Introduction
Patients with advanced cancer face a complex array of physical, social, psychological, and spiritual situations related to their disease and treatments (Allende-Pérez & Verástegui-Avilés, Reference Allende-Pérez and Verástegui-Avilés2013). These factors contribute to a significant number suffering from mental disorders; with adjustment disorders, delirium, major depression, and anxiety disorders being the most common (Chochinov, Reference Chochinov2001; Li, Fitzgerald, & Rodin, Reference Li, Fitzgerald and Rodin2012a; Mehta & Roth, Reference Mehta and Roth2015). In cancer settings, clinical depression is a treatable cause of serious additional suffering and distress in these patients (Chochinov, Reference Chochinov2001; Wilson et al., Reference Wilson, Chochinov and Skirko2007). There are two core symptoms of depression according to the Diagnostic and Statistical Manual of Mental Disorder, 5th ed (DSM-5; American Psychiatric Association, 2014), namely depressed mood and a marked loss of interest or pleasure in most or all activities. To qualify as a major depressive disorder (MDD), one of these two core symptoms must be present for at least 2 weeks, along with at least four other depressive symptoms. The other symptoms include appetite or sleep disturbance, psychomotor agitation or retardation, decreased energy, feelings of worthlessness or guilt, difficulties with memory or concentration, and suicidal ideation (American Psychiatric Association, 2014), which makes the diagnosis of depression a challenge in patients attending services with palliative care for cancer (Li et al., Reference Li, Fitzgerald and Rodin2012b).
Depression prevalence rates differ depending on the cancer population studied, the diagnostic criteria applied, and the timing and method of assessment (i.e., self-report vs. structured interviews (Carr et al., Reference Carr, Goudas and Lawrence2002; Chochinov, Reference Chochinov2001; Chochinov et al., Reference Chochinov, Wilson and Enns1994; Mitchell et al., Reference Mitchell, Chan and Bhatti2011, Reference Mitchell, Meader and Davies2012; Wilson et al., Reference Wilson, Chochinov and Skirko2007). Major depression has been found to occur in approximately 16.5%, 95% confidence interval = 13.1, 20.3, of patients in palliative care settings (Mitchell et al., Reference Mitchell, Chan and Bhatti2011). It has been associated with poorer health-related quality of life; lower performance status; reduced treatment adherence; more severe physical symptoms such as pain, fatigue, and drowsiness; and perhaps even increased mortality (Arrieta et al., Reference Arrieta, Angulo and Núñez-Valencia2013; Kroenke et al., Reference Kroenke, Theobald and Wu2010; Lloyd-Williams et al., Reference Lloyd-Williams, Dennis and Taylor2004; Wadih Rhondali et al., Reference Rhondali, Girard and Saltel2012; Wilson et al., Reference Wilson, Chochinov and Skirko2007). Furthermore, depressed cancer patients are more likely to have a prominent persistent desire for death (Wilson et al., Reference Wilson, Dalgleish and Chochinov2016).
Unfortunately, depression is sometimes viewed as being an appropriate reaction in cancer patients, so it is often overlooked and left untreated by professionals in palliative and nonpalliative settings (Chochinov, Reference Chochinov2001; Fallowfield et al., Reference Fallowfield, Ratcliffe and Jenkins2001; Sharpe et al., Reference Sharpe, Strong and Allen2004). To try and improve recognition, a first step could developing tools to help identify cases of depression in patients with advanced cancer (Lloyd-Williams et al., Reference Lloyd-Williams, Shiels and Dowrick2007). These tools must balance validity of assessment against brevity to avoid burdening frail patients (Chochinov et al., Reference Chochinov, Wilson and Enns1997).
The Brief Edinburgh Depression Scale (BEDS) was constructed from the Edinburgh Postnatal Depression Scale, a 10-item self- rating scale, as a case-finding tool for depression specifically in patients with advanced cancer (Cox et al., Reference Cox, Holden and Sagovsky1987; Lloyd-Williams et al., Reference Lloyd-Williams, Shiels and Dowrick2007). The abbreviated version consists of six items, each rated on a 4-point scale, and gives a sensitivity of 72% and specificity of 83% with a cutoff score of 6 of 18 and was designed for palliative care patients (Lloyd-Williams et al., Reference Lloyd-Williams, Shiels and Dowrick2007). The BEDS has been studied and used in the United Kingdom and other European countries (Lloyd-Williams et al., Reference Lloyd-Williams, Cobb and O'Connor2013; Mitchell et al., Reference Mitchell, Meader and Davies2012; Rayner et al., Reference Rayner, Price and Hotopf2011; Rhondali et al., Reference Rhondali, Girard and Saltel2012, Reference Rhondali, Chirac and Celles2014; Ziegler et al., Reference Ziegler, Hill and Neilly2011). Additionally, it has been translated and validated in French and Korean (Lee et al., Reference Lee, Kim and Ko2009; Rhondali et al., Reference Rhondali, Girard and Saltel2012).
In Mexico, there are no validated, brief instruments to identify cases of depression in Mexican palliative patients (Landa-Ramirez et al., Reference Landa-Ramirez, Cardenas-Lopez and Greer2014). There is an unmet need to improve detection practices so that patients with depression may get opportune access to care, especially because interventions have shown to have an impact in reducing depression severity and improving anxiety, quality of life, role functioning, and even survival (Prescott et al., Reference Prescott, Hull and Dionne-Odom2017; Walker et al., Reference Walker, Hansen and Martin2014).
The main goal of this study was to validate the Spanish-language Mexican version of the BEDS in Mexican population with advanced cancer in a palliative care service.
Methods
This study was approved by the local Institutional Review Board and Ethics Committee (registration numbers 017/004/CPI and CEI/1114/17, respectively). All patients gave written consent.
Subjects
Consecutive patients were recruited from September 2016 to March 2017 at the outpatient area of the palliative care service at the Instituto Nacional de Cancerología in Mexico City. The sample size was estimated based on the number of items on the scale, considering 10 participants for each component of the instrument (N = 60) (Nunnally & Bernstein, Reference Nunnally and Bernstein1995). Patients were eligible if they understood written and spoken Spanish; were age 18 and older; had an Eastern Cooperative Oncology Group Performance Status (Oken et al., Reference Oken, Creech and Tormey1982) score of 0, 1, or 2; Karnofsky (Yates et al., Reference Yates, Chalmer and McKegney1980) index of 50% or better; and clinical prediction of survival through the Palliative Prognosis Index (Morita et al., Reference Morita, Tsunoda and Inoue1999) A (>6 weeks) or B (4–6 weeks). Participants were required to complete the self-assessment scale and respond to an interview. Those with any uncontrolled physical symptom, cognitive impairment, delirium, psychosis, cerebral metastases, or current antidepressant treatment were excluded.
Procedure
Baseline information was obtained from medical records. The Spanish-language BEDS items were transcribed from the previously validated, Spanish translated Mexican version of the Edinburgh Postnatal Depression Scale (Alvarado-Esquivel et al., Reference Alvarado-Esquivel, Sifuentes-Alvarez and Salas-Martinez2006, Reference Alvarado-Esquivel, Sifuentes-Alvarez and Salas-Martinez2014a, Reference Alvarado-Esquivel, Sifuentes-Alvarez and Salas-Martinez2014b).
Participants completed the BEDS in the waiting room. Afterwards, they underwent a semistructured clinical interview according to the DSM-5 (American Psychiatric Association, 2014) criteria for MDD. The interview was conducted by a psychiatrist who was blind to the BEDS score at the time of the assessment. Depression was defined as a dichotomous variable based on the presence or absence of MDD.
Statistical analysis
Quantitative variables were reported as mean ± standard deviation when the distribution was normal and as median and range (minimum, maximum) when it was not. Categorical variables were expressed as absolute and relative frequencies. Cronbach's alpha coefficient was used to examine the internal reliability of the scale. To assess the accuracy of the instrument to discriminate between cases and not cases of depression, we performed receiver operating characteristic analysis. Sensitivity and specificity values were used to estimate the criterion validity of different threshold scores of BEDS compared with the reference standard (DSM-5 MDD diagnosis). Prevalence was also determined at various cutoff points (as a percentage). All analyses were performed using STATA, version 12.1, software (StataCorp, 2011).
Results
The sample comprised 70 participants, with a median age of 56.5 years (range, 20–85 years). Fifty (71%) were female and had 8 years (range, 0–22 years) of formal education. Most participants were married or cohabitating (55.7%, n = 39), had an Eastern Cooperative Oncology Group Performance Status score of 1 (55.7%), and a Palliative Prognosis Index survival clinical prediction of A (75.7%). The most frequent oncological diagnosis was gynecological (24.3%, n = 17), urological (20%, n = 14), and breast cancer (15.7%, n = 11). The clinical and sociodemographic characteristics of the sample are described in Table 1.
Clinical and sociodemographic characteristics of the participants (N = 70)
ECOG, Eastern Cooperative Oncology Group; KPS, Karnofsky Performance Scale; PPI, Palliative Prognostic Index.
The reliability of the translated BEDS, as assessed by Cronbach's alpha coefficient of internal consistency, ranged from 0.63 for item 3 to 0.73 for item 6, with an α = 0.71 result for the complete Mexican BEDS (Table 2).
BEDS Cronbach's alpha coefficient for internal consistency
The definition of case using ≥6 score has a sensitivity of 64.3% and a specificity of 75%. In contrast, lowering the cutoff point to ≥5 increases sensitivity to 85.7% with 62.5% specificity. Results of the different threshold scores identifying depression in BEDs are shown in Table 3.
Criterion validity and prevalence of Brief Edinburgh Depression Scale identified depression at different threshold scores
The prevalence of depression identified by the DSM-5 interview was 20% (n = 14). When the Mexican version of the BEDS was used, with a threshold score of 5, 20.5% of palliative advanced cancer patients were identified as cases of depression (Table 3).
Receiver operating characteristic analysis found the area under the curve for the scale to be 0.826, 95% confidence interval = 0.719, 0.933, p < .0001, which represents good accuracy to discriminate between cases and not cases of depression (Figure 1).
Receiver operating characteristic curve. Area under the curve = 0.8259.
Discussion
The development of psychometric scales is necessary for the evaluation of complex phenomena such as depression; however, most of these tools are created in other countries and precludes their usage in our language and population. To assess if the instrument works in the same way in different scenarios, it must be validated in the setting it is required to be used. This is especially important in populations with particular characteristics, such as patients with advanced cancer in palliative care, in whom the cancer symptomatology and physical symptoms of depression may overlap. In an effort to improve depression detection, there has been increasing interest in the use of screening and case-finding tools. The results in our study support the Mexican BEDS as a valid case-finding tool for depression in patients with advanced cancer in a palliative care unit.
In the original study, the cutoff score ≥6 gave a sensitivity of 72% and specificity of 83% (Mari Lloyd-Williams et al., Reference Lloyd-Williams, Shiels and Dowrick2007). In our results, that cutoff point yielded lower sensitivity and specificity values. The ≥7 score was the most stable, with 64.3% sensitivity and 87.5% specificity; however, considering the impact of depression on a frail population with multiple risk factors, an approach that increases diagnosis and, subsequently, proper treatment, may be more useful. A higher sensitivity could be better suited to reduce the proportion of missed cases at the expense of false positives. The cutoff score ≥5 had 85.7% sensitivity and 62.5% specificity, which is likely to be more useful for identifying patients who need a more in-depth assessment of their mood.
The prevalence of depression in Mexican palliative patients is unknown. In this study, we obtained a prevalence of MDD, according to the DSM-5 interview, of 20%; however, because of the nonprobabilistic sampling and sample size, this prevalence may not properly describe our population. Furthermore, we approached only palliative care patients who were cognitively lucid, with higher performance status, and able to tolerate an interview. Because cognitive impairment and disabling illness may be associated with depression, the actual prevalence could be higher than the one we found. When the Mexican BEDS was used with a threshold score of 5, it identified 20.5% of advanced cancer patients as cases of depression, which reflects good criterion validity.
The Spanish-translated BEDS had a Cronbach´s alpha coefficient of 0.71, which is considered acceptable. This result may be affected by the number of items on the scale; in a brief assessment, the value of alpha is reduced. The reliability of the translated version was similar to that reported by other authors (Lee et al., Reference Lee, Kim and Ko2009; Lloyd-Williams et al., Reference Lloyd-Williams, Shiels and Dowrick2007; Rhondali et al., Reference Rhondali, Perceau and Berthiller2012).
The Spanish-language Mexican BEDS is the first case-finding tool for depression validated in Mexican patients with advanced cancer in palliative care. We suggest using ≥5 as a cutoff score to improve detection of cases.
Acknowledgments
The authors would like to thank to Bruno Ocampo-Garcés, M.D., for his valuable collaboration in the development of this paper. Also, we would like to thank all the patients who participated in this study.
Conflicts of interest
Leticia Ascencio-Huertas, Emma Verástegui and Silvia Allende-Pérez are supported by National Council of Science and Technology (CONACyT) National System of Researchers and have not conflict of interest. The rest of the authors do not report actual or potential conflicts of interest.
