Use of Primary and Specialized Mental Health Care for a Major Depressive Episode in Spain by ESEMeD Respondents

A stratified, multistage, clustered-area, probability sample without replacement design was used to guarantee the representativeness of all the regions in the country. The target population was noninstitutionalized adults (aged 18 years and older), who were identified from census tract data. The final sample included 5,473 respondents (including a random selection of 323 spouses of the married main respondents, who were also interviewed); the final response rate was 78.6%. Data collection was conducted from September 2001 to September 2002. Individuals provided informed consent to participate in the interview. Data for the individuals were weighted to account for the different probabilities of selection within households and among hard-to-reach individuals, as well as to reflect age, gender, and Autonomous Community distribution of the Spanish general population.

The survey instrument was the World Mental Health (WMH) Survey version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI 3.0) (13), which is a fully structured diagnostic interview to assess disorders and treatment, as well as several additional variables, such as disease severity and disability. By means of computerized algorithms, the CIDI provides lifetime and 12-month diagnoses of mental disorders based on DSM-IV criteria (14). In establishing the diagnosis, CIDI rules about excluding organic mental disorders (direct physiological effects of a general medical condition, a drug, or a medication) were imposed. The CIDI was administered by trained lay interviewers using a computer-assisted personal interview.

Respondents were assessed for major depressive episode in the past 12 months. On the basis on researchers' consensus, individuals with the disorder were classified into three severity groups. The group with the most severe symptoms included those who had attempted suicide in the past 12 months, those who had a diagnosis of alcohol dependence with physiological dependence syndrome, and those who experienced severe role impairment (scores from 8 to 10) in at least two areas of the WMH-adapted version of the Sheehan Disability Scales (SDS). Individuals not classified as having severe symptoms were assigned to the moderate group if they had moderate role impairment in at least one domain of the SDS (a score of more than 4) or if they had alcohol dependence disorder without physiological dependence. All other individuals were classified as having mild symptoms.

All respondents were asked about use of services for their “emotions or mental health problems” in the 12 months before the interview. Those reporting use of services were asked to select from a list all providers whom they visited, including psychiatrist; psychologist; general practitioner, other physician, or other health care professional; human services (outpatient treatment with a religious or spiritual advisor or with a social worker or counselor in any setting other than a specialty mental health setting); and a complementary-alternative medical sector (outpatient treatment with any other type of healer, participation in an Internet support group, or participation in a self-help group). Psychiatrists and psychologists constituted the specialized mental health category; general practitioners, other physicians, and other health professionals constituted the general medical care category; and both specialized mental health and general medical care constituted the health care sector.

The type of treatment received by individuals who reported use of a service in the health care sector was assessed. Pharmacological treatment was defined as that prescribed by a psychiatrist, general practitioner, or any other physician or health professional in the previous 12 months. Psychotherapy was defined as having more than one session in the past 12 months of psychological counseling or therapy with a mental health professional (psychiatrist or psychologist) that lasted 15 minutes or longer. Individuals who received neither drug nor psychological treatment were classified as being in the group with no active treatment.

The definition of minimally adequate treatment followed previous research that was based on recommendations from clinical guidelines (15–17). It was defined as receipt of antidepressant pharmacotherapy for at least two months plus at least four visits with a psychiatrist, a general practitioner, or any other physician or health professional or at least eight sessions with a psychologist or a psychiatrist lasting an average of 15 minutes or longer. Human services and complementary-alternative medicine treatments were not considered to constitute adequate care because of the lack of experimental data documenting the effectiveness of such treatments for major depression (18–22).

The variable for chronic conditions included eight groups: musculoskeletal (arthritis), chronic pain (neck or back pain or other somatoform pain), digestive (stomach or intestinal ulcer), respiratory (asthma, allergies, or any other chronic lung disease), neurological problems, cancer, cardiovascular (stroke, heart attack, heart disease, or high blood pressure), and diabetes. The variable for mental disorders included mood disorders (major depressive episode or dysthymia), anxiety disorders (panic disorder, specific phobia or social phobia, generalized anxiety disorder, posttraumatic stress disorder, and agoraphobia without panic), and alcohol abuse or dependence.

All analyses were performed with SASTM software, version 9.1, of the SAS System for Windows as well as SUDAAN, version 9.0, a statistical package used to estimate standard errors of data obtained from surveys with complex designs. Data analyses were carried out at the IMIM-Hospital del Mar in Barcelona, Spain, and at Harvard University in Boston.

Our results should be interpreted within the context of some limitations. First, information about treatment was self-reported, which could have led to recall bias. Previous studies have found that self-reports of service use for some mental disorders may be higher than use documented in administrative records (23,24). The WMH Survey tried to reduce recall bias by excluding respondents who failed to say that they would think carefully and answer honestly. Furthermore, use of a 12-month time frame can further reduce the risk of recall bias. Second, some analyses consisted of small numbers of respondents, and this might have affected the reliability of results. Third, ESEMeD excluded several population groups, such as those who were institutionalized or homeless. Although service use patterns of these groups may be different, their relative size is small, and it could be argued that overall population estimates would not diverge considerably if they were included. Fourth, there is no consensus on criteria for defining minimally adequate treatment; however, to our knowledge, the relationship between various criteria and significant clinical outcomes has not been studied. Fifth, the Spanish National Health System comprises 17 regional systems whose management relies on each regional government. As a consequence, the level of development of the mental health network, the allocation of economic resources, and the population density of professionals vary across regions; these differences may influence the use of services. Because of limitations in sample size, it was not possible to perform a more detailed analysis that would offer a clearer picture of probable regional differences.

With these limitations in mind, we note that this is the first study that provides information about the use of services for a major depressive episode in Spain that is based on a nationally representative sample and that used a standardized methodology that allows for international comparisons. Previous research was restricted to specific regions and population subgroups (25,26) or clinical samples (27) or used mixed methodologies, which provided an incomplete picture of service use for depression.

Our findings show that almost half of respondents with a major depressive episode did not use services in the health care sector for their emotional problems in the year before the interview. Although concrete and easily comparable data are lacking, some studies have found lower use of health services for mental disorders in certain southern European countries, such as Spain and Italy, compared with other European countries, but the rates are still higher than those found in many low-income American samples or in Asian countries (6,28). Comparable data on service use by individuals with major depression are even scarcer, but studies have shown that use is higher in Spain than in countries such as Italy and Germany (29) and lower than in some northern European countries, such as Finland, Ireland, and the United Kingdom (30).

Although underrecognition and undertreatment of common mental disorders are frequent in most countries, substantive cross-national differences in treatment rates have been described. Treatment rates ranging from 50% to 79% in high-income countries have been reported, with rates typically lower in low-income regions (6). Reasons for these differences are not well known, but several factors have been proposed. They include factors that predispose patients to seek treatment, factors related to the type and amount of services available from the health care system, and factors related to health care professionals' ability to provide an adequate response to the need for treatment (6,31–35). The effect of social networks and support, which might be stronger in Spain than in other European countries (36), has also been discussed in the literature as either reducing service use by helping an individual to cope with stress or increasing service use by supporting treatment seeking (37). Although an extensive review of these factors is beyond the scope of this article, the relevant findings are summarized below.

Our results show that the proportion of respondents who sought service in the general medical care sector was only slightly higher than the proportion who visited mental health specialists. This finding is in contrast to other European countries and regions of the world, where general practitioners are clearly the professionals most frequently involved in the treatment of depression (6,44). This difference might be partly explained by the higher rates of referral to mental health specialists by general practitioners that have been described in Spain (44). These higher referral rates could be considered an attempt to improve the efficiency of a short-resourced health system. Additional factors, such as the relative delay in mental health care reform in Spain, may help explain why individuals with mild disorders were more likely to receive specialized mental health care.

The likelihood of use of services for depression has been linked to a variety of sociodemographic factors, such as middle age, female gender, higher education and income level, and not being married (6,16,53,54). However, our results support only a few of them. Of special concern is our finding that women were no more likely than men to seek treatment. This finding should receive attention because the higher prevalence of depression among women compared with men in Spain was found to be much higher than in many other countries (5,55,56).

Symptom severity is another major determinant of service use (30,57,58), and monotonic relationships between severity and the probability of service use have been previously described (6). Our results show that among individuals using any health sector services, those with a moderate disorder had higher odds of receiving treatment than individuals with mild disorders. Further research should be undertaken to confirm these findings—in particular, the role of severity and of gender in help seeking. Also, the finding that 31.7% of those with a severe disorder received care only from a general practitioner raises some concern about possible misallocations of scarce specialized resources, especially when the misallocations affect patients with higher risks of serious complications.

In our study 31.2% of the patients with a major depressive episode who received active treatment from general medical or mental health specialty care received minimally adequate treatment, a rate considerably lower than in other European countries (17). Consistent with previous studies, our study found low overall rates of treatment adequacy (6,59) and more adequate treatment in specialized settings (16,60). Reasons for the low rate of treatment adequacy are unclear, but they presumably involve provider factors, such as inadequate training and failure to adhere to guidelines (16). Inadequate treatment has significant health and economic consequences, because treatments that meet clinical guidelines are more cost-effective and have been shown to decrease the number of years lived with a disability (61).