Commonly diagnosed mental disorders in a general hospital system

Background Considering many patients receive care from general hospitals, these healthcare institutions are uniquely situated to address mental and physical health needs. Little is documented, however, on the common current mental disorders diagnosed in patients receiving care in general hospital settings, especially in Puerto Rico. The objective of this study was to characterize the five most common current DSM-5 mental disorder diagnoses made in patients receiving non-psychiatric medical and surgical care from a general hospital system in southern Puerto Rico between January 2015 and December 2019. Methods Our clinical health psychology team provides integrated psychology consultation-liaison services to select clinical units in general hospitals across the southwestern region of Puerto Rico. The clinical team conducted routine standardized psychological evaluations at patients' bedside, arrived at a current DSM-5 diagnosis if warranted, and documented the diagnosis and other select variables. A retrospective study of cross-sectional data generated from the clinical team’s standardized evaluations of 5494 medical patients was implemented. Multinomial logistic regression analyses were used to assess the odds of being diagnosed with a current DSM-5 mental disorder during hospitalization. Results Overall, 53% of the entire sample was diagnosed with a mental disorder during hospitalization. Major depressive, neurocognitive, anxiety, substance-related and schizophrenia-spectrum disorders were the most frequently diagnosed. Interestingly, females were 23% less likely to have been diagnosed with major depressive disorder than males (aOR: 0.769, CI [0.650, 0.909], p = 0.002). This is to say males evidenced 1.30 higher odds of being diagnosed with depression compared to their female counterpart. Age, biological sex, civil status, employment status, monthly household income, previous mental disorder and history substance use/abuse history was differentially associated with receiving a current DSM-5 disorder. Conclusion The integration of clinical health psychology services within a general hospital facilitated our team’s work of identifying and treating co-occurring mental disorders among hospitalized patients receiving medical and surgical care. Future studies examining the opportunities and barriers of integrating clinical health psychology services within a general hospital’s administrative and clinical infrastructure for rapid identification and treatment of co-occurring mental disorders among medical patients is encouraged. Supplementary Information The online version contains supplementary material available at 10.1186/s13033-021-00484-w.


Introduction
Mental disorders are among the top causes of morbidity and account for approximately 14% of all deaths globally [45] and approximately 11% of the world's population currently lives with a mental disorder [29]. Alarmingly, this estimate is higher in Puerto Rico, as revealed by findings from a recent behavioral health needs assessment conducted on the island. This assessment reported 12-month prevalence rates for any mental disorder to be approximately 24% among adults ages 18 to 64 [9]. Moreover, Puerto Ricans with a serious mental illness make up 7.3% of the adult population. It is estimated that nearly half (43.1%) of Puerto Rican adults with some type of mental disorder are in need of mental healthcare services [9]. These findings elucidate the magnitude of the mental health disparity crisis on the island, with inaccessible quality mental healthcare being a leading driver of such disparity [21,43]. Since 2008, there has been a consistent migratory stream from the island to the United States. Included in this efflux of people are specialty mental healthcare providers such as clinical psychologists and psychiatrists, which further worsens the conundrum of lack of access to quality mental healthcare services [41,42]. Thus, there is an urgent need to address this. Considering many patients receive medical and surgical care from general hospitals, these institutions are uniquely situated to facilitate access to mental healthcare services through rapid identification of mental disorders and initiation of appropriate interventions. In fact, a recent meta-analysis revealed that between 4 and 32% of inpatients receiving care in a non-psychiatric general hospital setting evidenced clinically significant levels of depression [46]. An earlier review of epidemiological studies revealed the prevalence of mental disorders in general hospital inpatients ranged from 41 to 46%, with organic brain syndrome, adjustment disorders with depressed mood and alcohol abuse being the three most commonly diagnosed [30]. Such a characterization has yet to be accomplished in Puerto Rico. Thus, the current study's research team implemented a retrospective observational study of cross-sectional data to assess the most frequently diagnosed mental disorders in medical and surgical inpatients of a general hospital system in southern Puerto Rico between 2015 and 2019.
In order to achieve such a characterization, the systems that facilitate rapid identification of mental disorders and the provision of mental health services for medical and surgical inpatients of a general hospital are of critical value and worth briefly mentioning. Two widely used systems to integrate mental and physical health care of medical patients are the collaborative care [40] and consultation-liaison models [39]. The collaborative care model is mostly used in outpatient, primary-care settings and have limited implications for use in general hospital settings. The consultation-liaison (C-L) model, however, is more relevant for this present study's aim and in fact, is the model which was used by our clinical team as detailed in the methods section below. Briefly, the C-L model entails two distinct yet overleaping components as the name implies: consultation and liaison. The consultation component is a process whereby attending physicians, surgeons, patients themselves or family members may request mental health services, and a clinical team member responds to this request by conducting an initial assessment, arriving at a clinical diagnostic impression, facilitating interventions as needed and making appropriate referrals if warranted. Liaison services are an especially crucial aspect in the provision of mental health care services to medical inpatients as this component requires a certain degree of integration at not only the service level (i.e., multidisciplinary teams) but at the organizational level as well [37]. Most attention in the literature, however, is given to the consultation aspect of the C-L model [14]. This selective attention is rather unfortunate as the liaison component, as overviewed in the discussion section below, is an aspect that warrants urgent attention and widespread implementation to achieve integrated care models in general hospital settings in an effort to curtail the mental health disparities that exist among medical populations such as in Puerto Rico. The success of detecting current mental disorders and facilitating interventions among the sample of medical and surgical inpatients of this present study is owed, in part, to the clinical health psychology C-L integration model implemented.

Methods
The research team conducted a retrospective study examining cross-sectional biopsychosocial variables obtained from a sample of 5494 inpatients that received medical or surgical care from a general hospital in Ponce, Puerto Rico between January 2015 and December 2019.

Standardized clinical evaluation
The standardized clinical evaluation entailed a mental status examination, exploration of psychological and emotional symptoms, a review of past physical and mental illness history, and a review of social functioning. Clinical inventories such as the Mini Mental Status Exam [15], Beck Depression Inventroy-II [6], and Beck Anxiety Inventory [7] were used to aid in the exploration of mental status and psychological-emotional symptomology. Additionally, a review of laboratory results, medications, and physician and nurse progress notes was conducted to aid in ruling out organic causes of possible mental aberrations. The CPSP clinical team performed the standardized clinical evaluation at the patient's bedside. If warranted, the clinical team made a current DSM-5 diagnosis and a corresponding brief treatment plan was implemented for the duration of the patient's hospital stay. For patients that did not evidence clinically significant symptoms as determined by the DSM-5 during evaluation, no diagnosis was made, and these patients' data was used as a comparison group. The clinical inventories mentioned above were not included in the data analysis process as these measures were not available in the digital database alongside the current DSM-5 diagnoses and sociodemographic variables.

Sociodemographic, medical diagnosis, and past mental health history
Questionnaires generated by the clinical team were used to collect the following biopsychosocial variables: age, biological sex, civil status, level of education, employment status, and monthly household income. Past history of mental disorder and history of substance use/abuse variables were also included in the study. ICD-10 medical diagnoses were obtained from chart reviews performed by the CPSP clinical team.

Data analysis
SPSS 27 (IBM Corp) was used to perform data analysis. Data preparation and exploration procedures were conducted, and data was examined for errors and quality. Missing data analysis was conducted, and missing variables were replaced with multiple imputation linear regression method, with 10 imputations utilized. [31]. All missing data (12-monthly household income variables and 23-civil status variables) in this dataset were less than 5%. Descriptive statistics were performed to assess measures of central tendencies, standard deviation, and confidence intervals. Chi-squared tests of independence were conducted to assess associations between the top five diagnosed mental disorders and select sociodemographic indicators. Multinomial logistic regressions were conducted to evaluate the odds of being diagnosed with a DSM-5 depressive, neurocognitive, anxiety, substancerelated, or schizophrenia spectrum disorder compared to no disorder among inpatients. The Benjamini-Hockberg procedure [8] was applied to control for false discovery rates due to multiple comparisons in the adjusted multinomial logistic regression model.

Results
Data from a total of 5494 inpatients was analyzed. The mean age of this patient sample was 62.2 years, 95% CI [61.75, 62.65]. There were more females (57.8%) than males and the majority of patients were unmarried (53.2%). Additionally, 83.6% were unemployed or retired and 26% reported having obtained less than a high school education. Overall, more than half (53%) of the entire sample was diagnosed with a current DSM-5 mental disorder during hospitalization. Major depressive disorders accounted for 42%, neurocognitive (33%), anxiety (11%), substance-related (8%), and schizophrenia spectrum (6%) of the five most frequently diagnoses made by the CPSP clinical team. Further, 97% of all patients evidenced a chronic physical condition. Ninety-nine per cent (99.4%) of male patients had a chronic physical condition compared to 95% of female patients. Fifty-nine percent (59%) of male patients had comorbid mental and physical conditions compared to 48% of females.

Specific DSM-5 diagnoses
Among major depressive disorders diagnoses (MDD), recurrent episodes (70%) were most frequently diagnosed and recurrent episodes with psychotic characteristics (6.9%), the least frequent. Within neurocognitive disorders (NCD), delirium due to another medical condition (24.8%) was most commonly diagnosed and unspecified delirium (6.8%) the least common. Generalized anxiety disorder (72.5%) was the most frequently diagnosed anxiety disorder and alcohol use disorder (48%) the most frequently diagnosed substance-related disorder. Schizophrenia and brief psychotic disorder were the most and least common diagnosed within the schizophrenia spectrum disorders, 34.6% and 6.7%, respectively.

Comorbid chronic physical conditions and diagnosed mental disorders
Overall, cardiovascular disease, orthopedic conditions, chronic respiratory, gastrointestinal and cancer diagnoses were the most prevalent comorbid medical conditions in this sample. Thirty-five percent (35%) of this sample were evaluated in the skilled nursing unit, 24% coronary care, 22% intensive care and 19% cardiovascular surgery unit. Among patients diagnosed with major depressive disorders, slightly more than one in five patients (24%) also lived with a co-occurring cardiovascular disease. The least frequent co-occurring chronic physical disease among patients with depressive disorders was immunologic disease, accounting for 0.2% of all ICD-10 diagnoses among this patient group. Chronic neurological diseases were the most common co-occurring physical conditions among patients diagnosed with a DSM-5 neurocognitive condition (22%), and chronic endocrine disorders (1%) were the least common. Slightly more than one out of every four patients (27%) diagnosed with an anxiety disorder also had a co-occurring cardiovascular disease. Chronic immunologic disorders were least common among this group of patients, accounting for 0.3% of ICD-10 diagnoses. Similarly, patients diagnosed with a substance-related disorder had higher cardiovascular disease comorbidity (18%) than any other chronic physical condition. Lastly, among patients diagnosed with a schizophrenia spectrum disorder, 5.8% also lived with a co-occurring cardiovascular disease. Routine standardized clinical evaluations as part of the administrative standing orders accounted for 84% of all DSM-5 diagnoses made and the remaining 16% were made from physician or patient requested consults. The two most common diagnoses resulting from the administrative standing orders (routine standardized clinical interviews-the liaison component) were neurocognitive and major depressive disorders, accounting for 20% and 19%, respectively. Diagnoses made from all consults, however, consisted primarily of major depressive disorders (37%), followed by substance use disorders (9%).

Association of biological sex and mental disorder diagnoses in a general hospital setting
An unadjusted multinomial logistic regression model revealed that biological sex is differentially associated with being diagnosed with a current mental disorder during hospitalization. The overall model was significant χ 2 (5) = 215.659, p < 0.001, and biological sex accounted for 4.1% of the variance found within the mental disorder diagnosis categories (Nagelkerke pseudo R 2 = 0.041). Overall, male patients had higher odds of being diagnosed with a current DSM-5 diagnosis compared to females. However, being diagnosed with an anxiety disorder was not significantly associated with biological sex. Female patients were 15.7% less likely to be diagnosed with MDD compared to male (OR: 0.842, 95% CI [0.732, 0.968], p = 0.01). This is to say males were 1.18 times more likely to have been diagnosed with a current MDD at bedside than females. Moreover, males evidenced 1.78

Discussion
After controlling for covariates in the multinomial model, spurious associations were discovered among anxiety, substance-related, and schizophrenia spectrum disorders. Covariates of major depressive and neurocognitive disorders, however, not only maintained significance, but increased in odds ratio from the unadjusted model. Thus, the following discussion will highlight select implications regarding major depressive and neurocognitive disorders diagnosed in inpatients receiving medical or surgical care in a general hospital. Overall, 53% of our sample of medical inpatients received a current DSM-5 mental disorder diagnosis. While seemingly high, it is comparable to other studies that have characterized mental illness prevalence among medical inpatients [5,18,33,34,46].

Major depressive disorders
Major depressive disorder (MDD) is a mood disorder that negatively impacts affective, cognitive, physiological, and behavioral systems. Hallmark features of MDD are sadness, inability to enjoy things that were once pleasurable, and a myriad of signs and symptoms ranging from lack difficulty concentrating and insomnia, to fluctuation in weight and changes in eating patterns. In the severest form, patients with MDD may evidence recurring thoughts of death, suicidal ideation, and even suicide attempts [3]. A recent epidemiological study revealed that among non-institutionalized community members, MDD was slightly higher in Puerto Rican adults living on the island (9.9%) compared to the U.S. adult population (8.5%) [10]. Traditionally, among U.S. adults, MDD rates are higher in females compared to males, 8.7% and 5.3%, respectively, and is most prevalent in people aged 18 to 25.
[26]. Sociodemographic factors such as low educational attainment, economic instability, and lack of social support are implicated in the etiology and course of MDD [4]. Traditionally, females are disproportionally affected by MDD [2,23,26,28]. However, this was not the case in this sample of general hospital inpatients. Male patients evidenced 1.30 higher odds of being diagnosed with a major depressive disorder compared to females. At least two plausible explanations can account for this finding. First, in this setting where evaluations and diagnoses were made (a general hospital), both males and females were equally likely to have received a psychological evaluation and subsequent diagnosis, if warranted. Yet, this is not the case in community and outpatient settings, where females are more likely to seek mental healthcare services than males [25,36,38]. It begs the question if higher MDD prevalence rates among women are a function of healthcare-seeking behavior or due to actual biological sex-related difference in development and course of depression. Secondly, considering that male patients evidenced higher proportions of mental and physical illness comorbidities than females, 59% and 48%, respectively, it is plausible that males engaged in relatively maladaptive coping mechanisms in managing their conditions, contributing to depressive affect and subsequent diagnosis during hospitalization. Nevertheless, this does not fully explain why male patients evidenced higher odds of being diagnosed with MDD during hospitalization compared to females.
Regarding the associations between sociodemographic variables spanning age, civil status, employment status, income level and a current diagnosis of depression among this present study's sample, our findings are consistent with existing literature describing non-medical and medical populations [1,4,27,32,33,46]. In short, relatively older age and being married is associated with a lower incidence of receiving a current depression diagnosis. Unemployment (including retirement), and lower income is associated with higher incidence of receiving a current depression diagnosis (see Additional file 1).

Neurocognitive disorders
Neurocognitive Disorders (NCD) are a group of disorders that are characterized by significant clinical dysfunction in areas of complex attention, executive functioning, perpetual motor skills, social cognition, and emotional reaction. Such dysfunction is readily recognizable from premorbid functioning and unfolds in a progressive manner [3]. Considering there are 10 million new cases of NCDs registered yearly, it is estimated that by 2030 there will be an alarming 82 million people suffering from a form of NCD worldwide [47]. NCDs have been among the most prevalent causes of death in the world. While much less is known about the prevalence rates of NCD in Puerto Rico, researchers have estimated that 5% of adults who are 45 years of age and older report cognitive decline. Of those 5%, the vast majority (87%) of people with subjective cognitive decline have at least one chronic physical condition [11].
In our sample, as expected, relatively older age was associated with receiving a current NCD diagnosis, specifically, a delirium diagnosis during hospitalization, where each one-year increase in age was associated with 1.05 higher odds of receiving an NCD diagnosis. Delirium disorders are common among elderly patients admitted to general hospitals, with estimates ranging from 14 and 56% of all elderly patients being affected during hospital stay [3,16]. Even though most patients that present with delirium fully recover, if not detected and treated promptly, delirium can lead to adverse consequences and even mortality, especially among older patients [19]. Interestingly and fortunately for the patients, in this present study, unlike major depressive disorders, NCDs were diagnosed more owing to the standing order routine evaluations compared to consults requested by attending physicians, suggesting that if it had not been for routine the evaluations in place, rapid identification of an NCD and initiation of appropriate psychosocial interventions would have possibly been forfeited.
In our study, the second most single common NCD was unspecified neurocognitive disorder, accounting for 14.8% of all NCD diagnoses. In terms of sociodemographic factors, increasing age, especially individuals over 60 years is associated with receiving a current NCD diagnosis [3]. While there are sex-related differences associated with various subtypes of NCDs, generally, females have higher rates of diagnosed NCDs compared to males. This difference is believed to be due to greater longevity in females [17]. Our findings differ slightly from existing literature as males were more likely to have received a current NCD diagnosis. In fact, female patients evidenced 45% lower odds of being diagnosed with an NCD than males. We believe these findings are a function of the healthcare seeking behavior specific to this sample as discussed in the section regarding MDD diagnoses.

Key implications
The vast majority (90%) of patients that were diagnosed with MDD and 87% that were diagnosed with NCD received an evaluation within 24 h of being admitted to corresponding hospital units or after physician or patient requested consultations. Swift detection of a current mental disorder during hospitalization and initiation of appropriate interventions is largely due to the integration of clinical health psychology services within a general hospital care delivery model. The CPSP is part of a process-based integration model [22]. In a processbased integration model, a set of methods and models are intertwined to ensure the alignment of financial, administrative, organizational, and clinical functions, with the intention of effectively combining the provision of two or more originally separate clinical services (eg. Coronary care and clinical health psychology). When done correctly, process-based integration results in a fluid, interdisciplinary approach that generates benefits for the whole person/patient and can be conceptualized as integrated care [22]. The implementation of process-based comprehensive care models can be located within one or more of the following general typologies: (a) organizational, (b) functional (c) service, and (d) clinical [24]. Organizational integration occurs when two or more separate organizations (for example, a general hospital and a graduate health sciences institution) join forces to some extent to align the delivery of care for a given patient population. Functional integration is the exchange of non-clinical support components, such as electronic medical records, within and between organizations and clinical units. Service integration is most easily identified as a team of professionals offering their perspective clinical services in a comprehensive and interconnected manner. Finally, clinical integration requires the provision of care by an interdisciplinary team through the use of shared guidelines or protocols (for example, a standing administrative order requiring that all patients admitted to the clinical unit undergo a screening) and standardized assessment. The model implemented by the CPSP incorporates, to some degree, organizational, functional, service, and clinical components. The program has two modalities of patient evaluation (standardized administrative orders and medical referrals/consults). There is organizational integration established through a collaborative agreement between Hospital Damas and Ponce Health Sciences University (the home institution of our clinical and research teams). Also, the CPSP is located in physical space within the hospital which allows for quick response times and direct contact with doctors and nurses. Contact with these health service providers helps in obtaining direct information on the medical conditions of patients, contributes to reaching an accurate diagnosis of mental disorders, and facilitates the implementation of comprehensive treatment. This approach has been shown to help improve health outcomes for patients and reduce costs related to healthcare expenses [12].The theoretical model of integrated care on which CPSP operates was designed and implemented by [20], the CPSP coordinator and senior researcher and author of our research team.
Interventions facilitated by the clinical team were implemented at the patient's bedside during the course of the initial clinical evaluation. Follow-up sessions were scheduled if warranted. Interventions were tailored to meet the patient's most pressing need during hospitalization and spanned cognitive and behavioral techniques primarily aimed at improving cognitive restructuring, emotion regulation, activation of physiologic relaxation response and solution focused techniques to enhance adherence to medical treatment regimens. While not detailed in this present study, the clinical team referred patients to specialty care services when necessary and implemented brief interventions with family and other healthcare workers including physicians, nurses and social workers as needed to enhance the clinical health psychology services provided. Additionally, patients that require out-patient behavioral health services (e.g., psychotherapy and pharmacotherapy) are referred to our community health center (Wellness Center of Ponce Health Sciences University).

Strengths, limitations, and future work
While the results from this study cannot support a causative link between sociodemographic factors, previous psychiatric history, and current mental disorder diagnosis, the relatively large sample size and applied research approach allow these findings to be generalized outside of this sample of patients. A possible limitation for this study, however, is the research team did not analyze quantitative psychological symptom inventories (e.g., BDI-II) as covariates. These inventories are administered at patient's bedside to aid in arriving at an accurate current diagnosis. However, these inventories were not available in the digital database as were the other variables analyzed in this study. Also, our clinical team did not implement measures to initiate patient follow-up post hospital discharge for symptom monitoring or continued psychotherapy. Due to the brief (average two-to-three sessions) and often time single session encounters that is common in C-L psychology in general hospital settings, it is challenging to gauge the effectiveness of the treatment interventions once patients are discharged. This is concerning as the ultimate goal of intervening in medical patients is to enhance overall health outcomes by addressing pressing mental health challenges that may be interfering with physical health and daily functioning. When there is no follow-up system in place, there is no way of knowing if patients actually benefited from the interventions of if they require more long-term therapy.
There is limited literature detailing the effectiveness of post-discharge C-L follow-up systems in identifying areas of patient need and continued therapy [13,35]. Thus, future clinical and research work related to designing and implementing post-discharge follow-up is a promising avenue.

Conclusion
Major depressive, neurocognitive, anxiety, substance use, and schizophrenia-spectrum disorders were the five most commonly made DSM-5 diagnoses in a general hospital system. Male patients were 1.30 times more likely to have been diagnosed with major depressive disorder compared to females, and delirium was the most common form of NCD among both males and females. The importance of integrating clinical health psychology services within the administrative and clinical structure of general hospitals was discussed. Future work concerning the opportunities, barriers and program evaluation methods are needed. Additionally, ways of continuing and or monitoring of psychological care post discharge is worth future clinical and research efforts.