Objective: Varying definitions of treatment-resistant depression (TRD) across studies make it difficult to estimate the size of the problem and to identify patients at increased risk. The aim of this cohort study was to examine the incidence of TRD, disease-related risk factors, and changes over time using different definitions of TRD.
Methods: From 1996 through 2014, all patients with a first-time hospital contact for depression (ICD-10 codes F32 and F33) were identified in Danish National Patient Registries. A total of 211,689 patients were followed for shifts in antidepressant treatment in the Danish Patient and Prescription Registries. TRD was defined at the second shift in treatment during the first 12 months after diagnosis. The associations of year and type of hospital contact, depression subtype, and severity of TRD were analyzed using Cox proportional hazard regression.
Results: A total of 14.0% of patients experienced a second shift in antidepressant treatment during the first year after admission. When applying 3 other common TRD definitions, the proportion varied from 13% to 31%. Psychiatric inpatients and patients with recurrent or severe depression had the highest incidence of TRD. The incidence of TRD was also slightly higher in patients diagnosed after 2001. All associations were replicated when data were reanalyzed using the alternative definitions of TRD.
Conclusions: About 14% of patients with depression developed TRD during the first year after first hospital contact. The incidence was highest in patients with severe depression and was relatively stable over time. Various definitions of TRD provided different estimates of the frequency of TRD but were all associated with disease severity.
Incidence of, Risk Factors for, and Changes Over Time in Treatment-Resistant Depression in Denmark:
A Register-Based Cohort Study
Frederikke Hordam Gronemann, MSca; Martin B. Jorgensen, DMScb; Merete Nordentoft, DMScb,c; Per K. Andersen, DMScd; and Merete Osler, DMSca,e,*
ABSTRACT
Objective: Varying definitions of treatment-resistant depression (TRD) across studies make it difficult to estimate the size of the problem and to identify patients at increased risk. The aim of this cohort study was to examine the incidence of TRD, disease-related risk factors, and changes over time using different definitions of TRD.
Methods: From 1996 through 2014, all patients with a first-time hospital contact for depression (ICD-10 codes F32 and F33) were identified in Danish National Patient Registries. A total of 211,689 patients were followed for shifts in antidepressant treatment in the Danish Patient and Prescription Registries. TRD was defined at the second shift in treatment during the first 12 months after diagnosis. The associations of year and type of hospital contact, depression subtype, and severity of TRD were analyzed using Cox proportional hazard regression.
Results: A total of 14.0% of patients experienced a second shift in antidepressant treatment during the first year after admission. When applying 3 other common TRD definitions, the proportion varied from 13% to 31%. Psychiatric inpatients and patients with recurrent or severe depression had the highest incidence of TRD. The incidence of TRD was also slightly higher in patients diagnosed after 2001. All associations were replicated when data were reanalyzed using the alternative definitions of TRD.
Conclusions: About 14% of patients with depression developed TRD during the first year after first hospital contact. The incidence was highest in patients with severe depression and was relatively stable over time. Various definitions of TRD provided different estimates of the frequency of TRD but were all associated with disease severity.
J Clin Psychiatry 2018;79(4):17m11845
To cite: Gronemann FH, Jorgensen MB, Nordentoft M, et al. Incidence of, risk factors for, and changes over time in treatment-resistant depression in Denmark: a register-based cohort study. J Clin Psychiatry. 2018;79(4):17m11845.
aCenter for Clinical Research and Prevention, Bispebjerg and Frederiksberg Hospitals, Frederiksberg, Denmark
bMental Health Centre Copenhagen, Copenhagen, Denmark
cLundbeck Foundation Initiative for Integrative Psychiatric Research, iPsych, Copenhagen and Aarhus, Denmark
dDepartment of Public Health, Section of Biostatistics, University of Copenhagen, Copenhagen, Denmark
eDepartment of Public Health, Section of Epidemiology, University of Copenhagen, Copenhagen, Denmark
*Corresponding author: Merete Osler, DMSc, Center for Clinical Research and Prevention, Bispebjerg and Frederiksberg Hospitals, Nordre Fasanvej 57, 2000 Frederiksberg, Denmark ([email protected]).
Depression is one of the leading causes of years lived with disability worldwide,1 and resistance to treatment is a common problem in clinical practice.2-6 It has been estimated that about one-third of patients treated for depression do not respond satisfactorily to at least one adequate treatment.2,5,7 These patients seem to have poorer prognostic outcomes,3,7 and treatment resistance appears to pose a substantial economic burden to the health care system.8,9 Treatment-resistant depression (TRD) has been defined as the absence of remission after treatment for depression,3 and during the past decade a consensus has emerged that at least 2 failed antidepressant treatments can be considered to be TRD,10 but this definition is not univocal.2,3,11 Varying definitions of TRD across studies have contributed to different estimates of TRD, with proportions ranging from 12% to 29%.9,12 However, these estimates are mainly derived from smaller clinical studies, and a diversity in how patients have been sampled hampers the possibilities of determining risk factors for TRD as well as prognostic outcomes.13 Together, these factors make it difficult to estimate the size of the problem and to identify patients at risk of TRD, who might be candidates for alternative treatment strategies.2-5,13 It has also been questioned whether TRD exists or simply reflects that depression consists of several disease entities with different etiology and therefore shows varying responses to standard antidepressant treatment.4 However, clinical studies that have investigated the association of depression subtype or severity or number of previous episodes with TRD have been inconsistent.13 Large-scale epidemiologic clinical data are required to determine what the rates are and whether there are differences in relation to setting and disease severity and over time.
The overall aim of this study was to determine the incidence of treatment resistance in patients with single-episode or recurrent depression and explore whether it associates with other disease-related risk factors and changes over time (1996-2014). We also explored if the findings could be replicated when data were reanalyzed using other common definitions of TRD. We hypothesized that at least 25% of all patients would be defined as resistant to first-line treatment and that this estimate would be independent of type of hospital contact, depression subtype, and severity. All Danish citizens have free access to psychiatric treatment, and though there were some changes in the available treatment during the study period, we expected that the frequency of TRD would be stable over time. We also expected that the aforementioned associations would be replicated using other definitions of TRD.
METHODS
Study Population
All citizens in Denmark with a first-time inpatient or outpatient hospital contact due to depression between January 1, 1996, and December 31, 2014, and no prior diagnosis of other affective disorders were included. In total, 254,250 inpatients or outpatients were identified in the Danish Psychiatric Central Research Registry (DPCR)14 and the Danish National Patient Registry (DNPR)15 using the International Classification of Diseases, Tenth Revision (ICD-10), codes F32.0-F32.9 and F33.0-F33.9. Both registers contain information on date, type of contact, and diagnosis for all in-hospital admissions in Denmark since 1969 (DPCR) and 1977 (DNPR). In 1995, the DPCR became an integrated part of DNPR, and data on outpatient contacts in hospital clinics as well as emergency department contacts were added.14 A total of 31,292 patients with a comorbid diagnosis of manic episode (ICD-10: F30; n=2,991), bipolar disorder (ICD-10: F31; n=9,375), persistent mood disorder (ICD-10: F34; n=8445), other or unspecified mood (affective) disorders (ICD-10: F38 and F39; n=3,416), or a comorbid schizophrenia spectrum disorder (ICD-10: F20; n=7,065) were excluded from the study. Further, 478 individuals not living in Denmark at the time of first hospital contact, with missing information on birth date or age below 10 years, were excluded. This left 222,480 patients for the analyses. The study was approved by the Danish Data Protection Agency. All data were retrieved from the Danish Health Data Authority and located in Statistics Denmark. Access to data can be obtained after approval by the Danish Data Protection Agency and Statistics Denmark.
Varying definitions of treatment-resistant depression (TRD) across studies make it difficult to estimate the size of the problem and to identify patients at increased risk.
At least 14% of depressed patients were defined as treatment resistant during the year after first hospital contact, and the risk was highest in inpatients with severe or recurrent depression.
Data
Each patient’s hospital records were linked to information from the National Prescription Registry and the Danish Civil Registration System using the unique Danish person identification (CPR) number.
Outcome: Treatment-Resistant Depression
TRD was defined a priori based on common treatment guidelines16-18 and a discussion with the participating psychiatric specialists as more than 2 shifts in first-line treatment, which was among the most restrictive definitions. An alteration of treatment was a shift in either antidepressant medication (from one to another group of chemical substances [see Supplementary Table 1]) or electroconvulsive therapy (ECT). The individual patients’ refills of prescriptions for antidepressant medication were identified by the Anatomic Therapeutic Chemical (ATC) classification system codes (N06A) in the Danish National Prescription Registry. This register contains information on all prescribed, redeemed drugs sold at Danish pharmacies.19 Data on use of ECT were extracted from DNPR, which also holds information on specific treatments provided to patients during hospitalization (Danish Health Classification System [SKS] codes: BRXA1* and BRTB1*).15 Pharmacologic treatment could have been initiated in primary care before the patients were referred to hospital. To distinguish between prevalent and incident TRD, patients were followed in a pre-index period starting 12 months prior to the date of being diagnosed and followed until 12 months after. Prevalent TRD was defined as having obtained 2 shifts in treatment (medication or ECT) at time of study entry (date of first hospital contact with depression), while incident TRD was defined at the second shift in treatment during follow-up. We also explored the frequency and associations of TRD using 3 of the definitions most commonly reported in the existing literature. Here TRD was defined as (1) as at least 1 failed antidepressant treatment,3,13,20 (2) as at least 2 failed antidepressant treatments,8,21-24 or (3) by a register-based treatment pattern algorithm including treatment with a monoamine oxidase inhibitor (MAOI) or ECT or at least 2 failed other antidepressant treatments.9,12,25 Some patients might shift treatment after only a few weeks due to side effects. In a series of sensitivity analyses, we explored the frequency of shift in antidepressant treatment within 4 weeks of initiating treatment and whether exclusion of such shifts influenced our findings. The same was done for treatment gaps of more than 6 weeks based on volume (defined daily dose dispensed) and number of refills. Further, inpatients might receive their antidepressant medication at the hospital, and this is not recorded in the Danish National Prescription Registry. Consequently, we examined whether length of hospitalization was associated with an unexpected lower rate of TRD.
Covariables
The ICD-10 codes provided information on depression subtypes (single [F32] or recurrent [F33]) and reflect disease severity.26 Disease severity was categorized as mild (ICD-10: F32.0, F32.8, F32.6, F33.0, F33.4, F33.8, F33.9), moderate (ICD-10: F32.1, F33.1), or severe (ICD-10: F32.2-F32.3, F33.2-F33.3). If multiple diagnoses were registered, the most severe was chosen. If both an F32 and an F33 diagnosis were registered, F33 was chosen. From DNPR, we also obtained information on whether the patients were diagnosed with depression at a psychiatric or medical ward as an inpatient (full or daytime), outpatient, or emergency contact. Since many of those diagnosed with depression in an emergency department were admitted as inpatients afterward, we dichotomized this variable as inpatients (full-time, daytime, or emergency department contact) or outpatients seen in hospital clinics. Further, information on comorbid anxiety (ICD-10: F41-F43) was included. We extracted the year of diagnosis from the data of first admission for each patient in DNPR. This information was collapsed into 3 periods; 1996-2001, 2002-2007, and 2008-2014. Finally, information on patients’ date of birth, sex, vital status, and migration status were obtained from the CPR.
Statistical Analysis
Incidence rates (IRs) of TRD and 95% confidence intervals (CIs) were calculated using time-to-event analysis. Individuals were followed from the date of first registered depression diagnosis until the date of complying with the TRD definition, emigration, death, or end of follow-up (defined as 12 months after study entry), whichever came first. The 12 months of follow-up was chosen as it was expected that treatment resistance would manifest itself within this period and we would avoid treatments initiated due to a new depressive episode. Individuals with prevalent TRD were excluded, as they were not at risk of developing TRD. We examined the association of disease severity, type of depressive episode, and year of diagnosis with incidence of TRD using Cox proportional hazard (hazard ratio [HR] and 95% CI) regression analysis. The proportional hazard assumption was examined graphically, which indicated a slight violation of the assumption for the variable year of diagnosis. However, splitting the model with year of diagnosis as covariable on follow-up time showed only minor differences in the HR for TRD with early (<6 months) and later (6-12 months) onset. We also completed all analyses for each of the 3 TRD definitions.
RESULTS
Frequency of TRD
Of 222,480 patients included in the study, 62.8% (n=139,622) were women, and the mean age was 50.4 years (range, 10-110 years) at time of study entry. Overall, 135,200 patients had a prescription filled for antidepressant medication before they were diagnosed with depression at a hospital ward, and 10,791 (8%) of these were defined as having prevalent TRD, which corresponded to 4.9% of the study population. Patients with prevalent TRD were excluded from further analysis, which left 211,689 patients at risk for developing incident TRD. Of these, 31,415 patients (14.8%) initiated no treatment with antidepressants or ECT during the pre-index period or the 12 months of follow-up (Supplementary Table 2). During follow-up, 14.0% (n=29,691) of the patients developed TRD, and the IR was 163.6 per 1,000 person years (95% CI, 161.8-165.6).
Disease Related Risk Factors
Table 1 shows the distribution of disease characteristics. Patients with recurrent depression had higher incidence of TRD compared to individuals with a single depressive episode (HR=1.28; 95% CI, 1.25-1.31). Patients with a moderate or severe depression diagnosis also had higher incidence of TRD when compared to patients diagnosed with mild depression. Patients diagnosed with depression at a medical ward had a much lower incidence of TRD compared to patients diagnosed within a psychiatric ward (HR=0.64; 95% CI, 0.62-0.66). Further, outpatients had lower incidence of TRD than inpatients (daytime or full-time) or those diagnosed at an emergency department contact (HR=0.78; 95% CI, 0.76-0.80). Patients with comorbid anxiety showed lower risk of TRD (HR=0.88; 95% CI, 0.85-0.91) when other covariables were included in the model.
Click figure to enlarge
Year of Diagnosis
The number of patients with a hospital contact for depression increased steadily over time from 6,260 (2.8%) in 1996 to 14,218 (6.4%) in 2014. The incidence of TRD was higher during both the second (HR=1.22; 95% CI, 1.18-1.26) and the third period (HR=1.14; 95% CI, 1.11-1.18) (Table 1).
Other Definitions of TRD
When the analyses were repeated using the 3 other common TRD definitions, the corresponding estimates of both prevalent and incident TRD were similar, apart from the estimates for the least restrictive definition of 1 failed treatment, for which the prevalence and incidence of nonresponse were higher (Table 2). Further, the associations between depression diagnosis and TRD were similar (Table 3), whereas the associations between year of diagnosis and TRD differed slightly. Thus, when 1 failed AD treatment was used as the definition of TRD, there was no difference in incidence of TRD over time (Table 3).
Click figure to enlarge
Click figure to enlarge
Sensitivity Analysis
Overall, 100,126 patients altered their first antidepressant treatment during follow-up. Of these, 20.4% (n=20,413) did so within 4 weeks after start of treatment. Further, 40,482 patients changed their second treatment, and 19.9% (n=8,050) did so within 4 weeks after initiation. When antidepressant treatments lasting less than 4 weeks or with insufficient coverage were excluded, the numbers of both prevalent and incident cases of TRD were smaller; however, this exclusion did not influence the aforementioned patterns of association (Supplementary Table 3). Further, long-term inpatients did not have lower TRD rates than those with a hospital stay of only a few days.
DISCUSSION
In this nationwide register-based study of 211,689 patients with a first-time hospital contact for depression in Denmark, we found that 14.0% of patients fulfilled our definition of TRD (IR=163.6 per 1,000 person years). When we used the other common definitions of TRD, the incidence ranged from 13% to 31%. Further, psychiatric inpatients and patients with recurrent or severe depression had the highest incidence of TRD as compared to the opposed patient groups. The incidence of TRD was slightly lower in patients diagnosed before 2001 than among those diagnosed from 2002 through 2014. Most associations were replicated when data were reanalyzed using the 3 alternative definitions of TRD.
The incidence of TRD was lower than we had expected. Previous studies8,9,12,23,25,27 have estimated the frequency of TRD between 12% and 29% among patients with depression. The incidence in our study was within but in the lower end of this range of variation, possibly due to the unselected patient sample and a relatively strict TRD definition. Contrary to our study, both Gibson et al12 and Kubitz et al8 included only patients with both a depression diagnosis and at least one prescription of antidepressants in their analyses. These patients may represent a subpopulation of depressed patients who are already at greater risk of developing TRD than patients being treated exclusively with psychotherapy or other treatment regimens not reflected when filled antidepressant prescriptions are used as inclusion criteria. Rizvi et al23 estimated the prevalence of TRD in primary care in Canada at 21.7%. In Denmark, up to 85% of all antidepressants are prescribed by general practitioners, and it could therefore be expected, as shown by the TRD prevalence in the present study, that some patients treated in this sector already had developed TRD. Unfortunately, a Danish Register for General Practice (the Danish Health Service register) and the Danish National Prescription Registry do not contain information on diagnosis. Identifying patients with depression through filled prescriptions for antidepressant medication would lead to misclassification since approximately 50% of all antidepressants are prescribed for purposes other than depression.28 However, we assumed that patients not responding to first-line antidepressant treatments in the primary sector would, as recommended in the clinical guidelines,16,17 be referred to the secondary health care sector for further evaluation and hence be included in our study population. As such, it could be expected that including all patients treated for less severe depression would entail a further reduction of the incidence. Another difference between ours and other studies was the inclusion only of patients with a first-time diagnosis of depression independent of disease severity.
Few studies have investigated the association of depression subtype with TRD, and most of these were clinical studies reporting no association.13 However, a European multicenter study of 702 patients29 found that the risk of TRD was associated with melancholic but not psychotic features. Similar to our results, a subsequent analysis of data from this multicenter study30 showed that TRD was associated with disease severity. It has been reported that patients not responding to their first antidepressant treatment might already be at increased risk of developing TRD.31 However, we also wondered whether treatment resistance exists or is merely a reflection of more severely ill patients, who have not yet been offered the appropriate treatment for their specific disease pattern or who do not respond to the standardized first-line treatment regimens for depression. The substantial symptom variation among patients diagnosed with depression indicates difference in etiology and might explain the difficulties in obtaining treatment effects.32 It is also likely that patients with severe symptoms are much more motivated to try different treatments compared to their counterparts and, as such, the association of TRD and disease severity becomes a self-fulfilling prophecy when TRD is defined by shifts in treatment.33 Severe cases could also show psychotic symptoms or suicidality, which might lead to reluctance to treatment. Consequently, it was not surprising that patients diagnosed with depression at medical wards had lower hazard ratios of TRD, since these patients had depression as a secondary diagnosis. Neither was it unexpected that inpatients had higher hazard ratios of TRD than outpatients, since the latter can be regarded as having less severe cases. In contrast to previous studies,34 we found that comorbid anxiety was associated with a lower risk of TRD after adjustment for other measures of depression severity.
Time trends in TRD rates have not been investigated previously, but our study indicated that the incidence of TRD was slightly higher after 2001. During the study period, the use of antidepressants changed from tricyclic antidepressants to selective serotonin reuptake inhibitors or serotonin-norepinephrine reuptake inhibitors, which might be less efficacious.6,11 Further, the introduction of more antidepressants might have led to more shifts. However, some of the increase might reflect that ECT was insufficiently registered before 2003 in the DNPR.35
The primary strength of this study was the use of nationwide population-based registers in a country with free access to health care, which provided us with a large, relatively unselected group of patients. Our study was restricted to patients with a first-time diagnosis based on clinical assessments, which reduces the risk of misclassification of index diagnosis. Thus, although the diagnoses might rely on assessments made by a variety of different clinicians, the information on depression in the DNPR has been shown to have high validity.36 However, as we based our study on patients with a first-time hospital contact for depression, it might not include the presumed milder cases of depression diagnosed in primary care in Denmark. A further advantage was the CPR number, which allowed us to link individual patient data with other registries and obtain complete follow-up on information for purchase of medicine, migration, and death, which reduces potential selection biases. The data were flexible and made it possible to study different definitions of TRD. We used common treatment guidelines and switches in first-line antidepressant treatment as primary outcome measures. The registration of purchased medicine and ECT is mandatory and has a high degree of completeness, reducing the risk of underreporting.18,35 It is, however, a limitation that our TRD measures are based solely on information from registers on shifts in antidepressant treatments, because use of ratings scales would have allowed a more precise assessment of symptom severity and treatment response. Further, we did not have information on reasons for changes in antidepressant treatment or measures of adherence to treatment, and the pharmacologic definitions of TRD do not incorporate psychotherapy, which is an important consideration.37 Combination and augmentation approaches are typically used when switching antidepressant medications has failed. Consequently, we considered this as the next step for patients defined as treatment resistant in our study and did not include medications such as lithium or quetiapine. Additionally, other factors such as the length of the episode are clinically significant but not captured by our outcome measure.
The use of various definitions of TRD in the literature has been criticized since it makes it difficult to compare findings and pool estimates for meta-analysis to improve knowledge of risk factors and patient outcomes. We chose to apply a definition of TRD reflecting the clinical guidelines and practices in Denmark, adding yet another definition to the pack. Reassuringly, we found nearly the same patterns of associations when we used other definitions of TRD. However, more studies on potential genetic and sociodemographic factors, comorbidities, and treatments that might increase the risk of TRD are needed to produce a better tool to predict patients at risk. Further, studies on short- and long-term consequences of TRD with sufficient adjustment for disease-related risk factors could shed light on the question of whether TRD is just a proxy measure for disease severity.
In conclusion, about 14% of patients with a first-time hospital contact for depression developed TRD during the first year after admission. The risk of TRD was highest in patients with the most severe depression. The rates of TRD have been relatively stable over the last decade—a period with relatively small changes in antidepressant treatment. The various definitions of TRD provided different estimates of the frequency of TRD but were all associated with disease severity.
Submitted: August 5, 2017; accepted November 1, 2017.
Published online: May 29, 2018.
Potential conflicts of interest: None.
Funding/support: The study was funded by the Danish Council for Independent Research, grant-ID: DFF-6110-00195.
Role of the sponsor: The funding source had no involvement in conduct of the research or in the preparation of the present manuscript.
1. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1545-1602. PubMedCrossRef
2. Fagiolini A, Kupfer DJ. Is treatment-resistant depression a unique subtype of depression? Biol Psychiatry. 2003;53(8):640-648. PubMedCrossRef
3. Fekadu A, Wooderson SC, Markopoulo K, et al. What happens to patients with treatment-resistant depression? a systematic review of medium to long term outcome studies. J Affect Disord. 2009;116(1-2):4-11. PubMedCrossRef
4. Ionescu DF, Rosenbaum JF, Alpert JE. Pharmacological approaches to the challenge of treatment-resistant depression. Dialogues Clin Neurosci. 2015;17(2):111-126. PubMed
5. Trevino K, McClintock SM, McDonald Fischer N, et al. Defining treatment-resistant depression: a comprehensive review of the literature. Ann Clin Psychiatry. 2014;26(3):222-232. PubMed
6. Thase ME, Rush AJ. When at first you don’ t succeed: sequential strategies for antidepressant nonresponders. J Clin Psychiatry. 1997;58(suppl 13):23-29. PubMed
7. Gaynes BN, Warden D, Trivedi MH, et al. What did STAR*D teach us? results from a large-scale, practical, clinical trial for patients with depression. Psychiatr Serv. 2009;60(11):1439-1445. PubMedCrossRef
8. Kubitz N, Mehra M, Potluri RC, et al. Characterization of treatment resistant depression episodes in a cohort of patients from a US commercial claims database. PLoS One. 2013;8(10):e76882. PubMedCrossRef
9. Corey-Lisle PK, Birnbaum HG, Greenberg PE, et al. Identification of a claims data "signature" and economic consequences for treatment-resistant depression. J Clin Psychiatry. 2002;63(8):717-726. PubMedCrossRef
10. Berlim MT, Turecki G. What is the meaning of treatment resistant/refractory major depression (TRD)? a systematic review of current randomized trials. Eur Neuropsychopharmacol. 2007;17(11):696-707. PubMedCrossRef
11. Dold M, Kasper S. Evidence-based pharmacotherapy of treatment-resistant unipolar depression. Int J Psychiatry Clin Pract. 2017;21(1):13-23. PubMedCrossRef
12. Gibson TB, Jing Y, Smith Carls G, et al. Cost burden of treatment resistance in patients with depression. Am J Manag Care. 2010;16(5):370-377. PubMed
13. De Carlo V, Calati R, Serretti A. Socio-demographic and clinical predictors of non-response/non-remission in treatment resistant depressed patients: a systematic review. Psychiatry Res. 2016;240:421-430. PubMedCrossRef
14. Mors O, Perto GP, Mortensen PB. The Danish Psychiatric Central Research Register. Scand J Public Health. 2011;39(suppl):54-57. PubMedCrossRef
15. Schmidt M, Schmidt SA, Sandegaard JL, et al. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol. 2015;7:449-490. PubMedCrossRef
16. National Institute for Health and Care Excellence. Depression in adults: recognition and management. NICE website. 2009. https://www.nice.org.uk/guidance/cg90.
19. Kildemoes HW, Sorensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health. 2011;39(suppl):38-41. PubMedCrossRef
20. Fava M. Diagnosis and definition of treatment-resistant depression. Biol Psychiatry. 2003;53(8):649-659. PubMedCrossRef
21. Mandelli L, Serretti A, Souery D, et al. High occupational level is associated with poor response to treatment of depression. Eur Neuropsychopharmacol. 2016;26(8):1320-1326. PubMedCrossRef
22. Rane LJ, Fekadu A, Wooderson S, et al. Discrepancy between subjective and objective severity in treatment-resistant depression: prediction of treatment outcome. J Psychiatr Res. 2010;44(15):1082-1087. PubMedCrossRef
23. Rizvi SJ, Grima E, Tan M, et al. Treatment-resistant depression in primary care across Canada. Can J Psychiatry. 2014;59(7):349-357. PubMedCrossRef
24. Souery D, Amsterdam J, de Montigny C, et al. Treatment resistant depression: methodological overview and operational criteria. Eur Neuropsychopharmacol. 1999;9(1-2):83-91. PubMedCrossRef
25. Scherrer JF, Chrusciel T, Garfield LD, et al. Treatment-resistant and insufficiently treated depression and all-cause mortality following myocardial infarction. Br J Psychiatry. 2012;200(2):137-142. PubMedCrossRef
26. Kessing LV. Severity of depressive episodes according to ICD-10: prediction of risk of relapse and suicide. Br J Psychiatry. 2004;184(02):153-156. PubMedCrossRef
27. Scherrer JF, Salas J, Sullivan MD, et al. The influence of prescription opioid use duration and dose on development of treatment resistant depression. Prev Med. 2016;91:110-116. PubMedCrossRef
28. Wong J, Motulsky A, Eguale T, et al. Treatment indications for antidepressants prescribed in primary care in Quebec, Canada, 2006-2015. JAMA. 2016;315(20):2230-2232. PubMedCrossRef
29. Souery D, Oswald P, Massat I, et al. Clinical factors associated with treatment resistance in major depressive disorder: results from a European multicenter study. J Clin Psychiatry. 2007;68(7):1062-1070. PubMedCrossRef
30. Balestri M, Calati R, Souery D, et al. Socio-demographic and clinical predictors of treatment resistant depression: a prospective European multicenter study. J Affect Disord. 2016;189:224-232. PubMedCrossRef
31. Schosser A, Serretti A, Souery D, et al. European Group for the Study of Resistant Depression (GSRD)—where have we gone so far: review of clinical and genetic findings. Eur Neuropsychopharmacol. 2012;22(7):453-468. PubMedCrossRef
32. Fried EI, Ness RM. Depression is not a consistent syndrome: an investigation of unique symptom patterns in the STAT*D study. J Affect Disord. 2015;172(2):96-102. PubMedCrossRef
33. Marcus SC, Hassan M, Olfson M. Antidepressant switching among adherent patients treated for depression. Psychiatr Serv. 2009;60(5):617-623. PubMedCrossRef
34. Fava M, Rush AJ, Alpert JE, et al. Difference in treatment outcomes in outpatients with anxious versus nonanxious depression: a STAR*D report. Am J Psychiatry. 2008;165(3):342-351. PubMedCrossRef
35. Hundrup E, Osler M, Jorgensen MB. Time trends and variations in electroconvulsive treatment in Denmark 2008 to 2014: a nationwide register-based study. J ECT. 2017;33(4):243-248. PubMedCrossRef
36. Kessing LV. Validity of diagnoses and other clinical register data in patients with affective disorder. Eur Psychiatry. 1998;13(8):392-398. PubMedCrossRef
37. Cowen PJ. Backing to the future: pharmacological approaches to the management of resistant depression. Psychol Med. 2017;47(15):2569-2577. PubMedCrossRef