Frequently Hospitalized Patients' Perceptions of Factors Contributing to High Hospital Use

A hybrid machine learning framework to improve prediction of all-cause rehospitalization among elderly patients in Hong Kong

BACKGROUND

Accurately estimating elderly patients' rehospitalisation risk benefits clinical decisions and service planning. However, research in rehospitalisation and repeated hospitalisation yielded only models with modest performance, and the model performance deteriorates rapidly as the prediction timeframe expands beyond 28 days and for older participants.

METHODS

A temporal zero-inflated Poisson (tZIP) regression model was developed and validated retrospectively and prospectively. The data of the electronic health records (EHRs) contain cohorts (aged 60+) in a major public hospital in Hong Kong. Two temporal offset functions accounted for the associations between exposure time and parameters corresponding to the zero-inflated logistic component and the Poisson distribution's expected count. tZIP was externally validated with a retrospective cohort's rehospitalisation events up to 12 months after the discharge date. Subsequently, tZIP was validated prospectively after piloting its implementation at the study hospital. Patients discharged within the pilot period were tagged, and the proposed model's prediction of their rehospitalisation was verified monthly. Using a hybrid machine learning (ML) approach, the tZIP-based risk estimator's marginal effect on 28-day rehospitalisation was further validated, competing with other factors representing different post-acute and clinical statuses.

RESULTS

The tZIP prediction of rehospitalisation from 28 days to 365 days was achieved at above 80% discrimination accuracy retrospectively and prospectively in two out-of-sample cohorts. With a large margin, it outperformed the Cox proportional and linear models built with the same predictors. The hybrid ML revealed that the risk estimator's contribution to 28-day rehospitalisation outweighed other features relevant to service utilisation and clinical status.

CONCLUSIONS

A novel rehospitalisation risk model was introduced, and its risk estimators, whose importance outweighed all other factors of diverse post-acute care and clinical conditions, were derived. The proposed approach relies on four easily accessible variables easily extracted from EHR. Thus, clinicians could visualise patients' rehospitalisation risk from 28 days to 365 days after discharge and screen high-risk older patients for follow-up care at the proper time.

An Intensive Intervention to Reduce Readmissions for Frequently Hospitalized Patients: the CHAMP Randomized Controlled Trial

BACKGROUND

A small number of patients are disproportionally readmitted to hospitals. The Complex High Admission Management Program (CHAMP) was established as a multidisciplinary program to improve continuity of care and reduce readmissions for frequently hospitalized patients.

OBJECTIVE

To compare hospital utilization metrics among patients enrolled in CHAMP and usual care.

DESIGN

Pragmatic randomized controlled trial.

PARTICIPANTS

Inclusion criteria were as follows: 3 or more, 30-day inpatient readmissions in the previous year; or 2 inpatient readmissions plus either a referral or 3 observation admissions in previous 6 months.

INTERVENTIONS

Patients randomized to CHAMP were managed by an interdisciplinary team including social work, physicians, and pharmacists. The CHAMP team used comprehensive care planning and inpatient, outpatient, and community visits to address both medical and social needs. Control patients were randomized to usual care and contacted 18 months after initial identification if still eligible.

MAIN MEASURES

Primary outcome was number of 30-day inpatient readmissions 180 days following enrollment. Secondary outcomes were number of hospital admissions, total hospital days, emergency department visits, and outpatient clinic visits 180 days after enrollment.

KEY RESULTS

There were 75 patients enrolled in CHAMP, 76 in control. Groups were similar in demographic characteristics and baseline readmissions. At 180 days following enrollment, CHAMP patients had more inpatient 30-day readmissions [CHAMP incidence rate 1.3 (95% CI 0.9-1.8) vs. control 0.8 (95% CI 0.5-1.1), p=0.04], though both groups had fewer readmissions compared to 180 days prior to enrollment. We found no differences in secondary outcomes.

CONCLUSIONS

Frequently hospitalized patients experienced reductions in utilization over time. Though most outcomes showed no difference, CHAMP was associated with higher readmissions compared to a control group, possibly due to consolidation of care at a single hospital. Future research should seek to identify subsets of patients with persistently high utilization for whom tailored interventions may be beneficial.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT03097640; https://clinicaltrials.gov/ct2/show/NCT03097640.

Patients' perceptions of frequent hospital admissions: a qualitative interview study with older people above 65 years of age

BACKGROUND

Although 'frequent flyer' hospital admissions represent barely 3 to 8% of the total patient population in a hospital, they are responsible for a disproportionately high percentage (12 to 28%) of all admissions. Moreover, hospital admissions are an important contributor to health care costs and overpopulation in various hospitals. The aim of this research is to obtain a deeper insight into the phenomenon of frequent flyer hospital admissions. Our objectives were to understand the patients' perspectives on the cause of their frequent hospital admissions and to identify the perceived consequences of the frequent flyer status.

METHODS

This qualitative study took place at the University Hospital of Leuven. The COREQ guidelines were followed to provide rigor to the study. Patients were included when they had at least four overnight admissions in the past 12 months, an age above 65 years and hospital admission at the time of the study. Data were collected via semi-structured interviews and encoded in NVivo.

RESULTS

Thirteen interviews were collected. A total of 17 perceived causes for frequent hospital admission were identified, which could be divided into the following six themes: patient, drugs, primary care, secondary care, home and family. Most of the causes were preventable or modifiable. The perceived consequences of being a frequent flyer were divided into the following six themes: body, daily life functioning, social participation, mental status and spiritual dimension. Negative experiences were linked to frequent flying and could be situated mainly in the categories of social participation, mental status and spiritual dimensions.

CONCLUSIONS

Frequent hospital admissions may be conceived as an indicator, i.e., a 'red flag', of patients' situations characterized by physical, mental, spiritual and social deprivation in their home situation.