Predicting hospital length of stay for geriatric patients with mood disorders

Influencing factors of length of stay among repeatedly hospitalized patients with mood disorders: a longitudinal study in China

BACKGROUND

Patients with mood disorders usually require repeated and prolonged hospitalization, resulting in a heavy burden on healthcare resources. This study aims to identify variables associated with length of stay(LOS) of repeatedly hospitalized patients with mood disorders and to provide information for optimizing psychiatry management and healthcare resource allocation.

METHODS

Electronic medical records (EMRs) of repeatedly hospitalized patients with mood disorders from January 2010 to December 2018 were collected and retrospectively analyzed. Chi-square and t-test were adopted to investigate the differences in characteristics between the two groups of short LOS and long LOS. Generalized estimating equation (GEE) was conducted to investigate potential factors influencing LOS.

RESULTS

A total of 2,009 repeatedly hospitalized patients with mood disorders were enrolled, of which 797 (39.7%) had a long LOS and 1,212 (60.3%) had a short LOS. Adverse effects of treatment, continuous clinical manifestation, chronic onset type, suicide attempt, comorbidity and use of antidepressants were positively associated with long LOS among all repeatedly hospitalized patients with mood disorders (P < 0.050). For patients with depression, factors associated with long LOS consisted of age, monthly income, adverse effects of treatment, continuous clinical manifestation, suicide attempt and comorbidity (P < 0.050). Whereas, for patients with bipolar disorder (BD), adverse effects of treatment, four or more hospitalizations and use of antidepressants contributed to the long LOS (P < 0.050). Influencing factors of LOS also vary among patients with different effectiveness of treatment.

CONCLUSION

The LOS in repeatedly hospitalized patients with mood disorders was influenced by multiple factors. There were discrepancies in the factors affecting LOS in patients with different diagnoses and effectiveness of treatment, and specific factors should be addressed when evaluating the LOS.

Predictors of long-stay in a tertiary care psychiatry hospital in India

BACKGROUND

Many psychiatric patients continue to stay for longer duration in psychiatric institutions. Community reintegration and rehabilitation of such patients needs to be explored to ensure appropriate bed occupancy rates and access to in-patient care for newer patients in need of the same.

OBJECTIVES

The aim is to identify the risk and protective factors leading to long-term admission of mentally ill patients in tertiary care hospital.

METHODOLOGY

A cross-sectional study was conducted with all patients in the long-stay ward between May 2018 and February 2023. A retrospective chart review followed by cross sectional assessment of risks and disability was carried out with all the patients living in a long-stay psychiatric ward (N = 101) at a tertiary hospital in Bangalore, India, between May 2018 and February 2023.

RESULTS

The hospital's average length of stay (LOS) was 5.70 ± 8.30 years. The risk and protective factors for LOS in psychiatric hospitals were analyzed using the Poisson Regression model. The results revealed that the male gender, diagnosis of schizophrenia or psychosis, clinicians' knowledge of family information, good clinical improvement, and higher participation in ward activities are the protective factors for a shorter hospital stay. While higher age, family history of mental illness, being married and employed, absence of children, and/or having family members rarely visiting the patient at the hospital were some factors that increased the risk of LOS.

CONCLUSION

This study highlighted the importance of possible LOS predictors in the tertiary care psychiatric hospital. The multi-disciplinary team may utilize risk and protective factors for delayed LOS to introduce comprehensive psychosocial interventions and policies that reduce the risk of delays or length of stay in mental health hospitals.

The prediction of hospital length of stay using unstructured data

Objective

This study aimed to assess the performance improvement for machine learning-based hospital length of stay (LOS) predictions when clinical signs written in text are accounted for and compared to the traditional approach of solely considering structured information such as age, gender and major ICD diagnosis.

Methods

This study was an observational retrospective cohort study and analyzed patient stays admitted between 1 January to 24 September 2019. For each stay, a patient was admitted through the Emergency Department (ED) and stayed for more than two days in the subsequent service. LOS was predicted using two random forest models. The first included unstructured text extracted from electronic health records (EHRs). A word-embedding algorithm based on UMLS terminology with exact matching restricted to patient-centric affirmation sentences was used to assess the EHR data. The second model was primarily based on structured data in the form of diagnoses coded from the International Classification of Disease 10th Edition (ICD-10) and triage codes (CCMU/GEMSA classifications). Variables common to both models were: age, gender, zip/postal code, LOS in the ED, recent visit flag, assigned patient ward after the ED stay and short-term ED activity. Models were trained on 80% of data and performance was evaluated by accuracy on the remaining 20% test data.

Results

The model using unstructured data had a 75.0% accuracy compared to 74.1% for the model containing structured data. The two models produced a similar prediction in 86.6% of cases. In a secondary analysis restricted to intensive care patients, the accuracy of both models was also similar (76.3% vs 75.0%).

Conclusions

LOS prediction using unstructured data had similar accuracy to using structured data and can be considered of use to accurately model LOS.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12911-021-01722-4.

Emergency hospital admissions among older adults living alone in the community

Background

Among older adults, living alone is often associated with higher risk of Emergency Department (ED) admissions. However, older adults living alone are very heterogeneous in terms of health. As more older adults choose to live independently, it remains unclear if the association between living alone and ED admissions is moderated by health status. We studied the association between living alone and ED admission outcomes (number of admissions, inpatient days and inpatient costs) among older adults with and without multimorbidity.

Methods

We used data from 16,785 individuals of the third follow-up of the Singapore Chinese Health Study, a population-based cohort of older Singapore Chinese (mean age: 73(61-96) years). Participants were interviewed face-to-face from 2014 to 2016 for sociodemographic/health factors and followed-up for one year on ED admission outcomes using Singapore Ministry of Health’s Mediclaim Database. We first applied multivariable logistic regression and two-part models to test if living alone is a risk factor for ED admission outcomes. We then ran stratified and joint effect analysis to examine if the associations between living alone and ED admission outcomes were moderated by multimorbidity.

Results

Compared to living with others, living alone was associated with higher odds of ED admission [Odds Ratio (OR) 1.28, 95 % Confidence Interval(CI) 1.08-1.51)], longer inpatient days (+0.61, 95 %CI 0.25-0.97) and higher inpatient costs (+322 USD, 95 %CI 54-591). The interaction effects of living arrangement and multimorbidity on ED admissions and inpatient costs were not statistically different, whereas the interaction between living arrangements and multimorbidity on inpatient days was borderline significant (p-value for interaction=0.050). Compared to those living with others and without multimorbidity, the relative mean increase was 1.13 inpatient days (95 %CI 0.39-1.86) for those living alone without multimorbidity, and 0.73 inpatient days ( 95 %CI 0.29-1.17) for those living alone with multimorbidity.

Conclusions

Older adults living alone were at higher risk of ED admission and higher inpatient costs regardless of multimorbidity, while those living alone without multimorbidity had the longest average inpatient days. To enable aging in place while avoiding ED admissions, interventions could provide instrumental support and regular health monitoring to older adults living alone, regardless of their health status.

Retrospective Analysis of Demographic, Psychiatric, and Physical Characteristics That Impact Length of Stay on an Inpatient Geriatric Psychiatric Unit

Currently, the aging adult population is rising fast and presenting multiple challenges for the US healthcare system. Older adults present unique challenges in their care of medical and psychiatric conditions. This study retrospectively examined characteristics that are associated with length of stay on an inpatient geriatric psychiatric unit in an urban located psychiatric hospital. A sample of 74 individuals was examined. Factors that influenced length of stay included commitment status and discharge to a different level of care. Reducing the length of stay for geriatric patients can help reduce costs and improve health outcomes.

Length of stay among multi-ethnic psychiatric inpatients in the United Kingdom

BACKGROUND

Disproportionate rates of psychiatric admissions and length of stay (LOS)¹ have been repeatedly observed among Black and Minority Ethnic (BME)² groups with severe mental illness (SMI), specifically those of Black Caribbean heritage. Commensurate understanding of the mechanisms underlying this observation remains poorly understood, warranting further attention.

METHOD

A quasi-experimental retrospective cohort design was used to examine 11,617 psychiatric inpatient admissions in London, U.K. Variations in LOS across four ethnic groups (i.e., Black Caribbean, Black African, White British, and Asian) were recorded. Multivariate logistic regression was then used to determine the extent to which demographic (age, marital status, gender, homeless at admission), clinical (primary diagnosis, substance misuse), and behavioral variables (violent behavior, deliberate self-harm, victimization) explained any observed differences across ethnic groups.

RESULTS

Univariate analyses revealed that while Black Caribbean and African minority group admissions were significantly more likely to be a longer (>28 days) than shorter (≤28 days) LOS, White British admissions demonstrated the opposite association. After adjusting for demographic, clinical, and behavioral factors, the multivariate logistic regression only partially explained the association between Black Caribbean and African admissions and longer LOS compared to White British admissions. The regression model explained approximately 18% of the observed variance in LOS.

CONCLUSIONS

These findings indicate that additional variables, such as treatment, system related, and discrimination-based factors, should be examined to better understand the complex relationship between BME groups and LOS in order to improve the experience and outcomes of these vulnerable psychiatric populations.

Prevalence of anxiety in patients with mild cognitive impairment: A systematic review and meta-analysis

BACKGROUND

Prevalence rates of anxiety in patients with mild cognitive impairment (MCI) varied widely across studies and may confer a higher likelihood of progression to dementia. Our aim was to estimate the prevalence of anxiety in MCI and identify reasons for heterogeneity in the reported results.

METHODS

A computerized search in PubMed, EMBASE, and Psyc INFO for studies on anxiety in MCI was performed up to March 2017. The overall prevalence of anxiety in patients with MCI was pooled using a random-effects model. Heterogeneity was explored using stratification (recruitment resource; method of anxiety diagnosis; method of MCI diagnosis; and region) and random-effects meta-regression.

RESULTS

Of 2494 unique abstracts, 290 were selected for full-text review, and 39 studies, representing 10,587 patients, met all inclusion criteria. The overall pooled prevalence of depression in patients with MCI was 21.0% (95% CI, 16.2-26.7) with significant heterogeneity present (I² = 97.2%, p < 0.001). When stratified by source, the prevalence of anxiety in patients with MCI in community-based samples was 14.3% (95% CI, 9.7-20.5) and was 31.2% (95% CI, 23.6-40.0) in clinic-based samples, which was significantly different (p < 0.01). The prevalence of anxiety estimates also differed significantly, when stratification by the method of anxiety diagnosis (p < 0.01). However, the criteria used for MCI diagnosis and geographical region did not significantly influence the prevalence estimate.

LIMITATIONS

We did not consider the usage of anti-anxiety drugs and language bias.

CONCLUSIONS

The prevalence of anxiety in patients with MCI was common and variability in prevalence of anxiety across studies can be attributed partly to the source of the sample and method of anxiety diagnosis. Further research is needed to identify sources of heterogeneity.

Socio-demographic characteristics and factors associated with hospitalization in psychiatry of old age patients: an international comparison between Ireland and Turkey

BACKGROUND

Taking predictors of hospitalization characteristics into consideration internationally would broaden our understanding of this population on a local basis.

AIMS

We aimed to examine and compare socio-demographic profiles along with hospitalization characteristics including length of hospital stay (LOS), reasons for admission and diagnoses among older adult inpatients hospitalized in Ireland and Turkey, and to assess factors predicting these features.

METHODS

The admission charts of 356 psychiatric inpatients over 65 years of age who were admitted to two different acute psychiatric hospitals (Sligo/Ireland and Istanbul/Turkey) were analysed by means of descriptive modalities and logistic regression.

RESULTS

There were significant differences in several domains of socio-demographics, reasons of admission and diagnoses. LOS was significantly longer in Ireland. Living alone was the only significant predictor for longer LOS in both countries, whereas in addition to living alone, younger age was also a contributor for longer LOS in Turkey.

DISCUSSION

Given that the only factor predicting LOS both in Turkey and Ireland was living alone, helping to identify more acceptable ways of providing social support for living arrangements constitutes an important service to shorten LOS in old age psychiatric population.

CONCLUSIONS

It is possible to infer that independent from the cultural diversities, living arrangement is a consistent entity to influence length of hospital stay in older adult population.

Validity and reliability of the Patient Centred Assessment Method for patient complexity and relationship with hospital length of stay: a prospective cohort study

OBJECTIVES

Several instruments for evaluating patient complexity have been developed from a biopsychosocial perspective. Although relationships between the results obtained by these instruments and the length of stay in hospital have been examined, many instruments are complicated and not easy to use. The Patient Centred Assessment Method (PCAM) is a candidate for practical use. This study aimed to test the validity and reliability of the PCAM and examine the correlations between length of hospital stay and PCAM scores in a regional secondary care hospital in Japan.

DESIGN

Prospective cohort study.

PARTICIPANTS AND SETTING

Two hundred and one patients admitted to Ouji Coop Hospital between July 2014 and September 2014.

MAIN PREDICTOR

PCAM total score in initial phase of hospital admission.

MAIN OUTCOME

Length of stay in hospital.

RESULTS

Among 201 patients (Female/Male=98/103) with mean (SD) age of 77.4±11.9 years, the mean PCAM score was 25±7.3 and mean (SD) length of stay in hospital (LOS) 34.1±40.9 days. Using exploratory factor analysis to examine construct validity, PCAM evidently has a two-factor structure, comprising medicine-oriented and patient-oriented complexity. The Spearman rank correlation coefficient for evaluating criterion-based validity between PCAM and INTERMED was 0.90. For reliability, Cronbach's alpha was 0.85. According to negative binomial regression analyses, PCAM scores are a statistically significant predictor (p<0.001) of LOS after adjusting for age, gender, Mini Nutritional Assessment Short-Form, Charlson Comorbidity Index, serum sodium concentration, total number of medications and whether public assistance was required. In another model, each factor in PCAM was independently correlated with length of stay in hospital after adjustment (medicine-oriented complexity: p=0.001, patient-oriented complexity: p=0.014).

CONCLUSION

PCAM is a reliable and valid measurement of patient complexity and PCAM scores have a significant correlation with hospital length of stay.

Moderate efficiency of clinicians' predictions decreased for blurred clinical conditions and benefits from the use of BRASS index. A longitudinal study on geriatric patients' outcomes

OBJECTIVES

Accurate prognosis is an essential aspect of good clinical practice and efficient health services, particularly for chronic and disabling diseases, as in geriatric populations. This study aims to examine the accuracy of clinical prognostic predictions and to devise prediction models combining clinical variables and clinicians' prognosis for a geriatric patient sample.

STUDY DESIGN AND SETTING

In a sample of 329 consecutive older patients admitted to 10 geriatric units, we evaluated the accuracy of clinicians' prognosis regarding three outcomes at discharge: global functioning, length of stay (LoS) in hospital, and destination at discharge (DD). A comprehensive set of sociodemographic, clinical, and treatment-related information were also collected.

RESULTS

Moderate predictive performance was found for all three outcomes: area under receiver operating characteristic curve of 0.79 and 0.78 for functioning and LoS, respectively, and moderate concordance, Cohen's K = 0.45, between predicted and observed DD. Predictive models found the Blaylock Risk Assessment Screening Score together with clinicians' judgment relevant to improve predictions for all outcomes (absolute improvement in adjusted and pseudo-R(2) up to 19%).

CONCLUSION

Although the clinicians' estimates were important factors in predicting global functioning, LoS, and DD, more research is needed regarding both methodological aspects and clinical measurements, to improve prognostic clinical indices.

Predictors in geriatric psychiatry hospital length of stay

This paper examined predictors of length of stay in a freestanding geriatric psychiatry hospital. Data on patient and treatment characteristics of geriatric inpatients (N = 1,593) were extracted from an archival administrative tracking database from Mary Starke Geriatric Harper Center. Five independent variables (length of time between last discharge and most recent admission, number of previous admissions, number of assaults, co-morbid medical condition, and admitting psychiatric diagnosis) were entered into a hierarchical regression model as potential predictors of length of stay in a geriatric psychiatry hospital. Number of assaults committed by the patient was the only significant predictor of length of stay, such that patients that had a greater number of assaults were more likely to have longer lengths of stay than those with fewer assaults. These findings highlight the importance of identifying patients at risk for assaultive behavior and developing effective interventions for aggression in geriatric psychiatry hospitals.

Length of stay of psychiatric admissions in a general hospital in Ethiopia: a retrospective study

BACKGROUND

In sub-Saharan Africa, the number of psychiatric beds per population is disproportionately low. Moreover, there is a lack of data regarding the patterns of psychiatric admissions and the factors leading to long psychiatric hospitalization in this region. This study aimed to investigate the average length of stay (LOS) and the factors associated with prolonged hospitalizations.

METHODS

A ten-year retrospective chart review of patients admitted to the psychiatric facility of Jimma University Specialized Hospital in southwest Ethiopia was conducted. The medical charts of 846 admissions spanning the period from January 2001 to December 2010 were reviewed. LOS greater than 21 days was considered as a cut-off point for lengthier stay. Bivariate and multivariable logistic regression analyses were conducted to identify factors independently associated with LOS.

RESULTS

The most common discharge diagnoses were schizophrenia and other psychotic disorders (27.6%), and bipolar disorder (23.4%). A global clinical rating taken on discharge showed 90.3% improved outcome. The median (25th, 75th percentiles) LOS was 22 (15, 36) days. Patients with major depressive disorder [aOR = 0.51 (0.32 - 0.81)] and brief psychotic disorder [aOR = 0.52 (0.33 - 0.84)] were less likely than patients with schizophrenia and other psychotic disorders to have long hospital stays. Presence of extrapyramidal side-effects and out of pocket expenditures predicted LOS.

CONCLUSIONS

Patients with psychoses and bipolar disorder have lengthier hospital stays burdening the cost of care of psychiatric treatment in a general hospital setting. Our findings call for identifying those cases quickly, attending to their needs with evidence-based efficient treatment and for improving and developing an aftercare system such that the utilization of acute inpatient beds, already a scarce resource, could achieve higher efficiency.

Predictors of length of stay in a psychiatric hospital

OBJECTIVE

This article reviews the available evidence for predicting length of stay in a psychiatric hospital based on several patient characteristics.

METHODS

An OVID Medline search was done for studies in this area with terms: length of stay, acute psychiatric hospital and predictors. Indexed articles were searched for additional references and subsequently cross referenced. The variables described in various studies were classified into demographic, diagnostic, treatment related and system-based variables ( Table 1 ).

RESULTS

Substance abuse has been consistently shown to be associated with a shorter length of stay. Mood and psychotic disorders have shown to lengthen the stay in an acute psychiatric hospital. Women tend to stay longer in an acute psychiatric hospital. Other demographic variables like age and education have had equivocal results, while a married marital status and employment are consistently related to a shorter length of stay. Use of restraints and ECT during hospitalization has shown to lengthen the stay in the hospital.

CONCLUSIONS

Many studies have explored the relationship between the demographics, diagnosis and treatment variables on length of stay in an acute psychiatric hospital. These studies have been able to account for about 10-37% of the variance in the length of stay. The relevant findings impacting the length of stay are shown in Table 2 .

Associations of medical comorbidity, psychosis, pain, and capacity with psychiatric hospital length of stay in geriatric inpatients with and without dementia

ABSTRACT Background: Geriatric psychiatry hospital beds are a limited resource. Our aim was to determine predictors of hospital length of stay (LOS) for geriatric patients with dementia admitted to inpatient psychiatric beds. Methods: Admission and discharge data from a large urban mental health center, from 2005 to 2010 inclusive, were retrospectively analyzed. Using the resident assessment instrument - mental health (RAI-MH), an assessment that is used to collect demographic and clinical information within 72 hours of hospital admission, 169 geriatric patients with dementia were compared with 308 geriatric patients without dementia. Predictors of hospital LOS were determined using a series of general linear models. Results: A diagnosis of dementia did not predict a longer LOS in this geriatric psychiatry inpatient population. The presence of multiple medical co-morbidities had an inverse relationship to length of hospital LOS - a greater number of co-morbidities predicted a shorter hospital LOS in the group of geriatric patients who had dementia compared to the without dementia study group. The presence of incapacity and positive psychotic symptoms predicted longer hospital LOS, irrespective of admission group (patients with dementia compared with those without). Conversely, pain on admission predicted shorter hospital LOS. Conclusions: Specific clinical characteristics generally determined at the time of admission are predictive of hospital LOS in geriatric psychiatry inpatients. Addressing these factors early on during admission and in the community may result in shorter hospital LOS and more optimal use of resources.

Predictors of length of hospital stay for patients with late-onset depression in Japan

BACKGROUND

The length of hospital stay (LOHS) for elderly patients tends to be long. We aimed to identify factors related to long hospitalization periods for elderly Japanese patients with late-onset depression.

METHODS

We retrospectively examined 133 patients with late-onset depression. Univariate and multivariate analyses were performed to confirm the relationship between each clinical feature and the duration of hospital stay.

RESULTS

The mean LOHS was 83.9 ± 55.6 days. On the basis of univariate analysis, we found that living as a housewife, death of a close relative or friend, recurrent depression, melancholic features, and treatment with electroconvulsive therapy, tricyclic or tetracyclic antidepressants or mood stabilizer were associated with a longer LOHS. Multivariate analysis showed that treatment with tricyclic or tetracyclic antidepressants, atypical antipsychotics were associated with prolonged LOHS.

CONCLUSIONS

These results suggest that job status, changes in household circumstances and the failure of initial treatment is responsible for the long LOHS in Japan.

Depression and anxiety among high-risk obstetric inpatients

OBJECTIVE

To assess the following among women hospitalized antenatally due to high-risk pregnancies: (1) rates of depression symptoms and anxiety symptoms, (2) changes in depression symptoms and anxiety symptoms and, (3) rates of mental health treatment.

METHODS

Sixty-two participants hospitalized for high-risk obstetrical complications completed the Edinburgh Postnatal Depression Scale (EPDS), Generalized Anxiety Disorder 7-item scale (GAD-7) and Short-Form 12 weekly until delivery or discharge, and once postpartum.

RESULTS

Average length of total hospital stay was 8.3 ± 7.6 days for women who completed an initial admission survey (n = 62) and 16.3 ± 8.9 (n = 34), 25.4 ± 10.2 (n = 17) and 35 ± 10.9 days (n = 9) for those who completed 2, 3 and 4 surveys, respectively. EPDS was ≥ 10 in 27% (n=17) and GAD-7 was ≥ 10 in 13% (n = 8) of participants at initial survey. Mean anxiety (4.2 ± 6.5 vs. 5.2 ± 5.1, p = .011) and depression (4.4 ± 5.6 vs. 6.9 ± 4.8, p = .011) scores were lower postpartum compared to initial survey. Past mental health diagnosis predicted depression symptoms [odds ratio (OR) = 4.54; 95% confidence interval (CI) 1.91-7.17] and anxiety symptoms (OR = 5.95; 95% CI 3.04-8.86) at initial survey; however, 21% (n = 10) with no diagnostic history had EPDS ≥ 10. Five percent (n = 3) received mental health treatment during pregnancy.

CONCLUSION

Hospitalized high-risk obstetrical patients may commonly experience depression symptoms and/or anxiety symptoms and not receive treatment. A history of mental health treatment or diagnosis was associated with depression symptoms or anxiety symptoms in pregnancy. Of women with an EPDS ≥ 10, > 50% did not report a past mental health diagnosis.

Depression and anxiety in Singaporean high-risk pregnancies - prevalence and screening

OBJECTIVE

Data on psychiatric morbidity in high-risk pregnant Singaporean women are limited. This study aimed to establish the prevalence of antenatal depression and anxiety in high-risk pregnancies, compare the prevalence of antenatal depression in high-risk pregnancies vs. pregnancies of unspecified obstetric risk and examine the Edinburgh Postnatal Depression Scale (EPDS) and State Trait Anxiety Inventory (STAI) as screening tools for these disorders.

METHOD

Two hundred high-risk pregnant inpatients at a national public maternity hospital were included. Three psychometric assessment tools were used to evaluate all participants: the diagnostic Mini International Neuropsychiatric Interview and the screening EPDS and STAI.

RESULTS

Rates of major depression, minor depression, anxiety disorder (agoraphobia, generalized anxiety disorder, panic disorder), and comorbid depression and anxiety were 11%, 7%, 12.5% and 5%, respectively. Major depression was more prevalent in high-risk pregnancies than in the historical cohort of unspecified obstetric risk (11% versus 4.3%). EPDS (cutoff 8/9) screens well for depression and anxiety in high-risk pregnancies (area under the receiver operating characteristic curve=0.82-0.87).

CONCLUSION

Antenatal depression and anxiety are highly prevalent in a sample of high-risk pregnant Singaporean women. EPDS performs well in screening for depression and anxiety in high-risk pregnant women, with further psychiatric assessment recommended for women with score ≥ 9.