Retrospective analysis of patients with advanced or metastatic gastric cancer in Colombia.

Aims: To assess patient and disease characteristics, treatment patterns, and associated costs in patients with advanced or metastatic gastric cancer (A/MGC) in Colombia, in both the public and private hospitals. Materials and methods: A total of 145 patients who had received first-line chemotherapy treatment (platinum analog and/or a fluoropyrimidine) and were followed for at least 3 months after the last administration of a first-line cytotoxic agent were eligible for inclusion. Case-report forms were elaborated based on the patients' medical records from three Colombian hospitals. Estimates of treatment costs were calculated using unit costs from the participating hospitals. Results: Of the 145 patients, more than half (64.83%) were male, 79.56% were diagnosed with metastatic stage IV disease (mean age = 58.14 years). Prior to MGC diagnosis, 31.71% of the patients being operated on received a total gastrectomy; 66.9% of the patients received a doublet therapy, of which 5-fluorouracil (5-FU) in combination with cisplatin was the standard treatment (14%), followed by combination with leucovorin (12%). Only around 10% of the patients responded to first-line treatment. Out of 41.38% of the patients who received a second-line treatment, 71.67% were still administered a platinum analog and/or fluoropyrimidine. During the follow-up period, 52% of the patients progressed and 20% achieved stable disease. Best supportive care mostly consisted of outpatient visits after last line-therapy (72.41%), palliative radiotherapy (18.6%), and surgery (37.2%). Limitations and conclusions: Gastric cancer is one of the main causes of cancer-related death in Colombia, as most of the patients are diagnosed at an advanced stage, when prognosis is poor. Treatment patterns are highly heterogeneous. Second-line treatments were mostly initiated with paclitaxel, capecitabine, irinotecan, or cisplatin.

Remote assessment of disease and relapse in major depressive disorder (RADAR-MDD): a multi-centre prospective cohort study protocol.

BACKGROUND: There is a growing body of literature highlighting the role that wearable and mobile remote measurement technology (RMT) can play in measuring symptoms of major depressive disorder (MDD). Outcomes assessment typically relies on self-report, which can be biased by dysfunctional perceptions and current symptom severity. Predictors of depressive relapse include disrupted sleep, reduced sociability, physical activity, changes in mood, prosody and cognitive function, which are all amenable to measurement via RMT. This study aims to: 1) determine the usability, feasibility and acceptability of RMT; 2) improve and refine clinical outcome measurement using RMT to identify current clinical state; 3) determine whether RMT can provide information predictive of depressive relapse and other critical outcomes. METHODS: RADAR-MDD is a multi-site prospective cohort study, aiming to recruit 600 participants with a history of depressive disorder across three sites: London, Amsterdam and Barcelona. Participants will be asked to wear a wrist-worn activity tracker and download several apps onto their smartphones. These apps will be used to either collect data passively from existing smartphone sensors, or to deliver questionnaires, cognitive tasks, and speech assessments. The wearable device, smartphone sensors and questionnaires will collect data for up to 2-years about participants' sleep, physical activity, stress, mood, sociability, speech patterns, and cognitive function. The primary outcome of interest is MDD relapse, defined via the Inventory of Depressive Symptomatology- Self-Report questionnaire (IDS-SR) and the World Health Organisation's self-reported Composite International Diagnostic Interview (CIDI-SF). DISCUSSION: This study aims to provide insight into the early predictors of major depressive relapse, measured unobtrusively via RMT. If found to be acceptable to patients and other key stakeholders and able to provide clinically useful information predictive of future deterioration, RMT has potential to change the way in which depression and other long-term conditions are measured and managed.

The association of religiosity with suicidal ideation and suicide attempts in the United Kingdom.

OBJECTIVE: The goal was to analyze the association of religiosity with suicidal ideation and suicide attempts in a UK nationally representative sample. METHODS: This study used cross-sectional data from 7403 people who participated in the 2007 Adult Psychiatric Morbidity Survey (APMS). Religion was assessed with the question 'Do you have a specific religion?' with 'yes' and 'no' answer options. Lifetime and past 12-month suicidal ideation and suicide attempts were assessed. The association between religiosity and suicidality was studied in multivariable logistic regression models adjusted for sociodemographic, behavioural, and psychopathological factors. RESULTS: Compared to those without a religion, the prevalence of past 12-month suicidal ideation (3.2% vs. 5.4%), past 12-month suicide attempts (0.4% vs. 0.9%), lifetime suicidal ideation (11.2% vs. 16.4%), and lifetime suicide attempts (3.6% vs. 6.0%) was lower among those with a religion. In the fully adjusted model, having a religion was significantly associated with lower odds for all types of suicidality except past 12-month suicide attempts: suicidal ideation (past 12-month: OR = 0.71, 95% CI = 0.51-0.99; lifetime: OR = 0.83, 95% CI = 0.69-0.99) and suicide attempts (past 12-month: OR = 0.71, 95% CI = 0.35-1.45; lifetime: OR = 0.69, 95% CI = 0.53-0.90). CONCLUSION: There is a negative association between religiosity and suicidality in the UK. Future studies should focus on the underlying mechanisms.

Socio-economic variations in the mental health treatment gap for people with anxiety, mood, and substance use disorders: results from the WHO World Mental Health (WMH) surveys.

BACKGROUND: The treatment gap between the number of people with mental disorders and the number treated represents a major public health challenge. We examine this gap by socio-economic status (SES; indicated by family income and respondent education) and service sector in a cross-national analysis of community epidemiological survey data. METHODS: Data come from 16 753 respondents with 12-month DSM-IV disorders from community surveys in 25 countries in the WHO World Mental Health Survey Initiative. DSM-IV anxiety, mood, or substance disorders and treatment of these disorders were assessed with the WHO Composite International Diagnostic Interview (CIDI). RESULTS: Only 13.7% of 12-month DSM-IV/CIDI cases in lower-middle-income countries, 22.0% in upper-middle-income countries, and 36.8% in high-income countries received treatment. Highest-SES respondents were somewhat more likely to receive treatment, but this was true mostly for specialty mental health treatment, where the association was positive with education (highest treatment among respondents with the highest education and a weak association of education with treatment among other respondents) but non-monotonic with income (somewhat lower treatment rates among middle-income respondents and equivalent among those with high and low incomes). CONCLUSIONS: The modest, but nonetheless stronger, an association of education than income with treatment raises questions about a financial barriers interpretation of the inverse association of SES with treatment, although future within-country analyses that consider contextual factors might document other important specifications. While beyond the scope of this report, such an expanded analysis could have important implications for designing interventions aimed at increasing mental disorder treatment among socio-economically disadvantaged people.

Suicidality and its relationship with depression, alcohol disorders and childhood experiences of violence: results from the ESEMeD study.

BACKGROUND: Suicidality constitutes a major health concern in many countries. The aim of the present paper was to analyse 10 of its risk factors and their interdependence. METHODS: Data on suicidality, mental disorders and experience of childhood violence was collected from 8796 respondents in the European Study of the Epidemiology of Mental Disorders (ESEMeD). The CIDI was used to assess mental disorders. Individuals were randomly divided into two subgroups. In one, a Graphical Markov model to predict suicidality was constructed, in the second, predictors were cross-validated. RESULTS: Lifetime suicidality was predicted mainly by lifetime depression and early experiences of violence, with a pseudo R-square of 12.8%. In addition, alcohol disorders predicted suicidality, but played a minor role compared with the other risk factors in this sample. CONCLUSION: In addition to depression, early experience of violence constitutes an important risk factor of suicidality. LIMITATIONS: This is a cross-sectional and retrospective study assessing risk factors for suicidality, not for suicide itself.

Analysis of burden in caregivers of people with Alzheimer's disease using self-report and supervision hours.

OBJECTIVES: This study aimed to describe the baseline characteristics of informal carers of community-living Alzheimer's disease (AD) patients by AD severity group and to identify factors associated with two measures of caregiver burden. DESIGN AND SETTING: GERAS is a prospective observational study in France, Germany, and the UK, designed to assess costs and resource use associated with AD, for patients and their caregivers, stratified by disease severity. PARTICIPANTS: 1497 community-dwelling AD patients and their primary caregivers. MEASUREMENTS: Subjective caregiver burden assessed using the Zarit Burden Interview [ZBI] and time spent supervising patients (an objective measure of burden recorded using the Resource Utilization in Dementia instrument) during the month before the baseline visit were recorded. Separate multiple linear regression analyses using ZBI total score and caregiver supervision time as dependent variables were performed to identify patient and caregiver factors independently associated with caregiver burden. RESULTS: Increasing AD severity was associated with both subjective caregiver burden (ZBI total score) and overall caregiver time, which includes supervision time (both p<0.001, ANOVA). Better patient functioning (on instrumental activities of daily living) was independently associated with both a lower ZBI total score and less supervision time, whereas higher levels of caregiver distress due to patient behavior were associated with greater caregiver burden. Other factors independently associated with an increased ZBI total score included younger caregiver age, caregiver self-reported depression, caring for a male patient, and longer time since AD diagnosis. Caregivers living with the patient, being a male caregiver, patient living in a rural location, higher patient behavioral problem subdomain scores for apathy and psychosis, more patient emergency room visits, not receiving food delivery and receiving financial support for caregiving were all associated with greater caregiver supervision time. CONCLUSION: Our results show that subjective caregiver burden and caregiver time are influenced by different factors, reinforcing the need to consider both aspects of caregiving when trying to minimize the burden of AD. However, interventions that minimize caregiver distress and improve patient functioning may impact on both subjective and objective burden.

Use of general practitioners versus mental health professionals in six European countries: the decisive role of the organization of mental health-care systems.

OBJECTIVE: To investigate patterns of use of general practitioners (GPs) and/or mental health professionals (MHPs) for mental health reasons, across six European countries, and the extent to which these patterns of use are associated with differences in mental health-care delivering systems. METHODS: Data are based on the European Study of the Epidemiology of Mental Disorders (ESEMeD): a cross-sectional survey of a representative sample of 8,796 non-institutionalized adults from six European countries, conducted between 2001 and 2003 using computer-assisted interviews with the CIDI-3.0. Countries were classified into: MHP- for countries where access to medical professionals tends to predominate (Belgium, France, Italy), and MHP+ where access to non-medical MHPs predominates (Germany, Spain, The Netherlands). RESULTS: Among respondents consulting GPs and/or MHPs in the past year (n = 1,019), respondents from the MHP- group more often consulted GPs (68 vs. 55 % in MHP+ group), GPs and psychiatrists (23 vs. 14 %). People from the MHP+ group more often used MHPs only (45 vs. 32 %), GPs and non-medical MHPs (16 vs. 8 %). GPs from the MHP+ group were more inclined to refer patients to MHPs. Factors associated with use of GPs versus MHPs were: being over 49 years, not highly educated, lower income and suffering from mood or severe mental disorders. CONCLUSIONS: Differences in the use of GPs versus MHPs are markedly linked to individual as well as organizational factors. Interventions are needed, in countries fostering medical access, to reimburse sessions with non-medical MHPs and improve cooperation between professionals to obtain better practice in access to care.

Cross-national differences in the prevalence and correlates of burden among older family caregivers in the World Health Organization World Mental Health (WMH) Surveys.

BACKGROUND: Current trends in population aging affect both recipients and providers of informal family caregiving, as the pool of family caregivers is shrinking while demand is increasing. Epidemiological research has not yet examined the implications of these trends for burdens experienced by aging family caregivers. Method Cross-sectional community surveys in 20 countries asked 13 892 respondents aged 50+ years about the objective (time, financial) and subjective (distress, embarrassment) burdens they experience in providing care to first-degree relatives with 12 broadly defined serious physical and mental conditions. Differential burden was examined by country income category, kinship status and type of condition. RESULTS: Among the 26.9-42.5% respondents in high-, upper-middle-, and low-/lower-middle-income countries reporting serious relative health conditions, 35.7-42.5% reported burden. Of those, 25.2-29.0% spent time and 13.5-19.4% money, while 24.4-30.6% felt distress and 6.4-21.7% embarrassment. Mean caregiving hours per week in those giving any time were 16.6-23.6 (169.9-205.8 h/week per 100 people aged 50+ years). Burden in low-/lower-middle-income countries was 2- to 3-fold higher than in higher-income countries, with any financial burden averaging 14.3% of median family income in high-, 17.7% in upper-middle-, and 39.8% in low-/lower-middle-income countries. Higher burden was reported by women than men and for conditions of spouses and children than parents or siblings. CONCLUSIONS: Uncompensated family caregiving is an important societal asset that offsets rising formal healthcare costs. However, the substantial burdens experienced by aging caregivers across multiple family health conditions and geographic regions threaten the continued integrity of their caregiving capacity. Initiatives supporting older family caregivers are consequently needed, especially in low-/lower-middle-income countries.

Including information about co-morbidity in estimates of disease burden: results from the World Health Organization World Mental Health Surveys.

BACKGROUND: The methodology commonly used to estimate disease burden, featuring ratings of severity of individual conditions, has been criticized for ignoring co-morbidity. A methodology that addresses this problem is proposed and illustrated here with data from the World Health Organization World Mental Health Surveys. Although the analysis is based on self-reports about one's own conditions in a community survey, the logic applies equally well to analysis of hypothetical vignettes describing co-morbid condition profiles. METHOD: Face-to-face interviews in 13 countries (six developing, nine developed; n=31 067; response rate=69.6%) assessed 10 classes of chronic physical and nine of mental conditions. A visual analog scale (VAS) was used to assess overall perceived health. Multiple regression analysis with interactions for co-morbidity was used to estimate associations of conditions with VAS. Simulation was used to estimate condition-specific effects. RESULTS: The best-fitting model included condition main effects and interactions of types by numbers of conditions. Neurological conditions, insomnia and major depression were rated most severe. Adjustment for co-morbidity reduced condition-specific estimates with substantial between-condition variation (0.24-0.70 ratios of condition-specific estimates with and without adjustment for co-morbidity). The societal-level burden rankings were quite different from the individual-level rankings, with the highest societal-level rankings associated with conditions having high prevalence rather than high individual-level severity. CONCLUSIONS: Plausible estimates of disorder-specific effects on VAS can be obtained using methods that adjust for co-morbidity. These adjustments substantially influence condition-specific ratings.

Assessment of medication adherence in a cohort of patients with bipolar disorder.

INTRODUCTION: This study aimed to identify factors associated with medication adherence in bipolar disorder (BPD) patients. METHODS: EMBLEM is a 2-year, prospective, observational study on the outcomes of BPD patients initiating or changing treatment for a manic/mixed episode. Data were collected at baseline, during the first 12 weeks of treatment (acute phase) and up to 24 months of follow-up (maintenance phase). Adherence was assessed by investigators at every visit. Repeated measures logistic regression analyses identified variables associated with adherence. RESULTS: Of 1,831 patients included in the analysis, 76.6% were adherent and 23.4% were non-adherent with their BPD medication during the maintenance phase. Patients were more likely to be adherent if they had insight into their illness at week 12. Patients were less likely to be adherent if they had cannabis abuse/dependence during the acute phase, work impairment or higher CGI hallucinations/delusions at baseline DISCUSSION: Psychotic symptoms, poor insight, cannabis abuse/dependence and work impairment are negatively related to medication adherence during maintenance therapy of bipolar disorder. Patients with these characteristics may need a different therapeutic approach.

Worldwide-Schizophrenia Outpatient Health Outcomes (W-SOHO): baseline characteristics of pan-regional observational data from more than 17,000 patients.

OBJECTIVE: To describe the Worldwide-Schizophrenia Outpatient Health Outcomes (W-SOHO) patient population at study entry, focusing on illness burden and prescribing practices across regions. METHODS: The SOHO study was a 3-year, prospective, observational study designed to assess costs and outcomes associated with antipsychotic use in outpatients initiating or changing antipsychotic (with an emphasis on olanzapine compared with other antipsychotics). SOHO was conducted in 10 European countries and 27 other countries as Intercontinental SOHO (IC-SOHO). Data from all countries have been pooled to produce the W-SOHO dataset. MAIN OUTCOMES MEASURES: Clinical Global Impression-Schizophrenia (CGI-SCH) severity scores, psychotropic medication use, adverse events, social interaction, housing and employment status, self-perceived health state (EuroQoL EQ-5D scale and Visual Analogue Scale, EQ-VAS), and reasons for initiation/change of antipsychotic. RESULTS: The W-SOHO database comprises 17,384 patients from six regions; East Asia (n = 1223), Central and Eastern Europe (n = 2175), Northern Europe (n = 4291), Southern Europe (n = 5788), Latin America (n = 2566), North Africa and the Middle East (n = 1341). Overall, patients were 38 +/- 13 years old (mean +/- SD), moderately ill (mean CGI-SCH overall score of 4.4 +/- 1.0) with a median duration of illness of 7 years (interquartile range 1-16 years); 43% were female, 10% were receiving antipsychotic medication for the first time. Adverse events were prevalent across all regions; on average, 50% (range 41-59%) of patients taking antipsychotics exhibited extrapyramidal symptoms at baseline, and 62% (34-67%) of patients reported sexual dysfunction in the previous month. On average, only 19% (16-23%) of patients were in paid employment and as many as 69% were living in dependent housing. CONCLUSIONS: Despite inherent diversity in these patients and the health care systems supporting them, there are striking cross-regional similarities in baseline characteristics for most measures. Not all countries are represented; regional comparisons may not be valid outside of the countries studied.

Mental disorders and termination of education in high-income and low- and middle-income countries: epidemiological study.

BACKGROUND: Studies of the impact of mental disorders on educational attainment are rare in both high-income and low- and middle-income (LAMI) countries. AIMS: To examine the association between early-onset mental disorder and subsequent termination of education. METHOD: Sixteen countries taking part in the World Health Organization World Mental Health Survey Initiative were surveyed with the Composite International Diagnostic Interview (n=41 688). Survival models were used to estimate associations between DSM-IV mental disorders and subsequent non-attainment of educational milestones. RESULTS: In high-income countries, prior substance use disorders were associated with non-completion at all stages of education (OR 1.4-15.2). Anxiety disorders (OR=1.3), mood disorders (OR=1.4) and impulse control disorders (OR=2.2) were associated with early termination of secondary education. In LAMI countries, impulse control disorders (OR=1.3) and substance use disorders (OR=1.5) were associated with early termination of secondary education. CONCLUSIONS: Onset of mental disorder and subsequent non-completion of education are consistently associated in both high-income and LAMI countries.

Validation in Spanish population of the family objective and subjective burden interview (ECFOS-II) for relatives of patients with schizophrenia.

INTRODUCTION: The family burden experienced by caregivers of people with schizophrenia is one of the most relevant consequences of this disorder. This paper aims to show the validity and reliability of the Spanish version of the Family Burden Interview Schedule (FBIS) designed to assess objective and subjective family burden for people with schizophrenia. DESIGN: cross sectional study. PARTICIPANTS: 356 patients fulfilling DSM-IV criteria for schizophrenia from four Spanish geographic areas (Barcelona, Madrid, Pamplona and Granada) and 205 main caregivers of these patients were assessed. MATERIAL: caregivers were assessed with the Family Burden Interview Schedule (FBIS) Spanish version (ECFOS-II), which assesses family burden in eight different modules: activities of daily living, disrupted behaviors restraint, expenses, caregiver's routine, concern, help, repercussions on health, and assessment of general burden. Patients were also assessed with PANSS, DAS-sv and GAF. STATISTICAL ANALYSIS: in order to assess internal consistency, parametrical tests of Cronbach's alpha were undertaken. To compute test-retest reliability Cohen's kappa and Weighted kappa were used. A principal component analysis was undertaken for assessing construct validity. Convergent validity was assessed with Spearman and Pearson correlation coefficients respectively, relating the instrument with the psychopathological (PANSS) and disability scale (DAS-sv) and general functioning (GAF). Moreover, a description of the viability of the ECFOS-II was described by a questionnaire especially designed for this purpose. RESULTS: Cronbach's alpha coefficient was 0.85 for the global assessment. Test-retest coefficients were very high, both for Cohen's kappa and for Weighted kappa, most values being between 0.61 and 1. Principal component analysis detected four factors that coincide with the modules of the original schedule. In the convergent validity we found that these factors are related with the symptom, disability and global functioning characteristics of the patients. CONCLUSIONS: ECFOS-II results in a valid and reliable instrument for assessing family burden experienced by caregivers of people with schizophrenia.

[Mental health care use in the Spanish general populations. Results of the ESEMeD-Spain study]. / Utilización de los servicios de salud mental en la población general española. Resultados del estudio ESEMeD-España.

INTRODUCTION: Between 17% and 30% of the world population suffers from a mental disorder in a given year and only a small proportion of them established contact with health services. This information is not available concerning the Spanish population. This study aims to describe the health services utilization related with mental disorders in Spain, as well as to identify factors related with such utilization. METHODS: The ESEMeD-Spain is an epidemiological study carried out on a representative sample of the Spanish adult general population, with a sample size of 5,473 individuals. We used the Composite International Diagnostic Interview (version 3.0) to collect information concerning mental disorders symptoms, health services utilization related with mental health, and administered treatments, with a 12 months time frame. RESULTS: 57% of those suffering from a mood disorder, 30.3 % of those suffering from an anxiety disorder and 71.7% of those suffering from both contacted health services in the year they suffered the disorder. Psychiatrists were the most frequently consulted professionals when a 12 month mental disorder was present, with the exception of anxiety disorders or comorbid mood-anxiety disorders. A 30.8% of those with any 12 month mental disorder who did consult were not prescribed with any treatment. Variables significantly related to 12 month health services utilization were age, gender, marital status, and mental disorders. CONCLUSIONS: A large proportion of individuals suffering from mental disorders do not use health services or, if indeed use them, do not receive any type of treatment.

Assessment of service use patterns in out-patients with schizophrenia: a Spanish study.

OBJECTIVE: The objective is to describe and characterize patterns of service use by out-patients with schizophrenia in Spain. METHOD: A representative treated prevalence sample of cases with schizophrenia was selected from four Spanish health areas. The evaluation included health service use, clinical severity, functioning and disability. Statistical analysis was based on hierarchical clustering methods. RESULTS: A total of 356 patients were included in the analysis. Five patterns of health service use were defined: heavy out-patient mental health users; mental health and general health service users; heavy hospital service users; nursing service users; low users of mental health services. Patients in each group showed differences in clinical and disability status. Patterns of health service use showed consistency, but also variability, among the geographical areas. CONCLUSION: Development and organization of mental health services should take into account the combinations of services patients most frequently use.

Cost-effectiveness of different clinical interventions for reducing the burden of schizophrenia in Spain.

OBJECTIVE: To estimate the cost-effectiveness of interventions for reducing the burden of schizophrenia in Spain. METHOD: The study examined the cost-effectiveness of seven different types of clinical interventions at the level of Spanish population: i) current situation; ii) older antipsychotics alone; iii) new antipsychotics alone (risperidone); iv) older antipsychotics plus psychosocial treatment; v) new antipsychotics plus psychosocial treatment; vi) older antipsychotics plus case management and psychosocial treatment; vii) new antipsychotics plus case management and psychosocial treatment. RESULTS: Interventions based on the combination of haloperidol with psychosocial treatment or psychosocial treatment plus case management proved to be the most efficient strategies. CONCLUSION: The relatively modest additional cost of concurrent psychosocial treatment has significant health gains, thereby making such a combined strategy for schizophrenia more cost-effective than pharmacology alone.

Cost-utility of selective serotonin reuptake inhibitors for depression in primary care in Catalonia.

OBJECTIVE: To determine the cost-utility of selective serotonin reuptake inhibitors (SSRIs) for treating depressive disorders prescribed in primary care (PC). METHOD: A total of 301 participants beginning antidepressant treatment with an SSRI were enrolled in a prospective 6-month follow-up naturalistic study. Incremental cost-utility ratios (ICUR) were obtained for several comparisons among different SSRIs. To address uncertainty in the ICUR's sampling distribution, non-parametric bootstrapping was carried out. RESULTS: Taking into account adjusted total costs and incremental quality of life gained, fluoxetine dominated paroxetine and citalopram with 63.4% and 79.3% of the bootstrap replications in the dominance quadrant, respectively. Additionally, fluoxetine was cost-effective over sertraline with 83.4% of the bootstrap replications below the threshold of 33,936 US$/quality-adjusted life year (30,000 euro/QALY). CONCLUSION: Fluoxetine seems to be a better cost-utility SSRI option for treating depressive disorders in PC.

Effectiveness and cost-effectiveness of antidepressant treatment in primary health care: a six-month randomised study comparing fluoxetine to imipramine.

BACKGROUND: Over the past decade, studies of the effectiveness of pharmacological treatment for depression have often been based on research designs intended to measure efficacy, and for this reason the results are of limited generalizability. Research is needed comparing the clinical and economic outcomes of antidepressants in day-to-day clinical practice. METHODS: A six-month randomised prospective naturalistic study comparing fluoxetine to imipramine carried out in three primary care health centres. Outcome measures were the Montgomery Asberg Depression Rating Scale (MADRS), direct costs, indirect costs and total costs. Subjects were evaluated at the beginning of treatment and at one, three and six months thereafter. RESULTS: Of the 103 patients, 38.8% (n = 40) were diagnosed with major depressive disorder, 14.6% (n = 15) with dysthymic disorder, and 46.6% (n = 48) with depressive disorder not otherwise specified. Patients with major depressive disorder or dysthymic disorder achieved similar clinical improvement in both treatment groups (mean MADRS ratings decrease in major depressive disorder from baseline to 6 months of 18.3 for imipramine and 18.8 for fluoxetine). For patients with major depressive disorder and dysthymic disorder, the imipramine group had fewer treatment-associated costs (imipramine 469.66 Euro versus fluoxetine 1,585.93 Euro in major depressive disorder, p < 0.05; imipramine 175.39 Euro versus fluoxetine 2,929.36 Euro in dysthymic disorder, p < 0.05). The group with depressive disorder not otherwise specified did not experience statistically significant differences in clinical and costs outcomes between treatment groups. LIMITATIONS: Exclusion criteria, participating physicians may not represent GPs. CONCLUSIONS: In a primary care context, imipramine may represent a more cost-effective treatment option than fluoxetine for treating major depressive disorder or dysthymic disorder. There were no differences in cost-effectiveness in the treatment of depressive disorder not otherwise specified.