Remember me
Introduction:
Persistent Postural-Perceptual Dizziness (PPPD) is a chronic functional vestibular disorder exacerbated by posture, movement, or visual stimuli. Widely used dizziness questionnaires lack specificity for PPPD symptoms. The Niigata PPPD Questionnaire (NPQ) and its German-translated and revised version (NPQ-R), including two additional subscales, were developed to address this gap. Its internal consistency, convergent validity, and test–retest reliability were found to be satisfactory. The aim of the present study was to examine the NPQ-R’s structure using confirmatory factor analysis (CFA) and Rasch item analysis.
Materials and methods:
We analysed data from 265 (135 female) patients (50.2 ± 16.8 years, dizziness duration 46.3 ± 76.6 months) who completed the NPQ-R. CFA was conducted using the robust maximum likelihood estimator in R (lavaan), and Rasch item analysis was performed for each subscale separately.
Results:
CFA revealed moderate-to-high covariances between the five latent variables (range: 0.59–0.97), with all items except one demonstrating significant standardised loadings. Rasch analysis indicated acceptable item fit for most items. Person separation reliability ranged from 0.63 to 0.75 across subscales. Item 2 (Visual Stimulation) exhibited misfit.
Discussion:
The NPQ-R demonstrates promising psychometric properties. While the Rasch analyses support reliability and internal coherence, the CFA results suggest that the overall five-factor model may require refinement. Conceptual clarity and wording of specific items should be re-examined. However, the NPQ-R remains suitable for PPPD assessment and severity determination in clinical and research contexts, while theoretical and empirical refinement of its factor structure is recommended.
1 IntroductionPersistent Postural-Perceptual Dizziness (PPPD) is a chronic functional vestibular disorder that has gained increasing clinical and scientific significance, as it is the most common chronic vestibular disorder in people aged 30 to 50 years (1, 2) and one of the most common forms of vertigo (3, 4). In 2017, the consensus committee of the Bárány Society defined PPPD as a persistent feeling of dizziness, unsteadiness, and non-spinning vertigo lasting at least for 3 months, with symptoms exacerbated by upright posture, self-or environment-induced motion, and exposure to complex visual stimuli (3). PPPD can develop either without (primary PPPD) or following an acute vestibular syndrome (secondary PPPD), such as an acute unilateral vestibulopathy, benign paroxysmal positional vertigo or vestibular migraine, or a period of heightened anxiety or psychological stress (3, 5, 6). PPPD is assumed to arise from maladaptive top-down control, characterized by a high-risk balance strategy and visual dependence (3, 7). Functional MRI studies have identified alterations in brain structure, function, and connectivity in patients with PPPD, including reduced activity of the vestibular multisensory networks and overactivity of the visual areas and other changes (8–10). The clinical presentation of PPPD is often complex and highly individual. Patients may have difficulty describing their symptoms precisely, which can contribute to underdiagnosis or misdiagnosis, particularly in primary care settings (4). Patients may use words like “lightness in the head” or “cloudiness” to explain their symptoms (3, 11). Additionally, PPPD is frequently associated with comorbid affective and anxiety disorders, including panic disorder and generalised anxiety disorder, which may predate or follow the onset of dizziness symptoms (5, 12). Notably, patients often get into a vicious cycle of emotional distress amplifying dizziness, which in turn increases anxiety and avoidance behaviour. This pattern is similar to that observed in functional neurological disorders and chronic pain conditions (5, 13).
Historically, instruments such as the Dizziness Handicap Inventory (DHI) (14, 15), the Vertigo Symptom Scale (VSS) (16, 17), and the Activities-specific Balance Confidence (ABC) Scale (18, 19) have been used to assess dizziness-related symptoms and impairment in patients with organic vestibular syndromes. While these tools provide valuable insights into the impact of dizziness on quality of life in these patient groups, they were not designed to capture the symptom profile specific to PPPD. As a result, their content validity for this disorder is limited (13, 20). To address this gap, Yagi et al. (21) developed the Niigata PPPD Questionnaire (NPQ), a Japanese self-report instrument based directly on the Bárány Society criteria. The NPQ assesses the severity of PPPD symptoms across various typical contexts, including visual stimulation, movement, and upright posture. As it was originally developed to assist in the clinical diagnosis of PPPD and to quantify symptom severity in a patient-reported format, its primary purpose is to capture the overall symptom profile and severity associated with PPPD rather than to differentiate between potential clinical subtypes of the disorder. In 2025, the Athens–Lübeck Questionnaire (ALQ) was introduced as an instrument specifically designed to assess potential PPPD subtypes (22). Accordingly, the NPQ should primarily be understood as a tool for diagnostic support and severity assessment. Its utility has since been demonstrated in both Japanese and Western populations (23, 24), and it has been adapted into several languages, including French (24), Spanish (23) and German (25). The German revised version of the instrument (NPQ-R) was developed through cognitive debriefing interviews with PPPD patients and an expert Delphi consensus process to enhance content validity and clinical relevance (26). This version introduced additional items and a more refined item structure to better reflect patients’ experiences. In a recent publication, Chételat et al. (25) presented the first German translations and validations of both the original NPQ and the NPQ-R, following internationally accepted guidelines for cross-cultural adaptation (27, 28). The study confirmed the internal consistency, convergent validity, and test–retest reliability of both instruments in a German-speaking clinical sample.
The availability of the culturally adapted and psychometrically validated German NPQ-R represents a major step forward for clinical and research applications. However, further psychometric evaluation is needed to confirm the underlying factor structure and assess model fit in a larger, more diverse sample. Thus, the aim of the present study was to examine confirmatory factor analysis and Rasch item analysis of the German revised version of the Niigata PPPD Questionnaire (NPQ-R) to support diagnostic procedures and monitoring of treatment in PPPD patients. We hypothesised confirmation of the five-factor structure, comprising the three original factors proposed by Yagi et al. (21) and the two additional factors introduced in the German revised version of the questionnaire by Behrendt et al. (26).
2 Materials and methods2.1 Study design, recruitment, and study patientsThis study is a continuation of the German adaptation and validation of the original and revised Niigata PPPD Questionnaire (NPQ and NPQ-R) conducted by Chételat et al. (25) and Behrendt et al. (26). As part of a larger project, the current work aimed to evaluate the factorial structure of the revised German NPQ-R in a sample of German-speaking patients with PPPD using confirmatory factor analysis. The five-factor model tested was based on both theoretical foundations and previous empirical findings (26).
This study included data from 265 patients with PPPD in Switzerland and Germany, recruited between September 2021 and December 2023. In Switzerland, patients were recruited at the Cantonal Hospitals of Lucerne, the University Hospitals of Zurich and Basel, Reha Rheinfelden, and various private practices. In Germany, all patients were recruited at the outpatient unit of the German Center for Vertigo and Balance Disorders (DSGZ) at the LMU University Hospital in Munich. Out of the 265 included patients, 140 (52.8%) were diagnosed with primary and 125 (47.2%) with secondary PPPD.
Patient inclusion criteria were a diagnosis of PPPD, age ≥ 18 years, good German language proficiency, and signed informed consent (25). Exclusion criteria were other forms of vestibular disorders as defined by the International Classification of Vestibular Disorders (ICVD). Detailed information on patient recruitment, diagnostics, and study methodology can be found in the recently published open access article by Chételat et al. (25).
2.2 Measures: NPQ-RThe NPQ-R is a 19-item self-report instrument developed to assess the core symptom dimensions of PPPD, whereas the subscales were defined as “Upright Posture,” “Movement,” “Visual Stimulation,” “Associated Symptoms” and “Symptom Behaviour” (26). Items are rated on a seven-point Likert scale ranging from 0 (“I have no complaints”/“Does not apply at all”) to 6 (“Is it unbearable”/“Fully applies”), with higher scores indicating greater symptom severity. Maximum total score of the NPQ-R is 114 points. The German NPQ-R was developed following cross-cultural adaptation guidelines by Beaton et al., 2000 (26) and validated for internal consistency, convergent validity, and test–retest reliability (25).
2.3 Statistical analysisThe dataset was compiled in Microsoft Excel and imported into R (Version 2025.05.1 + 513) for statistical analysis. The R packages lavaan, laavanPlot, dplyr, openxlsx, DiagrammeR and eRm were used for analysis. Level of significance was set at p ≤ 0.05.
As described in Chételat et al. (25), questionnaires with missing values exceeding 15% were excluded from analysis to mitigate potential selection bias (29). Therefore, two questionnaires containing more than 15% missing data were excluded from the analysis and the remaining questionnaires (≤15% missing data) were retained and analysed using full information maximum likelihood estimation. Demographics and test data were analysed using descriptive statistics in the preceding study conducted by Chételat et al. (25). Floor or ceiling effects are present when 15% of subjects achieve the highest or lowest possible score (30).
Data were screened for univariate normality by calculating skewness and kurtosis for each item. Items with skewness between −2 and +2 and kurtosis between −7 and +7 were considered approximately normal (31). These criteria were supported by visual inspection using histograms and Q-Q plots.
The confirmatory factor analysis was performed to test the previously proposed five-factor model of the NPQ-R, in which the 19 items were assigned to the following latent variables: (1) Upright Posture (Items: 4, 11, 12, 18), (2) Movement (Items: 1, 8, 15, 19), (3) Visual Stimulation (Items: 2, 6, 13, 16), (4) Associated Symptoms (Items: 3, 5, 9, 17), and (5) Symptom Behaviour (Items: 7, 10, 14). Due to deviations from normality in some items, the robust maximum likelihood estimator was used in lavaan to estimate model parameters. Missing data were handled using full information maximum likelihood, as implemented in the package.
The following fit indices were evaluated to assess model adequacy: comparative fit index: ≥ 0.95 (acceptable), ≥ 0.97 (good); Tucker–Lewis index: ≥ 0.90 (acceptable), ≥ 0.95 (excellent); root mean square error of approximation: ≤ 0.06; standardised root mean square residual: ≤ 0.08 (32).
Chi-square statistics and degrees of freedom were also computed. Fit indices were extracted using the fitMeasures() function, and parameter estimates (standardised and unstandardised) were obtained via the parameterEstimates() function in lavaan.
To further investigate the item properties of the NPQ-R, a Rasch analysis was conducted using the partial credit model for each of the five subscales individually. The Rasch analysis is a psychometric method designed to enhance the development, evaluation and refinement of psychological assessment instruments (33, 34). The focus was on evaluating item fit using outfit and infit mean square statistics. Reasonable item mean square ranges for infit and outfit was defined as 0.6–1.4 (35). Further, the person separation reliability was computed for each subscale. Associated t-statistics were not used due to their sensitivity to sample size. Given the size of our sample, t-statistics are liable to identify trivial misfit as statistically significant.
3 Results3.1 Study sampleThe demographic characteristics reported in this study are based on the analyses originally conducted by Chételat et al. (25) and were adopted here as a secondary analysis. The sample included 265 patients (mean age 50.2 years, SD 16.8; 130 males: 47.1, SD 17.2; 135 females: 53.1, SD 15.8; dizziness duration 46.3 months, SD 76.6). In total, 117 patients were recruited from different centres in Switzerland, and 148 patients were recruited from the DSGZ at the Ludwig-Maximilians-University Munich. The average duration of dizziness was 46.3 (SD: 76.6) months. For more detailed information regarding sex-specific analysis, please refer to Chételat et al. (25).
3.2 Confirmatory factor analysisFor the confirmatory factor analysis model estimation, data from all 265 patients were used. All 19 NPQ-R items were included in the analysis. Analysis of the NPQ (12-items) can be found as Table 1 and Figure 1. Descriptive checks showed that skewness values for the items ranged from −1.02 (PPPD_17) to 0.57 (PPPD_16), and kurtosis values ranged from 1.84 (PPPD_6) to 3.42 (PPPD_17). Histograms and Q–Q plots indicated minor deviations from normality. Based on these results, model estimation was performed using the robust maximum likelihood estimator with full information maximum likelihood for missing data.
SubscaleItemItem questionEstimateStandard errorz-valueP(>|z|)Standardized factor loadingUpright PosturePPPD_4Wenn ich in meinem eigenen Tempo zu Fuss gehe, dann …Confirmatory factor analysis factor loadings for the NPQ (N = 265).
The table presents the factor loadings for each item across the three latent subscales. Estimates were obtained using robust maximum likelihood (MLR) with full information maximum likelihood for missing data. NPQ, Niigata PPPD Questionnaire (12-items).
Path diagram of the confirmatory factor analysis model of the NPQ (N = 265). The figure the standardised factor loadings (outer paths) and standardised latent covariances (curved double-headed arrows) for each subscale. NPQ, Niigata PPPD Questionnaire (12-items).
Standardised covariances between latent variables ranged from 0.59 to 0.97. The covariance between Upright Posture and Movement was 0.97. Other covariances were:
Upright Posture – Visual Stimulation = 0.59,
Upright Posture – Associated Symptoms = 0.71,
Upright Posture – Symptom Behaviour = 0.96,
Movement – Visual Stimulation = 0.82,
Movement – Associated Symptoms = 0.75,
Movement – Symptom Behaviour = 0.96,
Visual Stimulation – Associated Symptoms = 0.66,
Visual Stimulation – Symptom Behaviour = 0.84, and
Associated Symptoms – Symptom Behaviour = 0.89.
Model fit indices were as follows: χ2 = 563.90, degrees of freedom = 142, p < 0.001; comparative fit index = 0.80; Tucker-Lewis index = 0.77; root mean square error of approximation = 0.11; standardised root mean square residual = 0.08.
All items showed significant standardised factor loadings (all p < 0.001) except PPPD_9 (p = 0.07). The standardised factor loadings can be seen in Table 2.
SubscaleItemItem questionEstimateStandard errorz-valueP(>|z|)Standardized factor loadingUpright PosturePPPD_4Wenn ich in meinem eigenen Tempo zu Fuss gehe, dann …Confirmatory factor analysis factor loadings for the NPQ-R (N = 265).
The table presents the factor loadings for each item across the five latent subscales. Estimates were obtained using robust maximum likelihood (MLR) with full information maximum likelihood for missing data. NPQ-R, Niigata PPPD Questionnaire Revised version (19-items).
Intercepts for the observed items ranged from 1.73 (PPPD_7) to 4.46 (PPPD_17). Standardised residual variances ranged from 0.34 (PPPD_13) to 0.76 (PPPD_7). All item residual variances were significant (p < 0.05). Latent variable variances were Upright Posture = 1.01 (p < 0.00), Movement = 0.99 (p < 0.00), Visual Stimulation = 0.89 (p < 0.00), Associated Symptoms = 0.97 (p = 0.05), and Symptom Behaviour = 0.47 (p < 0.00).
A graphical illustration of the confirmatory factor analysis, including standardised factor loadings and standardised covariances between latent variables, is provided in Figure 2.
Path diagram of the confirmatory factor analysis model of the NPQ-R. Legend: The figure shows the standardised factor loadings (outer paths) and standardised latent covariances (curved double-headed arrows) for each subscale. NPQ-R, Niigata PPPD Questionnaire Revised version (19-items).
3.3 Rasch analysisFor the Upright Posture subscale (PPPD_4, PPPD_11, PPPD_12, PPPD_18), outfit mean square statistic values ranged from 0.74 to 0.92, and infit mean square statistic values from 0.72 to 0.90. Person separation reliability was 0.75.
In the Movement subscale (PPPD_1, PPPD_8, PPPD_15, PPPD_19), outfit mean square statistics ranged from 0.69 to 0.93, and infit mean square statistics from 0.70 to 0.92. Person separation reliability was 0.73.
For the Visual Stimulation subscale (PPPD_2, PPPD_6, PPPD_13, PPPD_16), outfit mean square statistics ranged from 0.62 to 1.18 and infit mean square statistics from 0.62 to 1.13. One item (PPPD_2) had elevated outfit and infit mean square statistic values (1.18 and 1.13, respectively). Person separation reliability was 0.72.
The Associated Symptoms subscale (PPPD_3, PPPD_5, PPPD_9, PPPD_17) presented outfit mean square statistics from 0.64 to 1.05 and infit mean square statistics from 0.65 to 1.06. PPPD_9 was the only item with mean square statistics slightly above 1.0. Person separation reliability was 0.70.
In the Symptom Behaviour subscale (PPPD_7, PPPD_10, PPPD_14), outfit mean square statistics ranged from 0.67 to 0.78 and infit mean square statistics from 0.68 to 0.76. Person separation reliability was 0.63.
Person separation reliability and item fit statistics for all subscales of the NPQ-R are summarised in Table 3. Wright maps are presented in Figure 3.
SubscalePerson separation reliabilityItemItem questionOutfit mean square statisticsInfit mean square statisticsUpright Posture0.75PPPD_4Wenn ich in meinem eigenen Tempo zu Fuss gehe, dann …
Comments (0)