See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346170977

The Self‐Care of Heart Failure Index version 7.2: Further psychometric testing

Article in Research in Nursing & Health · November 2020

DOI: 10.1002/nur.22083

CITATIONS READS

26 2,993

9 authors, including:

Ercole Vellone Maddalena De Maria

University of Rome Tor Vergata Link Campus University
487 PUBLICATIONS 7,241 CITATIONS 76 PUBLICATIONS 552 CITATIONS

SEE PROFILE SEE PROFILE

Paolo Iovino Claudio Barbaranelli

University of Florence Sapienza University of Rome
55 PUBLICATIONS 352 CITATIONS 252 PUBLICATIONS 24,820 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Paolo Iovino on 26 November 2020.

The user has requested enhancement of the downloaded file.

Received: 3 May 2020 | Revised: 25 August 2020 | Accepted: 15 October 2020

DOI: 10.1002/nur.22083

RESEARCH ARTICLE

The Self‐Care of Heart Failure Index version 7.2: Further

psychometric testing

Ercole Vellone1 | Maddalena De Maria1 | Paolo Iovino1,2 |

Claudio Barbaranelli | Valentina Zeffiro | Gianluca Pucciarelli1 |
3 1

Angela Durante1 | Rosaria Alvaro1 | Barbara Riegel4

1
Department of Biomedicine and Prevention,
University of Rome Tor Vergata, Rome, Italy Abstract
2
School of Nursing, Australian Catholic Clinicians and researchers need valid and reliable instruments to evaluate heart
University, Melbourne, Australia
failure (HF) self‐care. The Self‐Care of Heart Failure Index (SCHFI) is a theoretically
3
Department of Psychology, Sapienza
University of Rome, Rome, Italy driven instrument developed for this purpose. The SCHFI measures self‐care with
4
School of Nursing, University of three scales: self‐care maintenance, measuring behaviors to maintain HF stability;
Pennsylvania, Philadelphia, Pennsylvania, USA symptom perception, measuring monitoring behaviors; and self‐care management,

Correspondence
assessing the response to symptoms. After the theory underpinning the SCHFI was
Ercole Vellone, Department of Biomedicine updated, the instrument was updated to version 7.2 but it was only tested in the
and Prevention, University of Rome Tor
United States. In this study we tested the psychometric characteristics (structural and
Vergata, Via Montpellier, 1, 00133 Rome,
Italy. construct validity, internal consistency, and test–retest reliability) of the SCHFI v.7.2
Email: [email protected]
in an Italian population of HF patients. We used a cross‐sectional design to study 280
HF patients with additional data collected after 2 weeks for test–retest reliability.
Funding information
Center of Excellence for Nursing Scholarship, Adults with HF (mean age 75.6 (±10.8); 70.8% in New York Heart Association [NYHA]
Rome, Italy classes II and III) were enrolled from six centers across Italy. Confirmatory factor
analysis showed supportive structural validity in the three SCHFI v.7.2 scales (CFI
from 0.94 to 0.95; RMSEA from 0.05 to 0.07). Internal consistency reliability esti-
mated with Cronbach's α and composite reliability ranged between .73 and .88;
test–retest reliability ranged between 0.73 and 0.92. Construct validity was sup-
ported with significant correlations between the SCHFI v.7.2 scale scores and quality
of life, brain natriuretic peptide levels and NYHA class. This study further supports
the psychometric characteristics of the SCHFI v.7.2, illustrating that it can be used in
clinical practice and research also in an Italian population.

KEYWORDS
heart failure, psychometrics, reliability, self‐care, validity

1 | INTRODUCTION result of the aging of the population and improved survival

(Danielsen et al., 2017; Gouveia et al., 2019). HF causes significant
Despite scientific advances, heart failure (HF) remains a major health symptom burden, high mortality and morbidity, and extensive use of
problem, with estimates of 26 million individuals affected worldwide health care services (von Lueder & Agewall, 2018).
(Savarese & Lund, 2017). Recent projections in Europe suggest that Self‐care of HF was defined as “a naturalistic decision‐making
in the next two decades, HF prevalence will increase 2–3‐fold as a process that influences actions that maintain physiologic stability,

640 | © 2020 Wiley Periodicals LLC wileyonlinelibrary.com/journal/nur Res Nurs Health. 2020;43:640–650.
VELLONE ET AL.
| 641

facilitate the perception of symptoms, and direct the management of those 2 | ME THO D S
symptoms” (Riegel et al., 2016, p. 226). Self‐care is recognized as
an effective approach for managing HF and is associated with a 2.1 | Design and settings
variety of positive patient outcomes including improved quality
of life, lower mortality, fewer hospitalizations, and decreased We used a descriptive multicenter cross‐sectional design with addi-
health care costs (Buck et al., 2012; Reilly et al., 2015; Ruppar tional data collected after 2 weeks for test–retest purpose. We en-
et al., 2016; Wang et al., 2011). Consequently, patients with HF rolled patients attending six HF outpatient centers in central Italy. To
are strongly encouraged to perform self‐care. For example, conduct this study, we followed the COSMIN guidelines (Mokkink
patients are taught to adhere to what is often a complex phar- et al., 2018).
macological treatment plan, monitor signs and symptoms, and
follow a specific diet and exercise regimen (Toukhsati et al.,
2015). However, despite its positive effects, self‐care is 2.2 | Sampling and data collection
frequently found to be poor in patients with HF (Cocchieri et al.,
2015; Seid et al., 2019). Poor self‐care may be associated with A convenience sample of adult patients (≥18 years) with a diagnosis
higher depression, lower self‐efficacy, older age, cognitive im- of HF (based on the guidelines of the European Society of Cardiol-
pairment, and poor sleep quality—all common in HF and all ogy; Ponikowski et al., 2016) was enrolled in the study. Considering
variables known to affect HF self‐care (Kessing et al., 2016; Ryou the SCHFI v.7.2 number of items, a sample size of at least 200
et al., 2020; Sedlar et al., 2017; Vellone et al., 2020). For these participants was considered adequate for our psychometric analysis
reasons, clinicians and researchers are actively studying ways to according to the recommendations reported in the COSMIN guide-
promote self‐care and need valid and reliable instruments for this lines (Mokkink et al., 2018) and those of Anthoine et al. (2014). To be
effort. In this regard, the COnsensus‐based Standards for the eligible to participate, patients had to understand and speak Italian,
selection of health Measurement Instruments (COSMIN) group and be cognitively intact, evaluated with a Six‐Item Screener
(www.cosmin.nl) has emphasized the importance of rigorous (Callahan et al., 2002) score ≥ 4. Anyone with a coronary event in the
and transparent procedures to test validity (e.g., structural and past 3 months was excluded because patients are limited in their
construct validity) and reliability (internal consistency and performance of self‐care early in the recovery period. Data were
test–retest) of research instruments. collected between March 2017 and January 2019 by nurses speci-
One of the most commonly used instruments worldwide to fically trained for this study. When a patient was eligible, the nurse
measure HF self‐care is the Self‐Care of Heart Failure Index described the study, specifying study objectives and procedures, and
(SCHFI; Barbaranelli et al., 2014; Riegel et al., 2004, 2019). The asked the patient to sign the informed consent form. Two weeks
SCHFI is a self‐report instrument based on the situation‐specific after data collection, the SCHFI v.7.2 was re‐administered by tele-
theory of HF self‐care (Riegel & Dickson, 2008). The SCHFI was phone for test–retest reliability purposes.
originally developed in the United States and was subsequently
translated into 22 languages (https://self-care-measures.com).
The SCHFI has undergone several modifications and refinements 2.3 | Measures
over the years (Riegel & Dickson, 2008; Riegel et al., 2004, 2009,
2019; Vellone et al., 2013). The aims of these modifications were The SCHFI v.7.2 Italian version, like the original U.S. version (Riegel
to assure that behaviors measured are evidence‐based, improve et al., 2019), is a 29‐item instrument composed of three scales mea-
its psychometric characteristics, and address its theoretical un- suring self‐care maintenance, symptom perception, and self‐care
derpinnings. The last modification was performed after updating management. Each scale uses a Likert‐type option for responses.
the situation‐specific theory of HF self‐care (Riegel et al., 2016), Specifically, the self‐care maintenance scale consists of 10 items
which introduced the new concept of symptom perception. The measuring behaviors aimed at maintaining HF stability (e.g., taking
SCHFI version 7.2 reflects this theoretical modification (Riegel medications as prescribed). Responses to each item range from “never”
et al., 2019). Now the SCHFI includes the prior components of (1) to “always” (5). The Symptom Perception scale includes
self‐care maintenance (10 items) and self‐care management 11 items, nine of which evaluate how often the patient performs
(eight items) and the new component of symptom perception specific monitoring behaviors (e.g., daily weight); the remaining two
(11 items). To date, the SCHFI v.7.2 has been tested only in a U.S. items assess how quickly patients recognized and interpreted
sample (Riegel et al., 2019). However, because the prior versions HF‐related symptoms the last time they occurred. For these two items
of the SCHFI are used widely in Europe, researchers and clin- patients can choose “not applicable” (if they did not have symptoms),
icians would benefit from testing of the SCHFI v.7.2 in a “Did not recognize symptoms” (0), or one of the 5‐point Likert options
European population. Therefore, the aim of this study was to test from “Not quickly” (1) to “Very quickly” (5). Lastly, the Self‐Care
the psychometric characteristics (structural, internal consistency Management scale has eight items: the first seven items measure how
and test–retest reliability, and construct validity) of the SCHFI likely the respondent would perform some commonly used behaviors
v.7.2 in an Italian population of patients with HF. to manage HF symptoms when they occur (e.g., contacting a
642 | VELLONE ET AL.

healthcare provider for guidance in case of shortness of breath); the clinical variables (e.g., Brain natriuretic peptide (BNP) levels, New York
eighth item assesses how sure the respondent is that the remedy Heart Association (NYHA) functional class) from their clinical records.
adopted the last time symptoms occurred helped with them to feel
better. The first seven items use 5‐point Likert options for responses
from “Not likely” (1) to “Very likely” (5); the eighth item uses these 2.4 | Ethical considerations
responses: “I did not do anything” (0) or “Not sure” (1) to “Very sure” (5).
In accordance with the COSMIN guidelines (www.cosmin.nl), the The study was approved by the Institutional Review Board of one
SCHFI v.7.2 Italian version was first translated from English into center where data were collected (approval number 2424/2016) and
Italian by two doctorally prepared nurses whose native language was all participating centers approved data collection. All participants
Italian. These nurses were expert in cardiovascular disease and self‐ provided informed consent. Patients were informed that they could
care. They independently translated the instrument and then com- withdraw from the study at any time without giving an explanation.
pared their translations to agree on one final Italian version of the
SCHFI v.7.2. This Italian translation was back‐translated into English
by a bilingual English teacher with experience in medical English. A 2.5 | Statistical approach
few issues were identified in the translation and these were dis-
cussed with the instrument author (B. Riegel). For example, one issue Descriptive statistics (mean, standard deviation [SD], range, fre-
was related to the best translation of the English verb “to monitor” quency, and percentage) were used to describe participants' socio-
(used in the symptom perception scale) which could be translated demographic and clinical characteristics, SCHFI v 7.2 items, and scale
into the Italian verb “monitorare.” However, as this translation may scores.
not be completely understood by older adults, we adopted the verb Confirmatory Factor Analysis (CFA) was performed to test the
“controllare” that in English could be translated “to check” but in structural validity (dimensionality) of the SCHFI v.7.2. The SCHFI 7.2
Italian it also means “to monitor.” We then conducted a series of scales were tested separately with three CFAs, considering, at first,
cognitive interviews with 10 HF patients to check the accuracy of the the same structure identified in the original U.S. population (Riegel
intended meaning of each item. The cognitive interviews demon- et al., 2019). Specifically, for the Self‐Care Maintenance scale, we
strated that the Italian SCHFI v.7.2 items were clearly understood by hypothesized a model with two factors: Consulting Behaviors (Items
the patients. The translation and back‐translation processes and the #1, #2, #4, #5, #7, #9, and #10) and Dietary Behaviors (Items #3, #6,
cognitive interview were conducted collaboratively with the instru- and #8). For the Symptom Perception scale, we hypothesized a two‐
ment developer in 2015. This was before the psychometric proper- factor model with Monitoring Behaviors (Items #11–#19) and
ties of the U.S. version of the SCHFI v.7.2 were published in 2019. Symptom Recognition factors (Items #20 and #21). Also, as for the
Three separate scores are calculated for the three SCHFI scales. U.S. model, we hypothesized an error covariance between Items #11
Each score is standardized from 0 to 100, with higher scores and #19 because these two items address symptoms. Finally, for the
indicating better self‐care. Based on the English SCHFI v.7.2 Self‐Care Management scale we hypothesized a two‐factor model
(Riegel et al., 2019) a score ≥70 is used to indicate adequate self‐care with Recommended Behaviors (items #22–#25) and Problem‐Solving
behavior on each SCHFI scale. Behaviors (items #26–#29).
We also administered valid and reliable measures to evaluate In prior studies (Barbaranelli et al., 2014; Vellone et al., 2013) we
SCHFI v.7.2 construct validity. The Short Form Health Status found a different factorial structure in the SCHFI v.6.2 between U.S.
Questionnaire‐12 (SF‐12) is a generic instrument used to assess and European populations. Thus, anticipating a possible misfit of the
mental and physical composite scores for quality of life (Ware et al., U.S. SCHFI v.7.2 model, we considered the factorial structure of the
1996). SF‐12 scores range from 0 to 100 with higher scores meaning Caregiver Contribution to Self‐Care of Heart Failure Index 2
better quality of life. The Kansas City Cardiomyopathy Ques- (CC‐SCHFI 2; Vellone et al., 2020) as an alternative model for SCHFI
tionnaire (KCCQ) is a 23‐item disease‐specific instrument used to v.7.2 testing. The CC‐SCHFI 2 is the caregiver version of the SCHFI
assess HF patients' perception of their health status over the last v.7.2 already tested for validity and reliability in an Italian popula-
2 weeks (Green et al., 2000). This instrument assesses physical and tion. The differences between the U.S. SCHFI 7.2 and the Italian
social limitations, symptom frequency, overall quality of life, changes CC‐SCHFI 2 models are in the self‐care maintenance and self‐care
in symptom status and self‐efficacy by means of individual subscales management scales. Specifically, in the self‐care maintenance scale of
(Green et al., 2000). Beyond the individual subscales, the KCCQ also the CC‐SCHFI 2, Item #3 measuring low salt diet cross‐loaded be-
contains three combined subscales to evaluate symptom frequency tween the Consulting Behaviors and Dietary Behaviors factors. In
and symptom burden, physical limitations, and the total number of the CC‐SCHFI 2 self‐care management scale, Item #25 loaded on the
symptoms. An overall summary score is also produced. All subscales Problem Solving Behaviors factor, and Item #29 loaded on the Re-
and the summary score range from 1 to 100, with higher values commended Behaviors factor, while in the U.S. SCHFI 7.2 model, the
indicating better HF‐related quality of life. opposite occurs.
In addition to these instruments, we collected patients' socio- Since the multivariate normal distribution of the SCHFI v.7.2 items
demographic and clinical characteristics (e.g., age) by self‐report, and was not tenable, we used the maximum likelihood robust estimator for
VELLONE ET AL.
| 643

parameter estimation. We used traditional Chi‐square (χ2) as a measure T A B L E 1 Sociodemographic and clinical characteristics of the
of goodness of fit, however, since this index can be inflated by sample participants (n = 280)
size (Kline, 2010), the following alternative fit indices were also con- Variables n (%) or Mean (SD)
sidered in the interpretation of fit (Hu & Bentler, 1998; Tanaka, 1993):
Age 75.6 (10.8)
Comparative Fit Index (CFI; values of 0.90–0.95 are considered ade-
Gender
quate, while values >0.95 are considered good fit); Tucker and Lewis
Male 152 (54.3%)
Index (TLI; values of 0.90–0.95 are considered adequate, while values.
Female 128 (45.7%)
>0.95 are considered good fit; Hu & Bentler, 1999), and Standardized
Root Mean Square Residual (SRMR; values ≤0.08 indicate a good fit; Educational level (years)

Byrne, 2006; Meade et al., 2008; Vandenberg & Lance, 2000). Factor ≤8 206 (73.5%)
9–13 57 (20.4%)
loadings >|.30| were considered adequate (Comrey & Lee, 1992;
≥14 17 (6.1%)
Tabachnick & Fidell, 2006).
Internal consistency reliability of the SCHFI v.7.2 scales was Marital status
estimated with Cronbach's α coefficients and composite reliability Single/never married 19 (6.8%)
(Fornell & Larcker, 1981) or omega coefficients (MCDonalds & Married/partnered 152 (54.3%)
Mahwah, 1999). All these reliability estimates are considered ade- Divorced/separated or widowed 109 (38.9%)

quate with a value ≥0.70. Test–retest reliability of the SCHFI v.7.2 Employment
was tested with the Intraclass Correlation Coefficient. Construct Employed 240 (85.7%)
validity was tested via hypothesis testing by correlating SCHFI v.7.2 Unemployed 40 (14.3%)
scale scores with other variables known to be associated with HF Income
self‐care (quality of life, measured with the SF‐12 and the KCCQ, the More than enough to make ends meet 62 (22.1%)
BNP and the NYHA class). For these purposes, we used Pearson's r Enough to make ends meet 198 (70.7%)
or Spearman's rank correlation as appropriate, with a significant Not have enough to make ends meet 20 (7.1%)
p value set at < .05. We hypothesized that higher scores for self‐care
Age 75.6 (10.8)
maintenance, symptom perception and self‐care management would
HF duration (months) 51.6 (50.4)
be associated with better quality of life, lower levels of BNP, and
better (lower) NYHA class. Analyses were done using SPSS 21 (IBM Ejection fraction 45.11 (10.75)
Corp.) and Mplus 8.2 (Muthen & Muthen, 2004). NYHA functional class
I 64 (23.1%)
II 106 (38.3%)
3 | RESULTS III 90 (32.5%)
IV 17 (6.1%)
3.1 | Sample description BNP (median and interquartile ranges) 433 (300–650)

HF etiology
We enrolled a sample of 280 HF patients. Sociodemographic char-
Ischemic 105 (37.9%)
acteristics of participants are reported in Table 1. In brief, the sample
Non‐ischemic 93 (33.6%)
was predominantly older (mean age = 75.6; SD, 10.8), male (54.3%),
Idiopathic 71 (25.6%)
married (54.3%), and with <8 years of education (74%). Patients re-
Other 8 (2.9%)
ported having HF for 51.6 months (SD 50.4) on average, with most of
a
Comorbidities
them (70.8%) in NYHA classes II and III. The most frequent cause of HF
Hypertension 210 (75.8%)
was ischemic (37.9%), followed by non‐ischemic (33.6%) and idiopathic
Prior myocardial infarction 106 (38.3%)
(25.6%). Comorbidities were frequent with more than half having hy-
Atrial fibrillation 91 (32.9%)
pertension (75.8%), and at least one‐third with prior myocardial in-
Diabetes 85 (30.7%)
farction (38.3%) or atrial fibrillation (32.9%). Non‐cardiovascular
COPD 80 (28.9%)
comorbidities were also high; diabetes was the most common (30.7%).
Peripheral vascular disease 72 (26.0%)
Other 81 (29.2%)

3.2 | Item descriptive analysis and SCHFI v.7.2 Abbreviations: BNP, brain natriuretic peptide; CCI, Charlson
Comorbidity Index; COPD, chronic obstructive pulmonary disease;
HF, heart failure; NYHA, New York Heart Association; SD, standard
Table 2 reports the descriptive analysis of the SCHFI v 7.2 items. In deviation.
the Self‐Care Maintenance scale, the item with the highest score was a
Patients could be affected by more than one comorbidity.
644 | VELLONE ET AL.

T A B L E 2 Item descriptive analysis and

Mean SD Skewness Kurtosis
scales scores of the SCHFI v.7.2 (n = 280)
Self‐Care Maintenance Scale
Listed below are behaviors that people with heart failure use to
help themselves. How often or routinely do you do the
following?
1. Try to avoid getting sick (e.g., wash your hands) 3.97 1.06 −0.75 −0.12
2. Get some exercise (e.g., take a brisk walk, use the stairs) 2.70 1.32 0.14 −1.00
3. Eat a low‐salt diet 3.56 1.39 −0.60 −0.80
4. See the healthcare provider for routine healthcare 4.06 1.13 −1.01 0.30
5. Take prescribed medicines without missing a dose 4.29 1.00 −1.38 1.35
6. Order low‐salt items when eating out 2.50 1.36 0.35 −1.04
7. Make sure to get a flu shot annually 3.78 1.52 −0.87 −0.77
8. Ask for low‐salt foods when visiting family and friends 2.46 1.31 0.37 −0.94
9. Use a system or method to help you remember to take 2.64 1.56 0.34 −1.38
your medicines
10. Ask your healthcare provider about medicines? 3.72 1.39 −0.71 −0.72

Symptom Perception Scale

Listed below are changes that people with heart failure
commonly monitor. How often do you do the following?
11. Monitor your weight daily 2.51 1.25 0.31 −0.76
12. Pay attention to changes in how you feel 3.83 1.13 −0.69 −0.14
13. Look for medicine side effects 3.43 1.29 −0.48 −0.66
14. Notice whether you tire more than usual doing normal 3.90 1.11 −0.73 −0.11
activities
15. Ask your healthcare provider how you're doing 3.88 1.24 −0.81 −0.30
16. Monitor closely for symptoms 3.90 1.15 −0.70 −0.36
17. Check your ankles for swelling 3.67 1.26 −0.60 −0.54
18. Check for shortness of breath with activity such as 3.89 1.17 −0.79 −0.19
bathing and dressing
19. Keep a record of symptoms 2.05 1.34 0.92 −0.47
The last time you had symptoms……
20. How quickly did you recognize that you had 2.22 1.49 0.19 −0.90
symptoms?
21. How quickly did you know that the symptom was due 1.97 1.45 0.38 −0.78
to heart failure?

Self‐Care Management Scale

Listed below are behaviors that people with heart failure use to
control their symptoms. When you have symptoms, how
likely are you to use one of these?
22. Further limit the salt you eat that day 2.88 1.45 0.07 −1.22
23. Reduce your fluid intake 2.85 1.41 1.00 −1.10
24. Take a medicine? 3.29 1.51 −0.32 −1.27
25. Call your healthcare provider for guidance 3.98 1.20 −0.89 −0.19
26. Ask a family member or friend for advice 3.67 1.37 −0.68 −0.68
27. Try to figure out why you have symptoms 3.91 1.18 −0.76 −0.36
28. Limit your activity until you feel better 4.00 1.19 −0.87 −0.26
29. Think of a treatment you used the last time you had 2.09 1.41 −0.13 −0.86
symptoms. Did the treatment you used make you feel
better?

Abbreviations: SCHFI, Self‐Care of Heart Failure Index; SD, standard deviation.

Item #5 “Take prescribed medicines without missing a dose”; the whether you tire more than usual doing normal activities” and Item
item with the lowest score was Item #8 “Ask for low‐salt foods when #16 “Monitor closely for symptoms.” The item with the lowest score
visiting family and friends.” Regarding the Symptom Perception scale, was Item #21 “How quickly did you know that the symptom was due
two items were equal in having the highest scores: Item #14 “Notice to heart failure?”. In the Self‐Care Management scale, the item with
VELLONE ET AL.
| 645

TABLE 3 CFA fit indices of the tested SCHFI v.7.2 models

χ2 (p value) df CFI TLI SRMR RMSEA (90% confidence interval) p (RMSEA)

Self‐Care Maintenance Scale

U.S. model 81.027 (<.001) 34 0.900 0.868 0.068 0.070 (0.051; 0.090) (.045)
CC‐SCHFI 2 model 57.061 (.006) 33 0.949 0.931 0.050 0.051 (0.027; 0.073) (.443)

Symptom Perception Scale

U.S. model 84.640 (<.001) 42 0.941 0.922 0.046 0.060 (0.041; 0.079) (.173)

Self‐Care Management Scale

U.S. model 145.56 (<.001) 19 0.759 0.644 0.118 0.154 (0.131; 0.178) (.001)
CC‐SCHFI 2 model 45.631 (<.001) 19 0.949 0.925 0.053 0.071 (0.045; 0.097) (.091)

Abbreviations: χ2, Chi square test; df, degree of freedom; CFA, Confirmatory Factor Analysis; CFI, Comparative Fit Index; RMSEA, root mean square error
of approximation; CC‐SCHFI, Caregiver Contribution to Self‐Care of Heart Failure Index; SRMR, standardized root mean square residual; TLI, Tucker and
Lewis Index.

the highest score was #28 “Limit your activity until you feel better”; goodness‐of‐fit indices (Table 3). As planned, the CC‐SCHFI 2 model
the item with the lowest score was Item #29 “Did the treatment you was used for testing, which resulted with supportive goodness‐of‐fit
used make you feel better?” Several items were non‐normally dis- indices (Table 3). The two factors, “consulting behaviors” and “dietary
tributed, with skewness and kurtosis > |1|. behaviors,” were positively correlated at 0.477. All factor loadings
were significant and > |0.30| except for Item #2 (“Get some ex-
ercise”) which had a low but significant loading of 0.147 (Figure 1).
3.3 | Psychometric analysis of the Self‐Care
Maintenance Scale
3.3.2 | Scale internal consistency reliability
3.3.1 | Structural validity (dimensionality)
Cronbach's α coefficient of the Self‐Care Maintenance Scale
Preliminary CFA testing of the factorial structure of the Self‐Care was .733. As Cronbach's α coefficient assumes a unidimensional
Maintenance Scale, adopting the U.S. model, had unsatisfactory structure and the Self‐care Maintenance scale has two dimensions,

TABLE 4 Construct validity of the SCHFI v.7.2 self‐care scales

Self‐Care Symptom Perception Self‐Care
Variables Maintenance Scale Scale Management Scale

SF‐12
PCS 0.054 (0.364) 0.143 (0.016) 0.183 (0.002)
MCS 0.134 (0.025) 0.156 (0.009) 0.237 (<0.001)

KCCQ
Symptom Frequency 0.204 (0.001) 0.261 (<0.001) 0.338 (<0.001)
Symptom Burden 0.102 (0.087) 0.107 (0.074) 0.338 (<0.001)
Total Symptom 0.159 (0.008) 0.189 (0.001) 0.364 (<0.001)
Physical Limitation 0.069 (0.251) 0.186 (0.002) 0.184 (<0.002)
Quality of Life 0.066 (0.274) 0.120 (0.045) 0.245 (<0.001)
Social Limitation 0.048 (0.422) 0.147 (0.014) 0.186 (0.002)
Self‐Efficacy Score 0.225 (<0.001) 0.291 (<0.001) 0.252 (<0.001)
Symptom Stability 0.073 (0.225) 0.049 (0.410) 0.167 (0.005)
Clinical Summary 0.120 (0.045) 0.202 (0.001) 0.289 (<0.001)
Overall Summary Score 0.091 (0.129) 0.177 (0.003) 0.266 (<0.001)

NYHA functional class −0.025 (0.679) 0.006 (0.927) −0.198 (0.001)

BNP −0.189 (0.025) −0.045 (0.598) −0.220 (0.008)

Note: All correlations were calculated with Pearson's r, with the exception of BNP and NYHA class correlations, which were calculated with Spearman's
rank coefficient. The numbers in parentheses after the correlation coefficients are the p values.
Abbreviations: BNP, brain natriuretic peptide; MCS, Mental Health Composite Scale; NYHA, New York Heart Association; PCS, Physical Health
Composite Scale; SF‐12, Short Form Health Questionnaire‐12.
646 | VELLONE ET AL.

FIGURE 1 Confirmatory factor analysis of the self‐care maintenance scale; i = item

we computed the composite reliability coefficient, which takes structure was tested with the CC‐SCHFI 2 model, the fit indices
multidimensionality into account. Composite reliability coefficient improved (Table 3). The correlation between the two factors was
was 0.809. 0.432 and all factor loadings were significant (Figure 3).

3.4 | Psychometric analysis of the Symptom 3.5.2 | Scale internal consistency reliability
Perception Scale
Cronbach's α coefficient for the Self‐Care Management scale
3.4.1 | Structural validity (dimensionality) was .802 and composite reliability coefficient was 0.888.

Preliminary CFA testing of the Symptom Perception scale with the

original CFA U.S. model, revealed supportive fit indices (Table 3). The 3.6 | Test–retest reliability
factors “Monitoring Behaviors” and “Symptom Recognition” were
significantly correlated at 0.498. All factor loadings were significant Test–retest reliability was tested with the intraclass correlation
and high, excepted for Item #19 (“Keep a record of symptoms”) coefficient and found to be 0.802, 0.734, and 0.916 for the Self‐Care
which had a low but significant factor loading of 0.196 (Figure 2). Maintenance, Symptom Perception, and Self‐Care Management
Scales, respectively.

3.4.2 | Scale internal consistency reliability

3.7 | Construct validity
Cronbach's α coefficients for the Symptom Perception Scale
was .830 and composite reliability coefficient was 0.882. Construct validity was supported by the presence of several corre-
lations between the SCHFI v.7.2 and other measures that were weak
but exhibited p < .05 (Table 3). The Self‐Care Maintenance scale
3.5 | Psychometric analysis of the Self‐Care scores were significantly correlated with the SF‐12 mental composite
Management Scale score, several dimensions of the KCCQ, and the BNP values. The
symptom perception scale scores were significantly correlated with
3.5.1 | Structural validity (dimensionality) the SF‐12 physical and mental composite scores, and almost all the
dimensions of the KCCQ. The self‐care management scale scores
Preliminary CFA testing of the self‐care management using the ori- were significantly correlated with the SF‐12 scores, all the dimen-
ginal U.S. model had inadequate fit (Table 3). When the factorial sions of the KCCQ, BNP, and NYHA class.
VELLONE ET AL.
| 647

FIGURE 2 Confirmatory factor analysis of the symptom perception scale; i = item

FIGURE 3 Confirmatory factor analysis of the self‐care management scale; i = item

648 | VELLONE ET AL.

4 | DI SCUSSION Such differences suggest that Italians may be more likely to involve
healthcare providers in problem‐solving than Americans. It is also
The aim of this study was to test the psychometric characteristics possible that the sociodemographic and clinical differences between
(structural and construct validity, internal consistency and the U.S. and Italian samples might have influenced these differences
test–retest reliability,) of the SCHFI v.7.2 in an Italian population of in self‐care management, since patients who are younger, better
patients with HF. We found that this instrument is valid and reliable educated, with higher income and a longer duration of HF might be
to measure HF self‐care in the Italian population. To our knowledge, better at problem‐solving.
this is only the second study testing of the SCHFI v.7.2 psychometric Results of this study supported good internal consistency and
characteristics and the first one conducted in a European population. test–retest reliability (stability) for the SCHFI v.7.2. We used both
Our findings provide additional support for validity and reliability of classic methods and those advocated for multidimensional scales
the SCHFI v.7.2. (Barbaranelli et al., 2015). Both approaches resulted in adequate
Regarding the SCHFI v.7.2 dimensionality, our tested model was reliability. Reliability pertains to instrument precision when mea-
in general, similar to the U.S. model, but with some minor albeit suring a variable and our psychometric analysis has shown that all
interesting differences. In the self‐care maintenance scale, Item #3 three SCHFI v.7.2 scales are precise in measuring self‐care main-
addressing tracking of a low salt diet also loaded on consulting be- tenance, symptom perception, and self‐care management. Precision
haviors factor. This result was not completely unexpected because in measurement has also been confirmed by testing stability with
we saw the same kind of response in earlier testing of the Italian test–retest methods, which were supportive as well. This aspect is
version of the SCHFI v.6.2. In that analysis, the item related to salt particularly useful for clinicians and researchers because an increase
loaded on the factor measuring “medical prescription.” This differ- or a decrease in SCHFI v.7.2 scale scores can be assumed to not
ence between the U.S. and Italian patients in performing self‐care occur by chance.
maintenance could be due to a difference in perceived salt restriction Our analysis also supported construct validity via hypothesis
in the diet; in the United States, salt restriction is a dietary choice, testing. We found significant correlations with generic and
while in Italy salt restriction is done in response to a healthcare disease‐specific quality of life dimensions on the KCCQ and with
provider's order. Indeed, it is very common in Italy that dietary salt NYHA functional class and BNP level. These correlations, even
reduction is “prescribed” by a provider. Another explanation for the though not very strong, support SCHFI v.7.2 validity because
difference between the U.S. and the Italian self‐care maintenance prior studies have shown that better self‐care is associated with
model might be related to sample differences in sociodemographic better patient quality of life (Seid, 2020; Vellone et al., 2017), a
and clinical variables. The U.S. sample was 10 years younger, better decrease in BNP level (Moon et al., 2018), and better functional
educated, with higher income and longer HF duration compared to class (lower NYHA; Vellone et al., 2017). Indirectly, the results of
the Italian sample. These factors may be associated with higher this study further support our premise that self‐care improves
autonomy in self‐care, including dietary salt restriction. patient outcomes.
In the symptom perception scale, testing the U.S. model resulted Several limitations should be acknowledged. First, we enrolled
in supportive fit indices. As in the U.S. model, we allowed the re- HF patients in only one European country and used convenience
siduals of Items #1 and #19 to covary because these two items have sampling. Consequently, we recommend that data on the factorial
something in common: Item #1 is related to monitoring body weight, structure of the SCHFI v.7.2 be generalized with caution to other
and Item #19 is related to keeping a record of symptoms. HF patients countries. However, the similarities we found between the Italian
are advised to monitor and keep a record of their weight daily, and the U.S. models and the construct validity are key elements
therefore the correlation between the residuals of these two items is supporting the psychometric characteristics of the SCHFI v.7.2
justified. The similarity between the U.S. and the Italian model in the across populations.
symptom perception scale is interesting because it suggests that, In conclusion, this study provides additional evidence that the
despite different cultural orientations, HF patients have the same SCHFI v.7.2 is a psychometrically sound instrument that has good
approach to monitoring and recognizing symptoms. psychometric characteristics of validity and reliability not only in
In our model, Item #25 of the self‐care management scale United States but also in Italy. The SCHFI v.7.2 may provide a useful
(measuring how likely the patient is to call the healthcare provider tool for clinicians to monitor HF self‐care and for researchers to
with symptoms) and #29 (evaluating the behavior used to treat the evaluate the effectiveness of interventions aimed at improving
symptoms) loaded on different factors from the original SCHFI v.7.2 self‐care.
U.S. model. This is interesting because it might illustrate two dif-
ferent ways in which Italian and U.S. HF patients behave in response ACKNOWLEDGM E NT
to symptoms of a HF exacerbation. In our tested model, treatment This study was funded by the Center of Excellence for Nursing
evaluation is considered a recommended behavior together with Scholarship, Rome.
limiting salt and fluids and taking a diuretic. In contrast, calling the
provider, asking for advice, trying to understand what is happening CO N FLI CT O F I N TER E S TS
and limiting activities are considered problem‐solving behaviors. The authors declare that there are no conflict of interests.
VELLONE ET AL.
| 649

ORCID review and meta‐analysis. Psychosomatic Medicine, 78(4), 412–431.

Ercole Vellone https://orcid.org/0000-0003-4673-7473 https://doi.org/10.1097/PSY.0000000000000270
Kline, R. (2010). Principles and Practice of Structural Equation Modeling. The
Paolo Iovino https://orcid.org/0000-0001-5952-881X
Guilford Press.
Barbara Riegel https://orcid.org/0000-0002-0970-136X McDonalds, R. J. I., & Mahwah, N. J. (1999). Test theory: A unified
treatment. In Test theory: A unified treatment (pp. xi, 485–xi, 485).
R EF E RE N C E S Lawrence Erlbaum Associates Publishers.
Meade, A. W., Johnson, E. C., & Braddy, P. W. (2008). Power and
Anthoine, E., Moret, L., Regnault, A., Sébille, V., & Hardouin, J. B. (2014).
sensitivity of alternative fit indices in tests of measurement
Sample size used to validate a scale: A review of publications on newly‐
invariance. Journal of Applied Psychology, 93(3), 568–592. https://
developed patient reported outcomes measures. Health and Quality of
doi.org/10.1037/0021-9010.93.3.568
Life Outcomes, 12, 176. https://doi.org/10.1186/s12955-014-0176-2
Mokkink, L. B., de Vet, H., Prinsen, C., Patrick, D. L., Alonso, J.,
Barbaranelli, C., Lee, C. S., Vellone, E., & Riegel, B. (2014). Dimensionality
Bouter, L. M., & Terwee, C. B. (2018). COSMIN Risk of Bias checklist
and reliability of the self‐care of heart failure index scales: Further
for systematic reviews of Patient‐Reported Outcome Measures.
evidence from confirmatory factor analysis. Research in Nursing and
Quality of Life Research, 27(5), 1171–1179. https://doi.org/10.1007/
Health, 37(6), 524–537. https://doi.org/10.1002/nur.21623
s11136-017-1765-4
Barbaranelli, C., Lee, C. S., Vellone, E., & Riegel, B. (2015). The problem
Moon, M. K., Yim, J., & Jeon, M. Y. (2018). The effect of a telephone‐based
with Cronbach's alpha: Comment on Sijtsma and van der Ark.
self‐management program led by nurses on self‐care behavior,
Nursing Research, 64(2), 140–145. https://doi.org/10.1097/NNR.
biological index for cardiac function, and depression in ambulatory
0000000000000079
heart failure patients. Asian Nursing Research, 12(4), 251–257.
Buck, H. G., Lee, C. S., Moser, D. K., Albert, N. M., Lennie, T., Bentley, B.,
https://doi.org/10.1016/j.anr.2018.10.001
Worrall‐Carter, L., & Riegel, B. (2012). Relationship between self‐
Muthen, L. K., & Muthen, B. O. (2004). Mplus User's Guide (8th ed). Muthén
care and health‐related quality of life in older adults with moderate
& Muthén.
to advanced heart failure. The Journal of Cardiovascular Nursing,
Ponikowski, P., Voors, A. A., Anker, S. D., Bueno, H., Cleland, J. G. F.,
27(1), 8–15. https://doi.org/10.1097/JCN.0b013e3182106299
Coats, A. J. S., Falk, V., González‐Juanatey, J. R., Harjola, V. P.,
Byrne, B. M. (2006). Structural Equation Modeling with EQS; Basic Concepts,
Jankowska, E. A., Jessup, M., Linde, C., Nihoyannopoulos, P.,
Applications, and Programming (2nd ed). Lawrence Erlbaum
Parissis, J. T., Pieske, B., Riley, J. P., Rosano, G. M. C., Ruilope, L. M.,
Associates Publishers.
Ruschitzka, F., … van der Meer, P. (2016). 2016 ESC Guidelines for
Callahan, C. M., Unverzagt, F. W., Hui, S. L., Perkins, A. J., & Hendrie, H. C.
the diagnosis and treatment of acute and chronic heart failure: The
(2002). Six‐item screener to identify cognitive impairment among
Task Force for the diagnosis and treatment of acute and chronic
potential subjects for clinical research. Medical Care, 40(9), 771–781.
heart failure of the European Society of Cardiology (ESC)Developed
https://doi.org/10.1097/00005650-200209000-00007
with the special contribution of the Heart Failure Association (HFA)
Cocchieri, A., Riegel, B., D'Agostino, F., Rocco, G., Fida, R., Alvaro, R., &
of the ESC. European Heart Journal, 37(27), 2129–2200. https://doi.
Vellone, E. (2015). Describing self‐care in Italian adults with heart
org/10.1093/eurheartj/ehw128
failure and identifying determinants of poor self‐care. European
Reilly, C. M., Butler, J., Culler, S. D., Gary, R. A., Higgins, M., Schindler, P.,
Journal of Cardiovascular Nursing, 14(2), 126–136. https://doi.org/10.
Butts, B., & Dunbar, S. B. (2015). An economic evaluation of a self‐
1177/1474515113518443
care intervention in persons with heart failure and diabetes. Journal
Comrey, A. L., & Lee, H. B. (1992). A First Course in Factor Analysis (2nd ed).
of Cardiac Failure, 21(9), 730–737. https://doi.org/10.1016/j.cardfail.
Lawrence Erlbaum Associates, Inc.
2015.06.382
Danielsen, R., Thorgeirsson, G., Einarsson, H., Ólafsson, Ö., Aspelund, T.,
Riegel, B., Barbaranelli, C., Carlson, B., Sethares, K. A., Daus, M.,
Harris, T. B., Launer, L., & Gudnason, V. (2017). Prevalence of heart
Moser, D. K., Miller, J., Osokpo, O. H., Lee, S., Brown, S., & Vellone, E.
failure in the elderly and future projections: the AGES‐Reykjavík
(2019). Psychometric testing of the Revised Self‐Care of Heart
study. Scandinavian Cardiovascular Journal, 51(4), 183–189. https://
Failure Index. The Journal of Cardiovascular Nursing, 34(2), 183–192.
doi.org/10.1080/14017431.2017.1311023
https://doi.org/10.1097/JCN.0000000000000543
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models
Riegel, B., Carlson, B., Moser, D. K., Sebern, M., Hicks, F. D., & Roland, V.
with unobservable variables and measurement error. Journal of
(2004). Psychometric testing of the self‐care of heart failure index.
Marketing Research, 18(1), 39–50. https://doi.org/10.2307/3151312
Journal of Cardiac Failure, 10(4), 350–360. https://doi.org/10.1016/j.
Gouveia, M., Ascenção, R., Fiorentino, F., Costa, J., Caldeira, D., Broeiro‐
cardfail.2003.12.001
Gonçalves, P., Fonseca, C., & Borges, M. (2019). The current and
Riegel, B., & Dickson, V. V. (2008). A situation‐specific theory of heart
future burden of heart failure in Portugal. ESC Heart Failure, 6(2),
failure self‐care. The Journal of Cardiovascular Nursing, 23(3),
254–261. https://doi.org/10.1002/ehf2.12399
190–196. https://doi.org/10.1097/01.JCN.0000305091.35259.85
Green, C. P., Porter, C. B., Bresnahan, D. R., & Spertus, J. A. (2000).
Riegel, B., Dickson, V. V., & Faulkner, K. M. (2016). The situation‐specific
Development and evaluation of the Kansas City Cardiomyopathy
theory of heart failure self‐care: Revised and updated. The Journal of
Questionnaire: A new health status measure for heart failure.
Cardiovascular Nursing, 31(3), 226–235. https://doi.org/10.1097/
Journal of the American College of Cardiology, 35(5), 1245–1255.
JCN.0000000000000244
https://doi.org/10.1016/s0735-1097(00)00531-3
Riegel, B., Lee, C. S., Dickson, V. V., & Carlson, B. (2009). An update on the
Hu, L.‐t, & Bentler, P. M. (1998). Fit indices in covariance structure
self‐care of heart failure index. The Journal of Cardiovascular Nursing,
modeling: Sensitivity to underparameterized model misspecification.
24(6), 485–497. https://doi.org/10.1097/JCN.0b013e3181b4baa0
Psychological Methods, 3(4), 424–453.
Ruppar, T. M., Cooper, P. S., Mehr, D. R., Delgado, J. M., & Dunbar‐
Hu, L.t, & Bentler, P. M. (1999). Cutoff criteria for fit indexes in
Jacob, J. M. (2016). Medication adherence interventions
covariance structure analysis: Conventional criteria versus new
improve heart failure mortality and readmission rates:
alternatives. Structural Equation Modeling: A Multidisciplinary Journal,
Systematic review and meta‐analysis of controlled trials.
6(1), 1–55. https://doi.org/10.1080/10705519909540118
Journal of the American Heart Association, 5(6). https://doi.org/
Kessing, D., Denollet, J., Widdershoven, J., & Kupper, N. (2016).
10.1161/JAHA.115.002606
Psychological determinants of heart failure self‐care: Systematic
650 | VELLONE ET AL.

Ryou, C., Kang, S. M., & Jang, Y. (2020). Factors associated with self‐care contribution to Self‐Care of Heart Failure Index Version 2. The
behaviours among Koreans with heart failure. European Journal of Journal of Cardiovascular Nursing, 35(3), 280–290. https://doi.org/10.
Cardiovascular Nursing. Advance online publication. https://doi.org/10. 1097/JCN.0000000000000655
1177/1474515120934060 Vellone, E., Fida, R., Ghezzi, V., D'Agostino, F., Biagioli, V., Paturzo, M.,
Savarese, G., & Lund, L. H. (2017). Global public health burden of heart failure. Strömberg, A., Alvaro, R., & Jaarsma, T. (2017). Patterns of self‐care
Cardiac Failure Review, 3(1), 7–11. https://doi.org/10.15420/cfr. in adults with heart failure and their associations with
2016:25:2 sociodemographic and clinical characteristics, quality of life, and
Sedlar, N., Lainscak, M., Mårtensson, J., Strömberg, A., Jaarsma, T., & hospitalizations: A cluster analysis. Journal of Cardiovascular Nursing,
Farkas, J. (2017). Factors related to self‐care behaviours in heart 32(2), 180–189. https://doi.org/10.1097/JCN.0000000000000325
failure: A systematic review of European Heart Failure Self‐Care Vellone, E., Riegel, B., Cocchieri, A., Barbaranelli, C., D'Agostino, F.,
Behaviour Scale studies. European Journal of Cardiovascular Nursing, Antonetti, G., Glaser, D., & Alvaro, R. (2013). Psychometric testing of
16(4), 272–282. https://doi.org/10.1177/1474515117691644 the Self‐Care of Heart Failure Index Version 6.2. Research in Nursing
Seid, M. A. (2020). Health‐related quality of life and extent of self‐care and Health, 36(5), 500–511. https://doi.org/10.1002/nur.21554
practice among heart failure patients in Ethiopia. Health and Quality von Lueder, T. G., & Agewall, S. (2018). The burden of heart failure in the
of Life Outcomes, 18(1), 27. https://doi.org/10.1186/s12955-020- general population: A clearer and more concerning picture. Journal of
01290-7 Thoracic Disease, 10(Suppl 17), S1934–S1937. https://doi.org/10.21037/
Seid, M. A., Abdela, O. A., & Zeleke, E. G. (2019). Adherence to self‐care jtd.2018.04.153
recommendations and associated factors among adult heart failure Wang, S. P., Lin, L. C., Lee, C. M., & Wu, S. C. (2011). Effectiveness of a self‐
patients. From the patients' point of view. PLOS One, 14(2): care program in improving symptom distress and quality of life in
e0211768. https://doi.org/10.1371/journal.pone.0211768 congestive heart failure patients: A preliminary study. The Journal of
Tabachnick, B. G., & Fidell, L. S. (2006). Using Multivariate Statistics (5th ed). Nursing Research, 19(4), 257–266. https://doi.org/10.1097/JNR.
Allyn & Bacon, Inc. 0b013e318237f08d
Tanaka, J. S. (1993). Multifaceted conceptions of fit in structural equation Ware, J., Jr., Kosinski, M., & Keller, S. D. (1996). A 12‐Item Short‐Form
models. In Bollen K. A., & Long J. S. (Eds.), Testing structural equation Health Survey: Construction of scales and preliminary tests of
models (Vol. 154, pp. 10–39). reliability and validity. Medical Care, 34(3), 220–233. https://doi.org/
Toukhsati, S. R., Driscoll, A., & Hare, D. L. (2015). Patient Self‐ 10.1097/00005650-199603000-00003
management in chronic heart failure—Establishing concordance
between guidelines and practice. Cardiac Failure Review, 1(2),
128–131. https://doi.org/10.15420/cfr.2015.1.2.128
Vandenberg, R., & Lance, C. (2000). A review and synthesis of the How to cite this article: Vellone E, De Maria M, Iovino P,
measurement invariance literature: Suggestions, practices, and et al. The Self‐Care of Heart Failure Index version 7.2:
recommendations for organizational research. Organizational Research
Further psychometric testing. Res Nurs Health. 2020;43:
Methods, 3, 4–69. https://doi.org/10.1177/109442810031002
Vellone, E., Barbaranelli, C., Pucciarelli, G., Zeffiro, V., Alvaro, R., & 640–650. https://doi.org/10.1002/nur.22083
Riegel, B. (2020). Validity and reliability of the caregiver

View publication stats