PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • Study Setting and Trial Methodology: Singapore is a multi-ethnic Asian population of 5.6 M people. Its population is one of the most rapidly ageing in Asia with an increasing chronic disease burden [18]. Care is fragmented and episodic and the lack of integration between the well-developed tertiary hospitals and less-developed primary and community care sectors is well acknowledged. 30-day readmission rates in the elderly are high at 19.0% [19] and only slightly lower than in the United States [20]. Healthcare expenditure is expected to triple from S$4 billion in 2011 to S$12 billion in 2020 [21] and driven mainly by inpatient cost. Concerns about long term sustainability of current trends is driving a concerted search for new models of care delivery which reduces the dependence on high cost hospitals care. There is great interest in developing transitional care programs to improve patient outcomes and reduce wasteful utilization of expensive hospital resources. Taking a population health approach, Singapore created six regional health systems (RHSs) in 2011, each being responsible to integrate care for a specific geographic region. Singapore General Hospital (SGH) is the largest tertiary hospital with 1597 beds, accounting for about 20% of the total public acute hospital beds in Singapore. SGH is the flagship hospital of the SingHealth regional health system. It has 36 specialist departments, including family medicine. SGH is wholly owned by the Ministry of Health Singapore. With a workforce numbering above 10,000, SGH admits over 1 million patients every year at its wards, emergency department and outpatient specialist clinics [22]. Since 2006, the hospital embarked on developing new models of care integration that will reduce care fragmentation and improve care continuity. As part of the hospital’s effort to innovate and deliver better transitional care models, we incorporated the IPU framework (Table 1) to a virtual ward care model to achieve closer administrative and physical integration of our multidisciplinary team that is responsible for care transitions. Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0168757.t001 10 Attributes of Integrated Practice Units (IPUs) as applied to our modified Virtual Ward Model. We conducted an open label, outcome assessor blinded, parallel-group randomized trial comparing the modified virtual ward model of care with usual care from October 2013 to December 2014 at the Singapore General Hospital. Inclusion Criteria and Exclusion Criteria: Patients were eligible if they were aged 21 years or older, being admitted to participating medical wards (general internal medicine, endocrinology, respiratory medicine, renal medicine, gastroenterology, neurology), at high risk of readmission (as determined by LACE score ≥ 10 [length of stay, acuity of the admission, charlson comorbidity score, and emergency department visits in the previous 6 months] score ≥10 [22] and at least ≥1 admission in the previous 90 days). The LACE index (Length of stay, Acuity of admission, Charlson comorbidity index, Emergency department visits in past 6 months) was a readmission predictive score derived in Ontario, Canada with a score ranging from 0 to 19. Patients with a score of ≥ 10 were found to have a five times higher risk of 30-day readmission in Singapore [22]. When a LACE score ≥ 10 was combined with one or admissions in the previous 90 days, patients had a 30% risk of an unplanned 30-day readmission [11]. Eligible case finding and risk stratification using the LACE index and prior admissions in the previous 90 days is performed electronically by the hospital’s enterprise analytics platform eHIntS (Electronic Health Intelligence System) on a daily basis. Potentially eligible patients were screened in the wards for further exclusion criteria by a research coordinator. Patients were excluded if they were critically ill at the time of screening: hemodynamically unstable as defined by requiring inotropes; acute respiratory support with ventilators or FiO2 more than 50%; acute dialysis support or care in the high dependency or intensive care unit. Other exclusion criteria are requiring surgical intervention at the time of screening; discharged or deceased before they could be screened for eligibility; non-resident; no telephone contact or local home address for post-discharge surveillance; impaired decision making capacity without a legal surrogate decision maker to provide consent; or if they did not wish to participate. Finally, patients who are admitted from a long term care facility were excluded as it was not possible to deliver our post-discharge interventions to this group of patients. Patients who met both inclusion and exclusion criteria were approached for written informed consent. For patients unable to provide consent, it was obtained from a legal substitute decision maker. Consented patients were randomized in a 1:1 ratio and blocks of six using a computer-generated randomization list to either our modified virtual ward model of care or standard care. Allocation concealment was effected via an off-site telephone service maintained by a hospital administrator. The generation of the random allocation sequence and enrolment of patients was performed by a research coordinator. Given the nature of the intervention, it was not possible to blind patients or clinicians. However, the outcome assessor and statistician were blinded to the treatment assignment. Ethics Approval and Trial Registration: Research ethics approval was granted by the Institutional Review Board of Singapore Health Services. Our trial was not registered prospectively due to a genuine administrative error by our administrator (Clinicaltrials.gov, no NCT02351648) and it was never our intention to publish only if we had positive findings. We had also published our earlier trial which did not have a significant result [11]. We noticed the error late and immediately registered our trial upon discovery of our error. The ethics committee approved the study on 11th October 2012. We recruited our first patient on the 24th October 2012 and completed the follow up for our last patient on the 6th January 2015. The authors confirm that all ongoing and related trials for this drug/intervention are registered. Trial Intervention, Control and Outcomes: 1. Integrated Practice Unit (IPU) Line-up and Roles and Responsibilities of team members: The intervention consisted of a multidisciplinary team of integrated care nurses, pharmacists, medical social workers organized into an IPU led by attending family physicians. The IPU comprised of an inpatient care team and an outpatient virtual ward (VW) team. The IPU is co-located in the same physical locality; share a common electronic patient record and a common mission to reduce avoidable readmissions (Table 1). An attending family physician, medical officer, a nurse case manager, and a part-time (0.1 Full Time Equivalent) pharmacist formed the inpatient team while the outpatient VW team comprised of an attending family physician and two nurse case managers. A medical social worker who spends 0.5 full time equivalent time in the team supports both the inpatient and outpatient teams in activating the appropriate community and social services that supports the patient in the home setting. Family physicians in SGH function as hospitalists providing general medical care to medical inpatients [23], including follow-up care of patients in the outpatient specialist clinic upon hospital discharge. The nurse case managers’ main function is to provide discharge planning and case management to patients who are admitted in the hospital or in the virtual wards. Patient education and activation are their part of their core activities. In addition they can provide short term nursing care when required. Pharmacists assist in medication reconciliation and advise on potentially harmful drug adverse events and interactions. The physiotherapist and occupational therapist assess and train patients and their caregivers in mobility and activities of daily living respectively. The speech therapist assess and manage swallowing and communication disorders including training to patients and their caregivers to minimize aspiration risk. 2. Pre-discharge Hospital phase: Upon assignment to the intervention group, patients had their hospital care transferred to the inpatient care team of the IPU. In addition to providing general medical care, the team systematically identified and addressed patient risk factors for readmission and focused on intensive discharge planning. Medications were reconciled and the appropriate community and social services necessary to support the patient’s transition to home were activated. A core component is patient education and coaching. The nurse case managers used standardized action plans for chronic diseases (examples include congestive cardiac failure, diabetes mellitus, asthma, chronic obstructive pulmonary disease, chronic kidney disease) for patient education. Other comorbid medical conditions including mental health conditions were managed holistically, including ensuring appropriate specialist follow up on discharge. Finally, an individualized care plan complete with written discharge instructions, patients’ appointments, medication changes and the contact information of the outpatient nurse case manager is provided to all patients on discharge. 3. Post-discharge Outpatient Phase: On the day of discharge, the inpatient nurse case manager introduces the patient to the outpatient nurse case manager and hands over care of the patient to the outpatient team. The outpatient nurse case manager follows up with a telephone call within 72 hours of discharge to assess the patient’s condition and ensure adherence to the prescribed care plans and successful activation of community services. A home assessment is performed by the outpatient nurse case manager within one week of discharge. During the home assessment, the nurse case manager assessed the patient’s medical condition and health literacy, competency of the care-giver, availability of nursing and home care equipment, adequacy of social support, safety of the home environment and adherence to medication. The nurse case manager addresses any identified areas of deficiency, with help from the multidisciplinary team. Referrals were made to activate social and financial support services when needed. Each nurse was responsible for an average of 30 patients at any time during the program. A multidisciplinary team meeting was conducted in the morning of every working day. New patients are brought up for discussion and care plans were reviewed. During the meeting, each nurse case manager will also report on the status of patients under her care. Patients with urgent problems are referred to the early review clinic. Patients who are doing well are reviewed less frequently at the team meetings. Throughout the intervention period, the nurse case managers are contactable during office hours. Scheduled calls to check in on the patients are made about once a week Patients were discharged at the end of the intervention period of three months to a primary care provider in the community with continuing specialist input as warranted. Patients in the control group received standard hospital care. On discharge, patients may be referred to primary care provider, specialists in the outpatient clinic and ambulatory community services as considered necessary by the medical team. Patients receive an abbreviated standardized patient copy of the hospital discharge summary listing their medical diagnoses and medications. For this study, there was no contact between patients in the control group and the study team throughout the 3-month interval. A scheduled telephone call was made at the end of 3 months when patients or their caregivers were assessed for various outcomes. The primary outcome was the unplanned readmission rate to any hospital within 30 days of discharge. An unplanned readmission is defined as an emergency admission from the emergency department or outpatient specialist clinic. Secondary outcomes included the unplanned readmission rate to any hospital within 90 days, 180 days of discharge, as well as emergency department (ED) attendance rate within 30 days, 90 days, 180 days of discharge and the probability without readmission or death up to 180 days. Index admission and post-discharge mortality rate at 90-day; Length of stay and number of outpatient specialist clinic visits at 90-day and 180-day were compared between the two treatment groups. For patients who died during the index admission, the count for the primary and secondary outcomes would be zero. For patients who died during the 90 days follow up, the count for primary and secondary outcomes would censored at the time of death. In addition, we performed cost analysis on the costs of intervention and hospital services utilization in both intervention and control groups. The cost of intervention is computed from the costs of additional components in the intervention, absent in standard hospital care. These included the additional manpower for the post-discharge virtual ward phase and the additional nurse case manager, social worker and pharmacist support for the hospital phase. We did not include physician costs for the hospital phase in both the intervention and control groups as they are primarily employed by the hospital to care for inpatients. We assumed that the hospitalization costs fully accounted for the salaries of physicians in both intervention and control groups. We estimated the potential cost savings to the patient from the difference in hospital bed days, emergency department attendances saved against the additional costs of outpatient specialist clinic visits at 90 days post-discharge. Outpatient specialist clinic visit cost, Emergency department visit cost and average bed cost per patient day in 2013 were S$75, S$216 and S$1075 respectively [24, 25]. Outcome data were objectively retrieved from the hospital’s electronic data repository and the National Electronic Health Records (NEHR) by an outcome assessor blinded to the group assignment. The hospital’s electronic data repository and the NEHR captured all admissions and ED attendances to SingHealth hospitals and all hospitals in Singapore respectively, and is available for all patients including those who were uncontactable, or transferred to specialist discipline during index admission, or transferred to long term care facility. We hypothesized that our integrated practice unit and modified virtual ward model could reduce readmissions by approximately 25% for patients enrolled in the intervention group. We estimated the baseline readmission rate of eligible patients to be 0.4 per patient within 30 days of discharge. Power was calculated at 80%, level of significance at 5%, and attrition at 10%, resulting in a required a sample size of 420 in each group. The statistical analysis was performed on an intention-to-treat basis. The baseline characteristics were presented as mean ± SD for continuous variables and frequency counts and percentages for categorical variables. A negative binomial regression model was used to obtain incidence rate ratios (IRRs) of readmission, ED visits and outpatient specialist clinic visits. A zero-inflated negative binomial regression model is used if there is a statistically significant excess-zero problem for any of the readmission, ED visit or outpatient specialist clinic visit outcomes. An exponential hurdle regression model was used to obtain the IRRs of length of stay within different pre-defined time periods. Kaplan-Meier method was used to study the survival distribution (based on time to readmission or death) of the intervention and control groups from the date of index discharge (day 1) to 180 days post-discharge. The hazard ratio on time to readmission or death is calculated using the Cox proportional hazard model. If patients did not experience the event (readmission or death) by the end of the follow-up period, they were censored at the end of the time period. All tests of significance used 95% level (p<0.05). All analyses were performed using STATA 14.0.
rdf:type