This study demonstrated the face validity and acceptability of the Swahili versions of the FPS-R and the NRS as single-item pain assessment tools for use in hospitalized Kenyan patients. Of the unidimensional pain assessment tools available, these scales were judged by our team to be most practical for use in our setting in western Kenya, and they also correspond to the Kenyan Hospice and Palliative Care Association’s recommendation of using a numerical scale and a faces scale for pain assessment. These tested translations use straightforward, non-idiomatic Swahili phrasing, which should make them usable in other Swahili-speaking countries of East Africa, as well as in immigrant populations worldwide.
By using comparisons of their current pain to their previous levels of pain as well as to the facial expressions of the FPS-R to select an answer, participants demonstrated rational decision-making processes in choosing responses to the FPS-R and the NRS. They showed good comprehension of the progression from the left-most anchor of no pain to the right-most anchor of the worst pain on both scales. Participants also found the scales easy to use and denied any confusion with the phrasing of instructions or the scales themselves. This comprehension and acceptability to patients is particularly important to demonstrate across cultures, especially in a setting where pain assessment is not routinely done.
There were a few suggestions for changes that could be made to the scales, though none of these were judged to qualify for revision, as they were determined to be subjective preferences (one boy wanted to put ears on the faces) or misunderstandings of the cognitive interview probes (one woman wanted to give the faces medicine to make them feel better). The outcomes of cognitive interviewing help define “tradeoffs” – the advantages and disadvantages of asking a question in a certain manner – rather than dictating a “correct” way to ask a question . In our case, the data generated from cognitive interviewing demonstrated the advantages of keeping the questions as they stand rather than making any changes.
Our team had been worried that the FPS-R would not be understood by Kenyan patients because they might be unfamiliar with “cartoon” faces. There was also a concern that, in this cultural context, patients would be too stoic or too afraid of being considered “weak” to admit to feeling pain. Given the wide range of pain ratings on both scales, including a considerable amount of high scores, and the overwhelmingly favorable acceptance of the faces scale, it seems that these concerns, while still possibly valid, were not the determining factors in our sample. Through the detailed process of cognitive interviewing, Kenyan patients demonstrated understanding and acceptance of both pain scales.
Participants both objectively and subjectively had more difficulties with the NRS than the FPS-R. Our finding of the effectively unanimous preference for the faces scale runs counter to the results of previous studies that have found that older children and adolescents generally prefer a numerical scale or a visual analogue scale to the faces scale . In our study, children and adults across all age ranges preferred the FPS-R. This may be attributable to lower levels of education and difficulties with numeracy. In 2006, Kenya’s adult literacy rate was 61.5%, and its adult numeracy rate was 64.5% . This may have implications for pain scales chosen for use in other populations with low educational levels. Alternatively, our findings may also be due to a cultural preference among this patient population. The subjects raised several questions about the NRS and had more difficulty correctly identifying the number that denotes “no pain,” suggesting that the NRS may be a more problematic scale for use in this population than the FPS-R.
These findings may serve to caution the recent movement towards an international standard for pain assessment in palliative care [21, 41]. Based on systematic literature reviews and expert opinion surveys, the NRS has emerged as the lead candidate for universal adoption . While we agree that establishing a universal standard for pain intensity measurement is desirable, it may not be feasible to dictate the use of a single tool that is optimized for all populations. In concert with other studies in non-Caucasian patient groups that have found a preference for the FPS-R, our results serve as a reminder that any efforts to establish a consensus must include representation from non-Western populations [42, 43].
Despite the concerns encountered during the cognitive interviewing process, the NRS still has strong potential for clinical use in our population. All the participants were able to choose a response to the NRS, and the FPS-R and NRS ratings were largely consistent subject-by-subject. Previous studies have reported the many advantages of the NRS, including ease of administration, straightforward scoring mechanism, patient preference, and ability to be used in parametric analyses [15, 21, 44, 45]. The NRS and FPS-R have different strengths, and these distinctions can help direct their use in this population in the future. For example, the FPS-R may be more suitable for day-to-day patient care, since subjects preferred this scale and found it easier to understand. On the other hand, the NRS may be more appropriate for research purposes, as it lends itself to statistical analysis and is becoming the international standard, thereby facilitating comparisons to be made with other populations.
There were some potential limitations to this study. We performed 15 cognitive interviews, and if we had conducted more, we may have found more subtle deficiencies in the pain scales. However, experts generally recommend doing a maximum of 12 to 15 cognitive interviews before analysis, as subsequent interviews reach a point of diminishing returns . Additionally, the cognitive interviewers are all part of the hospital’s palliative care team and are familiar with the topics of pain assessment and treatment. While this may have biased them towards overestimation of pain or may have led to a preconceived belief that the pain assessment tools are acceptable and needed, the use of standardized scripts for the cognitive interviewing process, and the process of translation and transcription of the interviews by a non-palliative care team linguist minimizes the influence of interviewer bias on our results.