Claudiane Salles Daltio Direct Medical Costs Associated with Schizophrenia Relapses in Health Care Services in the City of São Paulo

Custo direto médico-hospitalar de recaída em esquizofrenia em serviços de saúde na cidade de São Paulo ABSTRACT OBJECTIVE: To assess direct medical costs associated with schizophrenia relapses in mental health services. METHODS: The study was conducted in three health facilities in the city of São Paulo: a public state hospital; a Brazilian National Health System (SUS)-contracted hospital; and a community mental health center. Medical records of 90 patients with schizophrenia who received care in 2006 were reviewed. Information on inpatient expenditures was collected and used for cost estimates.

The mean incidence of schizophrenia is one to two cases per 10,000 people per year and its prevalence is 0.7% in adults worldwide. 17In Brazil, the literature shows inconsistent data, but schizophrenia prevalence is likely higher than that reported, and close to that reported in a meta-analysis by McGrath et al. 18,23 Schizophrenia has a chronic course with an estimated improvement of 40% after fi ve to six years of follow-up. 7 assess schizophrenia burden, Knapp et al (2004)  reviewed 62 studies and found three major studies on costs associated to schizophrenia from the United States, Canada, and England. 10The US study estimated a total cost of US$ 62.7 billion in 2002, of which US$ 22.7 billion in direct costs; US$ 7.6 billion in non-medical direct costs and US$ 32.4 billion in indirect costs. 27he Canadian study estimated a cost of schizophrenia of C$ 6.5 billion in 2004, of which C$ 2.02 billion in direct medical and non-medical costs and C$ 4.83 billion in indirect costs. 5The English study estimated total schizophrenia-related costs in 2004 and 2005 were £ 6.7 billion, £ 2 billion in direct and £ 4.7 billion in indirect costs. 14In all three studies the main component of indirect cost was loss of productivity.According to the Global Burden of Disease Assessment, schizophrenia causes a high degree of disability and accounts for 1.1% of all disability-adjusted life years (DALYs), and for 2.8% of all years lived with disability (YLDs). 26SUMO OBJETIVO: Avaliar o custo direto médico-hospitalar da recaída em esquizofrenia, em serviços em saúde mental.

INTRODUCTION
In Brazil, Leitão et al 12 assessed direct medical-hospital burden of schizophrenia in public health services in the state of São Paulo, Southeastern Brazil, in 1998 and found 81.5% of patients with schizophrenia were covered by the National Health System (SUS).Of these, 3.7% were temporarily admitted to hospitals, 2.3% were permanently admitted, 23% were treated on an outpatient basis and 71% received no regular care.Total direct cost burden of schizophrenia was US$ 191 million, accounting for 2.2% of total health care expenditures.Of this amount, 9.8% went to patients who did not receive regular care, 11% to outpatient care, 30.5% to permanent admissions and 48.7% to temporary admissions. 12ere are signifi cant differences in resource allocation to mental health between developing and developed countries.chizophrenia relapses are directly related to failure to comply with treatment. 8,22In fact, this is one of the most expensive aspects of the disease.In 2005, re-hospitalization costs in the United States resulting from treatment noncompliance were US$ 1,479 million dollars. 24,21A similar study in England showed that noncompliance accounted for an almost threefold increase in service costs. 9However, there is no consensus on the defi nition of relapse in schizophrenia.Lader claims that the patient's condition before disease onset, his or her level of functionality before a given episode, episode severity, development of new symptoms, duration and extent to which symptoms affect functioning should all be considered in the event of a new episode. 11A Delphi Consensus involving experts from the United Kingdom and Ireland did not manage to come to a unanimous agreement regarding the definition of schizophrenia relapse such as recurrence of positive symptoms. 2,4e objective of the present study was to assess direct medical costs of schizophrenia relapses in mental health services.

METHODS
Three mental health services from the city of São Paulo, Southeastern Brazil, were selected based on convenience and data accessibility: a public state hospital (PH); a SUS-contracted hospital (CH); and a community mental health center (CMHC).The differences between these three services are type of care provided (full admission in hospitals and partial admission in CMHC) and funding sources.The PH and CMHC are funded by the state while the psychiatric CH has a feefor-service contract with the state.The highest costs are expected to be seen in the PH.
Relapsed patients were those who required full-time hospitalization in any of the study hospitals or who were undergoing psychiatric intensive care at the CMHC.According to the Brazilian Ministry of Health, psychiatric intensive care involves 13 to 25 visits/ month.Inclusion criteria included patients who were diagnosed with schizophrenia regardless of subtype and treated between January 1 and December 31, 2006.CMHC patients were selected from fi rst quarter (2006) records since care authorization for high cost/high complexity procedures describing the type of treatment regimen are fi lled out on a quarterly basis.Exclusion criteria included: different admission and release diagnoses (90 CH); inconsistencies between admission diagnosis recorded in the database and medical chart (8 PH, 20 CH, 9 CMHC); associated neurological disease (2 PH, 5 CH), hospital stay longer than 90 days (6 PH, 27 CH); and hospital stays starting in 2005 and extending into 2006 (7 PH, 17 CH).The 90-day limit for hospital stay was set to exclude people living in the hospital and those admitted for social reasons.For CMHC patients we also excluded patients who had been hospitalized or dropped out (n = 4), as well as those for whom there were no notes in their medical records (n = 2) during the study period.
We randomly drew medical records for review as follows: 70 medical records from PH (only relapsed patients), 226 from CH (only SUS patients, 85% of all), and 47 from CMHC.We reviewed 53 medical records from PH, 58 from CH and 45 from CMHC until we had a total of 30 medical records for each service that met all inclusion and exclusion criteria.
We collected the following patient data from medical records: socioeconomic condition; clinical status and treatment prior to admission to the service; resources used by each patient in terms of number of psychiatric, psychological and occupational therapy visits; social worker and nursing resources used; clinical consultations; group and family therapy; outside activities; home visits; multidisciplinary team meetings; tests performed and medication taken.These costs where then evaluated along with general and emergency costs related to food, cleaning, safety, laundry, maintenance and administration.
Resource values were obtained from each service.The costs of state-funded standard drugs were obtained from Price Registration and State Health Department website.For non-standard drugs we used average selling price; for medical tests we used the offi cial SUS table.Drug cost was taken from the price list submitted by the seller, except for atypical psychotropic drug olanzapine that was obtained through the government program for high-cost drugs.Human resource costs were obtained from CMHC.These costs include personnel assigned to other locations but included in the CMHC payroll.For currency conversion purposes US$ 1 is approximately R$ 2.00 (R$ 2.16 on June 30, 2006).

Direct medical costs (DMC) were calculated as follows:
DMC = MCM + MCT + MCDR where: Mean cost of medication (MCM) = Total cost of medication/30 patients; mean cost of tests (MCT) = Total cost of testing/30 patients; mean cost of daily rates (MCDR) = (Human resource costs per month/total patients/30 days + monthly general costs/total patients/30 days + monthly emergency costs/total patients/30 days) x mean length of stay in the service (days) Since the study was conducted from the point of view of society at large, and because the study period was no longer than one year, no discount rate was applied.

RESULTS
Table 1 shows sociodemographic and clinical status of patients.We found similar age profi les in all three services: mean age of patients 36.2 years old in PH, 43.5 in CMHC, and 44.4 in CH patients.Most were white and single with no differences among services studied.In three services most of patients were out of work -unemployed/not active or retired.The PH only had beds for male patients and the CH had beds for females only; the CMHC served both males and females and 60% of its patients were male.Patients in the three services studied had chronic schizophrenia with mean disorder duration between 15 and 20 years.

DISCUSSION
The highest individual mean direct medical costs of schizophrenia relapses were seen for patients treated in the PH, followed by CMHC and CH.The major cost component was daily rate, which was higher in the PH.Drug costs varied according to the type of antipsychotic drug used.Low-cost typical antipsychotics were most often used in the CH and atypical antipsychotics were more commonly used in the CMHC.Relapses are often associated with patient failure to comply with maintenance treatment, and this risk is higher among those with prior history of noncompliance.A recent review by Marcus et al 15 found that 87,000 patients with schizophrenia are admitted to hospitals every year in the United States at a cost of US$ 806 million; most of these admissions are due to failure to comply with outpatient care. 15garding the use of resources during schizophrenia relapses, we found that, although patients stayed longer in the CMHC (90 days, almost twice the stay in the other two services), the use of professional/activity resources was notably lower.However, this may be explained by poor record keeping.Other possible explanation could be that community center patients have less severe disease compared to other inpatients or that these  patients receiving psychiatric intensive treatment were not actually experiencing a crisis.
The component that most infl uenced direct medical costs was daily rate charged per patient, with signifi cant variations between services.One possible explanation for these differences would be the effi ciency of public versus private management.Another explanation would be the fact that SUS pays the CH for a package of services rather than per procedure, encouraging the CH to keep its costs down.CMHC human resources and general expenses were similar to those in the CH, though lengths of stay in the former were longer, probably because this is not an inpatient situation and therefore there is no pressure to release patients.A study showed that treatment in a community mental health center is less expensive than conventional hospital care and can be more benefi cial to the patient's quality of life. 25dication was the second largest component of direct medical costs in all three mental health services, although these costs were much lower than those of daily rates in all cases.The main differences are attributed to the use of typical or atypical antipsychotic drugs and they explain varying costs of care found in the three services.There is no consensus about the superiority of atypical over typical antipsychotic drugs. 13onsistent with the literature, 19 none of the three services studied performed a signifi cant number of complementary laboratory tests and their associated costs were relatively insignifi cant.
The present study has some limitations.We could not confi rm the diagnosis of schizophrenia as it requires a personal interview with the patient, which was not feasible in this type of study.Also, we adopted a practical criterion to defi ne schizophrenia relapse.
It is also possible that the CMHC sample included relapsed patients but also non-relapsed patients who were institutionalized and remained under psychiatric intensive care.The fact that services are for one sex only can introduce a sampling bias as there are gender differences in schizophrenia prognosis. 3Another limitations are incomplete records, which may underestimate resource use and costs, and potential bias in cost spreadsheets supplied by the services studied, which were not audited.In addition, it was diffi cult to obtain certain information from the State Health Department such as cleaning and food costs and in some cases we had to use indirect means to assess costs or approximate costs.
The costs of treating schizophrenia relapses in community mental health centers are at a medium level between those of state and contracted hospitals, with the additional benefi t of not depriving patients of their family life.Besides, the use of drugs with fewer potential side effects (atypical antipsychotic) has a positive impact on their quality of life. 16rect medical costs associated with schizophrenia treatment support the investment in strategies that target improved treatment compliance and fewer relapses, as well as the development of antipsychotic drugs aiming at reducing the need for daily care, especially inpatient service.In addition, patients with schizophrenia should undergo routine medical assessment as they usually have concurrent conditions such as weight gain, metabolic syndromes and cardiovascular diseases. 20 conclusion, further prospective studies including interviews with patients and their family are needed for better estimating non-medical/hospital direct (e.g., expenditures with transportation, accommodation) and indirect costs (loss of productivity of patients and their family).

Table 1 .
The study was approved to the Research Ethics Committee at Universidade Federal de São Paulo (Protocol No. 1782/06, in 12/15/2006).Patient confi dentiality was ensured.Clinical and sociodemographic characteristics of relapsed schizophrenia patients by type of health care services.City of São Paulo, Southeastern Brazil, 2006.
PH: public state hospital; CH: Brazilian National Health System-contracted hospital; CMHC: community mental health center.

Table 2 .
Components of direct medical costs used in schizophrenia relapse by type of health care services.City of São Paulo, Southeastern Brazil, 2006.
PH: public state hospital; CH: Brazilian National Health System-contracted hospital; MHC: community mental health center.

Table 3 .
Antipsychotic drug components of direct medical costs used in schizophrenia relapse by type of health care services.City of São Paulo, Southeastern Brazil, 2006.
PH: public state hospital; CH: Brazilian National Health System-contracted hospital; CMHC: community mental health center; tb: tablet; amp: ampule.

Table 4 .
Price (in US$) of psychotropic drugs components of direct medical costs in schizophrenia relapse by type of health care services.City of São Paulo, Southeastern Brazil, 2006.
PH: public state hospital; CH: Brazilian National Health System-contracted hospital; CMHC: community mental health center.