Psychometric properties of the Brief Symptom Inventory-18 in a Spanish breast cancer sample

Abstract

Objective

The objective of this work was to study the psychometric and structural properties of the Brief Symptom Inventory-18 (BSI-18) in a sample of breast cancer patients (N=175).

Methods

Confirmatory factor analyses were conducted. Two models were tested: the theoretical model with the original structure (three-dimensional), and the empirical model (a four-factor structure) obtained through exploratory factor analysis initially performed by the authors of the BSI-18.

Results

The eligible structure was the original proposal consisting of three dimensions: somatization, depression, and anxiety scores. These measures also showed good internal consistency.

Conclusion

The results of this study support the reliability and structural validity of the BSI-18 as a standardized instrument for screening purposes in breast cancer patients, with the added benefits of simplicity and ease of application.

Introduction

One of the current primary objectives of comprehensive cancer care is the development of brief standardized instruments to screen for psychological distress, thereby identifying patients who may potentially experience significant difficulty in their attempts to cope with and adapt to their diagnoses and treatments [1].

In specific reference to cancer, distress has been defined as “a multifactorial unpleasant emotional experience of a psychological (cognitive, behavioral, emotional), social, and/or spiritual nature that interferes with the ability to cope effectively with cancer, its physical symptoms, and its treatment. Distress extends along a continuum, ranging from common feelings of vulnerability, sadness, and fears, to problems that can become disabling, such as depression, anxiety, panic, social isolation, and existential and spiritual crisis” [2].

The presence of emotional distress in cancer populations has been widely documented. Research has demonstrated that approximately one third of oncology patients experience clinically significant distress associated with diagnosis and treatment of cancer [3], [4], [5]. Nevertheless, less than 5–10% of cancer patients are referred for psychosocial care [5], [6]. Failure to detect and treat elevated levels of distress has been associated with greater nonadherence to treatment recommendations, lower satisfaction with care, and poorer quality of life [7], [8]. Patients who are distressed are also more likely to use community health services and emergency facilities [6], [9], in turn increasing health care costs.

Guidelines for psychosocial care developed for several agencies [2], [10] recommend routine screening for psychological distress. For screening purposes, the availability of a brief instrument that is reliable, valid, and easy to score has become an important aim [1].

The Brief Symptom Inventory-18 (BSI-18) provides normative data for cancer patients [11], and its utilization in this population is increasing [4], [10], [12], [13]. It has also been used as a benchmark measure to determine distress in validation studies of other screening instruments, such as the distress thermometer [7], [14]. The BSI-18 is the briefest and latest in an integrated series of instruments designed by Derogatis to measure psychological distress. Previous instruments in the series are the BSI [15] and the Symptom Checklist-90 Revised (SCL-90-R) [16]. In various reviews, the SCL-90-R and the BSI appear among the most utilized instruments for purposes of distress screening in oncology populations [1], [5], [17]. In fact, these two instruments, considered together, may be the most widely used among cancer patients [3].

The BSI-18 [11] was specifically developed as a highly sensitive and efficient screen for psychological distress and, according to its author, has two advantages over previous instruments. Firstly, its brevity (average completion time, 1–3 min) makes it easy to administer and score, which is a very important issue for screening purposes. Secondly, it has improved structural validity due to the fact that it is made up of three dimensions of symptoms that are conceptually and empirically more homogenous than others considered in previous instruments. The dimensions assessed by the BSI-18 are as follows: somatization (six items)—discomfort produced by the perception of corporal problems including cardiovascular, stomach, and muscular problems; depression (six items)—symptoms commonly related to depression, such as apathy, sadness, and thoughts of suicide; and anxiety (six items)—feelings of fear, general nervousness, and even panic. The BSI-18 also offers a measure of general distress (total score). This measure is highly correlated to the BSI general distress score (r>.90) and enables the comparison of results reached by both instruments. According to the author, this new inventory presents satisfactory reliability indexes (specifically Cronbach's α), both for the dimensions (ranging from .74 to .84) and for the general distress index (.89) [11]. Moreover, the author carried out a factor analysis (principal components analysis with Kaiser varimax rotation) throughout the study process of the structural validity of the instrument. The factor analysis, based on a community sample of 1134 subjects, yields a four-factor solution. Two of these factors (Factors I and II) contain the exact items belonging to the scales of somatization and depression. The other two factors are composed of the anxiety dimension items originally proposed: a set of three items related to distress and generalized nervousness integrates Factor III, and three items evaluating panic symptoms make up Factor IV. However, as Recklitis et al. [13] pointed out, Derogatis observed that support for a distinct fourth component in their analyses was tenuous; the eigenvalue for this fourth component was just at the conventional cutoff of 1.00. In fact, the author confirms the structural validity of the instrument because it can still be considered a single dimension of anxiety [11]. There are two reasons given for this: the theoretical relationship between panic symptoms (Factor IV) and other general manifestations of anxiety (Factor III), and the strong correlation between the last two factors obtained.

The BSI-18 has gained popularity as a screening instrument in cancer samples [4], [8], [12], [13]. However, there are few studies specifically supporting its use in this population.

Zabora et al. [8] analyzed the potential for the BSI-18 to be applied to cancer patients by studying its reliability and validity in an adult sample undergoing active cancer therapy (N=1543). The results revealed a satisfactory internal consistency (Cronbach's α) for a total score, and adequate sensitivity and specificity, using a new cutoff for the oncology population consisting of the 25th percentile (a total score of 13 for women and a total score of 10 for men). Moreover, the authors performed a principal component factor analysis to determine its structural validity. Four factors were identified: somatization, depression, general anxiety, and panic, which Zabora et al. [8] interpreted similarly to the empirical findings by Derogatis [11]. However, we believe that it is necessary to clarify this interpretation. The fourth factor (panic) obtained by Zabora et al. [8] only includes saturations above the cutoff (>.40) in a single item (Item 17: “thoughts of ending your life”); however, they take into account the other two items (Item 12: “spells of terror or panic”; Item 9: “suddenly scared for no reason”) and interpret them as a panic dimension. These two items show much higher saturations and are above the cutoff in the first factor, combining general anxiety and panic. In our opinion, the results indicate a three-factor structure for the BSI-18 integrating different anxiety symptoms into a single factor, similar to the one hypothesized by Derogatis [11]. In addition to this, the principal components analysis technique used in this study is not the most suitable technique for determining which latent factors can be measured by the BSI-18 items [18].

Recently, Recklitis et al. [13] confirmed the factor structure of the BSI-18 in a sample of 8945 adult survivors of childhood cancer. In this study, two steps were followed to establish the BSI-18 structure: firstly, an exploratory factor analysis was performed with a randomly chosen subsample, supporting a three-factor structure closely corresponding to the three BSI-18 subscales; secondly, a confirmatory factor analysis (CFA) with structural equation modeling validated this three-dimensional structure in a separate subsample. The results also showed the excellent internal consistency of the three dimensions and the total score of the BSI-18. Finally, the author indicated that, with these psychometric properties, the BSI-18 may be useful in assessing psychological distress in adult survivors of childhood cancer.

To date, studies on the structure of BSI-18 using oncology samples are still scarce, and initial studies of the distress dimensions have been limited to principal components analysis [8], rather than a factor analytic strategy that is more suited to latent structure research. Studies investigating the potential for the BSI-18 to be applied to different cancer patient samples and different cultural groups are essential for instrument validation in this population. Thus, the objectives of the present study are: (a) to determine whether the BSI-18 can be used as a reliable measure of psychological distress in a Spanish sample of women who had been recently diagnosed with breast cancer, and (b) to determine whether the BSI-18 demonstrates the hypothesized three-dimensional structure in this sample.