Investigating Symptom-Specific Effects of Psychotherapy and Antidepressant Medication in the Treatment of Major Depressive Disorder
F. V. O. Horneman S2352567 Master Thesis Clinical Psychology Supervisor: Dr. E. I. Fried Institute of Psychology Universiteit Leiden 31-07-2019
Abstract
Major Depressive Disorder (MDD) is a heterogeneous mental disorder which is the leading cause of disability across the world. Currently, the first-line treatment for MDD is based on Cognitive Behaviour Therapy and pharmacological treatment with antidepressant medications. Evidence suggests that both of these treatment modalities are effective, however, it is not well understood whether these treatments affect specific depressive symptoms differentially.
The present exploratory study investigated differential treatment effects of Cognitive Behaviour Therapy, antidepressant medication (paroxetine) and pill-placebo in a clinical sample by
reanalysing data from a previously published randomised controlled trial. The aim of this investigation was to explore whether support for previously reported findings concerning
differential symptom-level improvement would be found with a widely used MDD scale; the Beck Depression Inventory.
The American sample consisted of 192 outpatients who were randomly assigned to receive pill-placebo (N = 48), Cognitive Behaviour Therapy (N = 50) and paroxetine (N = 94) for eight weeks. A Multivariate Analysis of Variance was used to analyse the data.
Statistically significant symptom-level differences were not found between the three treatment conditions. The descriptive statistics revealed preliminary support for previous findings and they provide direction for future investigations in this area. Paroxetine was found to outperform (numerically) the other treatment conditions with respect to the majority of the depressive symptoms in the BDI.
Possible explanations for the results are discussed. Sufficiently powered future investigations are needed in this area to elucidate the symptom-specific effects of MDD treatments.
Introduction
Major Depressive Disorder (MDD) is a common recurrent mental disorder; it is the leading cause of disability worldwide and it comes with a great economic cost (World Health Organization, 2017). The detrimental effects of depression are linked to numerous global issues such as poor overall health, early mortality and low educational attainment (World Health Organization, 2017). Arguably, the overall burden of depression is compelling, which is why effective treatments are critically needed. MDD is characterised by an array of somatic and psychological symptoms such as sadness, diminished interest in activities, suicidal ideation, fatigue and cognitive dysfunction. Thus, MDD is an extremely heterogeneous disorder which poses considerable challenges to clinical research, diagnosis and treatment.
MDD is one of the most common mental health disorders across the world and while the etiology of MDD remains elusive, it has been well established that the cause of MDD relates to a complex interplay between biological, genetic and social factors (Saveanu & Nemeroff, 2012). A recent meta-analysis which evaluated the prevalence of depression by analysing data from one million individuals from 30 different countries between 1994 and 2014 reported a lifetime prevalence of 10.8% for depression (Lim et al., 2018). The WHO has estimated that 4.4% of the entire world’s population suffer from MDD. The prevalence of depression varies by region, for example, approximately 12% of the global population who suffer from depression are from Europe, whereas 21% are from the Western Pacific region and 27% are from the South-East Asia region (World Health Organization, 2017). Therefore, depression is a highly prevalent mental health disorder across the world albeit, prevalence rates vary greatly by geographical region.
MDD is a moderately heritable mental disorder, the genetic heritability of MDD is estimated to be close to 40% (Sullivan, Neale & Kendler, 2000). Moreover, there is evidence that some
dysfunction are more heritable than other depressive symptoms (Jang, Livesley, Taylor, Stein & Moon, 2004). MDD affects all age groups but it is more prevalent among older adults and
individuals who are in their twenties (American Psychiatric Association, 2013). Moreover, MDD is more common in females than in males (Albert, 2015). Currently, MDD can be treated effectively with antidepressant medications and psychotherapy, however, it is well established that all patients cannot be treated successfully with the available treatment modalities (Cuijpers, 2018). Thus, it is clear that future research in this field should aim to improve treatment for MDD. One way forward with regard to this matter is to improve MDD’s measurement, because better measurement may foster treatment efficacy and ultimately aid diagnosis for this heterogeneous disorder.
Issues Related to MDD Symptomatology
Currently, MDD diagnosis is based on clinical interviews and self-report inventories that measure depressive symptoms. Examples of measures that are often used in research are the Beck Depression Inventory-II (hereon BDI) (Beck, Steer, Ball & Ranieri,1996) and the Hamilton Depression Rating Scale (Hamilton, 1960). Both scales measure depressive symptoms, however, Beck’s scale is a self-report measure, whereas the Hamilton scale is administered by a clinician. These scales provide sum-scores and cut-off scores that represent depression severity in any given individual. The sum-score approach posits that all symptoms are a reflection of a single underlying construct, however, research has revealed that the two aforementioned scales do not measure such a construct for depression (Ballard et al., 2018; Fried et al., 2016).
The American Psychiatric Association (2013) has produced the current diagnostic criteria for MDD which are presented in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). The DSM-5 diagnosis for MDD requires an individual to experience five or more depressive symptoms during a two-week period nearly every day. Moreover, at least one of the symptoms is required to be either depressed mood or diminished interest in pleasurable
diagnosis. These symptoms are: significant weight loss or weight gain, insomnia or hypersomnia, loss of energy, psychomotor agitation or retardation, feelings of worthlessness or guilt, diminished ability to concentrate and recurrent thoughts of death. Notably, these symptoms cover a range of cognitive, physical and affective symptoms, some of which are the opposite extremes of the same symptom (e.g. insomnia and hypersomnia). Consequently, the current diagnostic criteria allow at least 227 different symptom profiles that can yield an MDD diagnosis (Fried & Nesse, 2015). It follows that symptom profiles that do not share any symptoms in common may lead to an MDD diagnosis. Moreover, researchers often treat MDD scales as if they were interchangeable and they often make general remarks about MDD based on a single outcome measure which is problematic given the collection of symptoms that contribute to MDD (Santor et al., 2006). Therefore, attempts to analyse individual depressive symptoms with different scales instead of placing focus solely on a single scale’s sum-score are justified and needed in clinical psychology when treatment efficacy is being evaluated.
Because depressive symptoms in the DSM-5 are diverse and derived from clinical
experience rather than research findings, concerns about these symptoms’ utility have been raised (Fried & Nesse, 2015). Researchers have also identified clinically meaningful symptoms of
depression that are not currently recognised by the DSM-5, such as anger and anxiety (Spielberger & Reheiser, 2009). The clinician inter-rater reliability for (adult) MDD diagnoses has been reported to be in the questionable range (kappa = 0.28) which means that clinicians cannot unanimously agree on who meets the diagnostic criteria for depression (Regier et al., 2013). It has also been reported that not all DSM-5 symptoms of depression are as good indicators of impairment as others (Fried & Nesse, 2014). For example, low mood, concentration difficulties, fatigue and loss of interest have been found to impair psychosocial functioning in depressed individuals to a greater extent than sleeping problems and changes in bodyweight (Fried & Nesse, 2014). Furthermore, a recent study utilising network analysis found that a web-based intervention for MDD directly
decreased sadness and indecision whereas, sleeping problems and weight problems remained unchanged post-intervention (Mullarkey, Stein, Pearson & Beevers, 2019).
To summarise, the presented evidence arguably implies that the depressive symptoms presented in the DSM-5 are not unanimously equal indicators of MDD and that symptoms may respond to treatment differentially. Together these findings provide a rationale for studying differential symptom improvement in MDD during treatment, because mounting evidence implies that sum-score analyses overlook important symptom-level information in a heterogeneous disorder like MDD.
Contemporary Treatment of MDD
Cuijpers (2018) has highlighted that MDD individuals form a diverse clinical population due to their differential responses to treatment. That is, some individuals benefit from treatment to a great extent while others respond poorly to it. Moreover, a considerable proportion of MDD
individuals experience recovery without any treatment. These notions advocate the need to study differences between patients and how different treatments alleviate specific symptoms. Research in this area has abundant practical significance for clinical psychology because a better understanding about which specific symptoms predict differential outcomes during treatment improve treatment efficacy and patient care. Furthermore, the establishment of a more well-defined diagnostic criteria in the forthcoming edition of the DSM is dependent on empirically supported research findings concerning MDD’s symptomatology.
Today, the first line of treatment for MDD is psychotherapy and antidepressant medication and research has shown that these treatment modalities are effective and widely used (Cuijpers, 2018). Treatment options depend on the severity of MDD; mild to moderate depression is often treated with counselling and psychotherapy, whereas antidepressant medications have been found to be most effective for moderate to severe depression (Price, Butler, Hatcher & Von Korff, 2007). A
combination of psychotherapy and antidepressant medication has been found to be more effective than psychotherapy or antidepressant medication alone (Cuijpers et al., 2014). Severe depression may also be treated with brain stimulation, such as transcranial direct current stimulation and electroconvulsive therapy (Bennabi & Haffen, 2018; Kho, van Vreeswijk, Simpson & Zwinderman, 2003). Recently, there has also been an increase in self-help and web-based interventions for mild to moderate depression with promising results (Andrews et al., 2018). Treatment modalities other than psychotherapy and antidepressant medication are beyond the scope of this research project and therefore discussed elsewhere, however, it can be concluded that contemporary approaches to MDD treatment are diverse. Below, the two most common treatments for MDD which are the subject of this research, are discussed.
CBT and Antidepressant Medications in the Treatment of MDD
Today, a type of psychotherapy called Cognitive Behaviour Therapy (CBT) is the gold standard in clinical practice for mood disorders (David, Cristea & Hofmann, 2018). CBT is an evidence-based psychotherapy which is widely used in the treatment of MDD (Gaudiano, 2008). Research on CBT and MDD has been extensive and a plethora of studies and randomised controlled trials have demonstrated its effectiveness (Barth et al., 2016; Cuijpers et al., 2013; Cuijpers, Cristea, Karyotaki, Reijnders & Huibers, 2016). When compared to waitlist or no treatment control groups CBT often produces moderate to large effect sizes whereas, when compared to other
psychotherapies effect sizes are favourable, albeit smaller (Hofmann, Asnaani, Vonk, Sawyer & Fang, 2012). According to some evidence, CBT and antidepressant medications seem to be equally efficacious in MDD treatment however, CBT seems to produce longer lasting effects (DeRubeis, Siegle & Hollon, 2008).
CBT’s theoretical framework was developed by Beck (1963) who proposed that
therefore targets for therapeutic interventions. With respect to MDD’s diagnostic criteria in the DSM-5, it is evident that CBT targets specifically, but not exclusively, cognitive symptoms such as feelings of worthlessness and excessive guilt. Because the theoretical background of CBT does not comprehensively elucidate the mechanism of which specific depressive symptoms improve during treatment, the present research aims to offer more insight into this inadequately understood topic. Because CBT treatment for mild to moderate MDD is typically delivered in 5 - 20 sessions (Wright, Brown, Thase & Basco, 2017) evidence for differential symptom improvement is anticipated to be found in the present research which analyses the results of an eight-week intervention.
A second main treatment option for MDD is pharmacological treatment with antidepressant medications. Today, they are widely used in the treatment of moderate to severe MDD (Trangle et al., 2012). Antidepressants’ popularity in MDD treatment is partly attributable to their availability; due to limited resources psychological treatments are inaccessible to many of those affected by MDD (Bower & Gilbody, 2005). An ongoing debate about antidepressant efficacy has prevailed in psychiatry and clinical psychology over the past years and the subject has received extensive research attention (Adli & Hegerl, 2014; Kirsch et al., 2008; Linde et al., 2015). Perhaps the most comprehensive evidence for antidepressant efficacy was presented in a recent network meta-analysis which reported that all 21 antidepressants that were included in the meta-meta-analysis were found to be more effective than placebo in the treatment of MDD (Cipriani et al., 2018). The summary effect sizes presented by Cipriani et al. (2018) were moderate (d = 0.3) and the results mirror other findings that have concluded that antidepressants are effective and that they should continue to be a first-line treatment for MDD alongside psychotherapy (Dunleavy, 2018).
These findings do not indicate that all antidepressants are equally effective; Cipriani et al. (2018) concluded that differences between antidepressants and their efficacy with respect to
different populations exist. The recent meta-analysis has also attracted criticism, and questions have been raised over the clinical significance of the reported effect sizes (Hengartner & Plöderl, 2018).
Additionally, antidepressants have not reached desirable efficacy when compared to placebo treatment and the literature on antidepressants is characterised by considerable publication bias (Pigott et al., 2010; Turner et al., 2008). Overall, whilst some researchers hold the viewpoint that antidepressants are effective, others respond to these notions by highlighting inadequate clinical significance and publication bias in their critique (Hengartner & Plöderl, 2018; Jakobsen et al., 2017; Turner et al., 2008). Thus, the debate surrounding antidepressant medications’ efficacy remains as contentious as ever which highlights the need for future research to study how MDD treatment could be improved.
Selective serotonin reuptake inhibitors (SSRIs) are a class of widely used antidepressants that aim to increase serotonin levels in the brain’s synapses (Berton & Nestler, 2006). SSRIs are not the only type of antidepressant medication that are currently used in MDD treatment, however, SSRIs dominate MDD’s pharmacological treatment because other types of antidepressant medication have been found to cause more severe side-effects (Breedlove & Watson, 2017). Paroxetine (also known by the name Paxil) is a common antidepressant and a SSRI medication which is used in the pharmacological treatment of MDD. According to meta-analytic evidence paroxetine has been found to be effective in the treatment of MDD with a modest effect size in comparison to placebo (Sugarman, Loree, Baltes, Grekin & Kirsch, 2014). In comparison to psychotherapy, paroxetine has been found to be as effective as CBT for moderate to severe MDD (DeRubeis et al., 2005).
A major issue concerning research on antidepressant efficacy relates to the fact that many of the antidepressant side-effects are also symptoms of MDD. For example, common side-effects of paroxetine include weight gain and insomnia, however, changes in bodyweight and sleeping problems are also part of the DSM-5 diagnostic criteria for MDD. Therefore, it is evident that antidepressant side-effects may obfuscate research findings because they cannot be easily distinguished from ‘pure’ depressive symptoms. Moreover, some symptoms may improve under
pharmacological treatment while others may worsen. For instance, it is well established that suicidal ideation (which is a symptom of MDD itself) may increase when treatment with antidepressants has commenced (Nevels et al., 2016). It has also been reported that SSRIs outperform placebo with regard to the mood symptom of MDD more consistently that has been previously reported (Hieronymus, Emilsson, Nilsson & Eriksson, 2016). However, literature on this topic is limited which is why further research is needed to investigate how certain antidepressant medications affect certain depressive symptoms.
Only a few studies have investigated paroxetine’s symptom specific effects on MDD. Schalet et al. (2016) found symptom specific effects favouring paroxetine when it was compared to a placebo, however, these effects were only significant with respect to MDD’s psychological symptoms on the Hamilton scale. That is, paroxetine did not outperform placebo with respect to the somatic symptoms of MDD. Fournier et al. (2013) studied differential treatment responses of CBT, paroxetine and pill-placebo in a sample of MDD outpatients. Fournier et al. (2013) reported that both paroxetine and CBT outperformed placebo by the eighth week of treatment with respect to cognitive / suicide symptoms (suicidal ideation, feelings of guilt, helplessness, hopelessness and worthlessness). Furthermore, CBT outperformed placebo and paroxetine with regard to the atypical vegetative symptoms (hypersomnia and weight gain / increased appetite) of MDD. These results were obtained by analysing patients’ symptoms with the Hamilton scale.
Overall, these preliminary findings imply that treatment with paroxetine and CBT may produce differential treatment response in moderate to severe MDD. These results also suggest that despite CBT’s emphasis on cognitive restructuring during treatment, CBT may not be superior to paroxetine in improving the cognitive symptoms of MDD. Due to the fact that Fournier et al. (2013) based their findings on five symptom clusters and a single outcome measure, symptom specific effects of paroxetine, placebo and CBT should be explored further with different scales and with an emphasis on specific symptoms instead of symptom clusters. Therefore, the present research
explores whether support for the reported preliminary findings concerning differential improvement of MDD symptoms can be found with another widely used measure of depression; the BDI.
The Aims of the Current Research
The aim of the current research is to expand the symptom-level research on MDD. This research will investigate the nature of symptom-level differences in MDD individuals after paroxetine, CBT and pill-placebo treatment by studying changes in depressive symptoms by reanalysing data from a previously published randomised controlled trial (DeRubeis et al., 2005). More specifically, this investigation explores whether previous findings concerning symptom clusters (Fournier et al., 2013) that have been obtained with the Hamilton scale concerning
differential improvement in MDD symptoms between the three treatment arms are supported when analyses are based on the 21-item BDI. Given the fact that the BDI and the Hamilton scale are not interchangeable measures of depression, the present research aims to offer important insight into the field of MDD research. It is anticipated that paroxetine and CBT will outperform pill-placebo with regard to the cognitive / suicide-related symptoms of MDD. Furthermore, it is anticipated that CBT will outperform paroxetine and placebo in reducing the atypical vegetative symptoms of MDD. Instead of focusing on symptom clusters the present research will analyse all 21 depressive symptoms presented in the BDI.
Methods Participants
An existing dataset from a placebo-controlled randomised trial was reanalysed for the purposes of this research. The original research (DeRubeis et al., 2005) was conducted between 1996 and 2003 at two sites in the United States. The participants were recruited from study clinics at the University of Pennsylvania, Philadelphia; and Vanderbilt University, Nashville, Tennessee. The original sample consisted of 240 participants. All participants were outpatients and they met the DSM-IV diagnostic criteria for moderate to severe MDD. Other inclusion criteria were: age
between the years of 18 and 70, fluency in English, and willingness and ability to give informed consent. Moreover, all participants achieved a score of 20 or higher on the Hamilton Depression Rating Scale on initial screening and during baseline measurement.
Exclusion criteria included the following: bipolar disorder, substance abuse or dependence requiring treatment, current or past psychosis, another axis I DSM-IV disorder requiring urgent treatment (e.g. eating disorder), antisocial, borderline or schizotypal personality disorder, acute suicidality requiring hospitalisation, medical condition that did not permit the use of paroxetine and nonresponse to paroxetine in the preceding year.
Design
In this randomised controlled trial participants were assigned to three different treatment arms which were antidepressant medication (paroxetine), Cognitive Behaviour Therapy (CBT) and pill-placebo. The antidepressant medication (N = 120) and CBT (N = 60) groups were studied for 16 weeks while the placebo group (N = 60) was studied for eight weeks (due to ethical
considerations). Therefore, comparisons between the two active treatments and the placebo group were restricted to results obtained after eight weeks of treatment in order to keep the time frame comparable. The antidepressant medication treatment arm included more participants than the CBT
and the placebo arm because participants in the antidepressant medication group were randomised to a subsequent study at the end of the trial.
Materials
In the current research the primary outcome measure was the 21-item BDI (Beck et al., 1996). The BDI (Appendix 1) is a self-report scale that evaluates MDD severity; it is a widely used measure in research and practice with clinical and non-clinical populations (Beck, Steer & Carbin, 1988). The BDI consists of 21-items that assess most of the major components of MDD such as the affective, cognitive and somatic symptoms. Patients are asked to assess the severity of their
experienced depressive symptoms during the preceding two week period by choosing an appropriate statement for each item. All items are measured on a four-point Likert scale which ranges from zero (symptom absent) to three (severe symptom). All four statements for each of the 21-items are presented from the respondent’s point of view (e.g. ‘I feel I am a total failure as a person’). All points are summed together to form a total score where a higher score indicates more severe MDD. The minimum BDI total score is zero and the maximum score is 63. In a clinical population a score in the 0-10 range indicates minimal depression, 10-18 indicate mild to moderate depression, 19-29 indicate moderate to severe depression and 30-63 indicate severe depression (Beck et al., 1988). The BDI is typically administered in approximately 10 minutes.
The BDI has been reported to demonstrate acceptable validity and reliability in various clinical trials (Wang & Gorenstein, 2013). In the current sample, the BDI demonstrated good internal reliability, Cronbach’s a = .89 (week one) and a = .92 (week eight).
Procedure
Access to the data was granted by researchers who were involved in the original study. All data were delivered to us in a deidentified form. Study approval for the clinical trial was obtained from the review boards of the University of Pennsylvania and Vanderbilt University in accordance with the Declaration of Helsinki. Participant recruitment was carried forward through referrals and media advertisements and informed consent was obtained from every participant before the
research commenced. Initially, 437 participants underwent assessment for this clinical trial. The screening process for exclusion criteria yielded a sample of 240 eligible participants. These 240 participants were randomly assigned to one of the study conditions (antidepressant medication, CBT and pill-placebo). Clinicians who completed the initial evaluations were blind to the objectives of the research. These evaluations were performed by using the Structured Clinical Interviews (Axis I and II) for DSM-IV.
Procedures in the Antidepressant Medication and the Pill-Placebo Treatment Arms Participants in the antidepressant treatment arm (N = 120) received paroxetine for eight weeks with a daily dose of 10 – 20 mg. This dose was raised during the course of the trial for those participants who responded poorly to the starter dose (maximum dose was 50 mg /day). Similarly, participants in the placebo group received pill-placebo for a total of eight weeks. Participants in the antidepressant medication group and the pill-placebo group received weekly treatment sessions during the first four weeks of the trial. After four weeks the sessions were delivered fortnightly. These 20 - 45 minute sessions included medication management and clinical management, however, clinicians refrained from using therapeutic techniques specific to CBT. Clinicians delivering the treatment and participants receiving the antidepressant medication and the pill-placebo were blind to the treatment condition during the first eight weeks of the trial. After eight weeks, the treatment condition was revealed to the participants and the clinicians, however, the
interviewer remained blind. Following the unblinding, individuals in the placebo group were offered subsequent treatment free of charge, while individuals in the antidepressant medication group continued treatment with paroxetine. Participants who did not respond to paroxetine during the eight week period received an add-on treatment with lithium carbonate or desipramine
hydrochloride. The trial was continued for another eight weeks without the placebo group following the unblinding.
Procedures in the CBT Treatment Arm
Treatment in the CBT group was delivered by six experienced therapists who adhered to standard CBT treatment practices for MDD and for comorbid personality disorders. During the first four weeks, 50-minute sessions were delivered twice a week. During the following eight weeks, sessions were delivered once or twice a week. For the last four weeks of the trial treatment was reduced to a single session per week. Therapists engaged in group supervision throughout the clinical trial. The BDI assessments were administered weekly to all participants in each of the treatment arms.
Statistical Analyses
Participants who had not provided both week one and eight data were excluded from the dataset. Those participants who had provided week two and eight data instead of week one and eight data were included in the analyses by treating week two data as week one data. Moreover, participants who had missing data for more than one BDI item were removed from the dataset. Participants who were removed from the dataset did not impact the three groups disproportionally. A flowchart describing the flow of the project is presented in Figure 1. Participant characteristics and descriptive statistics are presented in Tables 1 and 2.
Difference scores between week one and week eight assessment points were computed for each participant and these scores were subjected to statistical analyses. A MANOVA (Multivariate Analysis of Variance) was performed to study whether differential improvement occurred between any of the three treatment arms with regard to the 21 depressive symptoms. This exploratory approach was chosen due to the constraints of the relatively small sample size given the complex multivariate nature of the research question. All analyses were conducted in SPSS version 26 (IBM, 2019).
A post-hoc power analysis for MANOVA was performed in G*Power. With a total sample size of 192 participants, three groups and 21 response variables, alpha was set at .05 and the effect size (f2) at .252 = .0625 (medium). The obtained power was .66 (66%) indicating inadequate sample size for a multivariate test in the present study.
Figure 1. Flow of the Project Full sample (N = 240) Placebo (N = 120) CBT (N = 60) Antidepressant medication (N = 60)
Excluded (N = 44) due to missing week 1 and 8 BDI data
Antidepressant medication (N = 94) Week 2 data used because of missing week 1 data (N = 7) Participants
removed due to more than one missing BDI item (N = 3)
CBT (N = 50) Week 2 data used because of missing week 1 data (N = 3) Placebo (N = 48)
Week 2 data used because of missing week 1 data (N = 4) Participants removed due to more than one missing BDI item (N = 1)
Results Descriptive Statistics
As shown in Figure 1, the group sizes in the present study corresponded almost
proportionally to the number of participants in all treatment arms in the study by DeRubeis et al. (2005). All treatment arms had more female than male participants. Age and gender distributions are presented in Table 1.
Table 1
Characteristics of Study Participants Variable Total sample
(N = 192) Placebo (N = 48) CBT (N = 50) Antidepressant medication (N = 94) Age, M (SD) 40.76 (11.20) 42.25 (11.99) 41.38 (11.17) 39.67 (10.79) Female (%) 113 (59%) 29 (60%) 27 (54%) 57 (61%) Male (%) 79 (41%) 19 (40%) 23 (46%) 37 (39%)
A total of 192 participants were included in the statistical analyses. Week one and eight mean scores for all 21 BDI items are presented in Table 2. Moreover, Table 2 shows the mean difference scores for each of the BDI items between weeks one and eight. The three groups were slightly different with regard to specific depressive symptoms due to missing symptom-level data. One symptom-level datapoint was missing for BDI items ‘loss of pleasure’ (placebo group), ‘guilty feelings’ (antidepressant medication group) and ‘suicidal thoughts’ (antidepressant medication group). Three participants (two in the CBT group and one in the antidepressant medication group) had a missing datapoint for the BDI item ‘irritability’. As shown in Table 2, all 21 depressive symptoms decreased numerically in all treatment arms during the course of the trial.
Table 2
Means, Standard Deviations and Difference Scores of the 21 BDI Items Placebo (N = 48) M (SD) CBT (N = 50) M (SD) ADM (N = 94) M (SD)
BDI item Week 1 Week 8 D Week 1 Week 8 D Week 1 Week 8 D
Sadness 1.11 (.67) .77 (.59) -.34 (.67) 1.04 (.57) .71 (.71) -.33 (.78) 1.10 (.68) .38 (.61) -.71 (.82) Pessimism 1.36 (.82) 1.10 (.83) -.28 (.85) 1.04 (.61) .79 (.65) -.25 (.56) 1.22 (.75) .62 (.74) -.60 (.80) Past Failure 1.51 (.80) 1.21 (.77) -.32 (.86) 1.36 (.85) .98 (.76) -.38 (.67) 1.54 (.77) .78 (.81) -.76 (.91) Loss of Pleasure 1.68* (.73) 1.10 (.78) -.55 (.90) 1.44 (.79) .98 (.70) -.48 (.74) 1.61 (.82) .80 (.76) -.82 (.94) Guilty Feelings 1.40 (.90) .90 (.72) -.51 (.83) 1.16 (.84) .69 (.83) -.48 (.82) 1.22* (.81) .50 (.64) -.73 (.91) Punishment Feelings 1.02 (1.09) .65 (1.08) -.36 (1.01) .50 (.74) .42 (.79) -.10 (.66) .77 (1.11) .29 (.76) -.46 (.85) Self-Dislike 1.74 (.87) 1.27 (.96) -.45 (.95) 1.50 (.91) .94 (.86) -.52 (.85) 1.68 (.87) .91 (.98) -.79 (.95) Self-Criticalness 1.36 (.90) 1.13 (.87) -.21 (.93) 1.32 (.68) .83 (.66) -.50 (.65) 1.49 (.87) .71* (.81) -.79 (.98) Suicidal Thoughts .64 (.74) .35 (.60) -.28 (.45) .32 (.51) .15 (.36) -.15 (.50) .41 (.54) .17* (.38) -.22 (.51) Crying 1.11 (1.09) .79 (1.01) -.30 (1.04) 1.08 (1.10) .71 (1.07) -.40 (1.33) .97 (1.07) .41 (.89) -.56 (1.19) Agitation .85 (.88) .81 (.64) -.02 (.92) .98 (.74) .81 (.79) -.19 (.79) .92 (.73) .46 (.62) -.48 (.87) Loss of Interest 1.72 (.88) 1.29 (.90) -.40 (.97) 1.50 (.86) .98 (.79) -.54 (.92) 1.56 (.89) .76 (.76) -.81 (1.03) Indecisiveness 1.47 (.98) .98 (1.00) -.47 (1.04) 1.46 (.93) .90 (.86) -.58 (.77) 1.35 (.97) .67 (.84) -.68 (1.02) Worthlessness 1.43 (.90) .96 (.87) -.45 (.77) 1.08 (.80) .67 (.78) -.42 (.61) 1.31 (.87) .46 (.75) -.85 (.82) Loss of Energy 1.32 (.81) .96 (.74) -.34 (.89) 1.44 (.81) 1.06 (.79) -.38 (.73) 1.48 (.87) .86 (.74) -.65 (.94)
Table 1 (continued) Changes in Sleep .96 (1.08) .65 (.81) -.32 (1.09) 1.02 (1.08) .46 (.62) -.58 (1.11) .86 (1.11) .45 (.82) -.45 (1.32) Irritability 1.21 (.88) .88 (.73) -.32 (.81) 1.14 (.81) .94* (.86) -.21 (.90) .99 (.85) .42* (.65) -.56 (.92) Changes in Appetite .68 (.81) .40 (.74) -.28 (.74) .36 (.63) .29 (.62) -.08 (.82) .67 (.90) .39 (.65) -.27 (1.01) Concentration Difficulty 1.51 (.83) 1.15 (.90) -.34 (.92) 1.48 (.79) 1.10 (.78) -.38 (.70) 1.41 (.82) .72 (.76) -.70 (.91) Fatigue 1.28 (.88) 1.00 (.88) -.26 (.94) 1.48 (.84) 1.08 (.85) -.42 (.77) 1.46 (1.02) .79 (.78) -.68 (1.05) Loss of interest in Sex 1.26 (1.17) .98 (1.06) -.26 (1.11) 1.22 (1.08) 1.00 (1.03) -.23 (.69) 1.41 (1.08) .98 (1.07) -.45 (1.04) Note. ADM = Antidepressant medication, D = Difference score
*indicates missing item-level data
The descriptive statistics depicted in Table 2 show that the largest numerical decreases with regard to the difference scores between weeks one and eight were found in the antidepressant medication group for all symptoms, excluding ‘suicidal thoughts’, ‘changes in sleep’ and ‘changes in appetite’. Contrary to initial expectations the largest numerical decrease for ‘suicidal thoughts’ was found in the placebo group, however, all three groups showed only small numerical decreases for this symptom. As expected, the largest decrease for the symptom ‘changes in sleep’ was found in the CBT group. Contrary to the initial hypothesis the largest numerical decrease for the symptom ‘changes in appetite’ was found in the placebo group. As expected, paroxetine generally
outperformed (numerically) placebo with regard to the cognitive symptoms. However, contrary to initial expectations, numerical decrease in the CBT group with respect to the cognitive symptoms was only slightly greater relative to the placebo group (for some of the cognitive symptoms). In general, numerical differences across symptoms between the CBT group and the placebo group were small. Results from weeks one and eight, as well as the difference scores for each of the 21
symptoms are illustrated in Figures 2 – 22 in Appendix 2. The error bars represent two standard errors in each figure.
The overall mean changes in the 21 depressive symptoms between weeks one and eight with respect to the sum-scores were as follows, antidepressant medication group, M = -13.04, SD = 11.70, CBT group, M = -7.58, SD = 8.29, placebo group, M = -7.04, SD = 11.09. Figure 23 in Appendix 3 displays the change in the mean sum-scores for all treatment arms. The error bars represent two standard errors. A MANOVA was performed to study which specific symptoms were linked to statistically significant changes in improvement between the three treatment arms.
Inferential Statistics
A MANOVA was used to analyse the difference scores by entering the 21 dependent variables (depressive symptoms) in the same analysis with one factor which had three levels (three treatment conditions). Pillai’s trace was used as the multivariate test due to its robustness against statistical assumptions’ violations. Using Pillai’s trace, there was a non-significant effect of treatment condition on the improvement of the 21 depressive symptoms, V = 0.28, F(42, 328) = 1.28, p = 0.121. Therefore, there were no statistically significant differences in symptom
improvement between the three treatment arms meaning, that the null hypothesis which postulated that the change in symptoms was the same in all treatment conditions could not be rejected.
Discussion
The present study investigated the symptom specific improvement of moderate to severe MDD during treatment with pill-placebo, paroxetine and CBT by reanalysing data from an eight-week randomised controlled trial. Assessments were carried out at weeks one and eight in all three treatment arms, and the 21-item BDI was used as the outcome measure for all conditions. The aim of the present study was to expand the literature on symptom-level improvement in a clinical
sample and study whether support would be found for previously reported symptom specific effects of paroxetine, CBT and pill-placebo.
Support for statistically significant differences in improvement in the three treatment conditions was not found in the present study. The non-significant result highlights the key limitation of the present study, which relates to insufficient statistical power. Nevertheless, exploration of the obtained descriptive statistics provided preliminary support for the initial hypotheses, which were based on the findings of Fournier et al. (2013) who studied the same sample in a similar research by using the Hamilton scale as the outcome measure.
The results of the present study indicate that all symptoms decreased numerically in all treatment conditions. Contrary to initial expectations, paroxetine and CBT did not outperform placebo with respect to the symptom in the BDI that relates to suicidal ideation. In fact, very small numerical reductions in this symptom were found in all treatment arms. These results highlight the complex nature of suicidality and suicidal ideation which is well established in the literature (De Berardis, Martinotti & Di Giannantonio, 2018). A possible explanation for the results (meaning numerical changes in the difference scores) of the present study that are in disagreement with the results of Fournier et al. (2013) may be related to the fact that Fournier et al. combined symptoms into cluster symptoms (e.g. cognitive and suicide symptoms), which may have inflated effects for particular symptoms. Secondly, the Hamilton scale and the BDI are not interchangeable measures of depression and the BDI has more cognitive symptoms of depression than the Hamilton scale. Fournier et al. (2013) based their findings on the clinician-administered Hamilton scale which measures symptoms on scales that range from zero to four and zero to two, whereas, the BDI is a self-report measure that ranges from zero to three. Consequently, previous studies have reported that the Hamilton scale is more sensitive in detecting changes in depressive symptoms relative to the BDI (Edwards et al., 1984; Schneibel et al., 2012). Furthermore, the results of the clinician-administered Hamilton scale may be affected by clinician bias. The numerical changes in the mean
scores of specific MDD symptoms found in the present study support these notions when the results are compared to the results of Fournier et al. (2013).
As expected, the numerical changes in the symptoms’ mean scores indicated that the antidepressant medication group outperformed placebo with respect to the cognitive symptoms of MDD. However, contrary to initial expectations, CBT outperformed placebo (numerically) with respect to only two of the cognitive symptoms (‘self-dislike’ and ‘self-criticalness’) and all
differences between CBT and placebo were modest. Overall, CBT resulted in slightly advantageous numerical improvement relative to placebo with respect to individual symptoms that do not form a coherent class of depressive symptoms (‘fatigue’, ‘agitation’, ‘changes in sleep’, ‘loss of interest’, ‘self-dislike’, ‘self-criticalness’). These results provide only some support for the findings by Fournier et al. (2013) who reported CBT superiority relative to placebo with respect to cognitive / suicide and atypical vegetative symptoms (hypersomnia and increased appetite) of MDD. Thus, the fact that CBT was found to be numerically superior to placebo and paroxetine with respect to the symptom of ‘changes in sleep’ in the present study, cannot be regarded as conclusive evidence, because the direction of the change in sleep disturbance is unknown in the present study.
Overall, the reported results provide preliminary support for the notion that paroxetine may be superior to pill-placebo and CBT in the treatment of clinical MDD with respect to almost all BDI symptoms after eight weeks of treatment. Further, the results indicate that there may be modest advantages of CBT relative to placebo with relation to some of the depressive symptoms (physical, cognitive and sleep disturbance symptoms) that are associated with MDD. It should be emphasised that the results can only be generalised to patients who are diagnosed with moderate to severe MDD. Furthermore, clinical trial populations often form clinically ‘odd’ representations of patients due to the comprehensive exclusion criteria of these trials. Thus, remarks concerning the
All aforementioned effects are likely small if statistically significant support is found for them in future investigations. Nevertheless, these preliminary findings indicate that different treatments may produce differential improvement in depressive symptoms, and that these effects should be explored further with larger sample sizes in future work. Ultimately, research in this area has the potential to inform treatment decisions in the future by facilitating the establishment of customised treatments for MDD patients, who are known to form a very heterogenous clinical population. Furthermore, the results of the present study add to the existing evidence by suggesting that changes in cognitive symptoms of MDD are not exclusively produced by CBT. Further
investigations in this area could elucidate the working mechanisms of CBT and paroxetine. An interesting direction for future work on differential MDD symptom improvement is to compare symptom-level improvement of different antidepressant medications, because SSRIs differ with respect to their side-effects (Santarsieri & Schwartz, 2015). Moreover, it is known that both components of CBT (cognitive restructuring and behavioural strategies) are effective treatment modalities for MDD, however, it is not clear whether these elements produce differential relief in depressive symptoms (Ekers, Richards & Gilbody, 2008). Therefore, an interesting avenue for future work is to investigate whether the two main components of CBT produce differential symptom relief in MDD. Lastly, the results of the present study are restricted to results obtained after eight weeks of treatment. Future research should study whether potential differential effects are found in long-term studies, because there is evidence to suggest that CBT may have longer lasting effects than antidepressant medications (DeRubeis et al., 2008).
The strength of the present study relates to the fact that difference scores were obtained and calculated for all 21 depressive symptoms and that the outcome measure was a self-report scale which is free from clinician bias and has other advantages compared to clinician-administered scales (Zimmerman et al., 2018). As mentioned, the key limitation of the present study is linked to insufficient statistical power. The detection of small statistical effects requires considerable
statistical power which the current research was lacking. Furthermore, MANOVAs are often followed up with univariate ANOVAs if a significant overall effect is detected for the initial multivariate test. The number of the follow-up tests is dependent on the number of the outcome variables. The fact that the present research had 21 outcome variables highlights the issue relating to power in the kind of research that aims to study differential improvement with regard to multiple symptoms. Because the most important factor contributing to statistical power is the size of the sample, future research in this area should be conducted with large sample sizes that obtain sufficient statistical power.
Secondly, due to the aims of the original study (DeRubeis et al., 2005) the sample sizes of the three treatment arms were unequal in the present study. Unequal sample sizes violate the
assumptions of MANOVA which was another limitation in the present study. Additionally, unequal sample sizes are also related to problems with statistical power, because comparisons between groups that do not have equal sample sizes also differ with respect to statistical power.
In conclusion, although the results of the present exploratory study were not statistically significant, they offer modest preliminary support for the notion that treatment with placebo, paroxetine and CBT may be associated with differential improvement in clinical depression. The results suggest that paroxetine may be more effective than CBT and placebo with respect to almost all of the depressive symptoms presented in the BDI. In agreement with prior research in this area, the results of this study support the notion that future studies should employ the practice of using more than just one scale in MDD research.
References
Adli, M., & Hegerl, U. (2014). Do we underestimate the benefits of antidepressants? The Lancet, 383(9926), 1361-1362.
Albert, P. R. (2015). Why is depression more prevalent in women? Journal of Psychiatry & Neuroscience: JPN, 40(4), 219.
Andrews, G., Basu, A., Cuijpers, P., Craske, M. G., McEvoy, P., English, C. L., & Newby, J. M. (2018). Computer therapy for the anxiety and depression disorders is effective, acceptable and practical health care: an updated meta-analysis. Journal of Anxiety Disorders, 55, 70-78.
Ballard, E. D., Yarrington, J. S., Farmer, C. A., Lener, M. S., Kadriu, B., Lally, N., ... & Zarate Jr, C. A. (2018). Parsing the heterogeneity of depression: An exploratory factor analysis across commonly used depression rating scales. Journal of Affective Disorders, 231, 51-57.
Barth, J., Munder, T., Gerger, H., Nüesch, E., Trelle, S., Znoj, H., ... & Cuijpers, P. (2016). Comparative efficacy of seven psychotherapeutic interventions for patients with depression: A network meta-analysis. Focus, 14(2), 229-243.
Beck, A. T. (1963). Thinking and depression: I. Idiosyncratic content and cognitive distortions. Archives of General Psychiatry, 9(4), 324-333.
Beck, A. T., Steer, R. A., Ball, R., & Ranieri, W. F. (1996). Comparison of Beck Depression Inventories-IA and-II in psychiatric outpatients. Journal of Personality Assessment, 67(3), 588-597.
Beck, A. T., Steer, R. A., & Carbin, M. G. (1988). Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8(1), 77-100.
Bennabi, D., & Haffen, E. (2018). Transcranial direct current stimulation (tDCS): A promising treatment for major depressive disorder? Brain Sciences, 8(5), 81.
Berton, O., & Nestler, E. J. (2006). New approaches to antidepressant drug discovery: beyond monoamines. Nature Reviews Neuroscience, 7(2), 137.
Bower, P., & Gilbody, S. (2005). Stepped care in psychological therapies: access, effectiveness and efficiency. Narrative literature review. The British Journal of Psychiatry, 186, 11–17.
Breedlove, S. M., Watson, N. V. (2017). Behavioral neuroscience (8th ed.) Sunderland, MA, USA:
Sinauer Associates, Inc., Publishers.
Cipriani, A., Furukawa, T. A., Salanti, G., Chaimani, A., Atkinson, L. Z., Ogawa, Y., ... & Egger, M. (2018). Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: A systematic review and network meta-analysis. Focus, 16(4), 420-429.
Cuijpers, P. (2018). The challenges of improving treatments for depression. Jama, 320(24), 2529-2530.
Cuijpers, P., Berking, M., Andersson, G., Quigley, L., Kleiboer, A., & Dobson, K. S. (2013). A meta-analysis of cognitive-behavioural therapy for adult depression, alone and in
comparison with other treatments. The Canadian Journal of Psychiatry, 58(7), 376-385.
Cuijpers, P., Cristea, I. A., Karyotaki, E., Reijnders, M., & Huibers, M. J. (2016). How effective are cognitive behavior therapies for major depression and anxiety disorders? A meta‐ analytic update of the evidence. World Psychiatry, 15(3), 245-258.
Cuijpers, P., Sijbrandij, M., Koole, S. L., Andersson, G., Beekman, A. T., & Reynolds III, C. F. (2014). Adding psychotherapy to antidepressant medication in depression and anxiety disorders: A meta-analysis. Focus, 12(3), 347-358.
David, D., Cristea, I., & Hofmann, S. G. (2018). Why cognitive behavioral therapy is the current gold standard of psychotherapy. Frontiers in Psychiatry, 9, 4.
De Berardis, D., Martinotti, G., & Di Giannantonio, M. (2018). Understanding the complex phenomenon of suicide: From research to clinical practice. Frontiers in Psychiatry, 9, 61.
DeRubeis, R. J., Hollon, S. D., Amsterdam, J. D., Shelton, R. C., Young, P. R., Salomon, R. M., ... & Gallop, R. (2005). Cognitive therapy vs medications in the treatment of moderate to severe depression. Archives of General Psychiatry, 62(4), 409-416.
DeRubeis, R. J., Siegle, G. J., & Hollon, S. D. (2008). Cognitive therapy versus medication for depression: Treatment outcomes and neural mechanisms. Nature Reviews
Neuroscience, 9(10), 788.
Dunleavy, D. (2018). Rational antidepressant use. BJPsych Bulletin, 42(3), 131-131.
Edwards, B. C., Lambert, M. J., Moran, P. W., McCully, T., Smith, K. C., & Ellingson, A. G. (1984). A meta‐analytic comparison of the Beck Depression Inventory and the Hamilton Rating Scale for Depression as measures of treatment outcome. British Journal of Clinical Psychology, 23(2), 93-99.
Ekers, D., Richards, D., & Gilbody, S. (2008). A meta-analysis of randomized trials of behavioural treatment of depression. Psychological Medicine, 38(5), 611-623.
Fournier, J. C., DeRubeis, R. J., Hollon, S. D., Gallop, R., Shelton, R. C., & Amsterdam, J. D. (2013). Differential change in specific depressive symptoms during antidepressant medication or cognitive therapy. Behaviour Research and Therapy, 51(7), 392-398.
Fried, E. I., & Nesse, R. M. (2014). The impact of individual depressive symptoms on impairment of psychosocial functioning. PloS one, 9(2), e90311.
Fried, E. I., & Nesse, R. M. (2015). Depression is not a consistent syndrome: an investigation of unique symptom patterns in the STAR* D study. Journal of Affective Disorders, 172, 96-102.
Fried, E. I., van Borkulo, C. D., Epskamp, S., Schoevers, R. A., Tuerlinckx, F., & Borsboom, D. (2016). Measuring depression over time... Or not? Lack of unidimensionality and
longitudinal measurement invariance in four common rating scales of depression. Psychological Assessment, 28(11), 1354.
Gaudiano, B. A. (2008). Cognitive-behavioural therapies: Achievements and challenges. Evidence-Based Mental Health, 11(1), 5-7.
Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry, 23(1), 56.
Hengartner, M. P., & Plöderl, M. (2018). Statistically significant antidepressant-placebo differences on subjective symptom-rating scales do not prove that antidepressants work: Effect size and method bias matter! Frontiers in Psychiatry, 9, 517.
Hieronymus, F., Emilsson, J. F., Nilsson, S., & Eriksson, E. (2016). Consistent superiority of selective serotonin reuptake inhibitors over placebo in reducing depressed mood in patients with major depression. Molecular Psychiatry, 21(4), 523.
Hofmann, S. G., Asnaani, A., Vonk, I. J., Sawyer, A. T., & Fang, A. (2012). The efficacy of cognitive behavioral therapy: A review of meta-analyses. Cognitive Therapy and Research, 36(5), 427-440.
IBM Corp. Released 2019. IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY: IBM Corp.
Jakobsen, J. C., Katakam, K. K., Schou, A., Hellmuth, S. G., Stallknecht, S. E., Leth-Møller, K., ... & Krogh, J. (2017). Selective serotonin reuptake inhibitors versus placebo in patients with major depressive disorder. A systematic review with meta-analysis and Trial Sequential Analysis. BMC Psychiatry, 17(1), 58.
Jang, K. L., Livesley, W. J., Taylor, S., Stein, M. B., & Moon, E. C. (2004). Heritability of individual depressive symptoms. Journal of Affective Disorders, 80(2-3), 125-133.
Kho, K. H., van Vreeswijk, M. F., Simpson, S., & Zwinderman, A. H. (2003). A meta-analysis of electroconvulsive therapy efficacy in depression. The Journal of ECT, 19(3), 139-147.
Kirsch, I., Deacon, B. J., Huedo-Medina, T. B., Scoboria, A., Moore, T. J., & Johnson, B. T. (2008). Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Medicine, 5(2), e45.
Lim, G. Y., Tam, W. W., Lu, Y., Ho, C. S., Zhang, M. W., & Ho, R. C. (2018). Prevalence of depression in the community from 30 Countries between 1994 and 2014. Scientific Reports, 8(1), 2861.
Linde, K., Kriston, L., Rücker, G., Jamil, S., Schumann, I., Meissner, K., ... & Schneider, A. (2015). Efficacy and acceptability of pharmacological treatments for depressive disorders in primary care: Systematic review and network meta-analysis. The Annals of Family
Mullarkey, M. C., Stein, A., Pearson, R., & Beevers, C. G. (2019, January 17, Preprint). Network analyses reveal which symptoms improve (or not) following an Internet intervention
(Deprexis) for depression. Available from https://doi.org/10.31234/osf.io/ca5dg
Nevels, R. M., Gontkovsky, S. T., & Williams, B. E. (2016). Paroxetine—the antidepressant from hell? Probably not, but caution required. Psychopharmacology Bulletin, 46(1), 77.
Price, J., Butler, R., Hatcher, S., & Von Korff, M. (2007). Depression in adults: Psychological treatments and care pathways. BMJ Clinical Evidence, 2007.
Regier, D. A., Narrow, W. E., Clarke, D. E., Kraemer, H. C., Kuramoto, S. J., Kuhl, E. A., & Kupfer, D. J. (2013). DSM-5 field trials in the United States and Canada, Part II: Test-retest reliability of selected categorical diagnoses. American Journal of Psychiatry, 170(1), 59-70.
Santarsieri, D., & Schwartz, T. L. (2015). Antidepressant efficacy and side-effect burden: A quick guide for clinicians. Drugs In Context, 4.
Santor, D. A., Gregus, M., & Welch, A. (2006). Eight decades of measurement in
depression. Measurement: Interdisciplinary Research and Perspectives, 4(3), 135-155.
Schalet, B. D., Tang, T. Z., DeRubeis, R. J., Hollon, S. D., Amsterdam, J. D., & Shelton, R. C. (2016). Specific pharmacological effects of paroxetine comprise psychological but not somatic symptoms of depression. PloS One, 11(7), e0159647.
Schneibel, R., Brakemeier, E. L., Wilbertz, G., Dykierek, P., Zobel, I., & Schramm, E. (2012). Sensitivity to detect change and the correlation of clinical factors with the Hamilton Depression Rating Scale and the Beck Depression Inventory in depressed
inpatients. Psychiatry Research, 198(1), 62-67.
Spielberger, C. D., & Reheiser, E. C. (2009). Assessment of emotions: Anxiety, anger, depression, and curiosity. Applied Psychology: Health and Well‐Being, 1(3), 271-302.
Sugarman, M. A., Loree, A. M., Baltes, B. B., Grekin, E. R., & Kirsch, I. (2014). The efficacy of paroxetine and placebo in treating anxiety and depression: A meta-analysis of change on the Hamilton Rating Scales. PLoS One, 9(8), e106337.
Sullivan, P. F., Neale, M. C., & Kendler, K. S. (2000). Genetic epidemiology of major depression: Review and meta-analysis. American Journal of Psychiatry, 157(10), 1552-1562.
Trangle, M., Dieperink, B., Gabert, T., Haight, B., Lindvall, B., Mitchell, J., ... & Somers, K. (2012). Major depression in adults in primary care. Institute for Clinical Systems Improvement, 1-119.
Turner, E. H., Matthews, A. M., Linardatos, E., Tell, R. A., & Rosenthal, R. (2008). Selective publication of antidepressant trials and its influence on apparent efficacy. New England Journal of Medicine, 358(3), 252-260.
Wang, Y. P., & Gorenstein, C. (2013). Psychometric properties of the Beck Depression Inventory-II: A comprehensive review. Brazilian Journal of Psychiatry, 35(4), 416-431.
World Health Organization (2017). Depression and other common mental disorders, global health estimates. Retrieved from https://apps.who.int/iris/bitstream/handle/10665/254610/WHO-MSD-MER-2017.2-eng.pdf?sequence=1
Wright, J. H., Brown, G. K., Thase, M. E., & Basco, M. R. (2017). Learning cognitive-behavior therapy: An illustrated guide (2nd ed.) Arlington, VA: American Psychiatric Association
Publishing.
Zimmerman, M., Walsh, E., Friedman, M., Boerescu, D. A., & Attiullah, N. (2018). Are self-report scales as effective as clinician rating scales in measuring treatment response in routine clinical practice? Journal of Affective Disorders, 225, 449-452.
Appendices
Appendix 2. Figures 2 - 22 Depicting All 21 Depressive Symptoms
Figure 2. BDI Symptom Sadness
Figure 4. BDI Symptom Past Failure
Figure 6. BDI Symptom Guilty Feelings
Figure 8. BDI Symptom Self-Dislike
Figure 10. BDI Symptom Suicidal Thoughts
Figure 12. BDI Symptom Agitation
Figure 14. BDI Symptom Indecisiveness
Figure 16. BDI Symptom Loss of Energy
Figure 18. BDI Symptom Irritability
Figure 20. BDI Symptom Concentration Difficulty
Figure 22. BDI Symptom Loss of Interest in Sex Appendix 3. Figure 23.
SPSS output for the descriptive statistics that are presented in Table 2 are included in the supplementary SPSS output datafile (F.Horneman.descriptives.spv).
*Note. In the tables, BDI01 – BDI21 refer to week one data. BDI01_2 – BDI21_2 refer to week eight data. D.BDI.1 – D.BDI.21 refer to difference scores. Moreover, 0 = placebo group, 1 = antidepressant medication group, 2 = CBT group
