Literature

Searches for pain intensity measures were made in the COSMIN database, Google, and PubMed. In the Cosmin database (https://database.cosmin.nl/), the search was made using “Pain” (All fields) and Filters: Adult, patient reported outcome, physical symptom state, questionnaires/interviews/diaries/clinical rating scales, NRS/NRS-11/NRS-21/NRS-Child/PI-NRS (pain intensity NRS). Google.com was searched for images of “pain measure” and “smerteskala” (Danish, meaning pain scale). The PubMed.gov search strategy is given in Table 1. The articles, their references, and images were included if they contained pain intensity described in words, phrases, or visual cues such as color [18] and graphic presentation. Other ways of expressing pain intensity such as sound [19] or finger pressure [20] were excluded for lack of use in a web-based scale. Words describing pain intensity were explored further in English online dictionaries and thesauri to aid understanding. The resulting collection of words and phrases were qualitatively assessed by familiarization and inductively categorized.

Patient interviews

10 patients were recruited face-to-face in the Day Surgery Centre, Horsens Regional Hospital and interviewed after their surgery. Sample size was 10 based on expected saturation and more than 7 interviews being considered very good in the COSMIN checklist for content validity [6]. Patients were chosen based on availability and provided consent on the day of surgery. For a broader representation, maximum 2 patients for each NRS score were chosen. Interviews were conducted in Danish in the postoperative resting area. Interviews were semi-structured, based on an interview guide (translation in Supplementary data). Semi-structured interviews were chosen to allow for spontaneous use of pain descriptors without specific prompts from the interviewer. Patients were asked to rate their pain on the 0–10 pain scale and then prompted open-endedly to elaborate, followed by more specific questions regarding impact on sleep and activity.

Interviews were performed by KTB (MD, PhD) and/or CWGA (MD), experienced medical doctors int the department, trained for this study, but not involved in their treatment. All interviews were audio recorded and transcribed verbatim. The transcripts were analyzed inductively by thematic content analysis for categories of intensity descriptors. Data was managed in a framework (case- and theme-based matrix) including specific wording. Coding was done together for 5 transcripts and individually by CWGA and KTB for 5 transcripts, reaching consensus through a meeting comparing assigned codes.

Clinician interviews

10 clinicians were recruited from Horsens Regional Hospital and from the acute pain service at Aarhus University Hospital, a strategic convenience sample of male/female nurses/doctors working with pain assessment daily. Semi-structured interviews were conducted in Danish, following consent, individually in a private office setting by KTB using a guide (translation in Supplementary data). Transcription and analysis were conducted as for the patient interviews. Coding was done together for 1 transcript and individually by CWGA and KTB for 9 transcripts, reaching consensus through a meeting.

Patient telephone interviews

100 orthopedic day surgery patients were recruited for telephone interviews conducted 1 week postoperatively. Inclusion was consecutive and dependent on age above 18 years and written informed consent before discharge. 1 patient spoke English, but all other interviews were conducted in Danish. Questions included open-ended inquiries as to pain intensity during the first week, application of the NRS, and comments on the numeric scale. During these interviews, written notes were taken on the interview guide (translation in Supplementary data), as time restraints did not allow recording and transcription. Analysis was again made by coding done independently by CWGA and KTB and reaching consensus.

Patient on-site interviews

Interviews were conducted from November 2021 to March 2022. Patients were all literate and ethnically Danish, 3 males and 7 females, all interviewed postoperatively on the day of their surgery. Age was 33–74 years (mean 52.5 years), and they were a broad representation of orthopedic day surgery patients (hand, foot, hip, knee, and shoulder). At the time of interview, their NRS pain scores were: 0, 1, 3, 3, 4, 4, 5, 5, 5-or-6, and 7.

When asked what their pain intensity score meant to them, most patients described it in terms categorized as intensity (translated: It is not sore as such, I can feel it/moderate/it hurts), evaluative (tolerable/I can live with it), and discriminative (radiating/shooting/constricting). When asked directly, patients could describe how their pain would affect activity and their ability to sleep, and whether they needed additional analgesics. Only higher pain levels involved terms categorized as affective (NRS7: irritating, NRS5–6: It is not [intense enough to be] annoying), physical signs (It’s not like I’m lying here sweating), or with cognitive impact (inability to abstract from the pain).

A large variation in the 2 descriptions of the pain score “5” was found: 1 patient could easily accept, move, and possibly sleep with the pain, and another could do none of these. Also, 1 patient labelling her pain as NRS 3 described being pain free at rest, but with NRS 7 when moving.

Several found it easier to describe pain intensity in relation to previous experiences like stubbing a toe on a hard surface, giving birth, known chronic pain elsewhere (e.g., back pain) or the pain they had before surgery.

Overall, the categories induced from the patient interviews were: intensity, evaluative, discriminative, cognitive impact, sleep impact, activity impact, treatment, affective, exemplification, and physical signs. Furthermore, rest/movement, localization, timing (constant/varying), and recent pain medication were important variables to consider.

Clinician interviews

4 anesthesiologists (3 male) and 6 female nurses were interviewed from November 2021 to January 2022. Notably, in the clinician interviews the category of physical signs was elaborated (e.g., cold sweating/rising blood pressure and pulse/grimacing).

When asked to describe the NRS in words, all but 1 chose spontaneously to group the numbers by intensity (e.g., mild/moderate/severe) as describing each individual number was difficult.

No discriminative wording was used. Exemplification was used frequently. Differences between interviewed clinicians were smaller than between the interviewed patients, possibly reflecting similar training.

Several clinicians mentioned discrepancy between patients’ use of the NRS and clinical observations. For example, patients saying “10” when they can talk and move without difficulty. This discrepancy led the clinicians to suspect that the patients did not interpret the scale in the same way as they did.

Strengths

We have sampled the target population, so both timing (present pain or recalled 1 week) and surgery (orthopedic) are relevant for the intended use of the scale. Patients were not interviewed regarding hypothetical or much earlier recalled pain, which would make the collected data more reliant on imagination, previous experience, or memory. For the next phases, where the specific wording must be selected and the scale designed and validated, it is still essential to involve the specific target population defined by language and surgery as well as linguistic expertise. As Danish literature is scarce in comparison, English literature was chosen, and naturally interviews were in Danish. The two categories “exemplification” and “physical observations” obtained by interviews are not likely unique to Danish patients and clinicians. The quality of measurement is greatly dependent on the scale being understood as uniformly as possible by all trial patients. The categories are likely to be widely applicable, but upcoming studies will determine how broad a population can reliably deal with the same wording.

Limitations

A qualitative study depends on the ability of the researchers to interview comprehensively and catch relevant codes from the collected information. We optimized this by using interview guides, training for this study, and coding the data mostly separately, so our individual impressions of both the data and coding were applied and discussed. Also, we are as surgeons trained and experienced in communication and pain assessment, particularly with this patient group. The use of different sources (literature, patients, and clinicians) improves coverage. Repetitive findings and no new categories during the 100 telephone interviews are strong indications that we have sufficient data to catch all categories (thematic saturation). It is questionable whether all thinkable words or examples describing pain intensity can ever be listed, but based on our extensive material the identification of relevant wording in the upcoming selection and validation studies is quite certain. Second, we did not interview patients below 18 years due to rules of consent, and coverage for adolescents remains to be studied. Also, although we recruited clinicians of different sex, professions, and workplace, the narrow variation in answers could, in addition to similar training, indicate the need for wider sampling. Another limitation is that the relative importance of the findings is difficult to determine: it is not yet established which category, phrase, or graphic element is most useful.

Conclusion

Acute postoperative pain intensity is described in a large variety of terms collected and categorized in this study. Many graphic elements may also aid in improving standardized self-report, and we have identified variables important to consider when measuring pain.

Perspective

These results will be the basis for the next phase in PROM development, which is content validation: identification of the specific wording and graphics most relevant, understandable, and with sufficient coverage.