SSD

The SSD was originally developed to assess sleep disturbance in patients with atopic dermatitis and comprises 11 morning items and 4 evening items [24]. The morning items of the SSD were adapted to the PN context to assess the impact of PN-related symptoms on the sleep quality during the previous night; items were reworded from “atopic dermatitis” to “prurigo nodularis.” The evening items assess sleep-related impact during the day (i.e., daytime sleepiness), which are not directly related to sleep disturbance, and were therefore removed from the SSD for PN.

The morning SSD items were used to derive 10 sleep parameters: seven Consensus Sleep Diary-based parameters (sleep onset latency, wakefulness after sleep onset [WASO], terminal WASO, total awake time, time in bed, total sleep time, and sleep efficiency), two PN-specific parameters (PN-related WASO [WASO-PN] and number of WASO-PN), and one parameter on self-perceived sleep quality and restfulness after the previous night’s sleep (sleep quality/refresh) (Table S1). The parameter values range from 0 to 1 for sleep efficiency, from 0 to 10 for sleep quality/refresh, and from 0 to 24 for all other sleep parameters, except for the number of WASO-PN, which does not have a defined range.

Qualitative Interviews

Content validity of the SSD was evaluated through qualitative telephone interviews with a convenience sample of 21 US adults (aged ≥ 18 years) who had had PN for ≥ 6 months before the interview and scored ≥ 7 on the Peak Pruritus NRS (PP-NRS) and ≥ 4 on the SD-NRS at screening.

A semi-structured interview guide was used. During the interview, participants discussed the PN symptoms and PN-related impacts that were most important to them. For cognitive debriefing, patients were asked questions about their comprehension of the SSD. As a result of time limitations, interview fatigue, and the repetitive nature of the probing, not all participants were asked to answer every cognitive debriefing question for individual SSD items.

Interview transcripts were deidentified, coded, and analyzed using ATLAS.ti (ATLAS.ti Scientific Software Development, Berlin, Germany). A preliminary coding dictionary, including concept codes to capture participants’ experiences of PN symptoms and their impact on daily life, was generated based on the interview guide. Four researchers independently coded the first transcript, and the coding was compared for consistency. The coding dictionary was then adjusted based on the coding of this first transcript and was used in the coding of subsequent transcripts by the same four coders. A qualitative data manager monitored the quality and consistency of the coded data. Additional details regarding patient recruitment, interview procedures, and content analysis were reported previously [25].

Psychometric EvaluationStudy Design and Participants

The psychometric properties of the SSD were analyzed using data from the phase 3 OLYMPIA trials. Details of the patient inclusion/exclusion criteria and study design of the two trials are reported elsewhere [15, 22]. Briefly, both trials were conducted in Europe, North America, and Asia Pacific and enrolled adults (age ≥ 18 years) who had had PN for ≥ 6 months; an Investigator’s Global Assessment (IGA) score ≥ 3 at screening and baseline (day 1 of nemolizumab injection); and severe pruritus, defined as a PP-NRS score ≥ 7 within 24 h before screening and an average PP-NRS score ≥ 7 over the previous week before baseline. Eligible patients were randomized 2:1 to receive subcutaneous injection of nemolizumab or placebo once every 4 weeks for 24 weeks in OLYMPIA 1 and for 16 weeks in OLYMPIA 2. The primary endpoints of the trials were the proportion of patients achieving an IGA score of 0 or 1 plus a ≥ 2-point reduction from baseline at week 16 and the proportion of patients achieving a ≥ 4-point improvement from baseline at week 16 on the PP-NRS. The proportion of patients with a ≥ 4-point improvement from baseline at week 16 on the SD-NRS was a key secondary endpoint.

PRO and Clinician-Reported Outcome Assessments

Patients completed outcome assessments at home or before study drug administration during their clinic visits following prespecified schedules (Table S2).

The SSD was completed daily in the morning at home using a handheld device from screening through week 24 (OLYMPIA 1), week 16 (OLYMPIA 2), or early termination (both trials). The measure went through cultural adaptation and linguistic validation when translated to ensure accuracy and relevance across patient populations from different countries.

Additional PRO measures completed by patients at home included the Average Pruritus (AP) NRS [26], PP-NRS [27], SD-NRS [18], Patient Global Impression of Severity–Sleep Disturbance (PGIS-SD), and Patient Global Impression of Change–Sleep Disturbance (PGIC-SD). At each clinic visit, patients were asked to complete the Patient Global Assessment of Disease (PGAD), Dermatology Life Quality Index (DLQI) [28], EQ-5D-3L [29], and Hospital Anxiety and Depression Scale (HADS) [30] and report PN-associated pain frequency and intensity using an on-site tablet. Clinician-reported outcomes included the Prurigo Activity and Severity Score (PAS) [31] and a modified version of the IGA.

Psychometric Analysis

Data from OLYMPIA 1 were used for the primary psychometric analysis in this study, while those from OLYMPIA 2 were used for confirmatory purposes. Although the SSD was not a primary or key secondary endpoint of the OLYMPIA trials, the trials were designed to include appropriate measures that could serve as anchors in the psychometric analyses, supporting the validation of the SSD. The psychometric evaluation was performed following FDA guidance for validating an outcome measure [23]. A statistical analysis protocol outlining all the psychometric analyses was developed before receiving the trial data.

At each visit, average weekly values from the past 7 days were calculated for SSD sleep onset latency, WASO, total awake time, total sleep time, sleep efficiency, WASO-PN, and sleep quality/refresh. A sum of events within the same timeframe was used for number of WASO-PN. If fewer than 4 days of SSD data were available for a given week, that week’s data were considered missing. Data for the terminal WASO and time in bed parameters were not used in the psychometric analyses.

Test–retest reliability of the SSD was examined in patients who selected the same PGIS-SD rating at baseline and week 1; had a score of 3 (no change) on the PGIC-SD at week 1; or had a ≤ 1-point change from baseline in their PP-NRS average weekly scores at week 1. Paired t tests and intraclass correlation coefficients (ICCs) were computed for each SSD parameter. An ICC ≥ 0.70 indicates good test–retest reliability [32].

Convergent and divergent validity of the SSD were assessed at baseline and week 16 by calculating Spearman’s rank-order correlation coefficients between values for SSD parameters and other measures. Correlations were categorized as weak (|r|< 0.30), moderate (0.30 ≤ |r|< 0.50), and strong (|r| ≥ 0.50) [33]. The SSD was expected to be moderately or strongly correlated with PRO measures assessing similar constructs (SD-NRS) or proximal constructs (PP-NRS, AP NRS, PN-related pain frequency, PN-related pain intensity, and DLQI) and weakly correlated with those assessing distal constructs (EQ-5D index, EQ-5D Visual Analogue Scale [VAS], HADS anxiety, HADS depression, and PGAD). Correlations were expected to be weak to moderate between the SSD and the clinician-reported outcome measures (IGA and PAS items 4 [exact number of pruriginous lesions], 5a [percentage of pruriginous lesions with excoriations/crusts], and 5b [percentage of healed pruriginous lesions]).

To evaluate the known-groups validity of the SSD, mean SSD parameter values were compared between groups categorized based on PGIS-SD score, PGAD score, PP-NRS average weekly score, DLQI total score, and IGA score at baseline and week 16. Between-group differences were assessed using an analysis of variance adjusted for multiple comparisons based on the Scheffé method [34].

As a first step in the responsiveness analysis, Spearman’s rank-order correlation coefficients between SSD parameters and outcome measures were calculated for changes from baseline to week 16. Outcome measures that were moderately or strongly correlated with SSD parameters (|r| ≥ 0.30) were used as anchors in the responsiveness analysis. For each anchor, mean changes from baseline to week 16 for SSD parameters were compared between patient groups defined according to the score change on the anchor. Between-group differences were tested using an analysis of covariance model adjusted for baseline SSD parameter values.

Relationship Between SSD and SD-NRS

The relationship between the SD-NRS and SSD parameters was first examined by calculating correlation coefficients between the SD-NRS and SSD parameters for baseline values and for changes from baseline to week 16. The equipercentile linking method [35] was then used to link values for SD-NRS and SSD parameters that had the same percentile rank.

Ethical Approval

Ethics approval for the qualitative interview study was obtained from Advarra (Columbia, MD; reference number Pro00044847), and participants provided written informed consent before taking part. The OLYMPIA trials were conducted in accordance with the Helsinki Declaration of 1964 and its later amendments, good clinical practice guidelines, and International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use guidelines. Both trials received ethical approval from applicable regulatory authorities and the independent ethics committee or institutional review board at each participating site before study initiation. All participants provided written informed consent.

Statistical Analysis

SAS version 9.4 (SAS Institute Inc., Cary, NC) was used for all analyses. All statistical tests had a two-sided significance level of 0.05. Data collected during the qualitative interviews were included in the content validity analysis. Data from all randomized patients (intent-to-treat population) in the two OLYMPIA trials were analyzed separately for psychometric evaluation.