Development and Psychometric Validation of the Spanish Version of the Experience of Living With Chronic Heart Failure—University of Navarra Scale

ABSTRACT

Living with chronic heart failure (CHF) implies physical consequences but also alterations in personal identity, family and social roles, and spiritual needs. No patient‐reported outcome measures have been identified to know the meaning of living with CHF for the person. The aim of the study was to analyze the psychometric properties of the Spanish Version of the Experience of Living with CHF—University of Navarra Scale using a Spanish‐speaking population. An observational, multicenter, cross‐sectional study with retesting was conducted among 299 patients with CHF from five hospitals. Data quality, acceptability, and reliability were tested using techniques from classical test theory. Fit to the Rasch model, reliability, unidimensionality, item local independency, adequacy of response scale, and differential item functioning were tested following item response theory. The sample included 63% men, with a mean age of 69.34 years. The scale resulted in 30 items and five dimensions. Cronbach's alpha values ranged from 0.73 to 0.85, and the intraclass correlation coefficient from 0.79 to 0.87. Direct summative and linear scores could be calculated for each subscale. This is the first scale that favors knowledge on the experience of living with CHF.

Summary

• Tools are needed that help to understand the unitary vision of the person, how CHF influences the person and his or her daily life, and understand CHF as a health experience integrated into the person's life history.

• This is the first scale at an international level that allows understanding the experience of living with CHF.

• The Experience of Living With Chronic Heart Failure—University of Navarra Scale is a valid and reliable scale that is ready to be transferred to clinical practice.

1. Introduction

Chronic heart failure (CHF) is a global pandemic that affects approximately 64.3 million people worldwide (James et al. 2018). In the United States, approximately 6 million adults have CHF, and this prevalence is expected to increase to 3.0% by 2030 (Virani et al. 2021). In Europe, the average prevalence of CHF varies across countries, ranging from approximately 12 per 1000 people in Greece and Spain to more than 30 in Lithuania and Germany (Seferović et al. 2021).

CHF is characterized by poor responses to drug treatment, frequent hospital admissions, and numerous disabling symptoms (Nordfonn et al. 2019). Moreover, living with CHF implies not only physical consequences but also alterations in personal identity, family and social roles, and spiritual needs, such as thoughts about death and the need for hope (Olano‐Lizarraga et al. 2021). Despite this, medical care focuses mainly on the management of the disease and the prevention of complications, leaving aside the personal experience of living with CHF (Olano‐Lizarraga et al. 2021).

2. Background

There are very useful scales and questionnaires to understand people with CHF better. Several of these instruments focus on observing phenomena such as the quality of life (Garin et al. 2014), aspects related to symptoms and functional capacity (Asano et al. 2022; Freedland et al. 2019; Giannitsi et al. 2019; Waldréus et al. 2018), or self‐care (El Hadidi et al. 2018; Köberich et al. 2021; Matsuoka et al. 2016), but the impact of heart failure on people's lives goes beyond the presence or absence of symptoms and the restriction in performing specific activities of daily living. On the other side, a considerable number of qualitative studies and literature reviews have explored the experience of living with CHF in recent decades (Jeon et al. 2010; Olano‐Lizarraga et al. 2016; Welstand et al. 2009; Yu et al. 2008). However, no patient‐reported outcome measures (PROMs) have been identified that allow a more focused approach to such people's life experiences (Ekman et al. 2011; Martín et al. 2015). The validation of the scale that is presented in this study focuses on the construct of “the meaning of lived experience,” and for this research study, it has been conceptualized as a first‐person perspective that attempts to understand what it means for the person living with a process such as CHF and the meaning that the impact of this illness has on their life. Therefore, within the domains that compose this scale are the profound change that living with CHF entails for the person, the process of acceptance that they experience, the need to normalize their new situation, the need to live with hope, the thoughts about death and the influence of their situation on their family (Olano‐Lizarraga et al. 2021; Saracíbar‐Razquin et al. 2024). Knowing this experience will assist healthcare professionals in identifying the areas in which they should intervene to improve the well‐being of these persons, helping them to understand what is happening to them, integrating the illness process into their lives and enabling their personal development despite the illness.

Specifically, this scale is based on a theoretical framework that focuses on health experiences in which the person is conceived as a unitary, unique, and indivisible being who is in continuous growth and who is capable of projecting his or her future and developing his or her own potential for continuous improvement. Thus, health is conceived as a human experience integrated into people's lives, of which illness is also a part and constitutes a significant aspect of change for the person. Under this approach, health is seen as a value and a life experience according to the person's perspective and is related to his/her well‐being and the realization of his/her potential (Olano‐Lizarraga et al. 2021, 2022).

This focus is far from the orientation of other of the most used scales to assess the health‐related quality of life of patients with CHF, such as the Minnesota Living with Heart Failure questionnaire (MLWHFQ), Kansas City Cardiomyopathy Questionnaire (KCCQ), or Chronic Heart Failure Questionnaire (CHQ). MLWHFQ was developed to systematically and comprehensively evaluate the patient's perceptions of the effects of heart failure and its treatment on their daily life (Rector et al. 1993). In fact, MLWHFQ was originally used primarily as a valid measure of the therapeutic response of patients with heart failure to different drugs. The KCCQ was developed as a disease‐specific instrument to quantify physical limitations, symptom burden (including frequency, severity, and recent changes), quality of life, social interference, and self‐efficacy. It was intended to serve as an endpoint in clinical trials and to support disease management and quality improvement initiatives (Green et al. 2000). Finally, the CHQ was developed to assess dyspnea, fatigue, and emotional function in patients with CHF, and to be used in clinical trials to evaluate the impact of interventions on health status dimensions relevant to patients (Guyatt et al. 1989). While these instruments have significantly contributed to understanding the clinical impact of CHF, primarily by measuring symptoms, functional limitations, and treatment outcomes, they offer a predominantly biomedical and functional perspective. These tools, although valuable for clinical decision‐making and evaluating therapeutic efficacy, do not capture the deeper, subjective experience of living with the illness. As CHF increasingly affects individuals' identities, values, and life projects, there is a growing need for instruments that explore how patients make sense of their condition and its impact on their lives. The scale presented in this study responds to this gap by addressing the existential and experiential dimensions of illness, thus offering a complementary approach that can inform more person‐centered care practices.

New tools are needed that help to understand the unitary vision of the person, how CHF influences the person and his or her daily life, and understand CHF as a health experience integrated into the person's life history (Olano‐Lizarraga et al. 2021). The development of an instrument to capture what it means for the patient to live with a CHF could provide a systematic structure that would inform patients, professionals, and managers of the patients' situations, fostering the personalized care they need (Kraai et al. 2018). In addition, nursing professionals are well positioned to address these personal experiences, as they are the usual leaders in the follow‐up care of these patients (Oyanguren et al. 2016). Therefore, they need new tools that allow them to access and address this reality.

To respond to this need, a scale was designed, according to the DeVellis‐based theory of scale development (DeVellis 2017), to know the meaning that living with CHF has for the person: the “the Experience of Living with Chronic Heart Failure—University of Navarra Scale” (Experience CHF‐UNAV Scale). This scale, designed in Spanish, was based on a qualitative meta‐synthesis of the literature (Olano‐Lizarraga et al. 2016) and a hermeneutic phenomenological study (Olano‐Lizarraga et al. 2021). A panel of experts analyzed it, and later, it was evaluated by patients with CHF from outpatient units of three university hospitals in three Spanish regions (Saracíbar‐Razquin et al. 2024). This scale allows healthcare professionals to evaluate what living with CHF means to patients by focusing on their life experiences; these results will be helpful for both clinical practice and research.

This study aims to analyze the psychometric properties (validity and reliability) of the Experience CHF‐UNAV Scale among a broad Spanish‐speaking population.

3. Materials and Methods

3.1. Design

An observational, cross‐sectional, psychometric validation study with a retest was carried out in different healthcare from Spain.

3.2. Sample and Recruitment of Participants

The study was carried out between March 2021 and May 2022 in five university hospitals in Spain in outpatient heart failure units. Three hundred patients were recruited consecutively, following the psychometric recommendations of having no less than five patients per variable (Beavers et al. 2013). To avoid attrition, 335 patients (an addition of 25% subjects to the original sample size) were invited to participate in the study but 36 refused to participate, mainly due to lack of time to answer the questionnaires, leading to the final sample size of 299 participants. The selected participants met the following criteria: (1) over 18 years of age; (2) diagnosed with CHF with a functional class between II and IV according to the New York Heart Association Functional Class (NYHA‐FC) classification; (3) diagnosis of more than 6 months ago; (4) not hospitalized; (5) fluent in Spanish; (6) normal level of consciousness a present state of health allowing questionnaire completion; and (7) desire to participate in the study and freely‐given informed consent. Exclusion criteria were the following: (a) severe hearing impairment; and (b) alcohol or drug abuse.

Before starting the study, an information session was held with the nurses in the CHF units in the five hospitals who would be part of the research team. These nurses selected the participants according to the inclusion criteria. When any patients with CHF attended their scheduled check‐ups in the unit, the nurses offered them the opportunity to participate, obtained their informed consent, clarified their doubts and established with them the best time to collect data.

3.3. Data Collection

The data were collected between March 2021 and May 2022 through interviews conducted by a hospital nurse researcher in a room at the center. The questionnaires were completed directly in the electronic REDCap (Van Bulck et al. 2022) database by a nurse researcher.

First, the sociodemographic data (age, sex, marital status, educational level, and employment status) and some clinical data of the participants were collected, such as the presence of comorbidities (Charlson index), functional class (New York Heart Association Functional Class), time since diagnosis, and recent COVID diagnosis.

Subsequently, the Experience CHF‐UNAV Scale was administered to understand the experience of living with CHF. For the validation of the scale, the 53‐item version (Version 3) was administered. The original design of the scale and the development of this preliminary version were carried out in a previous phase of this study developed between 2018 and 2020 and published in 2024 by Sarcíbar‐Razquin et al. (Saracíbar‐Razquin et al. 2024). This version was obtained after completing the following steps in the design of a scale proposed by DeVellis (2017): (1) generating a list of items; (2) determining the format of the instrument; (3) reviewing the content validity of Version 1 of the scale with a panel of experts; (4) analyzing the results of the panel of experts and proposing Version 2 of the scale; (5) administering Version 2 of the scale to a sample of patients (cognitive interviews and pilot study with a group of 17 patients with CHF); (6) evaluating the items after the pilot test; and (7) optimizing the extension of the instrument, giving rise to Version 3 of the scale. See Sarcíbar‐Razquin et al. (Saracíbar‐Razquin et al. 2024) for more information.

This scale consisted of six domains: (1) Living with CHF involves profound personal changes (18 items); (2) People living with CHF have to accept their situation (six items); (3) People with CHF need to feel that their life is normal (nine items); (4) People with CHF need to live with hope (eight items); (5) People with CHF continually think about death (six items); and (6) People with CHF feel that their situation negatively influences their families (six items). For the response options, a 4‐point Likert‐type scale was used (1 = completely disagree—4 = completely agree). In addition, the scale contained an open‐ended question inviting the participants to comment on some aspect of their experience with the disease that had not yet been addressed.

For test–retest analysis, the Experience CHF‐UNAV Scale was administered for the second time to a consecutive subsample of patients who remained stable with respect to their health condition, without changes in NYHA functional class, occurrence of new symptoms, decompensations or admissions. The participants were selected randomly according to the aforementioned criteria only. This collection was carried out among 52 participants (Mokkink et al. 2023) 7–10 days after initial data collection. Most of these tests were performed in the CHF unit, and four were performed by telephone due to transportation difficulties of the participants.

Since there are no scales that measure the same construct as the Experience CHF‐UNAV Scale, two of the most widely used questionnaires for assessing quality of life in the care of people with CHF were administered. In this way, it was possible to verify the existence or not of correlations between the dimensions that compose the scales, reinforcing the originality of the new scale.

The following instruments were administered:

MLWHFQ (Garin et al. 2008): Measures the quality of life related to health among people with CHF. It is a self‐administered questionnaire with 21 items and two dimensions, physical (eight items) and emotional (five items). The response options range from 0 (no impact) to 5 (maximum impact), with a total sum score of 0–105. The Cronbach's alpha coefficient in the validation to Spanish was high in all dimensions, ranging from 0.817 to 0.915, and the intraclass correlation coefficient was > 0.7 for all three scores of the MLHFQ.

KCCQ (Comín‐Colet et al. 2011): Assesses health‐related quality of life and is specific to CHF. It is made up of 23 self‐completed items that make up seven dimensions: physical limitation; symptoms (stability, frequency, and severity); self‐care, quality of life; and social limitation. The response options for these items are Likert‐type scales from 1 to 5, 6, or 7 points; the sum score for each dimension was linearly transformed into a theoretical range of 0–100, with 100 being the best state. The Cronbach's alpha coefficient in the Spanish version of the tool was high for all scores, ranging from 0.7 to 0.96, and the intraclass correlation coefficient was > 0.7 for all scores, except for the one related to symptom stability (0.68).

3.4. Data Analysis

Adopting the Rasch model (Rasch 1993), the following metric properties were analyzed (Tennant and Conaghan 2007): fit of the data to the Rasch model with Bonferroni correction for the number of items; reliability (Person Separation Index, PSI; interpreted similarly as the Cronbach's alpha), unidimensionality; local independence of the items; adequacy of the response scale; person‐item distribution (graphic display of both person and item locations); differential item functioning (DIF) by sex, age (by means: less than 71 years; 71 years and more); and educational level (primary or lower vs. secondary or higher). There are excellent tutorials that explain Rasch analysis in lay terms (Tennant and Pallant 2007).

Fit to the Rasch model was considered adequate when there was a nonsignificant chi‐square and fit residuals ranged between −2.5 and 2.5 logits. Item local independency refers to a lack of association between two items, after the first Rasch factor is accounted for. It was assessed through the correlations among item fit residuals, which were expected to be similar. The adequacy of the response scale was indicated by ordered item thresholds, the point of equal response probability between two adjacent response options. In the case of disordered thresholds, adjacent categories collapsed. For unidimensionality (if items form a single dimension), a principal component of residuals was performed (Tennant and Pallant 2007). DIF assesses item bias by a group factor, and it was assessed using ANOVA. In the case of nonuniform DIF, the item in question was eliminated. Items that showed a poor fit with the model were also eliminated. These changes were made iteratively. Once a good fit was achieved, linear scores were generated on a 0–10 scale for each dimension.

The following psychometric attributes, according to classical test theory, were analyzed (Scientific Advisory Committee of the Medical Outcomes Trust 2002):

• –Data quality and acceptability, with the following standard criteria: missing data < 10%; floor and ceiling effects: < 15%; skewness: −1 to +1 (van der Linden et al. 2005).
• –Reliability, as internal consistency and temporal stability (test–retest reliability). For internal consistency, the following parameters were ascertained: Cronbach's alpha and McDonald's omega coefficients (standard criterion ≥ 0.70), item homogeneity index (standard criterion ≥ 0.15), and item‐total corrected correlation (standard criterion ≥ 0.20) (Scientific Advisory Committee of the Medical Outcomes Trust 2002). For test–retest reliability, the intraclass correlation coefficient was calculated (standard criterion: > 0.70) (Jenkinson et al. 2003).
• –Construct validity (convergent, discriminative, and internal). For convergent validity (Mokkink et al. 2010), Spearman's rank correlation coefficients were determined. A moderate‐to‐high correlation (r _S = 0.30–0.49 and r _S ≥ 0.50) between the Experience CHF‐UNAV Scale and the MLHFQ and the KCCQ scales was hypothesized (Juniper et al. 1996). For discriminative (known‐groups) validity, the differences in the Experience CHF‐UNAV subscale scores in the sample grouped according to variables of interest (sex, educational level, NYHA Functional Class, Charlson index scores) were calculated. Mann–Whitney and Kruskal–Wallis tests were applied. For internal validity, the intercorrelation among subscales was calculated, with a standard criterion of r _S = 30–0.70 (Hobart et al. 2001).

Data quality and acceptability, reliability, and internal validity were determined for the original version. After modifications to the scale according to Rasch analysis, the distribution of scores, internal consistency, and convergent and discriminative validity were calculated for the revised, abbreviated version. SPSS software was used for classical psychometric analysis, and RUMM2030 was used for Rasch analysis.

4. Results

Among the 299 participants recruited, 63% were men, with a mean age of 69.34 years (SD 13.85). Sixty‐two percent had a functional class II according to the NYHA, and 75% were retired. Clinical data showed that 37.1% had a Charlson morbidity index of 1–2 points, 35.8% had 3–4 points, and 27.1% had 5 or more points. According to the NYHA functional classification, the majority were in Class II (61.9%), with 35.8% in Class III and 2.3% in Class IV. Notably, 11.7% had a recent COVID‐19 infection. The duration of their disease varied from 0.5 to 56 years, with a mean duration of 8.56 years (SD = 8.68). The sociodemographic and clinical data of interest are presented in Table 1.

TABLE 1.

Sociodemographic and clinical characteristics of the sample.

Variables		N	%
Sample		299	100
Sex	Women	111	37.1
	Men	188	62.9
Marital status	Single	47	15.7
	Married. with partner	187	62.5
	Divorced	18	6.0
	Widow	47	15.7
Education level	No studies	16	5.4
	Primary studies	153	51.2
	Secondary studies	43	14.4
	Occupational training	40	13.4
	Universitary studies	47	15.7
Employment status a	Working	24	8.0
	Housekeeper	17	5.7
	Retired	225	75.3
	Not working (unemployed, disability leave)	32	11.1
Charlson morbidity index	1–2 points	111	37.1
	3–4 points	107	35.8
	≥ 5 points	81	27.1
New York Heart Association Functional Class (NYHA‐FC)	II	185	61.9
	III	107	35.8
	IV	7	2.3
Recent COVID‐19 contagion	Yes	35	11.7

	Range	Mean	SD
Age (years)	31–103	69.40	13.83
Disease duration (years)	0.5–56	8.56	8.68

Note: The red font indicates the changes made in R2 of the correction round requested by the reviewers.

Abbreviation: SD, standard deviation.

^a There was one missing value for employment status.

4.1. Original Version (53 Items)

The data quality, acceptability, and reliability results are displayed in Table 2. The Experience CHF‐UNAV Scale achieved a mean total score of 114.73 (median: 115.00; standard deviation, SD: 12.61; observed range: 56–157). No missing data were observed, and skewness (−0.22) and floor and ceiling effects (0.3 for both) of the total score were within the standard limits. For the subscales, skewness ranged from –0.04 to –0.48, with no floor and ceiling effects detected, except for the subscale Acceptance, which had a ceiling effect (16%).

TABLE 2.

Data quality, acceptability, and reliability of the original version (53 items) of the Experience CHF‐UNAV Scale.

	UNAV‐profound changes	UNAV‐acceptance	UNAV‐normality	UNAV‐hope	UNAV‐death	UNAV‐family	UNAV‐total
Mean	38.79	19.54	24.55	24.98	12.34	14.07	114.73
Median	39.00	20.00	25.00	26.00	12.00	14.00	115.00
SD	10.98	3.46	4.38	5.06	3.70	4.16	12.61
Skewness	−0.04	−0.47	−0.21	−0.48	0.30	0.05	−0.22
Observed range	16‐64	10–24	13–32	9–32	6–22	6–24	56–157
Floor and ceiling effects (%)	0.3–0.3	1.0–16.0	1.3–6.7	0.3–8.3	4.3–0.3	3.7–0.3	0.3–0.3
Cronbach's alpha	0.92	0.79	0.76	0.85	0.69	0.76	—
Item homogeneity index	0.41	0.40	0.29	0.44	0.26	0.34	—
Item total corrected correlation	0.35–0.79	0.40–0.71	0.21–0.59	0.41–0.78	0.20–0.64	0.43–0.58	—
Intraclass correlation coefficient (N = 52)	0.86	0.88	0.86	0.88	0.85	0.84	0.72

Regarding internal consistency, and considering that the Rasch analysis showed the scale to be multidimensional, Cronbach's alpha was calculated by subscales rather than as a total score. Alpha values ranged from 0.69 (subscale Death) to 0.92 (subscale Profound changes), with an item homogeneity index between 0.26 (subscale Death) and 0.44 (subscale Hope) (Table 2). Concerning test–retest reliability, the intraclass correlation coefficient values were 0.84–0.88 for the subscales and 0.72 for the total score.

The subscales correlated −0.33 (Hope and Family) to 0.74 (Acceptance and Hope) between them.

4.2. Rasch Analysis

As the initial scale with all the items showed a bad fit with the Rasch model, using multidimensionality, a Rasch analysis by dimension was performed. Almost all the items had disordered thresholds, so the response scale was recoded, collapsing the two intermediate categories whereby the response scale was left with 3 instead of 4 response categories for these items. Table 3 summarizes the goodness of fit statistics, and Table 4 presents the individual item fit for the revised version of the scale.

TABLE 3.

Goodness of fit statistics for Rasch analysis, revised version.

		Criteria	Dimension
			Profound changes	Acceptance	Normality	Hope	Death	Family
			(Seven items)	(Five items)	(Seven items)	(Five items)	(Three items)	(Six items)
Item fit residual	M	0	−0.125	0.027	0.173	−0.277	−1.007	0.126
	SD	1	1.031	0.955	0.876	1.153	1.220	0.960
Person fit residual	M	0	−0.456	−0.464	−0.295	−0.413	−0.388	−0.610
	SD	1	1.220	1.167	1.006	0.989	0.661	1.585
Item–trait interaction	χ 2	Low	41.099	18.555	49.032	16.885	24.784	22.433
	p	N.S	0.005	0.235	0.008	0.326	0.003	0.553
Person Separation Index		> 0.70	0.806	0.647	0.752	0.780	0.501	0.678
Unidimensionality: significant t tests	%	< 10%	3.33	1.31	1.67	2.67		5.33
	(95% CI)	(*)	(0.09;0.58)	(0.011;0.038)	(−0.08;0.041)	(0.02;0.051)	*	(0.029;0.078)

Note: The asterisk (*) indicates that the result for the unidimensionality assessment should be interpreted with caution. Although the observed percentage of significant t‐tests is below the recommended threshold of 10%, the 95% confidence interval includes values that approach or exceed this criterion and/or includes zero. According to Rasch measurement guidelines, unidimensionality can only be supported when both the point estimate and its confidence interval clearly fall below the 10% threshold.

TABLE 4.

Individual item fit of the Experience CHF‐UNAV Scale, revised version.

Item		Location	Standard deviation	Fit residual	Chi‐square	Probability
	Profound changes
Item 04	Loss of friends	1.319	0.123	1.611	14.242	0.003
Item 05	Feeling of loneliness	0.880	0.090	−0.558	3.256	0.354
Item 11	Sense of usefulness	−0.464	0.116	0.239	5.204	0.157
Item 13	Difficulties leaving home	−0.313	0.114	−1.732	8.275	0.041
Item 14	Carrying on with life	−0.792	0.119	−0.289	3.481	0.323
Item 16	Change in identity	−0.957	0.122	−0.535	4.426	0.219
Item 17	Body changes	0.328	0.088	0.390	2.214	0.529
	Acceptance
Item 19	Acceptance of changes	0.884	0.098	0.281	3.122	0.373
Item 20	Adapting to change	−0.787	0.124	0.825	1.933	0.586
Item 21	Enjoying life	0.110	0.139	−1.469	10.257	0.017
Item 23	Supporting environment	−0.275	0.130	0.801	0.411	0.938
Item 24	Fulfilling plans	0.069	0.104	−0.301	2.831	0.418
	Normality
Item 25	Perception of illness	1.075	0.077	1.317	7.333	0.119
Item 26	Normal life	1.076	0.084	−0.479	3.085	0.544
Item 27	Daily routine	−0.135	0.089	0.569	3.988	0.408
Item 28	Household tasks	0.541	0.081	0.867	0.518	0.972
Item 29	Attention to the family	−1.237	0.126	−0.021	6.640	0.156
Item 31	Family plans	−0.304	0.120	−1.307	17.742	0.001
Item 33	Rejection of the environment	−1.015	0.120	0.263	9.725	0.045
	Hope
Item 34	Overcoming difficulties	0.456	0.105	0.731	0.966	0.809
Item 35	Positive attitude	0.068	0.146	−1.995	6.663	0.083
Item 37	Motivation for living	−1.913	0.161	−0.897	6.959	0.073
Item 38	Sense of calm	0.508	0.106	0.192	0.733	0.865
Item 40	Achievable objectives	0.880	0.117	0.584	1.564	0.668
	Death
Item 43	Feeling of fear	−1.662	0.131	−0.871	4.750	0.191
Item 44	Anguish about dying	−0.571	0.111	−2.289	12.803	0.005
Item 46	Death wishes	2.234	0.155	0.139	7.231	0.065
Family
Item 48	The suffering of the family	−0.474	0.106	1.461	5.782	0.216
Item 49	Limiting family activities	0.234	0.102	0.223	2.943	0.567
Item 50	Feeling of burden	0.762	0.107	0.482	4.770	0.312
Item 51	Not worrying the family	−0.652	0.071	−0.469	2.636	0.620
Item 52	Hiding symptoms	−0.132	0.071	−1.373	2.424	0.658
Item 53	Avoiding family activities	0.262	0.108	0.430	3.879	0.423

In the Profound changes dimension, six items were deleted due to poor fit with the model (1–3, 9, 12, 18) and three due to nonuniform DIF (8, 10, 15). The resulting dimension with seven items presented a good fit with the Rasch model χ ² (21) = 41.099, p = 0.005, PSI = 0.81, unidimensionality, local independence of the items and a good person‐item distribution. Item 4 presented DIF by age, whereby older people overestimated the scores for this item on the same level of the construct. An adequate distribution of people and items was observed.

The Acceptance dimension, after eliminating Item 22 due to a poor fit with the model, presented PSI = 0.65, good fit with the model χ ² (15) = 18.55, p = 0.23, unidimensionality, and local independence of the items. Item 19 presented nonuniform DIF by studies, but it was preserved to prevent a bad fit with the global model of the scale. DIF was not observed by sex, age, or educational level for the other items. A ceiling effect of 25% was observed in the person‐item distribution.

Regarding the Normality dimension, after removing Item 32 due to a poor fit, the 7‐item scale presented a good fit with the model χ ² (28) = 49.03, p = 0.008, PSI = 0.75, one‐dimensionality, local independence of the items and DIF by age for Item 29, with overestimation of the scores by young people. The person‐item distribution showed an absence of items for the highest values of the construct.

The Hope dimension was reduced to five items, Items 36, 39, and 41 having been deleted due to a poor fit with the model. With a good fit with the model χ ² (15) = 16.88, p = 0.32, PSI = 0.78, unidimensionality and local independence of the items were obtained. Two items (37 and 38) presented DIF by age in opposite directions. In the person‐item distribution, a ceiling effect of 20% was observed.

The Family dimension, with six items, presented a good fit with the model χ ² (24) = 22.53, p = 0.55, PSI = 0.67, unidimensionality, local independence of the items, and DIF by age for Item 50 (older people they overestimate the scores for this item), and a relatively adequate item‐person distribution.

The Death dimension did not have a good fit with the model, with three items with poor fit and very low PSI (0.51). Finally, a Rasch analysis was performed using the five dimensions as superitems, which presented an inadequate fit with the Rasch model and item local dependency.

4.3. Abbreviated Scale (30 Items)

After applying the changes suggested by the Rasch analysis, the abbreviated scale was composed of the subscales Profound changes (seven items), Acceptance (five items), Normality (seven items), Hope (five items), and Family (six items). Direct summative and linear scores were calculated for each subscale (Table 5).

TABLE 5.

Conversion from initial scores to linear scores.

Initial score	Profound changes		Acceptance		Normality		Hope		Death		Family
	Logits	LM	Logits	LM	Logits	LM	Logits	LM	Logits	LM	Logits	LM
0	−4.35	0.00	−4.38	0.00	−4.15	0.00	−5.13	0.00	−4.30	0.00	−3.46	0.00
1	−3.38	1.14	−3.15	14.33	−3.13	12.38	−3.66	14.40	−2.69	19.83	−2.56	12.88
2	−2.60	2.06	−2.28	24.58	−2.39	21.34	−2.62	24.49	−0.96	41.04	−1.91	22.08
3	−1.98	2.81	−1.65	31.91	−1.85	27.82	−1.90	31.57	0.22	55.54	−1.44	28.65
4	−1.43	3.45	−1.14	37.85	−1.42	33.06	−1.29	37.55	1.22	67.83	−1.07	33.96
5	−0.95	4.02	−0.69	43.11	−1.05	37.54	−0.72	43.09	2.47	83.13	−0.75	38.57
6	−0.52	4.52	−0.26	48.12	−0.71	41.60	−0.17	48.44	3.84	100.00	−0.44	42.87
7	−0.14	4.98	0.18	53.22	−0.40	45.38	0.38	53.80			−0.14	47.10
8	0.23	5.41	0.65	58.76	−0.10	48.99	0.96	59.44			0.17	51.50
9	0.57	5.82	1.18	64.99	0.19	52.52	1.59	65.60			0.50	56.29
10	0.91	6.22	1.77	71.89	0.48	56.01	2.27	72.30			0.89	61.79
11	1.26	6.63	2.42	79.49	0.77	59.53	3.03	79.66			1.35	68.28
12	1.63	7.07	3.21	88.74	1.07	63.16	3.96	88.75			1.90	76.16
13	2.04	7.55	4.18	100.00	1.39	66.99	5.11	100.00			2.63	86.49
14	2.53	8.13			1.74	71.16					3.58	100.00
15	3.20	8.93			2.13	75.89
16	4.11	10.00			2.60	81.60
17					3.24	89.38
18					4.12	100.00

Abbreviation: LM, linear measure (range 0–100).

Internal consistency is also displayed in Table 5. Cronbach's alpha values ranged from 0.73 (subscale Family) to 0.87 (subscale Hope), while the item homogeneity index was between 0.31 (subscale Family) and 0.57 (subscale Hope). However, using McDonald's omega, the subscale Family did not reach the threshold. All items exceeded 0.20 in the item‐total corrected correlation coefficients. Regarding test–retest reliability (Table 6), the intraclass correlation coefficient was higher than 0.70 for all subscales.

TABLE 6.

Scores distribution, reliability, and convergent validity of the abbreviated version (30 items) of the Experience CHF‐UNAV Scale.

		UNAV‐profound changes	UNAV‐acceptance	UNAV‐normality	UNAV‐hope	UNAV‐family
Mean	Direct score	12.93	14.99	18.64	14.32	12.30
	Linear score	4.16	7.49	6.38	7.07	4.37
Median	Direct score	13.00	15.00	19.00	15.00	12.00
	Linear score	4.52	7.19	6.32	7.23	4.29
SD	Direct score	3.72	2.62	3.73	3.02	3.05
	Linear score	2.10	1.94	1.67	2.14	1.70
Observed range	Direct score	7–23	8–18	10–25	5–18	6–20
	Linear score	0–10	3.19–10	2.78–10	0–10	0–10

Reliability	UNAV‐profound changes	UNAV‐acceptance	UNAV‐normality	UNAV‐hope	UNAV‐family
Cronbach's alpha	0.84	0.81	0.78	0.85	0.71
McDonald's omega	0.85	0.81	0.80	0.86	0.66
Item homogeneity index	0.45	0.47	0.35	0.57	0.31
Item total corrected correlation	0.48–0.71	0.52–0.70	0.33–0.62	0.62–0.76	0.38–0.52
Intraclass correlation coefficient (N = 52)	0.87	0.89	0.86	0.79	0.83

Convergent validity a	UNAV‐profound changes	UNAV‐acceptance	UNAV‐normality	UNAV‐hope	UNAV‐family
MLHFQ physical	0.52	−0.55	−0.29	−0.39	0.45
MLHFQ emotional	0.69	−0.65	−0.46	−0.52	0.52
MLHFQ TOTAL	0.54	−0.59	−0.32	−0.41	0.45
KCCQ physical	−0.50	0.53	0.36	0.46	−0.34
KCCQ sympt freq	−0.40	0.41	0.28	0.38	−0.33
KCCQ sympt burden	−0.45	0.42	0.31	0.40	−0.36
KCCQ QoL	−0.67	0.69	0.49	0.61	−0.46
KCCQ social limit	−0.35	0.45	0.32	0.35	−0.26
KCCQ TOTAL	−0.56	0.60	0.43	0.54	−0.41

Note: The red font indicates the changes made in R2 of the correction round requested by the reviewers.

^a Spearman's rank correlation coefficients using linear scores of the Experience CHF‐UNAV scale, p < 0.01.

The Experience CHF‐UNAV Scale subscale linear scores correlated r _S = −0.32 and r _S = −0.59 with the MLHFQ total score and from r _S = −0.41 to r _S = −0.60 with the KCCQ total score. The subscales Profound changes and Family correlated r _S = 0.69 and r _S = 0.52 with the Emotional subscale of the MLHFQ, respectively; the subscales Normality, Acceptance, and Hope correlated between r _S = 0.49 and 0.69 with the KCCQ Quality of Life subscale.

Using linear scores, all subscales showed significant differences by sex (p < 0.001), except the Acceptance subscale, and by functional class (p < 0.001) (Table 7). The subscales of Profound changes, Acceptance, and Hope had significant differences by educational level (p < 0.05). No significant differences in the subscale scores were found by the Charlson Comorbidity Index score.

TABLE 7.

Experience CHF‐UNAV subscales linear scores by variables of interest (discriminative validity).

		UNAV‐profound changes	UNAV‐acceptance	UNAV‐normality	UNAV‐hope	UNAV‐family
Sex	Women (N = 111)	4.69 (2.02)	7.24 (1.94)	5.89 (1.42)	6.59 (2.08)	4.84 (1.75)
	Men (N = 189)	3.84 (2.09)	7.64 (1.94)	6.67 (1.75)	7.34 (2.13)	4.10 (1.61)
	p a	< 0.001	0.096	< 0.001	0.004	< 0.001
Educative level	No studies or primary (N = 169)	4.39 (1.99)	7.17 (1.96)	6.27 (1.67)	6.69 (2.19)	4.46 (1.63)
	Secondary or university (N = 131)	3.86 (2.20)	7.91 (1.85)	6.53 (1.67)	7.55 (1.98)	4.26 (1.78)
	p a	0.017	0.001	0.146	< 0.001	0.257
NYHA functional class	II (N = 186)	3.71 (2.12)	7.86 (1.86)	6.81 (1.64)	7.51 (2.03)	4.11 (1.81)
	III and IV (N = 114)	4.88 (1.86)	6.89 (1.94)	5.68 (1.48)	6.34 (2.14)	4.80 (1.41)
	p a	< 0.001	< 0.001	< 0.001	< 0.001	0.001
Charlson index	1–2 points (N = 112)	3.87 (2.19)	7.63 (1.88)	6.52 (1.68)	7.47 (1.87)	4.34 (2.01)
	3–4 points (N = 107)	4.28 (2.08)	7.49 (2.00)	6.28 (1.63)	6.89 (2.16)	4.29 (1.68)
	≥ 5 points (N = 81)	4.40 (1.97)	7.29 (1.96)	6.34 (1.72)	6.73 (2.40)	4.52 (1.19)
	p b	0.141	0.441	0.531	0.047	0.810

^a Mann–Whitney test.

^b Kruskal–Wallis test.

5. Discussion

This study presents the validation process of the first scale designed to ascertain the experience of living with CHF. The scale is conceptually composed of five domains: (1) Living with CHF involves profound personal changes (seven items); (2) People living with CHF have to accept their situation (five items); (3) People with CHF need to feel that their life is normal (seven items); (4) People with CHF need to live with hope (five items); and (5) People with CHF feel that their situation negatively influences their families (six items).

The original version of the “the Experience of Living with Chronic Heart Failure—University of Navarra Scale” (Experience CHF‐UNAV Scale) demonstrated satisfactory data quality and acceptability, with no missing data, skewness within the standard limits and an absence of floor and ceiling effects. This supports the use of the scale in routine clinical practice settings.

All reliability parameters (internal consistency and test–retest reliability) of the original version of the scale were also adequate, except for the subscale Death, which showed lower internal consistency. In fact, this subscale did not fit the Rasch model and obtained a lower PSI coefficient; thus, the revised version of the Experience CHF‐UNAV Scale does not include this subscale. A possible explanation for this could be that the conceptual design of the scale was based on the results of a qualitative study carried out by the research team (Olano‐Lizarraga et al. 2021); in a meta‐synthesis of the literature (Olano‐Lizarraga et al. 2016), studies were identified in which the mean age of the sample of patients with CHF was higher. A recent study among a population of CHF patients with a mean age of 74.6 years indicated that the proximity of death was a recurring theme among these people (Olano‐Lizarraga et al. 2022). However, the average age of the 299 participants in this validation study was 69.34 years, a fact that may have rendered the questions about death less appropriate for a younger sample.

After applying the changes suggested by the Rasch analysis results, that is, the deletion of missing items or those with DIF, the revised Experience CHF‐UNAV Scale is composed of 30 items grouped into five subscales. Direct summative and linear scores can be calculated for all the subscales. For research purposes, we recommend the use of linear scores according to the provided conversion table. However, for clinical use, there remains the option of using direct scores, which are easier to calculate. These direct scores have an important drawback: they are not on a linear scale, and the distances between their points are not the same, which hinders the calculation of change scores. The DIF analyses also suggested that five items were biased for age and one for education level. If this finding is confirmed in further studies, caution should be applied when interpreting differences by age.

Another change was modifying the response scale from 4 to 3 options for most items. This might be because respondents are not capable of distinguishing some response categories. Instead of changing the response scale, the Experience CHF‐UNAV Scale items may therefore be coded differently to accommodate this modification. Since the abbreviated scale provides linear, Rasch‐generated scores, the changes in the response scale codification scheme do not result in a different evaluation of the construct. This procedure has been suggested for other scales, such as the Nurses' Attitude Scale (Alfaro‐Díaz et al. 2023). Nevertheless, the abbreviated version of the scale has similar psychometric properties to the original one, with satisfactory internal consistency, adequate for group comparison, and temporal stability. This finding supports the use of the abbreviated version.

On the other hand, moderate to high correlations were identified between some of the subscales in the Experience CHF‐UNAV Scale and the dimensions of other scales for assessing health‐related quality of life among people with CHF. These findings are rational, as the Emotional subscale in the MLHFQ (Rector and Cohn 1992) contains four items that address concepts similar to some that are included in the Experience CHF‐UNAV Scale, for example, the perception of being a burden on one's family, the feeling of concern, and the feelings of depression and loss of control. Similarly, the Quality of Life subscale in the KCCQ (Green et al. 2000) reflects three items that have similarities with the scale designed above, for example, the ability to enjoy life, satisfaction with life, and feelings of discouragement or depression. These comparisons, in turn, affirm the originality of the Experience CHF‐UNAV Scale, as neither all five dimensions in the scale nor most of the items that compose them are included in the MLHFQ and KCCQ scales, which are the most used measures in the follow‐up care for people with CHF. This justifies its useful role in obtaining unique knowledge about the life experience of each patient. The Experience CHF‐UNAV Scale introduces a distinct conceptual approach, departing from these traditional tools, which mainly focus on symptom burden, physical limitations, and treatment outcomes. Instead, this scale is grounded in a humanistic framework that emphasizes the patient's inner experience and the personal significance of illness, thereby offering an additional perspective to existing clinical and functional assessments.

The results of this study have also shown that women with CHF experience a greater change in their life and perceive that they have a more negative impact on their family than men with CHF. According to Rasch analysis, these differences are not due to an item bias by sex. Other studies on the differences between men and women with CHF have also yielded results that show a greater impact of the disease upon various areas in women's lives. Specifically, they have identified that women have more symptoms, greater congestion, and more adverse reactions to treatments (Regitz‐Zagrosek 2020). They also show higher rates of depression, which are associated with worse clinical states and prognoses (Sbolli et al. 2020). All of this evidence signals the changes that these women experience in their lives. On the other hand, several studies indicate that women experience a greater impact on the family roles they consider crucial for the proper functioning of their family, that is, another member must fulfill their previous functions in addition to taking care of them, thus increasing their perception of negatively impacting their family (Checa et al. 2020).

5.1. Limitations

The sample consisted of 37% women, which could affect the representativeness of this group to some extent. However, the multicenter and multiregional study strengthened the representativeness of the situation of CHF patients in the country. In addition, the sample has been sufficient for the psychometric validation of the scale. In addition, although this study was carried out in a single country, the content of the items in the scale was notably based not only on a hermeneutical phenomenological study carried out in the same country but also on a systematic review of the literature, which included studies from multiple countries worldwide. Finally and importantly, this scale has been validated in a sample of patients with heart failure (NYHA II‐IV), indicating a moderate to severe impairment of the disease. Therefore, the tool may not be as valid for understanding the experiences of individuals with milder impairment (NYHA I).

6. Conclusion

The Experience CHF‐UNAV Scale (30 items) is the first scale built in the Spanish and international context that favors obtaining knowledge of the experience of living with CHF. This scale is valid and can reliably indicate what it means for a person to live with this disease. Thus, it can be applied to determine the aspects in which health professionals, especially nurses, could intervene via a person‐centered approach. The Experience CHF‐UNAV Scale is prepared and available for clinical use among a Spanish‐speaking population, but it would be of great interest to translate and validate this scale into other languages. Thus, part of this research team has started a project aimed at validating this scale in English with a sample of American patients.

7. Relevance for Clinical Practice

The Experience CHF‐UNAV Scale is a valid and reliable scale that is ready to be transferred to clinical practice. This scale allows healthcare professionals to evaluate what living with CHF means to patients by focusing on their life experiences and from there identify areas where intervention is needed to improve the well‐being of these patients.

Author Contributions

Maddi Olano‐Lizarraga: conceptualization, methodology, software, validation, formal analysis, investigation, writing – original draft, visualization, supervision, project administration, funding acquisition. Maribel Saracíbar‐Razquin: conceptualization, methodology, validation, investigation, visualization, supervision, project administration, funding acquisition. Amparo Zaragoza‐Salcedo: conceptualization, methodology, validation, investigation, visualization, funding acquisition. José Luis Cobo‐Sanchez: validation, investigation, resources, visualization, funding acquisition. Santiago Pérez‐García: validation, investigation, resources, visualization, funding acquisition. Aurora Simón‐Ricart: validation, investigation, resources, visualization, funding acquisition. Pilar Ara‐Lucea: validation, investigation, resources, visualization, funding acquisition. Leticia Jimeno‐San Martín: validation, investigation, resources, visualization, funding acquisition. Montserrat Ducay‐Eguillor: validation, investigation, resources, visualization, funding acquisition. Noelia De La Torre‐Lomas: validation, investigation, resources, visualization, funding acquisition. Jesica Pérez‐Herreros: validation, investigation, resources, visualization, funding acquisition. Nieves Moro‐Tejedor: validation, investigation, resources, visualization, funding acquisition. Luis Guerra‐Sánchez: validation, investigation, resources, visualization, funding acquisition. Laura Cambra‐Vidart: validation, investigation, resources, visualization, funding acquisition. Silvia Salinas‐Urra: validation, investigation, resources, visualization, funding acquisition. Carmen Rodriguez‐Blazquez: conceptualization, methodology, software, validation, formal analysis, investigation, writing – original draft, visualization, supervision. Maria João Forjaz: conceptualization, methodology, software, validation, formal analysis, investigation, writing – original draft, visualization, supervision. Alba Ayala: methodology, software, validation, formal analysis, investigation, writing – original draft, visualization. Jesús Martín‐Martín: conceptualization, methodology, software, validation, formal analysis, investigation, writing – original draft, visualization, supervision, project administration, funding acquisition.

Funding

This study was funded by the project PI17/00195, Integrated in Plan Estatal de I+D+I 2013–2016 (2013–2016 State Plan for Scientific and Technical Research and Innovation) and co‐financed by the Subdirectorate‐General for the Evaluation and Promotion of Research of the Carlos III Health Institute (Instituto de Salud Carlos III—ISCIII) and the European Regional Development Fund (Fondos Europeos de Desarrollo Regional—FEDER). This external agency did not intervene in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

Ethics Statement

This research was approved by the Research Ethics Committee of the University of Navarra (code 2017.125), the Clinical Research Ethics Committee of Cantabria (code 2017.252), the Ethics Committee of Drug Research of the University Hospital 12 de Octubre (code 17/473), the Ethics Committee for Research with medicines General University Hospital Gregorio Marañón (code 250/2018), and the Clinical Research Ethics Committee of Navarra (code 2018/41). Participants received oral and written information about the study when they came to the CHF unit for a scheduled routine review. All participants provided their informed consent before starting data collection.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1:

Olano‐Lizarraga, M. , Saracíbar‐Razquin M., Zaragoza‐Salcedo A., et al. 2026. “Development and Psychometric Validation of the Spanish Version of the Experience of Living With Chronic Heart Failure—University of Navarra Scale.” Nursing & Health Sciences 28, no. 1: e70285. 10.1111/nhs.70285.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.