Main results
Our results show that half of the immigrant patients who begin treatment with AD obtain medication for only two months, compared with a median of three months in the native group. The average risk of abandoning treatment during follow-up is 28% higher in immigrant patients than in the natives, adjusting for age and sex.
Of the patients who continued treatment beyond the first month, only 29.5% of immigrants presented good compliance, compared with 38.8% of natives (withdrawal of above 80% of prescribed medication over at least four months).
These rates fall well below the current recommendations of European and US scientific societies which indicate the initiation of ADT in moderate or severe episodes of depression, and prolongation of treatment during the acute phase (between four and eight weeks) with an additional period of four to nine months to prevent relapses and to recover normal psychosocial functions. Duration may be longer if the clinical history of the patient so requires. Overall, a minimum treatment duration of six months is recommended.
Comparison with other published studies
Early abandonment of ADT is a well-documented, persistent occurrence in spite of attempts to encourage adherence. The percentages of abandonment during the first month published in previous studies are similar to ours, with figures ranging between 33% and 42% [
1,
7‐
9]. As regards the duration of antidepressant treatment, a study performed with a similar methodology [
10] corroborates our data since it obtained a median treatment duration of three months, which the authors attributed to the high percentage of early abandonment. Other percentages of treatment abandonment published range between 50% during the first three months [
11], 62.4% at the end of the third month [
1] and 44% at six months [
12]. An earlier study in our region obtained a discontinuation rate of 56% during the first four months [
3].
Much less is known about the pattern of use of ADT by the immigrant population. Most of the data recorded are indirect and come from studies not specifically aimed at this population or studies that do not conduct comparisons with natives. Van Geffen et al. [
13] concluded that the risk of not starting treatment or of acquiring only one prescription is higher when the indication is not specific, and that the risk of early abandonment is five times higher in some immigrant groups than in natives. Van Dijk et al. [
14] found a relation between the risk of abandonment of ADT and immigrant status, the type of drug prescribed, and comorbidity. Hasen et al. [
7] associated early abandonment with foreign citizenship, low educational level, and low income, and Olfson et al. [
1] related abandonment in the first 30days with being Hispanic, low educational level, and low family income.
Our study provides evidence that immigrants of all origins present higher percentages of early discontinuation of ADT and lower median treatment durations than the native population in our health region, regardless of their geographic origin. These findings corroborate those of a previous study by our group that showed a lower exposure to ADT in immigrants than in the native population; this tendency was found in all the ethnic groups, but was particularly marked in the sub-Saharans [
15].
Although the study was not designed to analyse the causes of the differences between the two populations, the results oblige us to reflect on the factors that may contribute to this situation and to assess whether these differences indicate inequalities in health care.
Clearly, some of the factors associated with poor adherence will affect both immigrants and natives and therefore should be considered in the global strategy for preventing non-compliance in all patients. However, in the immigrant population, in which the economic, social and family situations tend to be less favourable, factors associated with socioeconomic status and environment are likely to predominate. The measures for preventing non-adherence, then, go beyond the jurisdiction of the health system and need to be addressed at the broader level of social and economic policy. Non-adherence due to the cost of medication, a common finding in studies in the US comparing white elderly with their black or Hispanic peers [
16], can be included in this group. In Spain the public health system offers pharmaceutical coverage for AD, under which patients pay only 10% of the total cost, a situation that minimizes (but does not eliminate entirely) this effect.
With respect to the factors associated with immigrant status, it might be hypothesized that the shorter duration of ADT and the lower rates of compliance were due in part to a diagnostic bias. Incorrect diagnosis might lead to ineffective prescription and to early drop-out due to the lack of improvement. We cannot test this hypothesis because we do not have information on the diagnoses made. Nonetheless, the difficulties involved in diagnosing psychiatric syndromes in populations from different cultures are well known, due both to the barriers in communication and to the possible unreliability of the standard diagnostic instruments and criteria such as the DSM-IV in these populations.
The beliefs, expectations and attitudes towards mental illness that form part of the cultural background of each ethnic group should also be taken into consideration. They may be very different from those of the host society and may well affect adherence to pharmacological and non-pharmacological treatments.
Most authors agree that patients attitude towards ADT is one of the strongest predictors of adherence. Aikens et al. found that subjects of advanced age, with more severe symptoms, greater acceptance of the biochemical cause of the illness and expectations of persistence of the symptoms were more likely to perceive the need for treatment; these authors also found that patients who had never received AD treatment and who were not certain of the cause of their symptoms were more likely to associate therapy with negative effects [
17]. Van Geffen et al. interviewed patients in the early stages of ADT; those who dropped out were less involved in decision-making, had less faith in their doctor, reported a lack of information and presented a negative attitude to ADT, and were not convinced of the necessity of taking it [
18]. Reviewing studies of the differences in treatment of depressive and anxiety disorders in ethnic minorities, Schraufnagel et al. detected several obstacles to appropriate treatment, such as the beliefs about treatment mentioned above, stigma, the inability of the doctor to motivate the patient and the biological differences that condition the treatment. Nonetheless, these authors conclude that if the treatment is adequate the results in minorities are comparable to those achieved in the native population [
19].
Finally, in relation to these results, it should be recalled that health professionals may lack the intercultural competence necessary to provide optimal treatment to their patients. Their unfamiliarity with their patients culture may present a serious obstacle to early diagnosis and to the development of the therapeutic alliance, both of which are indispensable for obtaining positive results.
All things considered, the cultural identities of immigrant patients and their previous experiences of mental illness are likely to affect their ability to understand, accept and believe in the benefits of treatment. This hurdle appears to be especially significant in certain groups such as the sub-Saharan Africans, the population with the lowest median treatment duration and the lowest overall exposure to ADT reported in the studies mentioned above, and a lower exposure to antipsychotics than the rest of the minorities [
20].
These findings have serious implications for the care of immigrant patients: implications for the use and provision of health services (i.e., the need to reach migrant groups considering the cultural specificity of each of them, especially in early stages of treatment of mental illness), implications for health policy (i.e., identifying the factors that limit access to mental health services) and implications for future research (i.e., analysing in depth the causes of the differences observed).
Limitations
The main limitations of the study are to do with the sample selection. The sample is not representative of the entire immigrant population in the region, only of the immigrants who appear in the municipal census. Furthermore, exposure to ADT is estimated on the basis of medication withdrawn from the pharmacies in the region, with the result that any medication obtained in another health region is not reflected; however, this situation would affect a minimal number of cases in both groups, as the active population recorded in the regional health register remained stable over the period of analysis. The amounts of ADT received over the counter were also considered to be minimal, as prescriptions for this kind of drug are obligatory.
We are aware that our use of the concept of immigrant is excessively general. Although we sought to distinguish between immigrant groups according to their geographical origin, this approach also ignores the existence of differences between countries, between ethnic groups inside the same country and between individuals in the same ethnic group. For this reason, on the basis of these general data it would be interesting to focus on those groups in which the greatest differences were found with respect to the general population.
The lack of diagnoses is a limitation inherent to many studies performed on the basis of claims databases, since this method does not identify the reason for the prescription or the reason for the ending of treatment (i.e., cure or abandonment). This means that the data should be interpreted with care. Similarly, the analysis of drugs dispensed does not indicate the use that patients make of this medication, but reflects the intention to comply with treatment; even so, most authors agree that dispensation is the best measurement of medication use in large populations in habitual clinical practice.
Another limitation is the fact that we were unable to determine primary non-compliance, that is, the number of patients who are prescribed ADT but do not acquire the first pack. This information should be added to the percentage of non-compliance detected, but at the time of the study we did not have the prescription data needed to carry out this estimation.
For the same reason, we were unable to check the dose of ADT prescribed. We assumed that the daily dose prescribed corresponded to the definitions of the WHO Collaborating Centre for Drug Statistics Methodology, and that the prescribing physicians followed the recommendations of this institution.