pain. Using the MPQ, participants were presented with a choice of 46 qualitative descriptors across 11 subclasses. Interestingly, the most frequently chosen MPQ words were consistent with the icon descriptors on Pain-QuILT. This relationship was independent of the order of tool assessment. Pain-QuILT icons and word descriptors have been iteratively refined based on patient interviews to ensure that they are representative of the pain experience. The observation that the icons correspond with the most frequently endorsed MPQ descriptors provides further evidence of validity. In terms of pain intensity scores, we examined correlations between Pain-QuILT (body site-specific pain scores) and the BPI (single global score for current pain). There were high correlations (r≥.70) observed between BPI score and (1) the calculated average pain score across all body sites, and (2) the single highest reported pain score across all body sites. There was also a moderate correlation (r=.55) observed between BPI score and the single lowest reported pain score across all body sites. Along with our previous pediatric study, which compared Pain-QuILT scores with a verbal NRS (r=.61), the current data provides further evidence of convergent validity. Importantly, in terms of clinical usefulness, we suggest that the greater level of detail elicited by Pain-QuILT may help inform pain management strategies (eg, observing how treatment affects pain quality and intensity scores within specific body sites) more than a single global intensity score.

The present clinical feasibility study was conducted at a single interdisciplinary pain management and rehabilitation clinic in Southwestern Ontario. Although the organization and treatment model of this site was consistent with other Canadian multidisciplinary pain treatment facilities [45], we acknowledge that future work is needed to evaluate clinical feasibility of Pain-QuILT in other settings. Further, given the interview component of this study, it was necessary for all participants to be able to speak and read English. Although 38% of participants spoke multiple languages, future work is needed to formally evaluate Pain-QuILT in non-English speaking groups. Given the visual nature of Pain-QuILT reporting, it could prove to enhance pain communication for individuals with limited verbal or cognitive skills.

Participants in this study completed only a single Pain-QuILT report. Future work will evaluate whether patient perceptions regarding ease of use and preferences, as well as time to complete, are affected by repeated usage.

The results of this clinical feasibility study in adults with chronic pain are consistent with our previously published pediatric findings [24]. Specifically, data indicate that Pain-QuILT is (1) easy to use, (2) quick to complete, (3) preferred by a majority of adults with chronic pain, and (4) correlated as expected with validated pain measures. As a digital, patient-friendly method of assessing and tracking pain, we conclude that Pain-QuILT has potential to add significant value as one standard component of chronic pain management.

The tool will be licensed for clinical use and research studies through the McMaster Industry Liaison Office [46,47]. Updated information on availability will be provided on the author website [47] and Twitter account (@PainQuILT).