Translation and Validation of the Self-Assessment Psoriasis Area Severity Index

Abstract Background The self-assessment psoriasis area severity index (SAPASI) is a patient-administered psoriasis assessment tool for which we present a validated translation from English to Swedish. Methods Validity was evaluated in this single-centre study using the psoriasis area severity index (PASI) as the standard. Test-retest reliability was assessed using repeated SAPASI measurements. Results Significant correlations (p < 0.0001) using Spearman’s correlation coefficient (r) were found between PASI and SAPASI scores (r = 0.60) for 51 participants (median baseline PASI 4.4, interquartile range [IQR]: 1.8–5.6) and repeated SAPASI measurements (r = 0.70) among 38 participants (median baseline SAPASI 4.0, IQR: 2.5–6.1). Bland-Altman plots showed generally higher SAPASI scores than PASI scores. Conclusion The translated version of SAPASI is valid and reliable, although patients generally tend to overrate their disease severity compared to PASI. Keeping this limitation in mind, SAPASI has the potential of being implemented as a time- and cost-efficient assessment tool in a Scandinavian context.


Introduction
The field of digital health is rapidly evolving, and teledermatology (TD) [1] is a developing field for providing dermatology care [2] well-accepted by patients and physicians [3].As the COVID-19 pandemic emerged, the need for rethinking traditional methods for evaluations in the clinical dermatology setting has become even more obvious and urgent [4].
The psoriasis area severity index (PASI) is the gold standard tool for the assessment of the severity of cutaneous manifestations of psoriasis [5].Administering the PASI requires trained health care professionals, which complicates remote assessment of psoriasis severity.The self-assessment psoriasis area severity index (SAPASI) is a self-assessment tool, originally developed in English, that does not require the involvement of trained personnel.The SAPASI is described in detail by Fleisher et al. [6].
Previous studies have suggested that the original SAPASI is valid and reliable in the assessment of psoriasis and has equivalent value to the PASI [7], although the correlations between the tools have been of varying strength [7][8][9].The SAPASI was recently translated and validated to Portuguese [10].
The aim of this investigation was to develop a tool useful for distance assessments of psoriasis in a Scandinavian dermatology setting.Therefore, we set out to perform a structured translation of the original English version of the SAPASI to Swedish and assess its reliability and validity, using the PASI as standard.In addition to only performing correlation analysis as in previous studies, the agreement between the tools was assessed using Bland-Altman plots, in accordance to current method comparison analysis recommendations [11].The study was a single-centre study conducted in a university hospital outpatient clinic.

Back-Translation and Cultural Adaptation
The translation and cultural adaptation process was performed in accordance with the principles of good practice stated by the ISPOR task force [12]: 1. Formal approval for the translation and validation of the SAPASI was received from the lead author of the SAPASI, Dr. Validation and Reliability of the SAPASI Score Patients presenting to the outpatient clinic at our department were recruited in the time period of September 2018 to September 2019.Written informed consent was obtained upon inclusion.The included patients had received a clinical diagnosis of psoriasis, were 18 years and older, and were literate in Swedish.The inclusion criteria allowed for new patients referred from other health care providers, as well as patients returning for follow-up visits, to be included in the study.
PASI measurements of included patients with psoriasis were performed by one or two trained health care personnel.The raters included four assistant nurses, six nurses, and one resident in dermatology (M.L.).The SAPASI was provided to the patients at the end of the visit.Patients had received no previous training in the use of the SAPASI tool nor in the assessment of psoriasis.No guidance or instructions were provided on how to fill out the SAPASI.The PASI raters were blinded to the SAPASI ratings, and patients were blinded to the PASI ratings.
To evaluate test-retest reliability, patients were asked to complete the SAPASI on two occasions: in connection with the visit at the dermatology clinic and at home, 1 week later.The latter assessment was returned by mail in a stamped addressed envelope provided to the subjects upon inclusion.The time interval was chosen to prevent memory bias but still within a short enough time span to prevent a potential change in the status of their skin disease.Where the repeated SAPASI had not been returned within 10 days, one reminder call was made.
Data entry of SAPASI scores was made by ML who was blinded to the corresponding PASI scores.Psoriasis subtype was obtained from patient charts in conjunction with data processing.The study was approved by the Regional Ethical Review Board in Gothenburg (DNR 552-18).

Data Analysis
The total SAPASI scores were compared with the total PASI scores.In addition, the subcomponents of the instruments were compared (body surface area [BSA], erythema, thickness, and scaliness).Since BSA is measured on the same scale in the SAPASI and PASI, a direct comparison could be made.However, since PASI measures the erythema, scale and indurated subcomponents in each of four body regions, whereas SAPASI assesses these for average lesions, averages for each of the subcomponents of the PASI had to be calculated.The calculations were made according to mathematical equations described elsewhere [8].Correlation analysis was performed using Spearman's rank correlation coefficient (r).Linear regression with confidence intervals was determined by robust regression analysis using MM-estimation.Bland-Altman plots with limits of agreement [11] were Swedish Validation of the SAPASI Dermatology 2023;239:794-801 DOI: 10.1159/000530045 constructed to evaluate the agreement between total PASI and SAPASI scores and repeated SAPASI measurements, respectively.Limits of agreement were only calculated if the criteria of normally distributed data and homoscedasticity were fulfilled.A subanalysis including only subjects with plaque psoriasis was performed.Any SAPASI scores >30 were considered as outliers and excluded from the analyses.All statistical analyses were performed using SAS v 9.4 software (SAS Institute, Cary, NC, USA).
The first SAPASI measurement systemically exceeded the second by an average of 0.58 ± 2.36 points, with limits of agreement ranging from −4.0 to 5.2 (Fig. 3).

Validity and Reliability of SAPASI Scores in Patients with Plaque Psoriasis
When including only individuals with plaque psoriasis (n = 39), the total PASI score was correlated with the SAPASI assessment (r = 0.51, p = 0.0008) (Fig. 4a).

Discussion
In the current study, we present a Swedish version of the SAPASI with statistically significant correlations between the total PASI and the total SAPASI scores.In previous studies, correlations have been of varying strength.The main authors of the original SAPASI first reported a correlation of r = 0.77 (p < 0.0001) [8].In a multi-centre study, the SAPASI scores reflected the PASIequivalent scores in a statistically significant way and indicated changes in disease severity.However, correlations were weaker than previously described (r = 0.54 and r = 0.33 [p = 0.0001] for the initial and final SAPASI measurements, respectively) [9].Measures of psoriasis severity were compared in a multi-centre setting, showing a strong correlation between SAPASI and PASI scores (r = 0.91, p < 0.00001) [7].The correlation in BSA between the PASI and SAPASI has also been of varying strength in previous studies [8,9].Consistent with previous reports, correlations were weaker when comparing the remaining subcomponents of the PASI with the corresponding subcomponents of the Swedish SAPASI [8,9].The correlation in the current study was the highest for erythema and the lowest for thickness, supporting previous findings [8,9].Considering that no statistically significant correlation was found between the PASI and SAPASI regarding thickness in patients with a Fig. 3. Test-retest reliability of the Swedish self-assessment psoriasis area severity index (SAPASI).Bland-Altman plot for the first and the repeated SAPASI assessment showing average differences between measurements plotted against the mean.n = 38.plaque psoriasis, caution should be exercised in the evaluation of the subcomponents of the SAPASI.
Based on correlation analyses, previous reports suggest that the SAPASI exhibits reproducibility and strong reliability upon repeated measurements (r = 0.82, p = 0.0001) [8].The current study showed a test-retest reliability of r = 0.70 p < 0.0001), supporting previous findings.

Interpretation of Correlation Coefficients
The correlation coefficient is often used in measurement comparison studies to determine agreement.However, the results should be interpreted with caution [13,14].Nevertheless, considering that previous studies have been performed using correlation coefficients, for the sake of comparison, we included the Spearman correlation as a complement to the Bland-Altman plots, which has been suggested as a more appropriate method to assess agreement [11,13].
What becomes evident in Figure 1b is that, even though there is a strong correlation between the SAPASI and the PASI, patients generally tend to have higher SAPASI than PASI scores, which is consistent with previous findings [15].Unsurprisingly, this difference increases the higher the score.Hence, the PASI cut-off scores categorizing psoriasis into mild, moderate, and severe disease cannot be directly transferred to SAPASI scores, which is important to keep in mind.Nevertheless, the SAPASI has previously been used in several studies for the assessment of psoriasis in Scandinavian settings [16][17][18][19][20], and disease severity has been classified by directly applying established PASI cut-off scores to SAPASI scores [20], which highlights the need for a validation.
The great advantage of using the PASI is its unbiased and objective nature, while the higher SAPASI scores may be a result of emotional involvement of the patient.It has been suggested that the SAPASI is responsive to changes in psoriasis severity and the use of photographs, illustrations, or other facilitating tools may be an effective way of further improving the agreement between the SAPASI and PASI [9].

Strengths and Limitations
The study design, including the structured method of back-translation and validation, is a strength of this study.Furthermore, all PASI assessments were made by trained health care personnel.However, considering that different staff members were involved in the PASI assessments, there may be a risk of inter-observer variation [21].
A longer time interval between the repeated SAPASI measurements compared with a previous report (7 days vs. 2 days) [8] was chosen to reduce the risk of memory bias.This could, however, increase the risk of significant clinical changes in the patient's psoriasis status, which may affect the reproducibility assessment.
A further weakness of the study is the single-centre design, where all subjects were patients at a single outpatient clinic.Finally, the median baseline PASI and SAPASI scores were relatively low (4.4 [IQR: 1.8-5.6]and 4.4 [IQR: 2.6-7.2],respectively), which limits the generalizability to psoriasis of greater severity.

Applications of the SAPASI
As the rapid digitalization process continues, access to valid and reliable tools for assessment of disease severity is fundamental.Although distance monitoring and management are less studied for inflammatory skin conditions [22], the assessment of these must not be compromised.The SAPASI offers a time-and costefficient assessment tool that has the potential of being implemented in a TD setting.

Conclusion
In the current study, the correlation between a validated Swedish translation of the SAPASI and the PASI was consistent with previous findings.However, patients tend to overvalue their disease severity using SAPASI compared with PASI assessments and there were weaker correlations between the subcomponents of the PASI and SAPASI.Hence, PASI scores cannot be directly transferred to SAPASI scores.The SAPASI may be useful as a time-and cost-efficient complement to monitoring disease severity in individual patients, especially in remote health care settings.

Key Message
The SAPASI could be implemented as a self-assessment tool for psoriasis monitoring in different languages.

Fig. 1 .
Fig. 1.Validity of the Swedish self-assessment psoriasis area severity index (SAPASI).a Linear regression with 95% confidence intervals for PASI and the first of two SAPASI scores.b Bland-Altman plot for PASI and the first SAPASI scores, showing average differences between measurements plotted against the mean.n = 51.PASI, psoriasis area severity index.

Fig. 2 .
Fig. 2. Validity of the subcomponents of the Swedish self-assessment psoriasis area severity index (SAPASI).Linear regression with 95% confidence intervals for subcomponents of PASI and the first of two SAPASI assessments of: (a) BSA scores, (b) average colour scores, (c) average thickness scores, (d) average scaliness scores; n = 51.BSA, body surface area; PASI, psoriasis area severity index.

Fig. 4 .
Fig. 4. Validity of the Swedish self-assessment psoriasis area severity index (SAPASI) in patients with plaque psoriasis.a Linear regression with 95% confidence intervals for PASI and the first of two SAPASI scores.b Bland-Altman plot for PASI and the first SAPASI score, showing average differences between measurements plotted against the mean.n = 39.PASI, psoriasis area severity index.