Validity and reliability of the Niigata PPPD Questionnaire in a Western population

Persistent postural-perceptual dizziness (PPPD), also known as functional dizziness, is a chronic vestibular disorder recognized by the International Classification of Vestibular Disorders [1] and is characterized by one or more symptoms of dizziness, unsteadiness or non-spinning vertigo present on most days for 3 months or more. This functional condition is one of the most frequent clinical entities; it is the first cause of chronic dizziness and is positioned as the first cause of consultation in adulthood between 19 and 64 years in specialized centers [2, 3]. Its prevalence peak appears between 30 and 50 years of age, and in terms of distribution by sex, this condition is more likely to affect women, with a female-to-male ratio of 3:1.

Different vestibular events, such as benign paroxysmal positional vertigo (BPPV), acute unilateral vestibulopathy/vestibular neuritis, Meniere's disease or central events, such as vestibular migraine or stroke, can act as precipitants of this condition, although their presence is not necessary or mandatory [4]. Other syndromes, diseases and functional conditions have been identified that can act as triggers [5], as well as a certain predisposition in the healthy population to the development of this health problem [6].

PPPD patients usually have dizziness that during the early stages tends to fluctuate with periods of exacerbation and even remission. Subsequently, dizziness, although it continues to fluctuate in terms of the intensity with which the subject perceives it, will remain persistent [4]. The symptoms can be exacerbated by different demanding sensory conditions, such as mobile visual environments (busy streets, road traffic, etc.), highly structured visual environments or when visual references are far apart [5]. In the same way, the patient's active own movement of both the head and the body as a whole, as well as passive movement (means of transport, escalators, etc.) and upright movement, may cause an increase in the patient's symptoms [4].

Within therapeutic management, multimodal approaches have shown efficacy in the medium term, such as the combination of serotonergic medication, cognitive–behavioural psychotherapy and physiotherapy [7, 8]. With regard to the approach through physiotherapy, the protocols are based on desensitization models through neurosensory stimulation of increasing intensity [9, 10]. Due to the lack of well-defined procedures and doses, the protocols are dependent on each work team [9, 10]. Thus, the identification of these exacerbating factors plays an important role in defining what type of action through physiotherapy is most adapted.

With this scenario, a health measurement instrument has recently been developed for this problem. The Niigata PPPD Questionnaire (NPQ) is the first instrument developed to measure the severity of PPPD. The NPQ includes 12 items that yield a global score. Furthermore, the items are organized into three four-item subscales that specifically assess the three exacerbating factors [11]. The original validation study included the measurement of two psychometric properties, internal consistency and diagnostic validity, through ROC curves and the establishment of cut-off points to predict the presence of PPPD. However, other characteristics, such as factorial validity, concordant validity with other similar measures or reproducibility, have not been studied, nor has the value of the standard error of measurement (SEM) or the minimum detectable change (MDC) been obtained.

The NPQ is the first questionnaire developed specifically for the measurement of health in patients with PPPD that is increasingly used in the country in which it was developed. However, its psychometric characteristics have been insufficiently studied. Although the authors of the original article present an English version of the questionnaire, it has not been evaluated outside the borders of Japan. For this reason, this study considered the analysis of the main psychometric properties of the NPQ in a Western population. This work required the cross-cultural adaptation of the questionnaire from the Japanese language to the Spanish language for its use for the first time in a European population.

Our work aims to validate the NPQ for use in Western populations. We also aim to analyze the psychometric properties of the NPQ in a population of patients with vestibular disorders. Specifically, this study aimed to measure the construct validity, internal consistency, reproducibility and concurrent validity of the NPQ as well as its ability to discriminate between patients with PPPD and patients with other conditions.

The study involved 140 subjects (Fig. 1), among whom 47 had a diagnosis of PPPD following the stabilized criteria. In the group of “other” vestibular conditions, the most frequent diagnosis was benign paroxysmal positional vertigo (BPPV). The main characteristics of the final sample are shown in Table 1. Additionally, in 31 patients, a second evaluation could be obtained approximately between one and two weeks after the first evaluation.

The factorial validity measured by principal components analysis showed a single factor structure (Fig. 2) that explained approximately 60% of the total variance. The KMO = 0.896 was satisfactory, and the Bartlett sphericity test was significant (X₂ = 1207,691, p < 0.001). The determinant of the matrix was very close to “zero”, which indicates that the correlation matrix between the NPQ items is very different from the identity matrix (absence of correlations between the items). All of these measurements indicated that the sample could be considered suitable for factor analysis.

The internal consistency analysis of the questionnaire taken globally yielded a Cronbach alpha coefficient = 0.938, which could indicate some redundancy of the items. Their analysis (Table 2) indicated that the elimination of any of them would slightly decrease the alpha value but without drastic changes. In the analysis of the subscales, the values can be considered good, with Cronbach’s alpha = 0.869 for the standing and gait subscale, Cronbach's alpha = 0.803 for the subscale of movements and Cronbach's alpha = 0.852 for the visual stimulation subscale. In practically all the cases, the scores that can be considered as good and not indicative of redundancy, the elimination of the items harmed the internal consistency of the subscale.

The reproducibility of the global score and subscales could be considered substantial except for the standing subscale, which could be considered moderate. Generally, the items showed moderate reproducibility except for items 4, 8 and 11, which could be considered substantial (Table 3). Taking into account the ICC found, as well as the pretest standard deviation, the SEM was 3.27, 2.41, 2.50 and 6.63, and the MDC was 6.40, 4.72, 4.91 and 12.99 for the standing, movement and visual stimuli subscales and the total score, respectively.

The concurrent validity analysis between NPQ scores and other variables of interest is shown in Table 4. The total score and subscale scores showed insignificant correlations with falls in the last 12 months and with the Physical Component Summary (PCS) of the SF-12. The correlation was moderate with the Mental Component Summary (MCS) of the SF-12, and there was a moderate correlation between the visual stimulation subscale and the measured central sensitization. The rest of the correlations were strong, showing good concurrent validity of the NPQ measures and DHI subscales, ABC scales, CSI and VVAS.

The diagnostic validity of the NPQ total score and subscales is shown in Table 5, and the ROC curves are shown in Fig. 3. In general, the diagnostic capacity was statistically significant but could be classified as low accuracy or in the range of low to good accuracy for the total and subscale scores. Sensitivity was high for the total score and visual stimulation subscale, and the specificity was better for the standing and movement subscales.

This study aimed to analyze the psychometric properties of the Niigata PPPD questionnaire for use in Western vestibular patients. The study was necessary for two fundamental reasons. First, there is an absence of specific instruments to determine the impact of the disease in patients with PPPD. Second, the psychometric properties of this instrument have never been studied, even in the original culture in which the questionnaire was validated. Our study provides NPQ data that have never been analyzed, such as concurrent validity with other measures of interest, and reproducibility with error of measurement and minimum detectable change.

In our study, the factorial analysis reported a unifactorial structure for the NPQ. These results are contradictory to those obtained by Chichiro Yagi et al. [29], which reported a structure in three factors that could be defined as a visual factor, Active-Motion Factor and Passive Motion/Standing Factor. The differences in the results of the two studies are surely due to the composition of the sample, since our factorial analysis was carried out in a population of different types of vertigo while Chichiro Yagi et al. had only patients with PPPDs. Future studies should investigate the structure of the NPQ in an adequate population of western subjects with PPPD.

Regarding the internal consistency analysis, our data are consistent with the original validation study that showed a Cronbach's alpha of 0.91 [11], similar to the value of 0.938 found by us. In both cases, it can be interpreted as the presence of a certain redundancy of the items taken globally [23], but whose effect is not observed in the subscales that present Cronbach's alpha values below 0.90 in both the Japanese version and ours.

With regard to diagnostic validity, our results are slightly worse than those of the original study [11]. Our findings suggest that the NPQ is very good at measuring the severity of the symptoms; however, its results as a PPPD diagnostic tool are poorer, which is secondary since it is not its main purpose. The reason for this observation could be the punctuation and the composition of the “control” group to compare, since our control group has more cases of migraine and vestibular migraine that have usually presented a high NPQ score and its subscales in a similar way than PPPD subjects. Nonetheless, Kitazawa et al. noted that the NPQ is the most useful tool for the diagnosis of PPPD when compared with the video head impulse test, posturography, rotatory chair test, the DHI, bithermal caloric test and cervical and ocular vestibular-evoked myogenic potential assessment [30].

Concerning concurrent validity, the NPQ showed a strong correlation with the ABC-16, ABC-6, VVAS and DHI, which are the main instruments used in the assessment of vestibular disorders. At this point and considering the high correlation between the NPQ and the CSI as well as the composition of our sample, we hypothesized that PPPD could follow a similar model to other central sensitization disorders. In this way, PPPD may be considered and treated as a functional disorder.

To our knowledge, since the publication of the scale in 2019, several studies have used the NPQ in the assessment of PPPD in Japanese patients. In addition to the study of Kitazawa [30] previously mentioned, Yamaguchi et al. used the NPQ to assess the effects of a treatment based on virtual reality in participants with PPPD, determining that participants improved their symptoms [31]. Similarly, Miwa and Kanemaru observed 15.5 points of improvement in the NPQ using traditional medicine in PPPD patients [32], which is greater than the MDC obtained by our study. Thus, these authors and future authors would benefit from our study, as we calculated the MDC for the NPQ, allowing a better and more precise interpretation of the results of every study that uses the NPQ. The same could be applied to the study of Eldøen et al., who assessed a website with vestibular rehabilitation videos for PPPD patients [33], and the study of Yagi et al., who considered gaze instability in PPPD [34]. Other studies used the NPQ to assess the relation between the presence of isolated otolith dysfunction and PPPD [35] or the presence of exacerbating factors for PPPD patients [36].

Regarding the use of the scale in clinical or research settings, our data show that the NPQ can be used both based on its overall score and the value of each subscale, which provides more specific information on the impact of the disease in different capacities of each patient. Although this impression is not justified by the factorial analysis that reports a one-dimensional item structure, it can be supported by the fact that the internal consistency analysis offers optimum values for each subscale. Additionally, the strong correlation with other scales that measure dysfunction due to dizziness or confidence in balance should give us the necessary confidence to use the NPQ as a specific instrument when working with subjects with PPPD.

Another key aspect of our study is obtaining the SEM and MCD values that report the magnitude of the change that clinicians or researchers can interpret as a simple measurement error or as a substantial change in the patient's state of health. Finally, our analysis of ROC Curves does not justify the use of the NPQ as a screening instrument, since poor predictive values have been obtained, so we recommend the diagnosis of PPPD based on the criteria established up to now.

This study has several limitations. First, the validity of our results can be applied to the European population that speaks Spanish without being properly extrapolated to other Western languages. Although the data have been extracted from a sufficiently diverse population of vestibular patients and with a valid total sample to obtain statistical results, the total number of subjects has been low for some categories, which prevents drawing conclusions for specific pathologies beyond PPPD. Finally, although our study analyses psychometric properties that had not been evaluated in the original version, some additional properties remain unanalysed, such as sensitivity to change or validity to discriminate between different vestibular pathologies.

In view of the results of our work, it can be concluded that the NPQ is a feasible instrument for use in the Western culture population. The NPQ presents a unifactorial structure with a high internal consistency that could indicate a certain redundancy of the items. However, the three subscales that make up the NPQ present good internal consistency but without redundancies between items. The test–retest reliability of the overall score and of the subscales can be considered substantial except for the standing subscale, which can be considered moderate. This reliability provides a measurement error of 6 points and a minimum detectable change of 13 points for the overall score. Both the global score and the NPQ subscales have strong concurrent validity with the magnitude of dizziness measured with the DHI and its subscales, balance confidence measured with the ABC-16 and ABC-6, with the central sensitization measure CSI, and with the vertigo visual analog scale. However, the correlation was moderate with the MCS of the SF-12 and negligible with falls in the last 12 months and with the PCS of the SF-12. The overall score and the NPQ subscales have low accuracy in discriminating between subjects with or without PPPD, possibly because the comparison group was made up of very diverse pathologies that correspond to high scores on the NPQ. Future research should analyze the values of the NPQ and its subscales in different subgroups of patients with vestibular and neurological pathologies. The sensitivity of this instrument to detect significant changes in the health of patients should also be analyzed.