A Hospital-at-Home Pilot in Singapore

1. Hypothesis We hypothesise that hospital-at-home programmes may reduce cost of delivering care with comparable clinical outcomes (readmissions, mortality, and hospital-acquired complications) and positive patient and staff experiences.
2. Specific Aims

The primary objective of this study is:

1. To compare the mean direct cost of care per day of index hospitalisation for patients receiving HaH care to similar patients receiving usual inpatient care.

The secondary objectives in this study are:

1. To evaluate the clinical effectiveness, in terms of 30-day readmissions, mortality and hospital-acquired complications) of HaH compared to usual inpatient care;
2. To evaluate impact of patient-reported outcomes (EQ-5D, Care Transitions Measure and patient satisfaction scores) of HaH compared to usual inpatient care, including impact on medication management;
3. To understand the experiences of patients receiving HaH care and their caregivers; and
4. To understand the experiences of clinicians providing HaH in providing such care.

Further exploratory objectives include:

1. To identify factors that potentially influence the decisions of patients to accept HaH care, such as patient demographics, family and social setup, or perspectives of home-based care;
2. To describe the use of telehealth platforms for vital signs monitoring and teleconsultation that was used within the intervention arm; and
3. Study population

3.1 Setting

We study will recruit 147 patients in the intervention group and 147 patients in the control group from 2 sites: National University Hospital (NUH) and Alexandra Hospital (AH). Patients will be recruited from both AH and NUH with no restrictions of proportion from each center. There are no restrictions based on race of the subject. We will exclude children as this is an adult medical service.

3.2 Recruitment

3.2.1 Recruitment Process

Study personnel will screen through the electronic health record of all patients newly admitted overnight every working weekday in the following settings:

• NUH EMD Lodgers
• NUH Acute Medical Unit (AMU)
• NUH General Medicine Wards
• NUH Cardiology Wards
• AH wards under General Medicine (GM)

Patients thought to be suitable by the research assistant will be confirmed by the NUHS@Home consultant for clinical criteria, who will subsequently discuss with the primary physician regarding eligibility.

Patients who are subsequently enrolled into the NUHS@Home programme will be approached for consent to be in the hospital-at-home cohort. The NUHS@Home team will be activated. Patients will be sent home via ambulance, where they will be reviewed by a NUHS@Home nurse. Intravenous therapies, pillboxes for medication, remote monitoring software, and telecommunication tools will be set up as indicated. NUHS@Home doctors will visit at least once daily, and nurses at least twice daily as required. Therapists will conduct home visits as necessary. The NUHS@Home team is available 24/7. When conventional discharge criteria are met, the patient will be discharged. The NUHS@Home clinical team will be under clinical governance of Division of Advanced Internal Medicine at NUH and patients will be considered 'inpatients' throughout the treatment period.

Patients who are suitable for NUHS@Home programme but the programme has no bed capacity will be approached for consent to be in the control cohort. There will be no change to the patient's clinical management.

Patients who were offered but declined enrolments inot the NUHS@Home programme will be approached for consent to be in the rejected cohort. There will be no change to the patient's clinical management.

3.2.2 Consent

If the patient has capacity, the study team will go through the study information with the patient and caregiver (which will cover the process for recruitment and the randomisation process) and take written informed consent. If the patient requests, their main spokespersons will be involved in the discussion. If the patient has no capacity to consent, the main spokesperson will represent them in this scenario.

3.3 Inclusion and exclusion criteria are entered elsewhere in this protocol

3.4 Withdrawal Criteria

Patients may discontinue the intervention if they are transferred back to hospital during the treatment period.

3.5 Subject Replacement

Subjects who drop out will not be replaced.

4 Methods and Assessments

4.1 Study Visits

4.2.1 Screening Encounter

Patients which are potentially eligible would be screened both using the electronic health record and in person using a screening checklist to confirm eligibility.

4.2.2 Study Visits

Patients or their proxies who consent will undergo a baseline questionnaire. This includes baseline demographics, Barthel's score, assessment of health literacy, mini-COG, EQ-5D-5L and EQ-VAS. If a proxy is used, the same proxy will need to complete all subsequent questionnaires.

4.2.3 Final Study Visits

Patients will undergo a final questionnaire 14 days from enrolment. If a proxy was used for initial questionnaires, the same proxy will be used to complete the final questionnaire. This includes EQ-5D-5L, EQ-VAS, perception of care transitions and patient satisfaction survey. The visit will either be done in the clinic (if patients are coming for a clinic appointment) or over a phone call. Patients and caregivers in the intervention group will undergo an additional survey.

Patients in the 'rejected cohort' will not undergo this questionnaire.

4.2.4 Post Study Follow up and Procedures

30 days from discharge, the study team will collect patient outcomes from all 3 cohorts from the electronic medical record. Every 3 months, the study team will extract billing data for a list of itemised consumables and their associated cost to the patient. Patients will not be involved in this data collection.

5 Measurements

5.1 Primary and secondary outcomes are detailed elsewhere in this protocol

5.2 Baseline Characteristics

The following baseline characteristics will be measured at enrolment:

• Demographics (age, sex, ethnicity)
• Socioeconomic factors (occupation, income group, marital status, residence type, highest education level)
• Diagnosis at enrolment
• Number of hospitalisations in the last 12 months
• Home environment (who patients live with, main caregiver, housing type)
• Functional status (Barthel's Index , I-ADLs, assistance to doctor)
• NEWS score at enrolment
• Charlson Comorbidity Index at enrolment
• Clinical Frailty Score (at baseline) rated by NUHS@Home physician
• Cognitive impairment using Mini-COG
• Health Literacy using screening questions
• EQ-5D-5L

5.4 Process Evaluation

5.4.1 Process Measures

The following processes will also be measured in both groups from chart review:

• Number of appointments reduced upon discharge
• Number of referrals to community services
• Number of handovers to transitional care
• Number of patient and family education sessions done
• Number of unplanned physician encounters ("call-to-see-patient" or CTSPs), during and after office hours
• Number of days that patients received:
• Vital signs monitoring
• BSL monitoring
• Wound dressing
• Intravenous medication
• Intravenous drip
• Nebuliser therapy
• Sliding scale insulin
• Phlebotomy
• Whether patients received the following after enrolment:
• Imaging
• Endoscopy
• Interventional radiology
• Referral to medical speciality
• Referral to non-medical specialty
• Discharged with written instructions or educational material
• Referral to advanced care planning
• Follow up in 72 hours from discharge

5.4.2 Disease specific quality indicators

Disease specific quality indicators will be measured for the common diagnoses that the investigators anticipate, and where established quality indicators exist.

• For heart failure :
• ACEI/ARB or ARNI at discharge
• Evidence Specific Beta blockers at discharge
• Measurement or plan for measurement of LV function
• Post-discharge appointment for heart failure
• Adult smoking cessation advice
• Influenza and pneumococcal vaccination
• 30-day readmission for heart failure
• For COPD :
• Initiation of long term oxygen therapy if appropriate
• Identification for pulmonary rehabilitation
• Antibiotics given
• Glucocorticoids given
• Appropriate prescription of relieving bronchodilators on discharge
• Inhaler technique assessed
• 30-day readmission for COPD exacerbation
• For urinary tract infection :
• Urine culture performed
• Prescribe empirical therapy according to hospital guidelines or previous cultures
• Change to pathogen directed therapy with culture results are available
• Change from intravenous to oral therapy in 48-72h on basis of clinical condition
• Change catheter for long term in dwelling catheter
• For pneumonia :
• Adult smoking cessation advice
• Influenza and pneumococcal vaccination
• Oxygen received if hypoxic
• Conversion from intravenous to oral therapy when clinically improving, haemodynamically stable and tolerating orally

5.4.3 Intervention-specific Process Measures

The following processes will also be measured to examine fidelity of the programme:

• Response time for out-of-hour home visits
• Number of unexpected telephone calls \& staff visits
• Number of medication reviews done with patient and family
• Number of patients transferred back to acute hospital
• Number of home visits by doctor and duration of each
• Number of home visits by nurse and duration of each
• Number of home visits by allied health and duration of each
• Number of home visits by caregivers and duration of each
• Number of virtual visits by doctor and duration of each
• Number of virtual visits by nurse and duration of each
• Number of virtual visits by allied health and duration of each

5.4.4 Stakeholder Perspectives Using a Qualitative Approach

Patients who are eligible and enrolled into the NUHS@Home Programme will be recruited for this study. They will also be asked if their family member/caregiver (excluding hired domestic help) might be interested to participate in the in-depth semi-structured interviews.

Study Design

The focused ethnography approach is adopted to elicit situated experiences of patients receiving care at home and the perceptions of caregivers and care providers in regard to facilitators and barriers to this new care model. Participant observations of care practices at home and interviews with the three parties including patients, caregivers and the staff will be undertaken. Analysis will be conducted using constant comparison technique to compare incidents applicable to each theme.

Observations will be carried out at participants' home by adopting 'observer as participant' role (Gold, 1958), where observations will be favored over participation. Interview guide will be adopted from the initial qualitative interviews and will focus on the following areas: Patients experiences of receiving the care at home: Caregivers and staff's perceptions of providing the care at home; Facilitators and barriers of receiving/providing care via HaH model and specific questions in regard to the rich points of social interaction observed in the field.

The following themes will be explored from the patient and their caregivers:

• Describe their experience of being admitted to the Hospital at Home programme.
• What, if any, are the benefits they have experienced during Hospital at Home?
• What, if any, are the disadvantages and drawbacks they have experienced during Hospital at Home?
• Did this program meet their expectations? Why or why not?
• Would they choose to be transferred to Home Hospital again if needed hospitalisation in the future? Why or why not?
• How did they feel about the vital signs monitoring setup at home?
• What, if any, is their experience of video or phone calls from the doctor during hospital at home?
• What, if any, is their experience with private doctors or nurses that attended to them during their stay?
• Any suggestions for the future development of Home Hospital?

Interviews will be audio-recorded and will last approximately 30-60 minutes. The interviews will focus on understanding the experiences of patients, caregivers and care providers on their perceptions of the HaH intervention program.

The total number of interviews required will depend on data saturation, estimated to involve 20-30 patients and 20-30 caregivers.