With the Declaration of Caracas in 1990, reforms of psychiatric services in Latin American countries received renewed interest (
1). Changes included the deinstitutionalization of people with mental illness and the inclusion of mental health as part of primary care. The number of beds in psychiatric hospitals was reduced, and psychiatric units in general hospitals, outpatient treatment centers, and residential services were created to promote the social inclusion of psychiatric patients and to support community care and treatments closer to home (
2,
3). Psychiatric reforms and policies recommended by the World Health Organization (WHO) were based on evidence from high-income countries in the global West (
4); however, these recommendations did not necessarily consider specific needs within the cultural and socioeconomic contexts of Latin American countries. More evidence is therefore needed on the changes in care associated with psychiatric and health system reforms in these countries.
According to the WHO report on mental health systems from 2013 (
5), the number of psychiatric beds per 100,000 people in general hospitals ranged from 2.5 in Central America and the Latin Caribbean to 4.6 in South America and in psychiatric hospitals from 12.0 to 16.4 in the same respective regions.
Previous cross-sectional research in Latin America has reported on the quantity and quality of therapeutic communities in five countries—Argentina, Brazil, Colombia, Mexico, and Peru—providing evidence that psychiatric facilities are available in this region (
6). However, variations in these facilities over time have not yet been reported in the literature for this region. Similarly, the number of available treatment slots in day hospitals (
7) and number of outpatient facilities—as well as trends in the availability of these services—have not yet been reported for the Latin American region. Moreover, the treatment gap for mental disorders in Latin America remains a major concern because more than three-quarters of individuals with mental disorders in the region do not have access to treatment. This situation is particularly critical for children, adolescents, and Indigenous populations (
8).
South America’s prison population has risen by 200% since 2000, and a similar phenomenon was observed in Central America, whereas the worldwide prison population has remained unchanged on average over this period (
9). A high prevalence of mental health and substance use problems has been reported in these populations (
7,
10–
12), and a lack of psychiatric care may have contributed to the increase in incarceration rates in Latin America. Previous research has reported an association between the reduction in the total number of psychiatric beds and the increase in prison population rates in the region (
13,
14). However, trends for aggregated total numbers of psychiatric beds can mask important shifts in the types of psychiatric beds available and the level of integration between psychiatric services and general medical systems of care. To date, among Latin American countries, disaggregated trends of the availability of psychiatric beds in settings such as general and psychiatric hospitals and for specific populations, along with other components of the mental health care system, have been reported only for Chile (
15). The present study aimed to assess the availability of service data for the entire Latin American region by detailing changes in the number of outpatient facilities and in the availability of psychiatric beds in psychiatric and general hospitals, psychiatric beds for children and adolescents, specialized forensic psychiatric beds, residential beds for substance use treatment, treatment slots in residential facilities, and treatment slots in day hospitals. Assessing and characterizing the nature and availability of mental health services in Latin America over time would enable comparisons across several countries in the region.
Methods
We conducted an exploratory retrospective observational study on the prevalence of psychiatric beds and outpatient psychiatric facilities and on trends in the availability of these care settings in Latin American countries between 1990 and 2020. Because we worked with administrative data of service indicators, no ethical approval was sought for this study.
Data Sources
Between April 2019 and September 2021, we formed an international network of researchers from Latin American countries. Collaborators were contacted on the basis of their authorship in scientific journals, authorship of WHO Assessment Instrument for Mental Health Systems reports, personal networks, and snowball recruitment. We also contacted the local ministries of health and related government institutions. We asked potential collaborators who could not participate whether they knew someone in the country or neighboring countries who was willing to participate. Communication in the research network was through e-mail in the Spanish language. A template was used to collect data for each country from 1990 to 2020. The template was based on the European Description and Evaluation of Services and Directories for Long-Term Care to longitudinally assess services and was modified to reflect the types of services and registries common in Latin America (
16) (see the
online supplement to this article). Contributors directly accessed national registries that had consistent collection methods and identical data definitions over time. Data were collected between April 2019 and July 2021. Researchers from the following 16 countries participated in the network: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Panama, Paraguay, Peru, and Uruguay. We contacted at least one interested collaborator in both Cuba and Nicaragua who could not obtain the data sought for this study or did not sustain contact. Venezuela was excluded from the analysis because of data scarcity.
Definition of Indicators
The prevalence of each psychiatric services indicator was calculated as the number of each indicator per 100,000 people for each year in the observation period (
17). Eight service indicators were assessed: psychiatric beds in psychiatric hospitals, psychiatric beds in general hospitals, psychiatric beds for children and adolescents, forensic psychiatric beds, residential beds for substance use treatment, treatment slots in residential facilities, treatment slots in day hospitals, and number of outpatient facilities.
Psychiatric beds in psychiatric hospitals were classified as short-stay (≤6 months) and long-stay (>6 months) beds. The total number of psychiatric beds in psychiatric hospitals was calculated when counts of both short- and long-stay beds were available. Psychiatric beds in general hospitals were included only as short-stay beds because long-stay beds were unusual in this setting and were reported only from Panama. Treatment slots in residential facilities included those in community-based nonhospital mental health facilities that provided overnight residence, primarily serving patients with stable mental illnesses. Outpatient facilities included all services that managed psychiatric disorders in an outpatient setting, including hospital- and community-based facilities.
When primary data were unavailable, the total number of psychiatric beds in psychiatric and general hospitals was retrieved from the Mental Health Atlas Project of the WHO (
18). Data for 2020 could be gathered for only three indicators: psychiatric beds for children and adolescents, treatment slots in residential facilities, and number of outpatient facilities. We excluded facilities specifically offering treatment for people with intellectual disabilities and any generic facility that was not explicitly intended for mental health needs (e.g., rest and nursing homes for older people). Because private psychiatric beds are available to only a small proportion of the population in Latin America, they were excluded from the main analyses, and results on this service are reported in the
online supplement.
Statistical Analysis
We worked with data from 16 countries collected over 31 years (i.e., 496 possible observations). We focused on the differences between the first and last data points, reporting median prevalence per 100,000 people and percentage changes. Of note, the median prevalence of service indicators was rounded to two decimals after calculating the differences between the first and last data points. Population estimates per year and country were obtained from the World Bank. We also investigated the trends in availability of the specified indicators in the countries with the largest increases or decreases in prevalence for each indicator. Additionally, we graphically represented the trends observed throughout the study period for the indicators studied (see figures in the online supplement). Because an extensive range of values needed to be plotted, we plotted the data on a semilogarithmic scale (i.e., a log scale on the y-axis for the widely differing prevalence of indicators and a linear scale on the x-axis for years). All analyses were conducted in Microsoft Excel, version 16.0. Data sets generated and analyzed for this study are available from the corresponding author on request.
Results
Tables 1 and
2 report the prevalence of each service indicator at the first and last available data points for each country. The tables also report the number of countries with available data for each indicator and the mean percentage of available data points during the observation period in the countries for which data were available. Semilogarithmic graphs show the prevalence of each indicator over the study period, by country (see the
online supplement).
Data for the total prevalence of psychiatric beds in psychiatric hospitals were available from 11 countries (Brazil, Chile, Costa Rica, Dominican Republic, Ecuador, Guatemala, Honduras, Mexico, Panama, Paraguay, and Peru), with an average availability of 43% of data points during the observation period. For five countries (Argentina, Bolivia, Colombia, El Salvador, and Uruguay), data were unavailable.
The median prevalence of psychiatric beds in psychiatric hospitals decreased from 5.1 to 3.0 per 100,000 people (−42% change) during the study period. The greatest relative decrease in the prevalence of psychiatric beds in psychiatric hospitals was observed in the Dominican Republic, with a reduction of 100% from the first to the last data point (
Table 1). No country reported increases in this indicator. The median prevalence of short-stay psychiatric beds in psychiatric hospitals also decreased, from 4.3 to 2.6 per 100,000 people (−40%). In most countries, a decrease in prevalence was observed for this indicator, with the most substantial decrease corresponding to the closure of all psychiatric hospitals in the Dominican Republic (100% reduction from the first to the last data point). However, an increase was observed in Peru (from 0.26 to 0.41 per 100,000 people). The median prevalence of long-stay psychiatric beds in psychiatric hospitals decreased from 2.0 to 1.2 (−41% change from first to last data point) over the study period. All countries reported decreases between the first and the last data point, with the strongest relative reduction again observed in the Dominican Republic (−100%).
The median prevalence of psychiatric beds in general hospitals decreased from 1.0 to 0.8 per 100,000 people (−24%). The most substantial decrease was observed in Ecuador, with a reduction of 63% from the first to the last data point. Chile had the most substantial increase (185%) for this indicator. The median prevalence of psychiatric beds for children and adolescents (0.18 beds per 100,000 people) was unchanged. We noted the greatest increase from 1990 to 2020 in Chile (0.11 to 1.01, 803%) and the greatest decrease in Mexico (0.14 to 0.07, −47%) (
Table 2).
The median prevalence of forensic psychiatric beds remained low at around 0.04 per 100,000 people during the study period. The greatest increase was reported from Chile (0.30 to 1.18, 292%), and the largest decrease was recorded in Brazil (3.38 to 1.55, −54%). The median prevalence of residential beds for substance use treatment rose from 0.40 to 0.57 per 100,000 people (43% increase). The most substantial decrease was observed in Peru, with a reduction of 16% from the first to the last data point (0.22 to 0.18). Honduras had the most substantial increase for this indicator, with a growth of 610%, from 0.09 to 0.64.
The median prevalence of available treatment slots in residential facilities increased from the first to the last data point (0.67 to 0.79, 17% increase). The strongest reduction was observed in Guatemala (−79%) and the most substantial increase in Brazil (4,998%). The median prevalence of treatment slots in day hospitals increased from 0.41 to 0.54 (32% increase). Chile had the greatest increase, from none to 5.17 per 100,000 people, and the largest reduction was observed in Panama (0.76 to 0.24, −69%). The median prevalence of psychiatric outpatient treatment facilities increased from 0.39 to 0.93 per 100,000 people (138% increase). The greatest increase was reported from Brazil (0.01 to 0.93, 11,449% increase), and the largest decrease was recorded in Costa Rica (0.03 to 0.02, −38%). The number of data points retrieved from the WHO Mental Health Atlas Project for each indicator is reported in the online supplement.
Discussion
Main Findings
The available service data revealed that the median prevalence of both short- and long-stay psychiatric beds in psychiatric hospitals and of psychiatric beds in general hospitals decreased in Latin America during the 1990–2020 period. Specialized psychiatric beds for children and adolescents and for forensic psychiatric populations had a very low prevalence, with negligible changes over time. In contrast, the prevalence of residential beds for substance use treatment and placements in residential facilities, in day hospitals, and in outpatient facilities increased throughout the study period.
Interpretation
During the past 30 years, amid political transformations, Latin American countries underwent changes in their capacity to deliver mental health services. We noted reductions in the prevalence of psychiatric beds in psychiatric hospitals and expansions in the capacities to deliver outpatient and community psychiatric care. The increase in outpatient and community psychiatric care in Brazil, which accounts for about a third of the population in Latin America, may guide developments in other health systems. Community-based inpatient treatment centers have shorter mean lengths of stay (e.g., 14 days in Brazil) compared with psychiatric hospitals (65 days) (
19), enabling treatment of more patients and faster turnover in general hospital and community settings than in psychiatric hospitals.
These changes in the services provided indicate a better integration of psychiatric care systems with general medical care, a more balanced geographic distribution, and improved integration with community and social care. However, the removal of psychiatric beds in psychiatric hospitals has not been offset by a corresponding increase in psychiatric beds in general hospitals. At the last available data point, the total prevalence of psychiatric beds was >15 times lower than recommended by international experts (
20,
21), suggesting a critical shortage of beds for acute care needs in the region (
21). The prevalence of total psychiatric beds in Latin America was lower than in Eastern Europe and Central Asia, representing other middle-income regions (
22), and only a little higher than the prevalence in sub-Saharan Africa (
23), which has a lower average per capita income. Low prevalence of psychiatric beds has been associated with higher occupancy rates and overcrowding (
24); increased readmission rates (
25); longer stays in emergency departments (
26,
27); treatments in other general medical departments (
28); increased rates of suicide, violent crime, early death, detention, and homelessness (
29–
32); and other adverse social and health outcomes (
33).
Availability of psychiatric beds for children and adolescents remained at inadequate levels across the Latin American region, posing a challenge to providing sufficient acute psychiatric care for this vulnerable population. Similarly, specialized forensic psychiatric beds remained unavailable in many Latin American countries. Such shortage can threaten the human rights of individuals with mental disorders and criminal legal involvement (
7). These two types of specialized psychiatric beds may need special attention during future service developments. An important finding of this study was that simple service data, for example, psychiatric bed counts, were unavailable in many places, including countries with relatively high per capita income, such as Argentina and Uruguay, where such data collections should be feasible.
Advances in service development have been achieved in residential facilities, therapeutic communities for the residential treatment of substance use disorders, and outpatient facilities. However, these service resources have remained scarce, with varying levels of availability among countries. Latin America presents particular challenges for developing mental health systems, indicating that the reforms of psychiatric systems recommended by the WHO (
34) are not easily transferable from high-income to low- and middle-income regions. Aspects of these reforms, such as removal of psychiatric beds, may have been applied without considering the specific national contexts and general scarcity of these resources in Latin American countries. Per capita mental health expenditures in Latin American countries are among the lowest globally (
34), which has implications for the development of mental health systems. Many residential beds for substance use treatment in Latin America are in nonmedical facilities run by people with lived experience of substance use disorders. These facilities may not always use evidence-based interventions, and their care quality is variable. Furthermore, service and epidemiological data have been limited (
35). In Argentina, data on most service indicators were unavailable during the observation period, and several other countries reported data for only 1 year. Health care systems were typically fragmented, and registries often did not cover whole countries but only specific areas, thereby complicating data access and analysis. In addition, definitions of service components require refinement and international consensus. The variety of definitions used worldwide and a lack of clear definitions of services limit international comparisons that are useful to plan mental health systems (
33,
36–
38), even among countries within Latin America.
Strengths and Limitations
Our study contributes to the literature by assessing and comparing components of mental health services in 16 Latin American countries—representing >600 million people—over three decades. An international Latin American network of researchers was crucial for synthesizing data from different national registries and contributing to diverse views on the psychiatric services provided in this region. We gained insights into changes and trends in psychiatric care in the region via comparisons among its nations that took into account the region’s dynamic history of psychiatric reforms. Documenting changes of services over time may help to address the lack of standardized service comparisons in Latin America and to gain a deeper understanding of the types of and changes in mental health services in this region, which is crucial for allocating resources at the local level.
Limitations of the study should also be considered when interpreting its results. For our analyses, we had to rely on aggregated administrative databases, which were often incomplete. Rates of missing data for psychiatric beds and outpatient services during the study period were high (approximately 50% and 70%, respectively, for countries where information could be retrieved). Also, the patterns of missing data across the period observed were inconsistent among countries, limiting national comparisons. Moreover, service data in the region were often reported for specific provinces and not for entire nations. Additionally, the data did not provide information on the degree of integration of mental health into primary care, and the COVID-19 pandemic imposed additional barriers to accessing some of the data registries.
Conclusions
Reforms to psychiatric services in Latin America have resulted in a more diverse set of services over the past three decades. Residential beds for substance use treatment, placements in residential facilities or day hospitals, and outpatient facilities are increasingly available in the region. These observations can be considered notable progress, and all Latin American countries should strive to achieve such levels. However, the prevalence of psychiatric beds decreased in general hospitals, a trend that should be reversed to provide sufficient and geographically balanced acute care. Specialized psychiatric beds for children and adolescents and forensic psychiatric beds are still lacking in large parts of the region, and their increased availability should be part of the agenda for additional service improvements. Further international standardization of data registries and public accessibility of data are needed to facilitate comparisons among regions and countries and to aid further developments (
39). The results of this international collaboration call for more and better data registries to monitor provision of psychiatric services in Latin America.