Understanding the emotions of patients with inadequate response to antidepressant treatments: results of an international online survey in patients with major depressive disorder

This survey was conducted in the United States, Canada, UK, Germany, France, and Spain between 15 March and 16 June, 2016. Respondents were recruited using consumer panels where all panelists have provided consent to participate in research. An online questionnaire was used to screen respondents and collect quantitative data. Access to the survey was secure, ensuring the anonymity of respondents.

All respondents were eligible for the survey if they met the following criteria: age 18 to 65 years, diagnosed with depression by a healthcare professional; no comorbid diagnoses of bipolar disorder, schizoaffective disorder, or schizophrenia; experiencing symptoms of depression most of the time over the past week; their depression has a moderate to extreme impact on overall functioning as measured by the Sheehan Disability Scale (SDS) [9]; has a psychiatrist or primary care physician (or, in the US, a nurse practitioner or physician assistant) who is primarily responsible for their treatment; currently on an ADT for at least 6 weeks at an adequate dose (see Appendix); they felt that their depression has not improved or only a little since the onset of the current episode (as measured by the Patient Global Impression of Improvement (PGI-I), and, finally, not working for a biasing employer (advertising agency or marketing research company or a company that manufactures, distributes or sells pharmaceuticals or health care products). Finally, per pharmacovigilance compliance regulations, all eligible respondents were required to provide consent for disclosure of any adverse-events reported within their responses.

Eligible survey participants all suffered from MDD with a current episode having moderate to extreme impact on overall functioning as measured by SDS, and to be on at least one ADT at an adequate dose for at least 6 weeks. Participants were considered as having an inadequate response to their current ADT if they indicated that their current symptoms had little to no improvement (based on the 7-item PGI-C, ranging from very much worse to very much improved).

The survey was conducted as an online 20-min questionnaire. All data collected was self-reported by participants and included both close-ended and open questions on the following topics: history of depression (onset of first episode, number of lifetime episodes, length of current episode), treatment history (hospitalizations, psychotherapy, treatment received for current episode, reason for discontinuation); change on current treatment (patients’ global impression of change, PGI-C), current symptoms of depression, other current symptoms, impact of depression on functioning (SDS), treatment goals, emotional feelings about their experience with medication, healthcare providers and overall healthcare system, including feelings of frustration, intensity of frustration and reasons for feeling frustrated. Here we present the results pertaining to their feeling towards medication and HCP.

Current symptoms of depression were assessed using a modified version of the Patient Health Questionnaire (PHQ-9) [10], excluding the question about suicidal ideation. The scale was the standard PHQ-9 where 0 = Not at all, 1 = Several days, 2 = More than half of the days and 3 = Nearly every day. Other (non-PHQ-9) current symptoms were elicited from a check list of clinical symptoms associated with depression (see list in Fig. 1). The lists of emotional feelings regarding their treatment and HCP, included in the survey were created using direct patient input from qualitative interviews conducted during the survey development (not shown). Reasons for frustration were asked using a voluntary open-end format and responses were manually coded. Coded responses were validated by a second coder and quality checked by a research analyst. A list of consequences of frustration was also included in the survey based on patient and physician input during pre-survey qualitative interviews. To gauge the patient’s perception on the overall level of impairment caused by their depression, the survey included the additional PHQ item 10 “How difficult have these problems made it for you to do your work, take care of things at home, or get along with other people?”, where the scale included was 1 = Not difficult at all, 2 = Somewhat difficult 3 = Very difficult and 4 = Extremely difficult.

Descriptive statistics included means and frequency of responses (%). A driver analyses was conducted using random forest analysis [11]. A random forest is a method where many decision trees (500 in our analysis) are constructed. The analysis selects a random sample of cases as the training set for each tree, and at each potential branching of the tree, a random subset of variables is evaluated as potential splits. The output of a random forest analysis provides ratio-level data of the influence that each characteristic has on the presence of dependent variable. Higher scores mean that, on average, the nodes created when the tree splits on this variable are more homogenous than nodes created by splitting on variables with lower scores – in other words they are considered more important. Patient characteristics entered as independent variables into the analysis included time since first episode, past hospitalization, past psychotherapy, number of episodes experienced, length of current episode, number of ADT failures, current treatment of monotherapy vs. combination therapy, change on current treatment (PGI-C), level of impairment (PHQ item 10) and disruption of daily living (SDS mean score). The dependent variable was the presence of moderate-to-severe frustration with their treatment. The random forest analysis was performed using the “randomForest” statistical package in R. Validity testing was done using a logistic regression to determine the explanatory power of the entered predictors.

A total of 18,083 adult (age 18–65) respondents diagnosed with MDD by a health care professional with no comorbid diagnoses of Bipolar Disorder, Schizoaffective or Schizophrenia were identified through our screening process using consumer panels (Fig. 1). Of the respondents with MDD, 11,466 were currently suffering from symptoms of depression and 93.5% of these (10,715) had at least a moderate impact on functioning. Approximately 56.3% (n = 6037) were being treated with an adequate dose of ADT for at least six weeks under the supervision of a psychiatrist, primary care physician (or nurse practitioner in the US). Of those adequately treated, 54% (3243) were experiencing an inadequate response to ADT. Finally, with drop-outs and quota caps at a country level, a total of 2096 respondents with inadequate response to ADT, completed the survey from six countries. These included the US (597, 28.5%), Canada (301, 14.4%), UK (300, 14.4%), Germany (299, 14.3%), France (299, 14.3%) and Spain (300, 14.4%).

The mean age of our survey respondents was 42.1 years and 56.7% were female (Table 1). While more than half of the respondents were working, 10% were on sick-leave or disability due to depression. The mean duration since onset of first episode was 10.5 years. Most respondents (91.1%) had suffered recurrent episodes and 39.6% had experienced over 10 episodes of depression. A third (34%) had been hospitalized for depression and 79% had received psychotherapy at some point.

Per the screening requirements, all respondents were suffering from a current episode of depression. Just over half (51.7%) had been experiencing their current episode of depression for a year or less. Just over a third (38.1%) were taking combination ADT (at least two classes), and 16.4% were taking an adjunctive antipsychotic (AP).

Despite ongoing AD treatment, nearly a quarter of patients (23.7%) indicated their depression had gotten worse since the onset of their current episode, while another 20.4% indicated no change, 48.7% indicated minimal improvement, and only 7.2% indicated moderate-great improvement (PGI-C). The SDS mean score was 6.9, with mean domain scores of 7.5, 6.9 and 6.5 in work/school, social life, and family life/home responsibilities, respectively. Respondents’ rated their overall level of impairment caused by depression as a mean of 2.8, with 60.6% indicating ratings of very or extremely difficult (PHQ item 10). Analysis of individual PHQ-9 items revealed that more than two-thirds of patients reported feeling down, depressed or hopeless, or having little interest or pleasure in doing things more than half the days (33.3 and 30.6%, respectively) or nearly every day (36.5 and 38.0%) (Additional file 1: Appendix). The most frequently reported ongoing symptoms other than those measured by PHQ-9, was lack of motivation (77.2%) followed by anxiousness, irritability and excessive worries, (67.1%, 58.1 and 57.4%).

Respondents were asked to describe how they felt about their prescription therapy and to select as many feelings that were relevant to their situation. Frustration was the emotion selected most often (29.8%) by respondents. Other emotions most often cited included, hopeless (27.4%), apprehensive/ nervous/anxious/scared (27.4%), resigned (26.9%), and dissatisfied (25.4%) (Fig. 2a). Of the respondents that indicated they were frustrated with their treatment, 85.9% rated their frustration as being moderate-to-severe (Fig. 2b). In addition, these respondents reported high levels of concomitant hopelessness (44.8% of those frustrated), anxiousness (48.2%) and dissatisfaction (44.8%) (Fig. 2c).

Patients with an inadequate response to their ADT were also asked how they felt about their HCP by selecting as many feelings that were relevant to their situation. The two most frequently chosen emotions were ‘feeling understood’ (31.6%) and ‘trusting/confident’ (28.8%). However, 19.2% of respondents reported frustration (Fig. 3a). Of those indicating they were frustrated with their HCP, 78% rated their frustration as being moderate-to-severe (Fig. 3b). Finally, of the 38.0% of patients who reported being frustrated with medication and/or their HCP, nearly a third reported frustration with both medication and their HCP (10.9% of all respondents).

Respondents who indicated they were moderately-to-severely frustrated with either their treatment and/or their HCP were asked why they were frustrated via an open-end question format. Responses were then coded for select themes and are shown in Fig. 4. The top reason for their frustration with medication was lack of efficacy (59.3%) followed by side-effect issues (19.7%). Other reasons included having to take medication daily, failing to meet their expectations, tapering effect, their relationship with their physician and onset of action (Fig. 4). Frustration with their HCP was driven by respondents feeling dismissed/HCP is not listening to them (22.4%), current treatment not being effective (13.5%) and lack of quality time spent with patient/too rushed (12.5%) among other reasons shown in Fig. 4b.

Random forest analyses indicated that for both frustration with medication and frustration with HCP the three largest drivers of frustration, of those entered in the model, were length of current episode, time since first episode and impact on function (SDS). Further, these three were twice as important as PGI-C and number of treatment failures (Fig. 5). The validity results however yielded a very small effect (Nagelkerke R-square = 0.074 for medication and 0.087 for HCP).

Patients moderately-to-severely frustrated with their treatment for depression were asked to identify how their frustration impacts their behaviors and decisions around their depression. The most frequently cited outcomes of frustration include patients considering alternative medication options (35.5%), other therapeutic approaches (33.8%) or simply quitting their medication all together (33.3%) (Fig. 6a). Another 14.7% admit that their frustration has resulted in them not taking their medication on a regular basis (Fig. 6a).

Those frustrated with their HCP suggested the frustration made them want to find a new HCP (40.5%), question the ability of their HCP (34.7%), share less information with their HCP (28.9%) and quit their medication altogether (27.3%) (Fig. 6b).