Measurement properties of the Polish version of the Cystic Fibrosis Questionnaire Revised 14+ in the adult population

Abstract

Measuring the quality of life in patients with cystic fibrosis is an important element of the patient care process. Many tools have been created for this measurement among adults. One of them is the Cystic Fibrosis Questionnaire-Revised 14+ (CFQ-R 14+). Its measurement properties have not been comprehensively assessed in the population of Polish adults. The aim of the study is to verify the construct validity, including structural and criterion validity, as well as internal consistency, of the Polish version of the CFQ-R 14+ in the population of adults with cystic fibrosis. We conducted a cross-sectional survey among adults with cystic fibrosis. After preparing the database, we performed a confirmatory factor analysis (CFA) followed by exploratory factor analysis (EFA) using the parallel analysis method principal axis factoring with Oblimin rotation. Intercorrelations of questionnaire factors and the occurrence of relationships among items for the general scale results was checked. We also presented basic descriptive statistics (mean, median, standard deviation, skewness, kurtosis, minimum and maximum values). The analyses included responses from 220 adult patients. CFA results did not show adequate goodness of fit (χ²(1025) = 2112.35, p < 0.001; CFI = 0.831; TLI = 0.814; RMSEA = 0.069; SRMR = 0.074). After EFA, 6 factors were extracted, considering 40 out of 50 questions of the CFQ-R 14+. CFQ-R 14+ may be useful in assessing the quality of life of Polish adult patients with cystic fibrosis. Our analysis demonstrates that the optimal factor structure of the tool in this population includes 6 scales.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-94184-x.

Introduction

Measuring the quality of life in people with cystic fibrosis is one of the key aspects in the ongoing assessment of the patient’s condition^1,2. Cystic fibrosis constitutes a major challenge for patients in terms of symptoms related to the respiratory and digestive systems^3,4. Depending on the severity of the course and the effectiveness of the treatment, the disease may also cause difficulties in performing everyday activities, including making it difficult to study or work, and limiting social contacts^5–7.

Several tools have been created to measure the quality of life in people with cystic fibrosis. Some of them are intended for use only in children, while others are designed for adults^8–10. In turn, the CFQ-R 14+ tool was created for people over 14 years of age, including adults. It was created in the early 2000s and has been translated into many different languages^10–12. It is still widely used in everyday practice, as well as in various types of analyses.

So far, no comprehensive study evaluating the measurement properties of the tool in the adult population has been conducted in Poland. This is a significant limitation, causing uncertainty as to whether the tool in the Polish version meets its basic objectives and if it demonstrates the quality of life of patients in the assumed scales. The case of the absence of a comprehensive analysis of the CFQ-R 14+, including the examination of structural validity is not an isolated one. The number of scientific publications from other countries analyzing issues related to the measurement properties of this tool is very limited^9,10.

The analysis of measurement properties of a tool translated into another language is very important. All tools are created with a certain, specific population in mind, at a specific time and with access to health care services at a specific moment. Intercultural differences may result in different perceptions of questions, therefore the evaluation of the tool in terms of maintaining the appropriate applicability of the questions is also an important aspect. The development of medicine has led to the implementation of new, effective, and safe therapies that significantly improve the condition of patients. The modulator therapy introduced for patients has revolutionized the treatment of cystic fibrosis and resulted in a huge increase in the quality of life^13–16. The CFQ-R 14+ was created long before the implementation of this medical technology. As a result, there may be several threats that reduce the usefulness of the tool, including the so-called ceiling effect, which means that the possible results would exceed the measurement capabilities of the tool.

Our study aims to determine the measurement properties (construct validity, including structural and criterion validity, as well as internal consistency), of the Polish version of the CFQ-R 14+ in the population of adults with cystic fibrosis.

Methods

Characteristics of the study tool

In our study, we applied the existing and translated version of the questionnaire, which was used in the 2009 Polish study¹⁷. We have also obtained consents to conduct the study from the authors of both the Polish and English versions of the tool. The CFQ-R 14+ consists of 50 questions, one of which is not included in any domain (conditional question related to the question preceding it) and several metric questions. The questions were originally divided into 12 domains: physical (8i), role (4i), vitality (4i), health perceptions (3i), emotion (5i), social (6i), body (3i), eat (3i), treatment burden (3i), weight (1i), respiratory (6i), and digestion (3i). The results for each scale are calculated separately according to the formula “(sum of responses − minimal possible sum)/ (maximum possible sum − minimal possible sum) × 100 = result”. For each domain, the answers are standardized to range from 0 to 100, where higher values ​​indicate a better level of quality of life. The tool uses a 4-point Likert scale (different categories of answers depending on the question asked) with a score from 1 to 4 points. Some questions are reverse scored, and their score must be changed to calculate the result. The original and modified distribution of factors, are available in the supplementary material.

In addition, to deepen the analyses related to the measurement properties of the CFQ-R, we decided to use the Patient Health Questionnaire-9 (PHQ-9) and General Anxiety Disorder-7 (GAD-7) tools. The PHQ-9 is a screening tool for depression. It consists of 9 items that can be answered on a 4-point Likert scale¹⁸. The GAD-7 is a screening tool used to determine feelings related to generalized anxiety disorder. It consists of 7 items that can also be answered on a 4-point Likert scale¹⁹. Both tools are recommended by the European Cystic Fibrosis Society for patients with cystic fibrosis²⁰.

Participants and data collection

We conducted a cross-sectional survey among adults with cystic fibrosis. The criteria for participation in the study included, in addition to age, fluent knowledge of Polish, enabling understanding and completing the questionnaire. We used Google forms to conduct the study from January 13 to May 30, 2024. We selected two main channels to reach patients, considering the small number of adults living in Poland with cystic fibrosis. Firstly, we contacted large patient treatment centers, asking them to send us the e-mail addresses of patients, who met the inclusion criteria or to directly forward links to the survey to their patients. There are several large, specialized cystic fibrosis treatment centers in Poland. Six centers (located in Warsaw, Gdańsk, Poznań, Karpacz, Białystok, Łódź) have granted consent for the study to be conducted. Due to personal data protection regulations, the facilities did not agree to provide us with patient email addresses. However, some agreed to provide information about the study directly to their patients (National Tuberculosis and Lung Diseases Research Institute in Warsaw, Polanki Hospital in Gdańsk, University Clinical Hospital in Białystok, Karpacz Medical Center, Voivodeship Multi-Specialist Center for Oncology and Traumatology in Łódź). The facilities received a previously prepared email message, ready-made to be sent for patients treated at a given center. Employees of these centers (medical staff or medical assistants/secretaries delegated by the facility’s management) sent such emails to patients treated at the center. When this channel of reaching patients was exhausted, we contacted representatives of 5 patient organizations associating adults with cystic fibrosis. Three have decided to participate in the study—MUKOHELP Foundation for Cystic Fibrosis Patients, Mukobohaterowie Foundation, Polish Cystic Fibrosis Association. As in the case of treatment centers, we asked them to distribute the link to the survey to patients, with whom the patient organization had contacts. At the same time, we asked both treatment centers and patient organizations to indicate the number of people to whom the questionnaire had been sent. According to the information obtained, the CF treatment centers distributed the request to fill out the questionnaires to 110 people, while patient organizations sent it to 340 people. To avoid including duplicates of the survey by the same people in the analysis, we decided to introduce additional security questions into the questionnaire (location of the treatment center, treatment used). Verification was also based on the metric questions from the questionnaire (date of birth, marital status, education). In the event of potential identification of identical answers, the record was to be removed from the database. We opted to select patients for the study in such a way in order to ensure the widest possible scope of the study, and to allow an analysis of factors, keeping in mind that the results would largely depend on the size of the population (the larger the better). The applied mixed method of patient selection allowed us to obtain by far the largest number of people with cystic fibrosis from studies conducted in Poland to date and to minimize the risk of selection bias.

The electronic questionnaire that we constructed, contained all the elements of the paper-based questionnaire. The order, layout of possible answers and all other elements were identical to the paper version. Additionally, at the very beginning of the form, information about the study was presented and organizational aspects related to the consent to participate in the study were outlined. The description also included information that in case of any doubts regarding the content of individual questions or the entire study, the study coordinator may be contacted via the provided e-mail address. The main part of the questionnaire was displayed continuously, without going to separate sections, which allowed for free return to previously completed questions and editing previously marked answers.

The exact number of adults suffering from cystic fibrosis in Poland is not known. According to the data based on healthcare services reported to the Polish National Health Fund in 2023, the number of adult patients with CF in Poland was estimated at around 1,000. As already mentioned, we wanted to include as many patients as possible in the study, and we assumed the minimum number at 200 persons (approximately 20% of the adult population suffering from cystic fibrosis in Poland).

Statistical analysis

After preparing the database, we performed a confirmatory factor analysis (CFA) for the 11-factor model of the CFQ-R 14+ tool (the Weight scale was treated separately because it consisted of 1 item). The following parameters were cited: chi2 test results, root mean square error of approximation (RMSEA), comparative fit index (CFI), standardized root mean square residual (SRMR), and Tucker–Lewis index (TLI).

Due to the poor fit of the model, we decided to conduct exploratory factor analysis (EFA) using the principal axis factoring (PAF) with the Oblimin rotation method. We performed the Kaiser-Mayer-Olkin (KMO) and Bartlett tests to check the adequacy of the factor analysis model. We based further analyses on the newly created scales. Intercorrelations of the factors of the modified CFQ-R tool and the occurrence of relationships among items for the general scale results was checked. The internal consistency was established by calculating Cronbach’s alpha coefficients. Basic descriptive statistics (mean, median, standard deviation, skewness, kurtosis, minimum and maximum value) were also presented, as well as floor and ceiling effects.

To check criterion validity, we conducted correlation analyses of the modified tool, using a nonparametric test (Spearman’s rho) or parametric test (Pearson’s r) with the PHQ-9 and GAD-7 tools.

We performed factor analyses utilizing the Jamovi program in version 2.3. Then we applied statistical analyses using the IBM SPSS Statistics package, version 29. It was used to analyze basic descriptive statistics and correlation analysis. The statistical significance level was set at α = 0.05.

Ethical considerations

Each participant in the study received written information about the purpose and method of conducting the study before completing the survey. It was also indicated that the information provided was anonymous and would be used only for scientific purposes, and that completing the surveys meant consent to the use of anonymized data in scientific work. Therefore, informed consent was obtained from all patients. After reading the information about the study, it was necessary to accept the procedure and express consent to the study by clicking the appropriate answer in the online form. Failure to express consent resulted in the end of the possibility of moving on to the next part of the form and the end of participation in the study.

This study was performed in line with the principles of the Declaration of Helsinki. This study was supported and approved by the Center of Postgraduate Medical Education of Warsaw. The Bioethics Committee operating at the Medical University of Warsaw committee also approved the study procedures and indicated that the study is in accordance with the principles of research ethics (AKBE/281/2023).

Results

Sample

Ultimately, 450 people were invited to participate in the study. We received responses from 231 people. We then verified responses for their content (patient age was checked, and identical records were eliminated). Finally, we included responses from 220 respondents with cystic fibrosis for analysis. Details regarding the study population are presented in Table 1.

Table 1.

Characteristics of the study population.

Variable	n	%
Gender
Female	130	59.1
Male	90	40.9
Age group
Up to 24 years	65	29.5
25–29 years	49	22.3
30–34 years	47	21.4
35 years and over	59	26.8
Marital status
Married	72	32.7
Divorced	5	2.3
Informal relationship	6	2.7
Single	137	62.3
Education
Primary school	10	4.6
Middle school	6	2.7
High school	103	46.8
University	101	45.9
Place of residence
Village	58	26.4
Town up to 50 K	39	17.7
Town 50-100 K	24	10.9
City 100-250 K	30	13.6
City above 250 K	69	31.4
Occupational situation
Attending school	25	11.4
Home schooling	7	3.2
Looking for a job	16	7.3
On site work	83	37.7
Hybrid on site and online work	24	10.9
Online work	19	8.6
Not attending school or work for health reasons	29	13.2
Not working—for other reasons	17	7.7

Confirmatory factor analysis (CFA)

We performed CFA to test the 11-factor CFQ-R model. Although Cronbach’s alpha is highly satisfactory for many of the indicators, and for each of them is greater than the acceptable threshold of 0.6, the model results do not indicate adequate goodness of fit: χ²(1025) = 2112.35, p < 0.001; CFI = 0.831; TLI = 0.814; RMSEA = 0.069; SRMR = 0.074. The Social scale is problematic, for which the Cronbach’s alpha is lower than the others—as are the factor loadings and total item correlations (Table 2). Before recoding the items, we presented descriptive statistics for the items.

Table 2.

Item descriptive statistics and CFA results for the CFQ-R.

Scale	Item	M	SD	Sk	Kurt	Item loadings	Item-Total Correlation	Cronbach’s Alpha
Physical	CFQ-R1	2.86	0.98	− 0.56	− 0.64	0.802	0.785	0.94
	CFQ-R2	3.30	0.89	− 1.06	0.16	0.873	0.831
	CFQ-R3	3.25	0.90	− 1.05	0.23	0.863	0.826
	CFQ-R4	3.38	0.84	− 1.31	1.04	0.871	0.825
	CFQ-R5	3.12	0.95	− 0.78	− 0.43	0.913	0.877
	CFQ-R13	1.59	0.73	1.18	1.18	− 0.792	0.749
	CFQ-R19	2.77	0.93	− 0.20	− 0.90	0.742	0.746
	CFQ-R20	2.69	1.08	− 0.13	− 1.30	0.740	0.742
Vitality	CFQ-R6	2.13	0.77	0.19	− 0.49	0.724	0.632	0.86
	CFQ-R9	2.47	0.80	− 0.15	− 0.47	− 0.842	0.783
	CFQ-R10	2.63	0.76	− 0.19	− 0.26	0.734	0.687
	CFQ-R11	2.72	0.85	− 0.34	− 0.42	− 0.822	0.714
Emotion	CFQ-R7	2.68	0.79	− 0.28	− 0.26	0.789	0.669	0.83
	CFQ-R8	3.30	0.89	− 1.05	0.10	0.835	0.702
	CFQ-R12	2.83	0.86	− 0.47	− 0.32	0.875	0.747
	CFQ-R31	2.74	1.09	− 0.29	− 1.22	0.666	0.655
	CFQ-R33	2.54	1.07	0.06	− 1.24	0.431	0.408
Eat	CFQ-R14	3.66	0.56	− 1.58	2.38	0.809	0.763	0.86
	CFQ-R21	3.35	0.91	− 1.11	0.04	0.877	0.784
	CFQ-R50	3.54	0.70	− 1.43	1.45	0.851	0.775
Treatment Burden	CFQ-R15	2.00	0.77	0.48	− 0.03	0.825	0.697	0.79
	CFQ-R16	2.94	0.81	− 0.51	− 0.08	− 0.780	0.644
	CFQ-R17	1.88	0.79	0.61	− 0.12	0.656	0.566
Health Perceptions	CFQ-R18	2.25	0.83	0.42	− 0.25	0.783	0.540	0.75
	CFQ-R32	2.47	1.00	0.16	− 1.05	0.669	0.602
	CFQ-R34	2.03	0.89	0.64	− 0.26	0.637	0.602
Social	CFQ-R22	3.00	1.04	− 0.56	− 1.00	0.665	0.427	0.64
	CFQ-R23	2.08	1.11	0.56	− 1.07	− 0.237	0.275
	CFQ-R27	3.22	0.86	− 0.82	− 0.22	0.435	0.343
	CFQ-R28	2.11	0.99	0.52	− 0.78	− 0.474	0.389
	CFQ-R29	2.64	1.05	− 0.07	− 1.23	− 0.351	0.420
	CFQ-R30	1.97	1.08	0.73	− 0.81	0.479	0.379
Body	CFQ-R24	3.09	1.03	− 0.70	− 0.83	0.479	0.378	0.66
	CFQ-R25	2.93	1.06	− 0.44	− 1.14	0.844	0.618
	CFQ-R26	2.69	1.04	− 0.09	− 1.22	0.616	0.424
Role	CFQ-R35	1.77	0.80	0.71	− 0.27	0.756	0.493	0.75
	CFQ-R36	3.15	0.70	− 0.38	− 0.30	− 0.590	0.526
	CFQ-R37	2.70	0.85	− 0.21	− 0.54	− 0.539	0.503
	CFQ-R38	3.17	0.78	− 0.65	− 0.06	− 0.681	0.659
Respiratory	CFQ-R40	2.95	0.94	− 0.46	− 0.78	0.766	0.705	0.89
	CFQ-R41	2.90	0.94	− 0.58	− 0.50	0.806	0.783
	CFQ-R42	2.95	0.95	− 0.60	− 0.53	0.740	0.723
	CFQ-R44	3.40	0.74	− 1.14	0.91	0.735	0.676
	CFQ-R45	3.46	0.71	− 1.09	0.39	0.788	0.705
	CFQ-R46	3.52	0.75	− 1.46	1.30	0.704	0.643
Digestion	CFQ-R47	2.88	0.91	− 0.43	− 0.62	0.609	0.494	0.71
	CFQ-R48	3.39	0.73	− 1.10	1.01	0.737	0.589
	CFQ-R49	3.33	0.74	− 0.88	0.34	0.701	0.524

M, mean; SD, standard deviation; Sk, skewness; Kurt, kurtosis.

Exploratory factor analysis (EFA)

We performed EFA due to the poor fit of the originally assumed model. The assumptions of EFA were met based on the statistically significant result of Bartlett’s sphericity test: χ²(780) = 6020.27; p < 0.001 and the appropriate KMO values (0.91). No item was included for which skewness exceeded |2| and kurtosis |7|²¹. We included all items, for which the loading was not less than 0.32²². Additionally, a threshold of 0.4 for commonalities and 0.3 for discriminatory power²³ was considered. The results of the analysis are presented in Table 3.

Table 3.

PAF results with Oblimin rotation for CFQ-R along with the reliability of the tool.

Item	Scale						Item rest correlation
	Physical	V&E	TB&HP	Respiratory	Eat, Weight, Body	Digestion
CFQ-R1	0.753	0.068	0.000	0.096	0.004	− 0.074	0.777
CFQ-R2	0.853	− 0.050	− 0.012	− 0.014	0.066	0.035	0.822
CFQ-R3	0.785	0.068	0.000	0.020	0.053	0.070	0.826
CFQ-R4	0.903	− 0.020	0.013	− 0.019	− 0.052	− 0.004	0.802
CFQ-R5	0.931	0.008	0.032	0.052	− 0.077	− 0.004	0.857
CFQ-R6	− 0.171	− 0.566	0.153	− 0.105	0.005	0.084	0.660
CFQ-R7	0.022	0.793	0.108	− 0.021	0.037	0.025	0.710
CFQ-R8	− 0.017	0.718	0.000	0.059	0.116	0.072	0.744
CFQ-R9	0.152	0.638	− 0.108	0.021	− 0.031	0.031	0.724
CFQ-R10	− 0.094	− 0.543	0.245	0.035	− 0.020	0.062	0.641
CFQ-R11	0.177	0.560	− 0.094	0.117	− 0.013	0.116	0.685
CFQ-R12	− 0.099	0.915	0.050	0.013	− 0.005	− 0.002	0.781
CFQ-R13	− 0.736	0.088	0.098	0.073	− 0.150	− 0.053	0.758
CFQ-R14	0.053	0.107	− 0.055	− 0.150	0.704	0.052	0.581
CFQ-R15	− 0.028	0.093	0.620	− 0.161	− 0.128	− 0.153	0.657
CFQ-R16	0.092	− 0.055	− 0.526	0.199	0.034	0.013	0.571
CFQ-R17	0.092	− 0.064	0.626	− 0.107	0.113	− 0.133	0.563
CFQ-R18	− 0.483	− 0.169	0.234	− 0.081	− 0.041	− 0.029	0.699
CFQ-R19	0.669	0.109	− 0.013	0.021	0.101	− 0.026	0.764
CFQ-R20	0.677	0.042	− 0.091	0.035	0.048	− 0.075	0.761
CFQ-R21	0.076	0.057	− 0.036	− 0.063	0.794	0.065	0.628
CFQ-R22	0.376	− 0.069	− 0.266	0.023	0.315	0.069	0.627
CFQ-R24	− 0.104	− 0.062	0.091	0.259	0.558	− 0.169	0.542
CFQ-R25	0.036	0.146	− 0.176	0.033	0.391	− 0.133	0.482
CFQ-R28	− 0.045	− 0.235	0.486	0.071	− 0.055	0.173	0.474
CFQ-R31	− 0.112	0.510	− 0.227	− 0.134	0.176	0.109	0.576
CFQ-R32	− 0.146	− 0.102	0.584	− 0.078	0.136	− 0.032	0.639
CFQ-R34	− 0.133	− 0.084	0.637	0.070	− 0.104	0.125	0.641
CFQ-R37	− 0.080	0.014	− 0.456	0.152	0.065	0.166	0.479
CFQ-R39	− 0.066	− 0.068	0.048	0.365	0.575	− 0.071	0.592
CFQ-R40	0.225	0.183	0.110	0.570	0.045	0.136	0.705
CFQ-R41	0.012	0.052	− 0.129	0.784	0.023	− 0.002	0.783
CFQ-R42	− 0.028	− 0.004	− 0.088	0.819	0.013	− 0.057	0.723
CFQ-R44	0.311	− 0.009	0.019	0.478	− 0.064	0.204	0.676
CFQ-R45	0.372	0.030	− 0.025	0.423	0.037	0.197	0.705
CFQ-R46	0.236	0.050	− 0.116	0.419	0.076	0.060	0.643
CFQ-R47	− 0.114	0.119	− 0.098	0.060	− 0.114	0.603	0.494
CFQ-R48	− 0.050	0.024	0.005	− 0.058	0.076	0.728	0.589
CFQ-R49	0.142	− 0.008	0.005	0.015	0.071	0.618	0.524
CFQ-R50	0.132	0.121	0.027	0.094	0.634	0.114	0.633
Eigenvalue	6.91	4.42	3.63	3.21	3.03	1.83	–
Explained variance	17.25%	11.05%	8.41%	8.03%	7.57%	4.58%	–
Alfa Cronbacha	0.94	0.90	0.83	0.89	0.80	0.71	–

V&E, Vitality & Emotion; TB & HP, Treatment Burden & Health Perceptions.

The model extracted a 6-factor approach, explaining a total of 56.89% of the variance. The first factor explained 17.25%, the second—11.05%, the third—8.41%, the fourth—8.03%, the fifth—7.57%, and the last 4.58%. Each of the factors had an alpha value greater than 0.7, which indicates satisfactory reliability of the scales.

In the next part, we checked how the factors were related to each other (Table 4).

Table 4.

Results of intercorrelation analyses of the modified CFQ-R tool factors.

	Scale	1	2	3	4	5	6
1	Physical	–
2	Vitality & Emotion	0.47***	–
3	Treatment Burden & Health Perceptions	0.55***	0.55***	–
4	Repiratory	0.64***	0.40***	0.51***	–
5	Eat, Weight, Bodya	0.40***	0.42***	0.35***	0.37***	–
6	Digestiona	0.27***	0.34***	0.25***	0.30***	0.12	–

^aAnalysis using nonparametric test (Spearman’s rho), otherwise Pearson’s r correlation analysis.

*** p < 0.001; **p < 0.01; *p < 0.05.

Significant values are in bold.

We observed the strongest correlations among the Physical, Vitality & Emotion, and Respiratory scales and the Treatment Burden & Health Perceptions scales. Furthermore, Respiratory is strongly correlated with and Treatment Burden & Health Perceptions. In turn, we observed the weakest correlations among Digestion and Physical, Treatment Burden & Health Perceptions and Respiratory. The results of these correlations indicate the appropriate structure of the tool, and the correct selection of rotation. Next, we checked the occurrence of correlations between items for the overall scale results (Table 5).

Table 5.

Results of correlation analyses of the modified CFQ-R items with the scales of the tool.

^aAnalysis using nonparametric test (Spearman’s rho), otherwise Pearson’s r correlation analysis.

V&E, Vitality & Emotion; TB & HP, Treatment Burden & Health Perceptions.

***p < 0.001; **p < 0.01; *p < 0.05.

Significant values are in bold.

The intra-scale item correlation analysis is strong or very strong for all factors. The Digestion scale is the weakest correlated with the items. The Physical, Vitality & Emotion, Treatment Burden & Health Perceptions, and Respiratory scale items are moderately correlated. This may indicate the formation of two groups of factors: one concerning clinical aspects (Physical, Vitality & Emotion, Treatment Burden & Health Perceptions and Respiratory scales) and the other concerning aspects related to nutrition and the digestive system (Eating, Weight, Body, and Digestion scales).

Basic descriptive statistics of the modified CFQ-R

In the next step, we checked the distributions of the factors of the modified CFQ-R. In addition, we evaluated the distributions of the PHQ-9 and GAD-7. In each case, the result of the Kolmogorov–Smirnov test was statistically significant, which indicates a violation of the assumption of a normal distribution. One value of skewness and kurtosis for the threshold |2| as acceptable for performing parametric tests was met for all variables²⁴. Nevertheless, outliers were noted within the Eat, Weight, Body, and Digestion scales, less than the third standard deviation (extremely low results). For this reason, a nonparametric test for the analyses of these two scales was applied. The exact values of all statistics are presented in Table 6.

Table 6.

Descriptive statistics for the modified CFQ-R scales and the PHQ-9 and GAD-7 scales.

Variable	M	Me	SD	Sk	Kurt	Min	Max	Floor	Ceiling
CFQ-R
Physical	68.44	76.67	24.95	− 0.81	− 0.09	0.00	100.00	1.4%	6.8%
Vitality & Emotion	58.26	60.42	21.85	− 0.40	− 0.34	4.17	100.00	0.0%	1.8%
Treatment Burden & Health Perceptions	62.62	61.90	20.43	− 0.34	− 0.44	9.52	100.00	2.3%	0.0%
Respiratory	73.23	77.78	22.50	− 0.70	− 0.47	11.11	100.00	0.0%	14.1%
Eat, Weight, Body	76.16	77.78	21.34	− 0.68	− 0.29	5.56	100.00	0.0%	22.7%
Digestion	73.28	77.78	21.09	− 0.80	0.55	0.00	100.00	0.5%	17.7%
PHQ-9	8.14	7.00	6.41	0.80	− 0.15	0.00	27.00	–	–
GAD-7	6.93	6.00	5.39	0.64	− 0.34	0.00	21.00	–	–

M, mean; Me, median; SD, standard deviation; Sk, skewness; Kurt, kurtosis; Min, minimal score; Max, maximum score.

No floor effect was noted. However, for the Eat, Weight, Body, and Digestion scales, there was a ceiling effect, exceeding the conventional limit of 15% for the maximum value²⁵. The Respiratory scale was also close to the limit, for which the distribution of the maximum value was 14.1%.

Criterion validity of the modified CFQ-R

To check the criterion validity of the tool, we performed correlation analysis with the PHQ-9 and GAD-7 tools (Table 7). We implemented a nonparametric analysis for the relationships with the Eat, Weight, Body, and Digestion scales, due to the presence of outliers, lower than the third standard deviation.

Table 7.

Results of correlation analyses between the CFQ-R scales and the PHQ-9 and GAD-7.

Scale	PHQ-9	GAD-7
Physical	− 0.42***	− 0.31***
Vitality & Emotion	− 0.78***	− 0.74***
Treatment Burden & Health Perceptions	− 0.46***	− 0.38***
Repiratory	− 0.33***	− 0.32***
Eat, Weight, Bodya	− 0.37***	− 0.26***
Digestiona	− 0.39***	− 0.37***

^aAnalysis using nonparametric test (Spearman’s rho), otherwise Pearson’s r correlation analysis.

***p < 0.001; **p < 0.01; *p < 0.05.

Significant values are in bold.

The analysis demonstrated statistically significant relationships between each CFQ-R scale and the patient’s health assessment and anxiety. Each of the observed correlations is negative, which indicates that with increasing anxiety and more symptoms of depression, the quality of life of people with CF decreases. The strongest relationships were noted with the Vitality & Emotion scale. The weakest relationships were reported for Eating, Weight, and Body (in the case of anxiety). The rest of the observed correlations were moderate.

Discussion

In this study, we analyzed the measurement properties of the Polish version of the CFQ-R 14+ in the adult population. To date, no study has been published that would analyze the measurement properties of this questionnaire in such a large population of Poland’s residents. No study has conducted an in-depth factor analysis. As a result, a tool has been in use with no certainty regarding the correctness of the parameters measured.

Confirmatory factor analysis revealed a lack of agreement in terms of the assumed factors. This means that the original division of factors is not optimal for measuring the quality of life in the population of Polish adults suffering from cystic fibrosis. The most problematic scale turned out to be the Social scale, which is not surprising. Many studies conducted in other countries so far, have indicated low-scale reliability^16,26–28. The general impression regarding the original division of factors indicates too much fragmentation of factors and too many questions in the tool. Due to the unfavorable CFA result, we conducted EFA, which demonstrated the presence of 6 factors: Physical (10i), Vitality & Emotion (8i), Treatment Burden & Health Perceptions (7i), Respiratory (6i), Eat, Weight, Body (6i), Digestion (3i). A total of 9 questions did not meet the criteria for inclusion in the scales (23. I feel comfortable discussing my illness with others; 26. I feel bad about my physical appearance; 27. People are afraid that I may be contagious; 29. I think my coughing bothers others; 30. I feel comfortable going out at night; 33. It is difficult to make plans for the future (for example, going to college, getting married, advancing in a job, etc.); 35. To what extent did you have trouble keeping up with your schoolwork, professional work, or other daily activities during the past two weeks?). One of the conditional questions that were not previously taken into account in calculating the results (43. Has your mucus been moist) is also not included in any of the scales after conducting the factor analysis. As one can see, some of the scales were combined and some other questions from the Role and Social scales were added to them. In the new model, the above-mentioned two scales were not distinguished (a significant number of questions did not meet the analysis criteria). This indicates major problems in their original construction and reveals that the questions contained in them did not measure the assumed constructs well enough. This is also evident from the analysis of the content of the questions contained in these scales. They are often imprecise and can be understood in different ways by the respondents. In terms of social functioning, they often no longer correspond to the current problems of people suffering from cystic fibrosis.

The newly created scales are characterized by satisfactory reliability (α = 0.71–0.94). The correlation analyses also indicate a good fit for the model. Confirmation of the good fit of the scales is also evident in the analysis with the PHQ-9 and GAD-7 tools. As expected, these tools correlate the most strongly with the Vitality & Emotion scale, which contains questions concerning mainly psychological aspects. Correlations also occur with the other scales, but they are much weaker. It is also worth noting that, compared to the original factor structure, the Respiratory scale is the most stable. This is important information demonstrating that this scale meets its measurement assumptions and can continue to be used without any changes.

The phenomenon that we expected was the possibility of ceiling effects^25,27. The CFQ-R was created at a completely different time in history. At the beginning of this century, there were no therapies as effective as those available today. Therefore, all previously designed tools had the risk of questions not being matched to the current clinical situation of patients. This effect occurs in two newly created scales: Eat, Weight, Body, and Digestion. This may indicate that the obtained results exceed the measurement capabilities of the tool. Currently, based on the analysis of the results obtained, it can be said that for a large group of patients, symptoms from the broadly defined nutritional area are not a problem. However, the ceiling effect does not allow for an unambiguous confirmation of such thesis. Further observations of these scales are necessary. Comparing the results to studies from other countries, while maintaining the original model of the tool, ceiling effects are still at a fairly low level. For example, in studies conducted in Denmark and Sweden, in the case of some scales, this value reached over 50%^25,27.

Based on the studies published so far, it seems that another tool, i.e. the Cystic Fibrosis Quality of Life Questionnaire (CFQoL), is characterized by better measurement properties^9,10. Similarly to the CFQ-R 14+, there were few of them^29–32. In the future, it would be worth conducting a study using this tool on a large population of Polish patients and conducting analyses similar to those performed in our article.

There are several limitations to our study. The first one is the method of conducting the measurement (cross-sectional study). A better solution would be to conduct a study with repeated measurements to assess other parameters, such as the stability of the tool over time. However, we wanted to include the largest possible population of patients to be able to conduct a reliable factor analysis. The selected method of reaching patients provided this possibility but prevented conducting other analyses related to repeated measurement. Another limitation is the lack of a possibility to analyze the patient’s medical records, to which we did not have access.

Conclusions

This study demonstrates that the CFQ-R 14+ can be used to measure quality of life among Polish adult patients with cystic fibrosis. However, the statistical analyses performed indicate that a different division of factors (6 scales) than the one originally proposed by the authors of the English version of the questionnaire (12 scales) may be better in the Polish population. It is important to conduct further studies that would confirm our findings.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CFA

Confirmatory factor analysis

CFI

Comparative fit index

CFQoL

Cystic Fibrosis Quality of Life Questionnaire

CFQ-R 14+

Cystic Fibrosis Questionnaire-Revised 14+

EFA

Exploratory factor analysis

GAD-7

General Anxiety Disorder-7

KMO

Kaiser–Meyer–Olkin test

PAF

Principal axis factoring

PHQ-9

Patient Health Questionnaire-9

RMSEA

Root Mean square error of approximation

SRMR

Standardized root mean square residual

TLI

Tucker–Lewis index

Author contributions

Conceptualization: D.S-P., J.Ś. and A.A.; Methodology: D.S-P. and J.Ś.; Validation: D.S-P. and J.Ś.; Formal Analysis: J.Ś. and A.A.; Investigation: D.S-P. and A.A.; Writing–Original Draft Preparation: D.S-P., J.Ś.; J.J.F. and A.A.; Writing–Review & Editing D.S-P., J.Ś.; J.J.F. and A.A.; Visualization: J.Ś.; Supervision: A.A. All authors have read and agreed to the published version of the manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Consent for publication

Not applicable.