Identifying cut‐off scores for interpretation of the Heart Failure Impact Questionnaire

Abstract Aims Heart failure (HF) influences health‐related quality of life. However, the factors that contribute to health‐related quality of life remain unclear in Taiwan. We aim to identify the factors influencing health‐related quality of life in HF patients. Methods Hospitalized HF (N = 225) patients were included from April 2011 to April 2014. Health‐related quality of life was assessed by using the 36‐Item Short‐Form Health Survey (SF‐36) and the Minnesota Living with Heart Failure Questionnaire. A new cut‐off was conducted based on the combination of SF‐36 and Minnesota Living with Heart Failure questionnaire. Results There were significant differences between good and poor quality groups on age, gender, education levels, occupational classification caregiver, New York Heart Association classes, and the numbers of comorbidities. The logistic regression analysis showed that the number of comorbidities was more than three and New York Heart Association class IV were significantly associated with health‐related quality of life.

. The measurements of quality of life (QOL) in most of the studies have been performed based on the hospital setting and discharge follow-up (McMurray et al., 2012;Naveiro-Rilo et al., 2010;Santos, Plewka, & Brofman, 2009).
Health-related quality of life (HRQOL) is defined as a multidimensional concept that has an impact on the daily lives activity performances of patients with chronic diseases including patients' functional capabilities, symptoms, and psychosocial perceptions on overall well-being (Heo, Doering, Widener, & Moser, 2008;Jaarsma et al., 2000;Yu, Lee, & Woo, 2004).
In HF patients, HF disease scenario is chronic and prognostic situations. Therefore, while evaluating QOL, there are two important types of QOL score system questionnaires, one is the Minnesota Living with Heart Failure questionnaire (MLFHQ) and other is the 36-Item Short-Form Health Survey (SF-36). Among the two scoring system, the most frequently used for both generic and disease-specific measures. The MLHFQ subscales screening for physical dimension and emotional dimension to give more specific information about a special group and reveals more sensitive results (Behlouli et al., 2009;Bilbao, Escobar, García-Perez, Navarro, & Quirós, 2016). The other instrument is the SF-36, which is the generic measure of QOL and gives validated, reliable and multidimensional results. The SF-36 consists of eight domains; the scores of these subscales can be combined to create two higher order summary scores: the physical component summary (PCS) and mental component summary (MCS). These results can be compared with those of a general population to give a more information on general health status (Huber, Oldridge, & Hofer, 2016;Ware & Sherbourne, 1992;Wylie, Beckmann, Granger, & Tashjian, 2014). The weighing of other QOL domains for combining two scores include psychosomatic symptomatology and emotional interference.
Correlations between all scores were calculated in previous studies.
Numerical values allow evaluation of patient change. The classification of QOL scores may be helpful to take a multifaceted decision-making for the implementation of treatment. However, the question is: which QOL measuring instrument is more accurate? Is it possible to define a new cut-off point using the combination of MLHFQ and SF-36?
To solve this question, in this study, we aimed to: (1) assess the combination of SF-36 and MLHFQ, and describe the sensitivity and specificity of cut-off scores in screening for HF; and (2) study the diagnostic properties and diagnostic values of SF-36 and MLHFQ in predicting HF patients QOL score.

| Design
The present cross-sectional study was collected data by face-to-face interviews with the participants at clinical sites. In total, 225 HF patients were enrolled from Cathay General Hospital in Northern Taiwan.
Inclusion criteria were the following: (1) a diagnosis of HF (both systolic and diastolic failure) by a physician and assessed based on New York Heart Association (NYHA) class II-IV heart disease for at least 3 months; (2) abnormalities of focal ventricular motion, abnormal left ventricular end-diastolic dimension, systolic dysfunction, or valves abnormalities detected by echocardiography; (3) left-ventricular ejection fraction (LVEF) <40%; (4) hospitalized at least twice due to HF; and (5) ability to engage in conscious and coherent verbal communication with the interviewer. Patients who: (1) were diagnosed with mental disorders; (2) were bedridden for >3 months and unable to ambulate; (3) had severe visual or hearing impairment; and (4) refused to participate were excluded from this study.

| Method
The MLHFQ is specifically designed for evaluating HF patients' to understand their disease status as well as their QOL within 1 month after the completion of primary treatment (Heo, Moser, Riegel, Hall, & Christman, 2005). It is composed of 21 items which cover HF-related physical, psychological and social impairments. The questions are calculated on a Likert-type scale that ranges from 0 to 5 and can be summarized to a total score of highest 105. Lower scores indicate better HRQOL. The content validity index was 0.98. The construct validity was supported by exploratory factor analysis in a Chinese version.
The SF-36 measures perceived health status in eight dimensions: physical function, role limitations due to physical problems, body pain, general health, vitality, social function, role limitations due to emotional problems, and general mental health. The scores were summarized into two component summary scores of physical and mental health. Scores range from 0 (worst)-100 (best).

| The combination of SF-36 and MLHFQ
To construct the new cut-off, the new definition for good HRQOL was as follows: (1) MLHFQ score <24; (2) patients with MLHFQ score <45 and SF-36 score ≥60. The new definition for poor HRQOL was as follows: (1) MLHFQ score ≥45; (2) MLHFQ score ≥24 and SF-36 score <60 (Figure 1). The weighted HRQOL index was verified by the area under the receiver operating characteristic (ROC) curve.

| Analysis
Descriptive statistics were expressed as the mean ± standard deviation (SD) for continuous variables, and as the frequencies and percentages for categorical variables. Chi-square test and Student's t test were used to test the differences between the two groups. The SF-36 and MLHFQ were analysed for the continuous sociodemographic and clinical characteristics by Pearson's correlation. Logistic regression analysis was used to investigate the associations between HRQOL measures and clinical characteristics.
A two-tailed test with α < 0.05 was considered to be statistically significant. All analyses were conducted by using SPSS statistical software (version 18.0; SPSS Inc., Chicago, IL, USA).

| Ethics
This study was approved by the Institutional Review Board of Cathay General Hospital. All participants agreed to participate in the research and signed an informed consent form.

| Patient characteristics
Baseline characteristics and clinical data in all participants with HF were collected. Most of the patients were retired or unemployed; the mean age was 70.88 years and 65% were male. Most the patients (70.18%) were in NYHA functional class III or IV. Only 56% of participates were married, but most (85%) of them lived with their family and had a low educational level. The mean body mass index (BMI) was 25.60 kg/m 2 (SD 4.11) and all patients had at least two comorbid illnesses. Patients with MLHFQ scores 24-45 as moderate were younger, had higher education, and low NYHA class as compared with those with MLHFQ >45 (Table 1). Similarly, patients with SF-36 scores ≥60 were younger, had higher education, higher probability of not having caregivers, lower NYHA class, and fewer comorbidities compared with those with SF-36 <60 (Table 2).

| MLHFQ and SF-36
The total score of MLHFQ had positive association with old age, more severe NYHA classification, higher number of comorbidities, while it had negative association with higher educational level and higher hospitalization frequency. In addition, there were positive association between SF-36 scores and higher BMI. However, SF-36 scores had negative association with higher age, more severe NYHA class, and higher number of comorbidities (Table 3).
In relation to discriminative validity, the MLHFQ total score and SF-36 dimensions were able to distinguish characteristics of patients

| D ISCUSS I ON
The results from our study showed that by using the combination of SF-36 and MLHFQ to divide HF patients into good HRQOL and poor HRQOL groups, the number of factors associated with HRQOL was more than individual questionnaire, suggested to be a new classification for HRQOL. In addition, comorbidities were the most powerful   (Table 4). These diseases are highly related with HF. To our knowledge, this is the first study to describe and compare two instruments measure to predict HF patients' HRQOL to explore risk factors. From this analysis, newly identified questionnaires cutoff scores were assigned to the "good" and "poor" QOL category.
It is important that the range of scores questionnaires overlapped, and therefore, they were merged to form a new score system. From this analysis, we believe this study has direct clinical importance. The data form SF-36 in our participants were able to distinguish different levels of HF, mostly HF patients with high score by 72% (163) and form MLHFQ high score by 60% (136), we considering the HF patients in the "poor" QOL category overall dimension. However, there was a significant change between the two clusters. The decision was focus to on the assessment tools necessary to give more information on the evaluation of patients with HF; this was based on the need to identify the disease influences on the patients' QOL, and more clearly, on different factors that may have an impact on QOL.
Furthermore, the studies would follow-up the relationship between these two tools in evaluating the QOL in HF patients and studies should focus on the validation of the combination of these two questionnaires by conducting a study which has larger sample size.

| Limitations
The following are the limitations of this study.  TA B L E 4 (Continued) (Continues) and SF-36 scales. Healthcare systems should understand the multifactorial of HF patients of progressive to achieve more effective management. By screening for patients, we achieve early detection and disclose the risk factors to help HF patients to improve outcome and delay the process to irreversible HF and mortality.

| CON CLUS ION
Moreover, assessing HF disease severity and HRQOL is also important to monitor patient care. HF is a multifactorial problem and its management requires a combination of intervention strategies.

CO N FLI C T O F I NTE R E S T
No conflict of interest has been declared by the authors.

AUTH O R CO NTR I B UTI O N S
All authors have agreed on the final version and meet at least one of the following criteria [recommended by the ICMJE (https://www. icmje.org/recommendations/)]: • substantial contributions to conception and design, acquisition of data or analysis and interpretation of data; • drafting the article or revising it critically for important intellectual content.