1. Introduction

The delivery of high-quality inpatient psychiatric services remains a central challenge for mental-health systems worldwide. Individuals admitted for acute psychiatric care often face complex psychosocial risks, and the transition from hospital care to the community is recognized as a particularly vulnerable period [1–3]. Studies consistently show that patients discharged from inpatient psychiatric units exhibit elevated rates of emergency department (ED) visits, hospital readmissions, suicidality, and other adverse outcomes [4–7]. Repeated emergency department visits after an inpatient discharge places a substantial financial and operational burden on hospitals. Psychiatric presentations require specialized assessments and frequently lead to prolonged ED stays due to challenges with placement and bed shortages. This cycle contributes to ED congestion, increases staff workload, and drives up health system costs [8–10]. Moreover, outcomes such as ED visits have been linked to a range of clinical, sociodemographic and service factors (e.g., substance use comorbidity, homelessness at discharge) [5,11]. While much of the available literature focuses on measurable outcomes (e.g., readmission, ED utilization, mortality), there is growing recognition that patient‐reported experience particularly inpatient satisfaction with hospital services, may serve as an upstream indicator of inpatient care quality and may influence post-discharge outcomes [12–14]. In general medical settings, patient satisfaction has been associated with better care engagement, adherence to discharge instructions, reduced acute care use, and improved outcomes. Accordingly, the concept of inpatient satisfaction reflecting patients’ perceptions of the care environment, the therapeutic alliance, discharge processes, and information provision may be particularly salient for psychiatric care, where engagement, trust, and continuity are critical [15,16]. Existing literature reveals that the post-discharge period is characterized by elevated risk, including limited social support, inadequate access to outpatient mental health services, financial inadequacy, and unstable housing status, which contribute to their increased utilization of emergency mental health care services [17,18]. Factors contributing to increased ED utilization following discharge encompass a range of sociodemographic and clinical variables. Patient characteristics such as age, ethnicity, employment status, and housing status have been consistently linked to a higher likelihood of return visits to the ED. A study conducted in Ontario found that being homeless at discharge was associated with nearly double the hazard of an ED visit relative to non-homeless peers [19]. Clinical factors, including primary psychiatric diagnoses, symptom severity at the time of discharge, and prior patterns of healthcare usage, further elevate this risk [20–22]. This study examines predictors of psychiatric emergency department (ED) visits within 12 months following discharge from acute psychiatric inpatient units among participants enrolled in a broader stepped-wedge study evaluating transitional care interventions. The analysis considered multiple potential predictors simultaneously, including study intervention exposure [treatment as usual (TAU), supportive text messaging (SMS), and supportive text messaging combined with peer support (SMS + PS)], inpatient satisfaction, sociodemographic characteristics, prior healthcare utilization, and clinical factors. In addition to exploring whether transitional care interventions were associated with subsequent ED utilization, this study specifically aimed to examine whether patient satisfaction during hospitalization independently predicted post-discharge psychiatric ED visits. The findings are expected to inform healthcare providers and policymakers in refining transitional care strategies, with the goal of reducing pressures on emergency psychiatric services and enhancing patient-centered recovery outcomes.

2. Methods

3. Results

Table 1 presents the univariate analysis examining factors associated with participants’ ED visit in the 12 months before their index admission. Across the full sample of 1,070 participants, most were between 25 years old or younger (395, 36.9%), with slightly more females than males (586 54.8%). The majority identified as Caucasian (586, 54.8%), held a high school diploma (554, 51.8%), and were single at the time of data collection (633, 59.2%). Over half of the participants were unemployed (568, 53.1%), and most lived with family or friends or in rented accommodation, with a smaller proportion owning their homes (211, 19.7%). More than half of the sample demonstrated low resilience (603, 56.4%). Most the participants were very satisfied or satisfied with their prior inpatient experience (831, 88.9%). Wellbeing, depression, and anxiety symptom categories were relatively evenly distributed. Overall, 40% of participants recorded an ED visit within 12 months post–discharge, and participants were distributed across the three study clusters (SMS, SMS + PS, and TAU), with the largest proportion assigned to TAU (432, 39.2%).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Distribution of demographic and clinical characteristics of the participants against the 12- month prior ED visit before index admission.

https://doi.org/10.1371/journal.pone.0351753.t001

Table 2 illustrates the results of a binary logistic regression analysis predicting the likelihood of ED visits within twelve months following discharge. Fourteen independent variables were entered into the model, including age category, gender, ethnicity, education, current relationship status, employment status, current housing status, inpatient satisfaction, mental health diagnosis, Brief Resilience Scale (BRS) category, WHO-5 wellbeing category, PHQ-9 depression category, GAD-7 anxiety category, study cluster, and admission status twelve months prior to the index admission. The overall regression model was statistically significant, Χ²(31, N = 929) = 82.24, p = < .001, indicating that the predictors, taken together, reliably distinguished between participants who were readmitted and those who were not. The model explained between 8.5% (Cox & Snell R²) and 11.4% (Nagelkerke R²) of the variance in ED visits status and correctly classified 76.0% of cases. In this analysis, none of the key participant characteristics, such as study group, satisfaction level, age, gender, ethnicity, education, relationship status, resilience, wellbeing, depression, or anxiety, were significantly associated with ED visits within the following 12 months. The confidence intervals for all odds ratios crossed unity, indicating that these factors did not independently contribute to predicting ED visits 12 months post-discharge in the adjusted model. In contrast, employment and housing status emerged as significant predictors of future ED utilization. Unemployed Participants had significantly higher odds of returning to the ED (OR = 1.57, 95% CI: 1.11–2.21) compared to those who were employed. Housing instability showed a similarly strong association; individuals living in shelters, couch-surfing situations, or on the street were more than twice as likely to visit the ED (OR = 2.60, 95% CI: 1.29–5.22) relative to homeowners. Prior ED utilization stood out as the strongest predictor in the model. Participants who had an ED visit in the six months before the index admission had more than triple the odds of returning within the year following their hospital discharge (OR = 3.72, 95% CI: 1.82–7.63).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Multivariate logistic regression model predicting the likelihood of the ED visits within 12 months post-discharge among study participants.

https://doi.org/10.1371/journal.pone.0351753.t002

4. Discussion

The present observational analysis in this study examined the association between multiple factors, including intervention exposure, inpatient satisfaction, sociodemographic characteristics, prior healthcare utilization and clinical variables, and psychiatric ED visits within 12 months following discharge from psychiatric inpatient care. The findings suggest that socioeconomic and structural factors had a greater influence on post-discharge ED utilization than most clinical or demographic characteristics which demonstrated little predictive power. The multivariate model revealed that only unemployment, housing instability, and previous ED use independently predicted future ED visits within a year post-discharge, despite the univariate analysis showing variability in ED visits across age, ethnicity, employment, and housing categories. This pattern highlights the complex relationship between healthcare utilization and social vulnerability, which is in line with a large body of research highlighting the importance of social determinants in influencing mental health outcomes [36–41]. Unemployment emerged as a significant predictor of subsequent ED visits, aligning with previous research demonstrating that economic constraint is strongly associated with increased acute mental-health service use [39,42,43]. Economic instability may exacerbate psychiatric symptoms, limit access to primary or community-based services, and diminish individuals’ capacity to engage in preventive care [43,44]. Similarly, housing instability was one of the strongest predictors of ED return visits in the present study. This is consistent with international and Canadian evidence showing that individuals who are homeless or unstably housed experience disproportionately higher rates of ED visits and hospitalizations [37,45–47]. An Ontario cohort study found that psychiatric patients who were homeless at discharge had nearly double the hazard of a mental health-related ED visit within 30 days compared with non-homeless patients, suggesting that housing instability may reduce treatment continuity, intensify stressors, and limit the ability to adhere to discharge plans, thereby increasing the likelihood of crisis presentations [19]. Prior ED use was the strongest predictor of subsequent ED utilization, suggesting a pattern of recurrent service use that may reflect unresolved clinical needs, inadequate transition support, or unmet psychosocial challenges. This finding is widely supported in the literature, with past ED use consistently identified as one of the most reliable predictors of future acute-care episodes across psychiatric and general medical populations [48–51]. Such patterns reinforce the importance of targeted interventions for high-utilizing groups, potentially including intensive case management, strengthened outpatient follow-up, and enhanced community supports. The present study evaluated psychiatric ED utilization across a 12-month follow-up period to capture broader patterns of post-discharge acute-care use. However, recurrent ED presentations may be concentrated within the earlier months following discharge, which could represent a period of heightened clinical and social vulnerability. Future studies examining earlier timing and distribution of ED presentations, may help identify critical windows for targeted transitional interventions and relapse prevention strategies. Contrary to expectations, neither psychiatric diagnosis, clinical characteristics (depression, anxiety, wellbeing), resilience, nor inpatient satisfaction predicted post-discharge ED visits in the adjusted model. While patient-reported experience has been hypothesized to influence post-discharge engagement and outcomes, evidence remains mixed. Some studies have found associations between poor satisfaction and increased ED visits risk [52–54], whereas others, including a large multicenter prospective study, report no independent effect after adjusting for clinical and sociodemographic variables [55,56]. Our findings add to the latter group, suggesting that socioeconomic disadvantages may predominate the influence of individual perceptions and clinical characteristics in driving acute care utilization among psychiatric populations. This study adds to the existing literature by simultaneously examining inpatient satisfaction, intervention exposure, sociodemographic factors, prior healthcare utilization, and clinical variables as predictors of psychiatric ED visits following discharge. While factors such as unemployment, housing instability, and prior ED use have been previously identified, fewer studies have evaluated these variables alongside patient-reported inpatient satisfaction within the same adjusted model and within a real-world transitional care setting. Additionally, the nonsignificant association between inpatient satisfaction and subsequent ED visits contributes to the limited and mixed evidence regarding the independent role of patient experience in predicting acute psychiatric service utilization after discharge. However, the identified predictors reflect modifiable social conditions that may be addressed through policymakers and service interventions. For example, enhancing access to stable housing, supported employment, and coordinated transitional care services may reduce repeated ED presentations and improve continuity of care [37,57–60]. All together, these findings underscore the need for mental health systems that prioritize structural and social supports in transitional care planning. While clinical and demographic factors contribute to understanding patient needs, it is the broader social environment, particularly housing stability, economic security, and prior care patterns, that appears to govern post-discharge ED visit patterns. Interventions that address these determinants may offer more substantial and sustained reductions in acute-care utilization than those focused solely on clinical or patient-experience variables. Although this study contributes to the international literature on psychiatric emergency department utilization, its findings should be interpreted within the context of the Canadian healthcare system. Canada’s publicly funded healthcare model reduces financial barriers to accessing both inpatient and emergency psychiatric services, which may influence patterns of ED utilization compared to countries with privatized or insurance-based systems. In such settings, cost-related barriers may delay care-seeking or shift utilization toward alternative services, potentially altering the relative importance of predictors identified in this study.

5. Conclusion

The current study highlights that socioeconomic factors and prior psychiatric healthcare utilization are the primary determinants of emergency department visits following psychiatric admission. Unemployment, unstable housing, and a history of recent ED visits significantly increased the likelihood of recurrent ED visits within 12 months post-discharge. In contrast, sociodemographic and clinical characteristics, including age, gender, ethnicity, education, relationship status, psychiatric diagnosis, symptom severity, resilience, wellbeing, and inpatient satisfaction, did not independently predict ED use post-discharge from psychiatric units. These findings reveal the important role of social determinants and prior service use in shaping post-discharge ED utilization and suggest that interventions targeting economic stability, housing support, and continuity of care may be crucial in reducing recurrent ED use among psychiatric patients.

Supporting information

Acknowledgments

The data collection instrument was designed by the Decision Support Services of the Alberta Health Services Edmonton Zone Addiction and Mental Health Program. The staff of Alberta Health Services provided varied support in conducting this study.

Institutional Review Board Statement: The Health Research Ethics Board of the University of Alberta (Ref # Pro00111459) provided the required ethical clearance for this study. Additional operational approval was obtained from the regional health authority.

Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.