Psychometric properties of Quality-of-Life Index for Vietnamese women with breast cancer three weeks postmastectomy

Definition of Quality of Life of Patients Postmastectomy

Quality of life (QOL) is the primary goal that most people attain during their daily life. Since this concept has been recognized, QOL is not separate from health because it is considered as the person’s sense of well-being that stems from the satisfaction and dissatisfaction with aspects of life (Ferrans, 1990) or functional capacity, symptoms (physical and psychological) and perceptions of health (Mccorkle et al., 1989). The quality of life in the nursing context is related to a specific illness, and it can be considered similar meaning with health-related quality of life. A concept analysis of nursing, based on the guiding theory of Peplau’s, Rogers’, Leininger’s, King’s, and Parse, is defined as a contextual, intangible, subjective perception of one’s lived experience (Plummer & Molzahn, 2009). Cella (1994) identified four dimensions of quality of life in the context of cancer that encompasses physical well-being, functional well-being, emotional well-being, and social well-being. The concept of Cella (1994) and its four significant domains of quality of life help investigate the concept multi-dimensionally. Besides, (Padilla & Grant, 1985) describe the quality of life as five dimensions: physical well-being, social concerns, body image concerns, psychological well-being, and diagnosis/treatment response. In breast cancer, the concept of QOL describes the impact of breast cancer on the domains of physical, social, psychological well-being, and spiritual well-being (Ferrell et al., 1998).

Receiving a mastectomy also raises concerns about body image, uncertainty in the situation of illness, surgical symptoms, lacking nursing care or social support, and poor patient-physician communication impacting the patient’s quality of life (Denieffe et al., 2014; Mandelblatt et al., 2003; Wronska et al., 2007). Thus, in patients with breast cancer postmastectomy, the concept of quality of life should be more specific, clearly describing the situation, which occurs among this group. That is why the definition of QOL defined by Padilla and Grant (1985) as physical well-being, social concerns, body image concerns, psychological well-being, and diagnosis/treatment response remains the most suitable application for QOL postmastectomy.

Following Padilla and Grant (1985), physical well-being can be considered a strength, fatigue, ability to work, current health, and perceived usefulness. Psychological well-being implies happiness, satisfaction, fun, general QOL, pleasure in eating and sleep. The body image concerns mean the ability to look at the changes in one’s body, the tendency to worry, and the ability to adjust and live with body changes. The social concerns focus on social rejection, social contact, or the need for privacy. Diagnosis/treatment response relates to surgical treatment symptoms, which are defined as the ability to have sufficient sexual activity, nutrition, weight, pain, and severity of pain, nausea, and vomiting (Padilla & Grant, 1985). In postmastectomy patients, the attributes of physical, psychological, and social concerns of QOL might be the same as other cancers; however, the defining attributes of body image and treatment response might differ. The body image in breast cancer patient postmastectomy relates to the ability to look at the changes of the body, worry over scarring, perceived femininity, and how easy it is to live with anybody changes (Barolia, 2008; Denford et al., 2011; Fobair et al., 2006; Lindwall & Bergbom, 2009; Toriy et al., 2013). The treatment response of mastectomy patients focuses on symptoms around the hand and shoulder such as the ability to raise the hand, any swelling of the arm, the sensitivity of the breast incision, sufficient nutrition, weight, as well as the severity and frequency of pain (Champion et al., 2014; Janz et al., 2007; Taghian et al., 2014). Operationally, the concept of quality of life on postmastectomy patients is defined as the perception of life experienced based on five domains: physical well-being, psychological well-being, body image concerns, social concerns, and treatment responses. Defining attributes of QOL consist of physical well-being (strength, fatigue, ability to work, current health and perceived usefulness), psychological well-being (happiness, satisfaction, fun/hobbies, eating pleasure and sleep), body image concern (look at the body, scare of scarring, perceived femininity, ability to live with losing a breast, the worry of future living without a breast), social concerns (family, friends or healthcare giver staff contact, social rejection, and privacy needs and treatment responses (ability raising the hand, swelling of the arm, sensitive of destroying breast, nutrition sufficient, weight, severity, and frequency of pain)

Existing Instruments

Most of the effective existing instruments measuring the quality of life for patients with breast cancer are all well-known instruments that have been used to examine QOL in many stages of breast cancer (Perry et al., 2007). Among those, FACT-B and EORTC-BR23 are specific for patients during chemotherapy treatment. QOL-BR23 focuses on physical function, whereas FACT-B emphasizes emotional well-being (Nguyen et al., 2015). EORTC-QLQ30 and SLDS-BC or QOLI are suitable for QOL in general. Interestingly, the QOLI of Padilla and Grant (1985) is based on the concept of QOL across a range of cancers in women, though sharing similar circumstances to breast cancer patients, such as cervical cancer, colorectal cancer, and hysteric cancer post-surgery. The original QOLI of Padilla and Grant (1985) identifies 14 factors and has been validated in many studies measuring QOL; hence it has proven validity and reliability (Rukholm et al., 1998). Over time, the QOLI has been modified for colostomy patients by adding nine items focusing on some aspects of symptoms post-surgery. The dimensions of QOLI would seem to be closest to the definition of quality-of-life postmastectomy with five domains of physical well-being, psychological well-being, social concerns, body image concerns, and treatment/diagnosis response. The length of 23 items with self-administer base on the visual line for the most concern in the past four weeks. Summarily, with the same aspects of colostomy and mastectomy on colorectal and breast cancer patients, the QOLI of Padilla and Grant (1985) covers most aspects of the operational definition of quality of life postmastectomy. Thus, this instrument will be selected to test the psychometric properties in the breast cancer population postmastectomy. The summary of the comparison of the tools measuring QOL is presented in Table 1.

Study Design

The descriptive cross-sectional study was designed to analyze the psychometric properties of a Vietnamese version of the modified Quality of Life Index (QOLI-V) on patients three weeks postmastectomy. The modified process was conducted by researchers after granting permission, acceptance, and consultation of the original authors.

Sample and Setting

The population of this study was the patients three weeks postmastectomy at the Breast Surgical Oncology Ward in the Oncology Hospital in Ho Chi Minh City, South of Vietnam. Convenient sampling was used to select the respondents. The inclusion criteria of the respondents were aged 30-60, could read and write Vietnamese, no other diseases, and normal surgical recovery process at seven days.

The literature suggests the estimated sample size of CFA should not be less than 200 to avoid violating the thumb rule of “too few degrees of freedom” (Hair et al., 2010). Other assumptions requested that the sample was > 200 for the theoretical model or ≥ 300 for the population model for CFA in physical health care. A systematic review also proposed that the number of subjects should be equal to the number of items multiplied by 10 in the nursing field (Watson & Thompson, 2006). It is estimated that 265 patients were included to test the psychometric properties of QOLI_V, with 26 items modified from QOLI (Padilla & Grant, 1985) combined with the five domains.

Instrument Validation

The demographic form and the modified quality of life index Vietnamese version (QOLI-V) were used to collect data in this study. The demographic form was developed by the researchers asking about the characteristics of the respondents, such as age, marital status, occupation, education, income, and mastectomy type.

The QOLI_V was a 26-item questionnaire composed of five domains: physical well-being, psychological well-being, social concerns, body image concerns, and treatment response. Data were indicated by marking an X on the visual line equal from 0 to 10 score. Scores were presented as numeric rating scales. QOL was calculated by the sum of the scores divided by the sum of items with a low score indicating a low QOL. The original QOLI with 23 items retained with five domains. In reference to the concept of QOL in a mastectomy group, four items related to the symptoms of patients with breast cancer postmastectomy, including swollen arms, the ability to raise hands, the sensitivity of breast incision was added to the section on treatment response and perceived femininity was added to the section on body image concerns.

Then the 27-item QOLI was sent to five experts for testing its content validity index following the recommendation of Polit et al. (2007): two surgeons with ten years of experience in the mastectomy process, two Ph.D. nursing lecturers, and one head nurse in the Breast Surgical Department. The results showed that the lowest I-CVI was .80 and the highest was 1.00; S-CVI/Ave was .95, and S-CVI/UA was .76, which implied good validity for this instrument (Osanloo & Grant, 2016; Polit et al., 2007). The CVI testing of 5 experts confirmed that for 26 items, most of all item was rated from 3 (relevant) to 4 (very relevant). The sum agreements of each item related to the quality-of-life postmastectomy were calculated. The result confirmed that most of the items correlated well with the quality-of-life postmastectomy, except the item of sufficient sexual satisfaction (.40). Experts rated this item with a lower score of relevancy and recommended researchers consider the meaning of this item on Vietnamese culture.

Instrument Translation

The 26 item-modified QOLI was translated into Vietnamese using Brislin’s model. It was translated from English into Vietnamese and back-translated by two different bilingual experts at the Language Center, University of Medicine and Pharmacy, Ho Chi Minh City, Vietnam. Two translated versions were reviewed by a Vietnamese nurse responsible for teaching English to nursing students in the university, identifying ambiguous words and confirming the symmetry. The Quality-of-Life Index Vietnamese version (QOLI-V) was then assessed for its intelligibility in the Vietnamese context and culture with 5 cases of patients in the Surgical Oncology Ward. The piloting of QOLI-V also showed that most patients skipped the question asking about sufficient sexual satisfaction after mastectomy. When researchers discussed the reason with patients, the answer was that mastectomy treatment was terrifying and tiring. They and their partner did not want to have sex, or it was of no meaning in the postmastectomy period. Based on the Vietnamese culture, women often feel shame and become uncomfortable when asked about sexual activities, or they could not express the meaning of sexual satisfaction. In the stage of three weeks postmastectomy, sufficient sexual satisfaction was felt completely irrelevant to ask because patients were usually concerned with other aspects of their life than sexual satisfaction. Therefore, this item was deleted from the questionnaire.

Ethical Consideration

This study was approved by the Board of Ethical in Biomedical Research at the University of Medicine and Pharmacy and the Research Ethical Board of Oncology Hospital. This study was an instrument development part of the Dissertation project for PhD education in the Faculty of Nursing, Chulalongkorn University, Thailand. After IRB approval, the researchers met patients and presented the objectives, procedures to collect data, and approximate length of time for data collection at the Surgical Oncological Ward. Patients who matched the inclusion criteria and wished to volunteer signed the consent form to prove that they agreed to answer the whole questionnaire.

Data Analysis

The input data were checked for errors before entering the analysis tests. The data were analyzed for the assumption of normal distribution and descriptive demographic data and reliability of the measurement by the SPSS program version 16.0. Then the confirmatory factor analysis was conducted by using the AMOS version 20.0. The process of conducting the CFA conducted by following standardized recommendations: (a) Initial data analysis to identify any problem of missing data or input errors, (b) Fix one-factor loading on each sub-construct to a specific value as equal to 1, (c) Factor loading higher .5 to .7, (d) Construct reliability .6 was accepted, (e) Standardize residual accepted (from 2.5 to 4.0), (f) Not using modification indices to adjust the model fit (Hair et al., 2010).

Characteristics of the Respondents

The demographic data of the respondents showed that most of them were in middle age (47%). Many of the women were married and living together as a family (78.9%). However, the respondents had low education (primary and secondary school, 60.5%), and income from the main family members was still low (<3 million and/month, 51.7%). The majority of the respondents were housewives (42.3%), farmers, or doing small business at home (19.6%). Only one-third of them were office workers or executives; 64.5% of the respondents received radical mastectomy (see Table 2).

Reliability of the QOLI-V

The reliability of the revised translation version, 26-item QOLI-V, was tested for its reliability with 265 Vietnamese patients three weeks postmastectomy. Cronbach's alpha of QOLI-V was .84, which was considered acceptable for the modified instrument (Polit & Beck, 2003). Most of the 26 items featured the correct item-total correlation .3 to .6. There were only two items correlated with the total scale at .18. Regarding the last item, "general quality of life," the total correlation was only .06, and Cronbach's alpha increased when it was deleted. Thus, it was considered that this item should be deleted from the item set or not (see Table 3).

Construct Validity - Confirmatory Factors Analysis

The construct validity of the instrument was tested using the confirmatory factor analysis (CFA). The model validity is assessed based on exact test fit, with Chi-Square/df <2.0 is considered good and <5.0 is acceptable, Root Mean Square Error of Approximation (RMSEA ≤ .08), Standardized Root Mean Square Residual (SRMR ≤ .05), Comparative Fit Index (CFI ≥ .90) (Hair et al., 2010). The researchers also used other evidence to concern the appropriate model fit.

The initial model 1 was drawn up in the AMOS graphic program and run CFA with the data set. The first analysis showed that with 26 items based on the construct of 5 dimensions, model 1 was not an ideal fit with the data. The findings in detail were reported as Chi-Square/df 623/289 = 2.15, CFI =.815, RMSEA =.066. The model was presented in (Figure 1).

For most items, the standardized estimation (factor loading) was from .50 to .66. There is no estimation indicated the cross-loading factor. However, there were three items that the general quality of life, weight, and breast incision sensitivity were lower than .50, with the standardized regression weight estimated as .045, .48, .48, respectively. The residual estimation of 26 items ranked from 1.1 to 3.8 was acceptable based on the standardization rule. However, the residual estimate of e6 (general quality of life) exceeded the accepted level with the result at 4.2. The construct reliability of the measurement was high and exceeded the level of .6.

As for the modification indices, the general quality of life item is considered the cross-loading item. The regression weight of these items was adjusted for the par change in every item or latent variable of the model. Therefore, this item was considered for deletion from the model.

Following the empirical evidence of the CFA in model 1, the general quality of life was deleted, and the CFA of model 2 was conducted. The findings showed that the model fit improved, with the criteria Chi-Square/df =2.269, CFI=.814, and RMSEA=.069. The construct reliability of the Psychological dimension was improved after deleted one item. The model was presented in Figure 2.