Development and Validation of Nasal Polyposis Quality of Life Questionnaire (NPQ)

BACKGROUND: To date, no disease-specic tool is available to assess the impact of Chronic Rhinosinusitis with Nasal Polyps (CRSwNP) on Health Related Quality of Life (HRQoL). Therefore, the purpose of this study was to develop and validate a questionnaire speci�cally designed to this aim: the Nasal Polyposis Quality of Life questionnaire –NPQ. METHODS: According to the current guidelines, the development and validation of the NPQ occurred in two separate steps involving different groups of patients. RESULTS: In the development process of NPQ an initial list of items of 40 items was given to 60 patients with CRSwNP; the 27 most signi�cant items were selected and converted into questions. The validation procedure involved 107 patients (mean age 52.9±12.4). NPQ revealed a ve-dimensional structure and high levels of internal consistency (Cronbach's alpha 0.95). Convergent validity (Spearman’ coe�cient r=0.75; p< 0.01), discriminant validity (sensitivity to VAS score), reliability in a sample of patients with a stable health status (Interclass Coe�cient 0.882) were satisfactory. Responsiveness to clinical changes was accomplished. The minimal important difference was 7. CONCLUSIONS: NPQ is the �rst questionnaire for the assessment of HRQoL in CRSwNP. Our results provide that the new tool is valid, reliable, and sensitive to individual changes.


Background
Chronic Rhinosinusitis with Nasal Polyps (CRSwNP) is a chronic in ammatory disease of the paranasal sinuses affecting 2 -4% of the general population 1 .It is the most severe subtype of CRS, characterized by symptoms often lasting for many years.Management of CRSwNP is di cult and recurrences are frequent, despite medical treatment and surgery approaches.As a consequence, CRSwNP has a considerable impact on' health related quality of life (HRQoL).This expression refers to the impact of an illness and its therapy upon a patient, as perceived by the patient himself 2,3 .The burden of troublesome symptoms (nasal blockage, loss of smell, rhinorrhea, and sneezing), the presence of comorbid diseases (chronic rhinosinusitis, asthma, aspirin sensitivity), the necessity of long term medical therapies, the need of surgical treatments, the changes to habits and lifestyle, all negatively impact physical, emotional and social aspects of daily life.
Despite the literature in this eld is not rich, available data con rm the clinical ndings 4 .Some studies explored the subjective burden of CRSwNP by means of the Short Form (36) Health Survey (SF-36) 5 , a generic measure that permit to assess health status in patients and healthy subject.Compared to general population, patients with CRSwNP had worse scores in all SF-36 domains except for physical functioning 6 .The disease burden has been detected also comparing CRSwNP with other chronic diseases, such as obstructive pulmonary disease 7 , asthma 8 and coronary artery disease 9 .No correlation was found between SF-36 scores and age, gender, nasal symptoms, CT scan, and polyp size 10 .HRQoL has been also assessed by mean of the Sinonasal Outcome Test (SNOT-22) 11 a specialityspeci c questionnaire that covers a broad range of rhinologic and general health issues.This widely-used tool has is not speci c for the phenotype with NP and for its characteristics has been used to assess the presence and the severity of sino-nasal disorders in clinical conditions really different from CRSwNP: smell dysfunction 12 , sino-nasal symptoms in cystic brosis 13 , allergic rhinitis 14 , sleep apnea 15 , chronic obstructive pulmonary disease (COPD) 16 , hereditary haemorrhagic telangiectasia 17 , Wegener's granulomatosis 18 .
The need of a speci c questionnaire to assess HRQoL in patients seems to be justi ed by several reasons: -a speci c questionnaire that encompasses all relevant aspects of HRQoL in CRSwNP does not exist; -the use of both generic and speci c tools to assess HRQoL is recommended 2 ; -the use of generic or speciality-speci c instruments is insu cient in capturing the impact of CRSwNP on patient's life and the changes of HRQoL.
The aim of the study was to develop and validate a speci c questionnaire to assess HRQoL in patients affected by CRSwNP.

Methods
The development and validation of the new questionnaire occurred in two separate steps involving different groups of patients.The method used for the two phases is described in detail below.
Consecutive patients who visited the Otorhinolaryngology and Personalized Medicine, Asthma and Allergy units from Istituto Clinico Humanitas between September 2018 and May 2020, were invited to participate in the study.
The Ethics Committee of the Humanitas University (Milan) approved the study protocol (approval no.P.R. 1920).The protocol complies with the general principles of Good Clinical Practice and the Declaration of Helsinki as amended in Edinburgh in 2000.Participation was voluntary and anonymous, and informed consent was obtained from all patients before study entry.
The inclusion criteria were as follows: con rmed diagnosis of CRSwNP; age ≥ 18 years; comprehension of spoken and written Italian language; availability and willingness to participate in the study.
Participants were excluded in case of the presence of other ear-nose-throat disorders.

Development process
In order to make certain that the questionnaire included items appropriate and relevant for CRSwNP patients, items generation and selection was conducted on the basis of current guidelines 19-21 : -Item generation.
The rst step had the aim to collect potentially relevant and troublesome problems related to CRSwNP on the basis of the following sources: (i) literature review of the available HRQL questionnaires used with CRSwNP patients; (ii) round-tables with ENT specialists and pulmonologists; (iii) unstructured interviews to 10 adult outpatients with CRSwNP.This resultant list included practical, emotional, social and physical aspects of daily life that could be in uenced by CRSwNP.
The second step was comprised of an item importance ranking, in order to identify the most relevant problems related to CRSwNP.The questions found during the item generation procedure, were randomly listed and administered to patients who were asked to indicate: a) which of the items they experienced as consequence of CRSwNP; the response options were yes/no; b) how relevant each of the identi ed items was, by a 5-points response option, indicating the degree of importance related to each item (1 = not important, 4 = very important) In this rst phase, a sample of 60 consecutive outpatients with CRSwNP has been accrued during a 2month period.On the basis of collected data we calculated: 1. the percentage of patients who identi ed each item as a consequence of CRSwNP (frequency range: 0-100); 2. the mean importance attributed to each item (range: 0-4); 3. the overall impact of each item, calculated as the product of the frequency and the mean importance divided by 100 (range: 0-4).
Selected items have been converted to questions where patients had to indicate how much they had been troubled by each problem during the last 2 weeks on a 5-point Likert scale (1 = not at all, 5 = very much).
This format of the questionnaire has been administered to a different group of patients for the validation process.Patients were selected using a convenience sampling method.The aim was to include almost 100 patients.The name of the new questionnaire is Nasal Polyposis Quality of Life (NPQ) questionnaire.

Validation process
Patients were assessed twice with a 4-week interval between visits.
At both visits, a physician collected a complete and accurate medical history reporting the ongoing therapy and patients lled in the NPQ along with the following tools: -Visual analogue scale (VAS): patients were asked to indicate on a horizontal line measuring 10 cm the degree of CRSwNP severity, giving a score from 0 to 10 (worse).The score obtained can be divided into mild (VAS 0-3), moderate (VAS 3-7) and severe (VAS > 7) 1 .
-The SNOT-22 (11) encompassess 22 items scored from 0 (meaning no problem reported) to 5 (as bad as it can be) giving a score to maximum 110 points; where, the higher the score the worse is the patient's QoL related to the disease.It has been adapted and validated in several languages and it is now available also in Italian 22 .
At Visit 2, patients lled the same questionnaires of the Visit 1 and a Global Rating Scale (GRS) to assess any change in health status.
The psychometric properties of the NPQ were tested as following: -Construct validity was evaluated by mean of factorial analysis; the principal component method with Varimax rotation was adopted.
-Convergent validity was calculated by Spearman correlations to examine the relationships between the new questionnaire and an established measure (SNOT-22).Convergent validity is con rmed with correlations ranging from 0.4 to 0.8.Two instruments are considered too similar if the correlation is 0.8 or more (the tested instrument has no added value) (23) .
-Discriminant validity was evaluated comparing patients according their VAS score by using ANOVA (Fischer's test) -Internal consistency was estimated using Chronbach's correlation coe cient on the extracted factors.
Measures with reliability of 0.50-0.70 or greater have been recommended for the purpose of comparing group 24 .
-Reliability was evaluated by means of the Intraclass Correlation Coe cient (ICC) in the subsample of patients with a stable health status (GRS = 0).An ICC of > 0.75 indicates excellent reproducibility while an ICC between 0.4 and 0.75 indicates a good reproducibility 24 .
-Responsiveness was assessed, analyzing the correlation between changes in the score of the new questionnaire and changes in GRS (GRS ≠ 0) and VAS by means of a non-parametric test (Spearman correlation coe cient).
-Clinical signi cance was explored by assessing the minimal important difference (MID).The receiver operating characteristics (ROC) curve method was applied 25 .The entire cohort for one dichotomization point (i.e., 'no change' vs 'any improvement or deterioration') was adopted.
The possible effect of age (Spearman's correlation coe cient), gender, smoking habits and comorbid asthma (Fisher's ANOVA) on patients' answers was also tested.The frequency distribution of the answers was calculated to evaluate whether patients used the entire answer scale and whether all possible scores were obtained.

Development process
Sixty patients completed the development-phase questionnaire of 40 items.Most of these patients (63.3%) were female, and the mean (Standard Deviation, SD) age was 41.4 (8.3) years, ranging from 18 to 74 years.
On the basis of patients' answers, items included in the questionnaire were those that scored highest in impact.Where an arbitrary cut-off value of 1.5 was used for impact, 13 items were excluded.Table 1 summarizes the results of this rst phase, indicating the items selected due to the total importance.Validation process 107 subjects were enrolled in the study.The mean age was 52.9 with a SD of 12.4; the majority were male (61.7%) and non-smoker (92.5%).
Regarding atopy (as at least one allergen sensitization via skin prick test), 54 (50.5%) were found positive.Acetyl salicylic acid (ASA) intolerance, meaning patients reporting some kind of respiratory symptoms upon aspirin or any nonsteroidal anti-in ammatory drugs (NSAIDS) intake, was found in 14 (13.1%)patients; 5 subjects out of 107 (4.7%) were affected by Samter's triad.
-Construct validity: the factorial analysis with eigenvalue > 1 extracted ve factors which explaine up to 66.97% of the total variance.Items belonging to each factor are listed in Table 2.
-Convergent validity: Spearman' correlations between NPQ scores and SNOT-22 were signi cant (r = 0.75; p < 0.01) -Discriminant validity: the group of patients with VAS > 7 had NPQ scores signi cantly higher than patients with VAS ≤ 7 (81.88-Reliability: Interclass Coe cient (ICC) was 0.882, exceeding the cut-off of 0.75 indicating an excellent test reliability. -Responsiveness: the assessment of a subsample of 44 patients reporting an improvement or deterioration in health status (GRS ≠ 0) demonstrate that a signi cant Spearman correlation between the variation of NPQ Total Score between the two visits and the change in VAS score (0.628 p< 0.001) and in GRS (-0.528 p < 0.001) (Fig. 1).

Discussion
HRQoL has become a crucial outcome in chronic conditions, allowing to capture the patient's perspective about disease and treatment.
The availability of generic and rhinologic-speci c questionnaires allowed to highlight that CRSwNP signi canltly affects patients HRQoL.However there is no speci c validated tool to assess HRQoL impairment of patients suffering from CRSwNP, that account approximately for 25-30% of CRS cases 26 .
Recently it has been shown that nasal polyposis might have a variable impact on HRQoL 27 and that patients with CRSwNP present a different HRQoL pro le compared to those with CRSsNP 28 .
To address this gap we developed and validated the rst disease speci c tool to detect HRQoL impairment in patients with CRSwNP, by following the established methodological guidelines and a recognized framework of questionnaire design [19][20][21] .The procedure we adopted provides evidences that the new instruments appropriately re ects HRQoL of patients suffering from CRSwNP.In fact, the development process guarantees that the item selection has been determined by the patients on the basis of their experience.
The new questionnaire consists of 27 items, that can be summed up to a total score and to ve factorial scores.As expected, a moderate, signi cant correlation was obtained between NPQ and SNOT-22.
Discriminant validity was demonstrated through the tool's ability to discriminate between groups de ned according to the VAS.
NPQ was shown to be an internally reliable tool as indicated by very high Cronbach α coe cients (0.95).
It was also a reliable questionnaire as supported by satisfactory ICC in stable patients (0.882).High responsiveness to changes were con rmed by a signi cant correlation between the change of NPQ.Total score between the two visits and the change in VAS score and in GRS.The ROC analysis indicates that 7 point is the smallest change that patients perceive as an improvement or deterioration The new questionnaire has several advantages: it is simple to complete and to score; it owns the necessary psychometric properties; the cutoff MID makes it easy to determine the clinical signi cance of the results and changes over time.Moreover, answers were not in uenced by socio-demographic characteristics, thus enabling the NPQ to be used regardless of the patient's sex, age and education.
Because of these characteristics, NPQ is appealing as an instrument to assess the patient experience of CRSwNP.It is also potentially useful to monitor the impact of both disease and treatment from the patient's perspective owing to its satisfactory responsiveness to changes.
Although we reached the primary aim of our study by providing evidence to support the validity, reliability, and responsiveness of NPQ, our ndings should be considered in the light of the following potential limitations.
First, the generalizability of the results should be limited because the sample was nonrandomized and the patients were enrolled in one specialistic center.Second, no objective measures of disease control and severity, besides patient's reported outcomes, were adopted to determine the reliability and the sensitivity to change.Third, the acceptability of the new tool for both patients and physicians has not been evaluated.However these limitations may be faced through further studies.

Conclusions
In conclusion, NPQ is the rst questionnaire for the assessment of HRQoL in CRSwNP.It is valid, reliable, and sensitive to individual changes.It is able to detect the speci c burden of CRSwNP on HRQoL, This tool should yield data to improve our ability to effectively monitor the burden of disease and treatment on patients with CRSwNP.

Table 2
Factors identi ed using principal components analysis on full data set of the ROC analyses are presented in Table3.A 7-point change in RAPP maximizes sensitivity, speci city, and the number of individuals correctly classi ed., identifying the MID.
Bold typeface shows the component upon which each item loaded most highly: 1 -Daily life impact; 2 -Mouth problems; 3 -Embarrassment; 3 -Treatment impact; 4 -Loss of smell results