Adherence to Guidelines and Treatment Compliance in the Chilean National Program for First-Episode Schizophrenia

The sampling frame was taken from SIGGES, a national register of all patients entering the GES program. We used data from July to December 2004. A two-stage sampling process was employed. First, 12 of 29 catchment areas were purposively selected by the Ministry of Health. These areas cover 58% of the population using the public system. Criteria were aimed at selecting a sample that reflected regional diversity and the distribution of sociodemographic characteristics across the country: urban population, 87%; percentage of males, 49.3%; population over age 15 year, 75%; and people living in poverty, 14%. Second, in each catchment area, the first 20 patients who were referred to the GES program as having first-episode schizophrenia were included in the study. When fewer than 20 patients were referred in the specified time, all of them were included.

The GES clinical guideline for first-episode schizophrenia, developed by the Ministry of Health (25), was reviewed, and a tool was designed to evaluate its implementation. The tool was designed to collect information on the characteristics of interventions delivered. Data were collected by specially trained physicians and nurses from patient register and case notes; when necessary the data were supplemented by information obtained from health personnel at the clinic where the patient was being treated. Treatment status of the patient was assessed 12 months after entry into the register. The tool was piloted in Iquique, Chile, and small adjustments were made. Fieldwork was carried out from January to May 2006.

The clinical guideline for first-episode schizophrenia was published by the Ministry of Health in 2005, but its pilot version was made available to health service providers in 2004. The guideline for first-episode schizophrenia recommends that the following diagnostic procedures be carried out during the first 30 days after referral to the psychiatric outpatient facility: minimum of two clinical interviews by a psychiatrist based on ICD-10 diagnostic criteria; assessment of the degree of disability using the multiaxial ICD-10 disability axis; assessment of the person's psychosocial problems; routine laboratory tests (blood count, glucose, urea, alkaline phosphatase, bilirubin, and electroencephalography) for all patients and other laboratory tests and computer tomography and electrocardiogram as needed; and psychometric diagnostic testing, consultation with other specialists, and consideration of other diagnoses when necessary.

To evaluate clinical progress at six months and annually thereafter, the guideline requires the use of the following symptom scales, in addition to clinical assessment. These include the Positive and Negative Syndrome Scale (PANSS), the Global Assessment of Functioning (GAF), and the Clinical Global Impression (CGI).

The guideline recommends initiating pharmacological treatment with risperidone and adding benzodiazepines for anxiety if necessary. The second-line option is to switch to another second-generation antipsychotic—quetiapine, olanzapine, or ziprasidone. The third-line option is switch to another of these three second-generation antipsychotics. The fourth-line option is to switch to a first-generation antipsychotic. If the patient cannot be stabilized with these treatments, clozapine is recommended. The following psychosocial interventions are recommended: individual and family psychoeducation and family and group psychosocial intervention.

There is no consensus on the definition of adherence to guidelines (5), although efforts have been made to develop a common set of indicators for compliance to pharmacotherapy guidelines (27). In this study, we considered adherence to first- and second-line medications, provision of psychosocial interventions, and use of diagnostic measures to be the most important features for evaluation.

Of the 248 patients eligible for inclusion in the sample, no follow-up data were available for 18: nine patients moved to another region of the country after program entry, one died, and eight could not provide information for other reasons. In addition, 21 clinical records could not be found, and for 11 patients the case notes were too sparse to extract data. Only charts with all the necessary information were included in the study. The final sample consisted of 198 patients who were referred to the GES program as experiencing a first episode of schizophrenia. Of these, 110 (56%) were confirmed as having a first episode of schizophrenia, and they constitute the final sample. By health service region, the proportion of patients confirmed as experiencing a first episode of schizophrenia varied from 15% to 100% (Table 1).

In the final sample of patients confirmed as having a first episode of schizophrenia, 70 (63%) were men and 40 (36%) were women. The proportion of men was higher in the group with confirmed first-episode schizophrenia compared with the group for whom the diagnosis was not confirmed (64% and 57%, respectively); however, the difference was not statistically significant. In addition, the mean age of patients in the group with a confirmed diagnosis was significantly less than the mean age of patients for whom the diagnosis was not confirmed (25.4±10.4 and 29.7±14.5 years; p=.048).

Most patients received basic laboratory tests. Imaging studies were rarely used; only 28 patients (25%) receiving computer tomography of the brain, and only one patient underwent magnetic resonance imaging (Table 2).

Symptom rating scales were rarely used: the PANSS was used for 29 patients (26%), the CGI was used for 24 patients (22%), and the GAF was used for 22 patients (20%). Multiaxial ICD-10 diagnostics were used for 82 patients (75%).

As shown in Table 3, 98% of patients received pharmacological treatment with an antipsychotic, and most of them (86%) received risperidone. Of the 110 patients, 68 (62%) continued with the first-line antipsychotic until the end of follow-up or until they discontinued treatment, 35 (32%) switched to another antipsychotic, and seven (6%) discontinued the first-line antipsychotic without a prescription for a new one.

As Table 4 shows, 71% of patients received a psychosocial intervention, most commonly psychoeducation, which was received by 55% of the patients. Of these, 32 (53%) received individual psychoeducation, nine (15%) received group psychoeducation, and 19 (32%) received a combination of both. The median number of sessions was six (interquartile range [IQR]=3–13).

Family psychoeducation was received by 49% of the sample. Of the 54 families to whom this intervention was given, 37 (69%) received it as an individual family, 13 (24%) received it in a group, and four (7%) received a combination of the two formats. The median number of sessions was four (IQR=2–8).

Individual psychotherapy was received by 36% of patients, usually conducted by a psychologist (N=35, 90%) and only rarely by a psychiatrist (N=4, 10%). The median number of therapy sessions was four (IQR=2–10).

Forty-seven percent of patients received a family intervention other than psychoeducation. These interventions included family psychotherapy, an intervention to reduce expressed emotion, and other interventions designed to reduce the burden on the family or caretakers. Other psychosocial interventions were rarely used.

Forty-three percent of patients were hospitalized, and their mean±SD length of stay was 42.4±37.3 days. For the 8% of patients who received day hospital treatment, the mean stay was significantly longer (98.6±86.3 days). Six percent of patients had an inpatient stay and also received day hospital treatment; in this group the mean length of hospital stay was 28.8±31.3 days, and the mean length of stay in the day hospital was 73.8±56.8 days.

Overall, every patient had at least one appointment with a psychiatrist, and the mean number of visits per patient was 7.9±12.9. Half of the patients (49%) also had at least one appointment with a psychologist, including appointments for psychotherapy, and the mean number of visits per patient was 3.7±3.0.

Table 5 shows treatment status at 12 months for 98 patients (12 had moved to another area). Sixty-seven percent of patients followed the treatment for the first 12 months without any periods of dropout (defined as missing appointments for two months or more). A total of 31 patients (32%) had at least one period of dropout during the first 12 months. At follow-up, 19% had dropped out of treatment, and 13% had dropped out but had returned to treatment.

Not only is there no consensus on the definition of adherence to guidelines, a gold standard for adequate adherence is also lacking (5). Many studies have found considerably lower rates of adherence than those in this study. In the Netherlands, for example, only 40% of patients with depression or anxiety disorders were treated according to clinical guideline (14). In a review of adherence to mental health clinical guidelines, only 27% of nonintervention studies found adequate adherence rates (5).

One reason for Chilean clinicians' remarkably good adherence to medication guidelines may be that this national program made second-generation antipsychotics widely available for the first time in public mental health services. Furthermore, psychiatric treatment in Chile is focused on pharmacologic therapies, and health professionals may have perceived this guideline recommendation as the most important. Clinicians did not record reasons for choosing another first-line treatment for some patients, but it is possible that risperidone had not yet arrived in all the clinics.

However, when moving to second-line antipsychotics, clozapine was used for 20% of patients even though the guidelines recommend it as the last alternative after a trial of four different antipsychotics. The rate of clozapine use may have been influenced by the fact that clozapine was the first second-generation antipsychotic introduced in Chile in the 1990s, and even though its use was restricted and required rigorous follow-up, clinicians were more accustomed to prescribing clozapine than other second-generation antipsychotics and were convinced of its effectiveness.

Other recommendations in the guidelines, including those on diagnostic procedures and psychosocial interventions, were not as rigorously followed. Lack of training in the use of symptom scales may partly explain their low utilization, but it is important to note that health professionals are not accustomed to using such scales or to evaluating outcomes in general and do not perceive the scales as important clinical tools.

About half of the sample received patient and family psychoeducation, but provision of other psychosocial interventions, particularly educational and occupational assessments, was low. It is possible that health professionals are not trained to conduct all the interventions recommended in the guideline.

Similar to our study, a study of U.S. patients with schizophrenia found that rates of adherence to recommendations for psychosocial treatment were much lower than for psychopharmacologic treatment, ranging from 0 to 43% (28).

In Chile, laboratory tests are available without restrictions, but their low use may have been influenced by clinicians' lack of experience. Imaging studies, however, may not have been available for all patients, nor were they recommended by the guidelines for everyone.

Finally, it should be noted that this evaluation was carried out in the pilot phase just after the guidelines were launched, and it is possible that improvements have occurred since then.

It is interesting that only 56% of patients referred as experiencing a first episode of schizophrenia were confirmed to have this diagnosis. Differences between catchment areas were notable, ranging from 15% to 100% confirmed. This finding is noteworthy for two reasons. First, it may be problematic for the GES program that some patients without this diagnosis are using resources intended to ensure quick diagnosis and treatment of patients with first-episode schizophrenia. Second, the finding reflects problems of access to specialized psychiatric treatment via other routes. Because priority is given to first-episode schizophrenia, patients who have other conditions face longer waiting times to access specialized psychiatric care. Furthermore, availability of second-generation antipsychotics is restricted to conditions covered by the GES program. It is understandable that both primary care clinicians and psychiatrists may prefer to register patients as having suspected first-episode schizophrenia to guarantee faster access to evaluation and to second-generation antipsychotic medication, even though the patient may have another diagnosis. Anecdotal evidence of such practices exists.

Only 19% of patients had completely dropped out of the program at 12 months. This is an important achievement, and the rate is low compared with those in other countries. For example, in Italy and Spain, dropout rates ranging from 33% to 82% have been noted among patients in first-time psychiatric treatment (29,30). In Singapore, before the establishment of an early-psychosis program, the dropout rate was higher than 50% but fell to 19% for patients in the program (31). Other studies have found rates ranging from 17% to 22% among patients with various mental disorders, including psychosis (32,33).

The Chilean program specifies measures to retrieve patients who drop out, such as home visits and telephone calls to patients who miss an appointment. Such measures were used for some patients in this study, which may explain why nearly half of the patients who dropped out returned to the program.

Purposive selection of study areas can be seen as a limitation. However, since the areas represent 58% of the users of public services and these users have sociodemographic characteristics similar to those in the national population, the areas are considered to be representative of Chilean public health services. In addition, charts were reviewed only once, and reliability of adherence ratings was not assessed. It was considered sufficient to provide thorough training to the researchers and to exclude charts with incomplete information.