Secondary Logo

Journal Logo

Construct Validity of the Chinese Version of the Self-care of Heart Failure Index Determined Using Structural Equation Modeling

Kang, Xiaofeng PhD, RN; Dennison Himmelfarb, Cheryl R. PhD, RN, ANP, FAHA; Li, Zheng PhD, RN; Zhang, Jian MD; Lv, Rong RN; Guo, Jinyu RN

Journal of Cardiovascular Nursing: May/June 2015 - Volume 30 - Issue 3 - p 222–228
doi: 10.1097/JCN.0000000000000134
ARTICLES
Free

Background: The Self-care of Heart Failure Index (SCHFI) is an empirically tested instrument for measuring the self-care of patients with heart failure.

Objective: The aim of this study was to develop a simplified Chinese version of the SCHFI and provide evidence for its construct validity.

Methods: A total of 182 Chinese with heart failure were surveyed. A 2-step structural equation modeling procedure was applied to test construct validity.

Results: Factor analysis showed 3 factors explaining 43% of the variance. Structural equation model confirmed that self-care maintenance, self-care management, and self-care confidence are indeed indicators of self-care, and self-care confidence was a positive and equally strong predictor of self-care maintenance and self-care management. Moreover, self-care scores were correlated with the Partners in Health Scale, indicating satisfactory concurrent validity.

Conclusion: The Chinese version of the SCHFI is a theory-based instrument for assessing self-care of Chinese patients with heart failure.

Xiaofeng Kang, PhD, RN Nurse Researcher, School of Nursing, Peking Union Medical College, Beijing, China.

Cheryl R. Dennison Himmelfarb, PhD, RN, ANP, FAHA Associate Professor, School of Nursing and School of Medicine, Johns Hopkins University, Baltimore, Maryland.

Zheng Li, PhD, RN Professor, School of Nursing, Peking Union Medical College, Beijing, China.

Jian Zhang, MD Director, Cardiovascular Institute and Fuwai Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences, Beijing, China.

Rong Lv, RN Head Nurse, Cardiovascular Institute and Fuwai Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences, Beijing, China.

Jinyu Guo, RN Registered Nurse, Mphil student, School of Nursing, Peking Union Medical College, Beijing, China.

This research was supported by a grant from China Medical Board (no. 04-797) and Youth Foundation, Peking Union Medical College (no. 2009001).

The authors have no conflicts of interest to disclose.

Correspondence Zheng Li, PhD, Peking Union Medical College School of Nursing, 9 Dong Dan San Tiao, Beijing, 100730, China (zhengli@hotmail.com).

Heart failure is an increasing public health problem in the western world.1 In 2001, the prevalence of heart failure in China was approximately 0.9%,2 and the prevalence is increasing. Self-care has been described as a nonpharmacologic strategy in the management of chronic heart failure (CHF).3 To evaluate the efficacy of interventions that promote self-care skill development, valid and reliable instruments measuring self-care are needed in practice and research. The Self-care of Heart Failure Index (SCHFI) has been used to specifically measure CHF self-care.4 It assesses multidimensional components of self-care: self-care maintenance (behaviors undertaken to maintain health), self-care management (the decision-making process with regard to symptom changes), and self-care confidence (confidence in managing symptoms and evaluating any actions implemented).5 The SCHFI was selected for its properties as a useful instrument for assessing the decision-making process of CHF patients.

The original SCHFI is a self-report index comprising 15 four-point items,6 and it was revised to a 22-item instrument in 2009.5 The SCHFI v. 6 is available in many languages (English, Italian, Spanish, Dutch, Portuguese, Japanese, and Persian). The purpose of this research project was to translate the SCHFI v. 6 into simplified Chinese (C-SCHFI) and to test the validity and reliability of the C-SCHFI in people with CHF who live in mainland China.

Back to Top | Article Outline

Methods

Design, Settings, and Sample

A descriptive, cross-sectional study was conducted at 3 units of Fuwai cardiovascular hospital (the National Center for Cardiovascular Diseases, China). Patients were recruited if they had CHF confirmed by a cardiologist. Patients who were younger than 18 years or who had a history of psychiatric diseases, acute heart failure, or severe cerebral vascular diseases were excluded.

Back to Top | Article Outline

Instruments

Self-care of Heart Failure Index

The C-SCHFI was translated from the revised 22-item English version.5 The 22 items are divided into 3 scales: (1) self-care maintenance (10 items), (2) self-care management (6 items), and (3) self-care self-confidence (6 items). Each scale score is standardized to a range of 0 to 100 and a score of 70 or greater can be used as the cut-point to judge self-care adequacy. Reliability of the SCHFI was adequate. Construct validity was supported with satisfactory model fit on confirmatory factor analysis (CFA) (root-mean-square error of approximation [RMSEA], 0.07; comparative fit index [CFI], 0.73).

The written language of Chinese can be categorized as traditional or simplified. In mainland China, simplified Chinese is used, and traditional Chinese is common in Taiwan and Hong Kong. The 18-item SCHFI in traditional Chinese was tested in Hong Kong.7 In the current project, the 22-item SCHFI was translated to simplified Chinese for users living in mainland China.

With the agreement of the original developer of the index, the C-SCHFI was developed via forward- and back-translation according to standard translation processes.8 Consequently, the following items were modified to suit the healthcare circumstance in mainland China. Item 5 was changed from “Keep your doctor or nurse appointments” to “Keep follow-up appointments” and item 15 was changed from “Call your doctor or nurse for guidance” to “Consult a doctor or nurse for guidance.”

After forward- and back-translation, 10 experts reviewed the instrument for grammar, syntax, organization, and appropriateness. These experts included 4 cardiologists, 4 nurses-in-charge who worked at heart failure units, 1 nurse educator, and 1 professor from the School of Nursing. They confirmed that the instrument appeared to flow logically and indicated that it would be useful in identifying decision-making difficulties with self-care and potential interventions, which suggested good face validity.

The C-SCHFI then was evaluated for comprehensibility and acceptability by 10 CHF patients. The mean (SD) age of participants was 59.2 (12.6) years, and 7 were men. The subjects reported having no difficulty in answering or interpreting the questions. There were no items missing in the answer scale. The time taken to complete the instrument was about 8 minutes.

To confirm content validity, the 10 experts rated all items for the expert validity index analysis. Item rating was typically on a 4-point ordinal scale: 1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, 4 = highly relevant. For each item, the content validity index for items (I-CVI) was computed as the number of experts giving a rating of either 3 or 4, divided by the total number of experts. Content validity index for scales (S-CVI), which places the focus on average item quality rather than on average performance by the experts, was calculated as the average I-CVI value.9 Based on Lynn’s10 criteria (I-CVI = 1.00 with 3 to 5 experts and a minimum I-CVI of 0.78 for 6 to 10 experts), acceptable level of interrater agreement was set at S-CVI of 0.90 or higher.9 All C-SCHFI items with I-CVI ranged 0.80 or higher, including 11 items (50%) yielding 100% agreement, and the S-CVI was 0.94, indicating that the content of the simplified Chinese version of the SCHFI was acceptable.11

Back to Top | Article Outline

Partners in Health Scale

To confirm the criterion-related validity of the C-SCHFI, the Partners in Health (PIH) Scale was used in evaluating the concept linked with self-management theoretically. The PIH Scale was developed using the 6 domains of self-management of chronic conditions by Flinders University. The PIH Scale consists of 12 items with a 9-point Likert-type response scale, where 0 means “strongly satisfied” and 8 means “strongly dissatisfied.” Internal consistency, assessed as Cronbach’s α, was .82. Factor analysis revealed 4 key factors with an explained variance of 66.79%.12 From these results, the PIH Scale was considered to be a valid and reliable scale. Scores on the PIH Scale were hypothesized to be moderately related to the SCHFI scores.

Back to Top | Article Outline

Minnesota Living With Heart Failure Questionnaire

To further evaluate the discriminant validity, we used the Minnesota Living With Heart Failure Questionnaire (MLHFQ) concerning “patient’s perceptions of the effects of heart failure and its treatment on his or her daily life.” This questionnaire is a 21-item, disease-specific measure of the quality of life for patients with heart failure, and psychometric properties of the Chinese version MLHFQ were excellent.13 If the C-SCHFI measure of self-care has discriminant validity, then it would not be related to the measure of the quality of life.

Back to Top | Article Outline

Data Collection Procedures

Institutional review board approval was obtained from the university and participating hospital. Nurses working in cardiovascular wards introduced the principal researcher to potential subjects. All subjects were given information about the survey, and written informed consent was obtained before enrollment. Patients who met the criteria were selected and asked to answer the self-administered C-SCHFI, PIH Scale, MLHFQ, and demographic questionnaire that contained information about age, gender, education level, and diagnosis. Clinical characteristics of the subjects were collected from their medical records. The time required to complete the questionnaires was 15 to 20 minutes.

Back to Top | Article Outline

Data Analysis

Reliability

Internal consistency was assessed with the Cronbach’s α coefficient. The estimates of internal consistency should be above .70 but below .90.14 Item analysis was performed by evaluating the item-total correlations. In a reliable instrument, items with an item-total correlation below 0.10 or above 0.70 should be evaluated, revised, or deleted.15 An item-total correlation of greater than 0.30 suggests that the item contributes to the scale.16

Back to Top | Article Outline

Validity

Construct validity was estimated through both exploratory factor analysis (EFA) and structural equation modeling (SEM). That factors might be correlated is a reasonable assumption in the health sciences. There are 3 factors in the model of heart failure self-care process, which, although distinct, are also correlated to some extent. Therefore, EFA was done using principal axis factoring with direct oblimin rotation. Factor loadings without extensive cross-loading on other factors were retained in the factor structure. The criterion of an eigenvalue greater than 1.0 was used to extract factors.

In agreement with Kline,17 a 2-step SEM procedure was applied. In the first step, a CFA was conducted to determine whether the intended construct was indeed measured by the underlying latent variables.18 The first-order factor model assumed each manifest variable to be a distinct indicator of an underlying latent construct, whereby different constructs were permitted to be intercorrelated. The second-order factor model was used to validate all first-order factors as indicators of an underlying latent construct. The appropriateness of a specific CFA model was determined with global fit measures.

Second, a corresponding SEM was based on the conceptual model of heart failure self-care that the self-care process included the following: self-care maintenance focused on symptom monitoring and treatment adherence (stage 1), self-care management focused on recognizing and responding to symptoms (stage 2), and confidence influenced the self-care process.19 The significance of the relationships between the exogenous latent variable (self-care confidence) and endogenous latent variables (self-care maintenance and self-care management) and the amount of variance explained in the endogenous variables were examined.

The SEM was set according to modification indices suggestion. Information produced from the LISREL 8.70 (Scientific Software International, Chicago, Illinois) output provided modifications to the SEM to improve the fit between the proposed model and the data. Once modifications were made, improvements to the proposed model were assessed by widely accepted fit measures, which included χ2 (χ2/df: <3 acceptable, <2 excellent), RMSEA (<0.05 excellent, <0.08 acceptable), CFI (>0.90 acceptable, >0.95 excellent), and goodness-of-fit index (GFI; >0.90 acceptable, >0.95 excellent).20

To further evaluate construct validity, Pearson rank correlation coefficient between the scores from C-SCHFI and the MLHFQ were calculated. Criterion-related validity was established by Pearson rank correlation coefficient between the scores from the C-SCHFI and the PIH Scale. The PIH Scale score was hypothesized to be moderately and negatively (because they are reversed) related to the C-SCHFI scores.

Back to Top | Article Outline

Results

A total of 200 patients were enrolled between June and September 2011, and 18 asymptomatic patients have no self-care management scores. Therefore, data were available on 182 patients who were symptomatic in the month before enrollment.

The sample had a mean (SD) age of 55.8 (15.0) years, 71% were men, 74% had less than a middle school education, 31% were employed, and 79% were New York Heart Association classification III/IV. The mean (SD) number of years since diagnosis of heart failure was 3.3 (4.7).

Back to Top | Article Outline

Internal Consistency and Item-Total Correlations

The internal consistency reliability for the C-SCHFI was adequate: self-care maintenance, α = .66; self-care management, α = .74; and self-care self-confidence, α = .87. The Cronbach’s α if-item deleted ranged from .82 to .85, indicating that no item was predicted to significantly raise the α coefficient if deleted. Corrected item-total correlations of 22 items were above 0.30, indicating that the items were adequately homogeneous. The corrected item-total correlations for the first 10 items that loaded on the self-care maintenance ranged from 0.30 to 0.58. The next 6 items on self-care management were 0.50 to 0.79, and the last 6 items on self-care confidence were 0.55 to 0.86.

Back to Top | Article Outline

Construct Validity

The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.811, indicating that the sample size was suitable for factor analysis.21 Bartlett test of sphericity was significant (χ2 = 1613.130, df = 231, P = .000), indicating that factor analysis was appropriate.

Three factors were extracted with an explained variance of 43%. The factor pattern matrix for the C-SCHFI is presented in Table 1. The factor loadings of this matrix were sorted by size. All items except items 4 and 8 loaded strongly and significantly on the original internal structure, with a factor loading of greater than 0.30. Item 16 loaded greater than 0.40 on multiple factors. The goal was to group items together so that the factor on which they were placed represented a consistent content area. Item 16 had a better fit with the items that loaded on self-care confidence. These content areas were consistent with the original version of the instrument.

TABLE 1

TABLE 1

According to the expert content index and the theoretical model, all items were retained. To test the 1-dimensionality of the C-SCHFI, CFA was performed. Results from the first-order CFA revealed that 3 latent constructs (self-care maintenance, self-care management, and self-care confidence) were moderately intercorrelated (r = 0.46, 0.59, and 0.64), indicating that they were measuring the same construct. The second-order construct showed that all items were indicators of the self-care construct, and the factor loadings of 3 latent constructs on the self-care were 0.65, 0.91, and 0.70, respectively, which demonstrated the relative contributions of the first-order actors to the global construct. The overall model was supported by χ2/df of x2.23, with other acceptable fit indices (CFI, 0.90; GFI, 0.81; RMSEA, 0.082).

According to the recommendations produced from the LISREL output, further modification of the structure in the Figure was made and the fit indices indicated an acceptable fit of the model. An excellent fit between the data and the proposed model was revealed with χ2/df of 1.90, CFI of 0.93, GFI of 0.84, and RMSEA of 0.070, and all standardized parameter estimates were significant (P < .01).

FIGURE

FIGURE

Back to Top | Article Outline

Criterion-Related Validity

Mean (SD) C-SCHFI scale scores were 43.5 (16.7) for maintenance, 51.4 (21.6) for management, and 52.0 (21.1) for confidence. There were statistically significant correlations among the most relevant domains of both questionnaires (Table 2). Moderate correlations (−0.35 to −0.60) were demonstrated between the scales of C-SCHFI and the PIH Scale.

TABLE 2

TABLE 2

Back to Top | Article Outline

Discriminant Validity

Discriminant validity was supported by the correlation analysis, which showed that the C-SCHFI scales were not correlated with the MLHFQ. There were only weak, nonsignificant correlations between the scores (Table 2). We concluded that the 2 instruments measure different concepts.

Back to Top | Article Outline

Discussion

Measurement of patient-centered outcomes is a central issue in nursing research, but cultural and linguistic variations between English and Chinese make translating instruments difficult. Not only semantic but also cultural equivalence must be considered in the translation process. The most widely used and accepted method for developing equivalence between the original language and the target language is Brislin’s back-translation.8 The C-SCHFI was developed by conscientious application of this process. Except for the items on healthcare system, all items of C-SCHFI were culturally relevant to the Chinese living in mainland China.

The item-total correlations and the α coefficients supported good internal consistency of the C-SCHFI. Our main findings were similar to those of the English version in patients with CHF.5 Internal consistency of the self-care maintenance scale was lower than desired, which was consistent with the English version. The original author mentioned 2 reasons: (1) health behaviors are independent of each other, controlled by different motivators, and (2) including behaviors such as activity ensures a low internal consistency of this subscale because factors other than HF greatly influence these behaviors.6

Construct validity is the degree to which an instrument measures the construct it is intended to measure.22 Exploratory factor analysis revealed a 3-factor structure, quite similar to the original instrument with the exception of 4 items. In the EFA model, all items loaded substantively with the exception of 3 items. Item 4 (Do some physical activity) was weakly correlated with self-care maintenance. The inconsistent result might be due to healthcare system differences. In China, cardiac rehabilitation programs have not been common, and home telemonitoring is not available yet. Worrying about high risk of death, many patients with heart failure seldom do physical exercise. Item 8 (Forget to take 1 of your medicines) did not load significantly on any of the factors, although it is an important contributor to the content of self-care and we determined that it should not be eliminated. Item 8 was a reverse scored item, which may have been confusing to respondents. Patients may have had trouble reversing the scale in their minds to give the appropriate response.5 However, the full omission of negatively worded items may be too restrictive of a practice because of the inability to identify acquiescent response patterns. Acquiescence is a response behavior such that respondents will endorse a position without cognitively processing the meaning of the item. Item 16 (How sure were you that the remedy helped or did not help), measuring treatment evaluation, was found to cross-load with self-care management and self-care confidence. These items should be improved in future research.

Valid instruments that are grounded in theory can be used not only to measure phenomena within the theory in question. They may also be a means to test the theory itself. The SCHFI was built on situation-specific theory of heart failure self-care.19 In situation-specific theory of heart failure, self-care is defined as a naturalistic decision-making process involving the choice of behaviors that maintain physiologic stability (maintenance) and the response to symptoms when they occur (management). Confidence in self-care is thought to moderate and/or mediate the effect of self-care on various outcomes. Results from this study verified that self-care confidence was a positive and equally strong predictor of self-care maintenance and self-care management, and self-care maintenance was a moderate predictor of self-care management. Significant and positive correlations (P < .001) were noted between 2 sets of residuals associated with the measurement of self-care confidence. These items included “Evaluate the importance of your symptoms” and “Recognize changes in your health if they occur,” indicating that symptoms are the major changes in patient’s health. Meanwhile, the correlated residuals were associated with the item’s impact of “Evaluate how well a remedy works” and “Do something that will relieve your symptoms.”

Several psychometric properties could not be evaluated in this study and thus remain undefined. These include assessment of item difficulty level and the instruments’ predictive value (eg, of patients’ outcomes). In addition, patients were recruited from 1 hospital with convenience sampling. Thus, it is essential to investigate further application value of the scale by increasing the sample size and conducting stratified sampling of patients from different regions and cultural backgrounds.

This study has a limitation. The patients were limited to those who were hospitalized or having severe heart failure at a single university hospital in Beijing. Self-care might vary under different circumstances or in different groups of patients. This point must be taken into consideration when interpreting the results. Further research on outpatients is necessary.

Back to Top | Article Outline

Conclusion

The C-SCHFI is a theory-based and empirically tested instrument for assessing self-care of Chinese with heart failure. Furthermore, findings from the SEM suggest a modified structure for the C-SCHFI that requires confirmation by future studies.

Back to Top | Article Outline

What’s New and Important

  • To measure self-care among cultures, a valid, reliable, sensitive, and cross-cultural instrument is required.
  • Both semantic and cultural equivalence should be considered in the translation process.
  • The Chinese version of the SCHFI is a reliable and valid instrument for patients with heart failure.
Back to Top | Article Outline

Acknowledgment

Special gratitude to Dr Barbara Riegel for her generous support for us to use the SCHFI.

Back to Top | Article Outline

REFERENCES

1. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics–2011 update: a report from the American Heart Association. Circulation. 2011; 123 (4): e18–e209.
2. Gu Dongfeng HGHJ, MacMahon S, Whelton PK. Investigation of prevalence and distributing feature of chronic heart failure in Chinese adult population. Chin J Cardiol. 2003; 31 (1): 3–6.
3. Riegel B, Moser DK, Anker SD, et al. State of the science: promoting self-care in persons with heart failure: a scientific statement from the American Heart Association. Circulation. 2009; 120 (12): 1141–1163.
4. Cameron J, Worrall-Carter L, Driscoll A, Stewart S. Measuring self-care in chronic heart failure: a review of the psychometric properties of clinical instruments. J Cardiovasc Nurs. 2009; 24 (6): E10–E22.
5. Riegel B, Lee CS, Dickson VV, Carlson B. An update on the Self-care of Heart Failure Index. J Cardiovasc Nurs. 2009; 24 (6): 485–497.
6. Riegel B, Carlson B, Moser DK, Sebern M, Hicks FD, Roland V. Psychometric testing of the Self-care of Heart Failure Index. J Card Fail. 2004; 10 (4): 350–360.
7. Yu DS, Lee DT, Thompson DR, Woo J, Leung E. Assessing self-care behaviour of heart failure patients: cross-cultural adaptation of two heart failure self-care instruments. Hong Kong Med J. 2010; 16 (Suppl 3): 13–16.
8. Brislin RW. Back-translation for cross-cultural research. J Cross Cultur Psychol. 1970; 1 (3): 185–216.
9. Polit DF, Beck CT. The Content Validity Index: are you sure you know what’s being reported? Critique and recommendations. Res Nurs Health. 2006; 29 (5): 489–497.
10. Lynn MR. Determination and quantification of content validity. Nurs Res. 1986; 35 (6): 382–385.
11. DeVon HA, Block ME, Moyle-Wright P, et al. A psychometric toolbox for testing validity and reliability. J Nurs Scholarsh. 2007; 39 (2): 155–164.
12. Petkov J, Harvey P, Battersby M. The internal consistency and construct validity of the Partners in Health Scale: validation of a patient rated chronic condition self-management measure. Qual Life Res. 2010; 19 (7): 1079–1085.
13. Ho CC, Clochesy JM, Madigan E, Liu CC. Psychometric evaluation of the Chinese version of the Minnesota Living With Heart Failure Questionnaire. Nurs Res. 2007; 56 (6): 441–448.
14. Nunally JC, Bernstein IH. Psychometric Theory. 2nd ed. New York, NY: McGraw-Hill; 1994.
15. Polit DF, Beck CT. Nursing Research: Principles and Methods. Philadelphia, PA: Lippincott; 2004.
16. Ferketich S. Focus on psychometrics: aspects of item analysis. Res Nurs Health. 1991; 14 (2): 165–168.
17. Kline RB. Principles and Practice of Structural Equation Modeling. New York, NY: Guilford; 2005.
18. Gering WD, Anderson AJ. An updated paradigm for scale development incorporating unidimensionality and its assessment. J Mark Res. 1988; 15 (1988): 33–57.
19. Riegel B, Dickson VV. A situation-specific theory of heart failure self-care. J Cardiovasc Nurs. 2008; 23 (3): 190–196.
20. Raykov T, Tomer A, Nesselroade JR. Reporting structural equation modeling results in psychology and aging: some proposed guidelines. Psychol Aging. 1991; 6 (4): 499–503.
21. Hair J, Anderson R, Tatham R, Black W. Multivariate Analysis. Upper Saddle River, NJ: Prentice Hall; 1998.
22. Cronbach LJ, Meehl PE. Construct validity in psychological tests. Psychol Bull. 1955; 52 (4): 281–302.
Keywords:

construct validity; instrument; psychometric; self-management; structural equation modeling

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved