Differences in Lifetime Use of Services for Mental Health Problems in Six European Countries

The questionnaire was based on the CIDI (version 3.0), which was developed and adapted by the Coordinating Committee of the World Health Organization's World Mental Health 2000 Initiative ( 24 ). This instrument, which has proved feasible, reliable, and valid for use in different cultures and settings ( 24, 25, 26 ), provides, by means of computerized algorithms, lifetime and current diagnoses of mental disorders according to DSM-IV and ICD-10 criteria. The following disorders were assessed: mood disorders (major depression and dysthymia), anxiety disorders (social phobia, specific phobia, generalized anxiety disorder, agoraphobia with or without panic disorder, and posttraumatic stress disorder), and alcohol use disorder (dependence and abuse). Other disorders included on the CIDI questionnaire, such as obsessive-compulsive disorder, eating disorders, drug use, childhood disorders, and psychotic disorders, were not considered in this study because they were assessed in only a small subsample of respondents.

All participants were asked to delineate lifetime use of any service as a result of their "emotions or mental health problems." Individuals reporting use of services were then asked to select the type of provider they had seen from a list of formal health care providers (including psychiatrists; psychologists, also called "nonmedical mental health providers" in the text, which included psychotherapists, social workers, and counselors; nurses; general practitioners; and other medical doctors) and of informal providers (such as religious advisors and other healers). In this study, any services, formal and informal, were considered.

Data analyses with Stata 8.1 software were weighted for sample design and nonresponse in order to represent the adult population of each country ( 22 ) for lifetime diagnoses of mental disorders and lifetime use of care. The chi square test was used to compare ratios. Multiple logistic regression analyses were used to assess the influence of demographic variables, clinical variables, and country on the use of services.

Participants were typically between 25 and 64 years old (67.9%), married (66.8%), and living in towns that had more than 10,000 inhabitants (66.8%). Approximately half were women (51.8%). A minority (34.6%) had a low education level.

All countries had a consistent pattern of consultation ( Table 2 ), with the highest lifetime rates observed among respondents with mood disorders and the lowest among those without any mental disorder. The presence of comorbid disorders was associated with increased consultation rates. In contrast, consultation rates among those with alcohol-related disorders were relatively low compared with those found among respondents with other disorders. When all countries were compared, significant intercountry differences were observed for all disorders. The most striking differences observed were between Italy, with the lowest consultation rates for almost all disorders, and the Netherlands, with the highest. Participants with a mood disorder in the Netherlands were twice as likely as their Italian counterparts to have sought professional help (71.0% versus 37.0%; p<.001). A similar difference between these two countries was found among those with an anxiety disorder (63.6% versus 33.1%; p=.001) and alcohol use disorder (45.9% versus 16.5%; p= .020). Spain had a rather low consultation rate among respondents without any mental disorder but relatively high rates among those fulfilling diagnostic criteria for mental disorders (with the exception of alcohol use disorder).

Multiple logistic regression analyses ( Table 3 ) showed that, in the overall sample, women were more likely to consult than men. Similarly, people who had been divorced, those with a higher education level, and those living in an urban area were more likely to seek services. Respondents in the youngest age group (18-24 years) and in the oldest one (65 years and older) were about 50% less likely to have sought professional help than those 25-64 years old. The probability of seeking help was increased by the presence of a lifetime mental disorder, even more so when it presented with a co-occurring mental disorder. Multiple logistic regression analyses also confirmed a lower level of consultation in Italy and Spain than in Belgium, France, Germany, or the Netherlands.

Observed consultation rates were generally consistent with the availability of professionals, especially of those who specialized in mental health. The lowest consultation rates were observed in the countries with the lowest availability. However, the reverse seemed not to be necessarily true: the Netherlands had a high consultation rate consistent with a high density, especially of psychologists; on the other hand, Belgium, France and Germany, which have a relatively high number of health providers, showed an intermediate consultation rate.

When the type of professional consulted was examined among all participants who had already used services for mental health problems ( Figure 1 ), we found that general practitioners were the main providers in all countries. Lifetime consultation rates ranged from 54.6% in Italy to 72.7% in Belgium and France (p<.001). The second most consulted professional was, in general, the psychologist, with the highest consultation rate in the Netherlands (64.8%) and the lowest in France (23.3%; p<.001), but a wide variation was observed. The consultation rate for psychiatrists ranged from 25.5% in the Netherlands to 43.8% in Spain (p=.03). The use of other doctors was highest in Italy (29.1%; p<.001), as was the consultation of religious advisors (12.3%; p<.001).

Among users of services, observed consultation rates for general practitioners were generally consistent with the number of professionals available in the different countries—highest in Belgium and France and nearly the lowest in Spain. However, the Netherlands, which has the lowest number of general practitioners, showed a high contact rate, whereas Germany and Italy had a relatively lower consultation rate than would be expected from the availability of general practitioners in those countries. Psychiatrist use was remarkably similar in countries with very different levels of availability and was notably high in Spain, which has the fewest psychiatrists. Use of nonmedical mental health providers also differed from patterns of availability, with a very high level of use in the Netherlands, which has half as many professionals as Germany has.

By selecting all participants who had already consulted either a mental health professional or a general practitioner and using multiple logistic regression analyses ( Table 4 ), we found that older individuals (65 years and older) and those with a lower education level were about half as likely to have consulted a mental health professional for their problems with or without consulting a general practitioner first, instead of consulting only a general practitioner. On the other hand, single persons were more likely than married or divorced persons to consult a mental health professional. Concerning mental health status, the presence of lifetime mental disorders increased the probability of consulting mental health professionals, especially when a comorbid condition was present. Belgians and the French were less likely than persons from the other four countries to have consulted a mental health specialist. This finding reflects the high density of general practitioners in these two countries.

When all participants who had already consulted a general practitioner were asked if they had been referred by the practitioner to a mental health specialist, a very diverse pattern was found. The referral rate was the highest in Italy, followed by the Netherlands and Spain for almost all the categories and the lowest in France, but it was low also in Belgium and Germany, even though the latter two had relatively higher referral rates. For example, French clients were referred to mental health specialists by their general practitioner in only 22.2% of cases, as opposed to 55.1% in Italy, 52.8% in the Netherlands, and 40.4% in Spain (p<.001). The same trend was found for individuals with a mood disorder, an anxiety disorder, or both (p<.001 for all comparisons).

Observed referral rates were fairly consistent with the availability of general practitioners, being highest in the Netherlands and Spain—countries with a low density of professionals—and lowest in Belgium and France, where there are many general practitioners. However, this relationship did not hold in Germany and Italy, which have a similar density of general practitioners but show quite different patterns of referral.

Our results show some similarities with studies performed in the United Kingdom ( 11, 12 ), which reported that major determinants of treatment access for mental problems are symptom severity, reflected in this study by comorbidity of mental disorders, and age. Furthermore, the Netherlands Mental Health Survey and Incidence Study reported that patients with mood disorders were more likely to enlist professional help than those with alcohol- and drug-related disorders ( 10 ).

Nonetheless, comparisons of previous reports on mental health provision in individual European countries are limited by the different sampling techniques and the different diagnostic instruments used, with six different instruments being used in 11 recent studies ( 27 ). There have been only two reports that used comparable assessments of the mental health of adults in more than one European country: the European Outcome of Depression International Network ( 28 ) and the Depression Research in the European Society ( 29, 30 ) studies. However, these studies were limited in that the former corresponded to comparisons between local community or primary care registration samples that were not representative of the general population of the participating countries, whereas the latter focused entirely on depressive disorders with some methodological limitations in the sampling design and achieved fieldwork.

On the other hand, a study across Europe of the mental health status of the European population in the previous 12 months (Eurobarometer) ( 31 ) reported patterns of use similar to this study. For example, the likelihood of seeking any help for a mental health problem was found to be the highest in the Netherlands and the lowest in Italy and Spain.

The main hypothesis tested in the study was that countries with more resources for psychiatric or psychological treatment would show a higher service utilization rate. However, the patterns of health care use observed among countries did not, on the whole, always correspond to actual resource availability, although the lowest consultation rates generally corresponded to the lowest resources. For instance, even though Belgium has nearly three times the number of general practitioners per capita as the Netherlands, the lifetime consultation rates for these professionals were practically identical in the two countries. Similarly, the country with the lowest density of psychiatrists, Spain, had a higher use of psychiatrists (43.8%) than Belgium (36.2%) or France (34.7%), countries with a high density of these specialists.

These findings suggest that use of resources appears to be mediated by other factors, such as the nature of professionals, referral practice, and sociocultural factors, which have been studied in depth elsewhere ( 6, 7, 8, 9 ). For example, Italians are less likely than their other European counterparts to consult formal health care providers for their mental health problems. In contrast, Italians seek relatively more help from religious advisors. In addition, how resources are financed might be relevant to explaining patterns of health care use. The different levels of health care insurance and the requirements for out-of-pocket expenditures, such as copayments ( 3, 5 ), could explain intercountry differences in utilization of services to a certain extent. For example, in France costs of psychotherapy are reimbursed if practiced by a psychiatrist but not when provided by a psychotherapist who is not a physician; in the Netherlands, on the other hand, psychotherapists are integrated into the health system, and their services are free of charge.

This study has several limitations. First, visits to health providers were self-reported. Reporting on past events is likely to be particularly prone to recall bias. In addition, many users of services for mental health problems might not have considered they had a mental health problem and might not have been captured in this section of the survey interview, which could lead to an underestimate of use of services. The extent of this phenomenon could vary between countries in an unidentifiable way.

Second, results are related to lifetime prevalence of psychiatric disorders. This period is useful in that it maximizes exploitable information by increasing the number of participants considered, and it gives an idea of the overall access to care for people in need. On the other hand, the disadvantages of using lifetime measures are recall bias and the existence of some age-related bias resulting from the inevitable higher probability of consulting for older individuals.

Third, certain disorders included on the CIDI questionnaire, such as obsessive-compulsive disorder, eating disorders, drug use, childhood disorders, and psychotic disorders, were not considered in this study because they were assessed in only a small subsample of respondents.

Fourth, survey participation rates were low in some countries. Nevertheless, nonresponse could have been overestimated occasionally, for example in France, where nonrespondents included individuals who never answered the telephone ( 27 ). Nonrespondents may use health services differently from respondents. This difference could introduce a potential bias, especially in the case of France and Belgium, which had the lowest participation rates. However, this difference should not have an impact on the results of our intercountry comparison because higher participation would be expected to have a significant impact only on the trends observed in extreme scenarios where nonrespondents differ drastically from respondents.

Fifth, we did not consider the number of consultations in our analysis. It might be that differences between countries change when considering intensity of use.

Sixth, certain resources (for example, mental health expenditure and number of psychotherapists, social workers, or counselors) were not examined because of the absence of available data.

Finally, only availability of services was compared between countries, and other relevant issues, such as organization and financing, were not addressed.

ESEMeD is, however, the largest comparative study of mental disorders in the EU to date in terms of size of the sample and the populations represented, the range of mental disorders assessed, and the information collected ( 32 ). This study provides a first attempt at assessing intercountry differences in access to mental health care, which needs to be refined by taking into account other important determinants, such as quality of care and outcome.