Patient experiences and outcomes in a South Australian stand-alone Hospital in the Home program
Timothy J. Schultz
A * , Candice Oster B , Aubyn Pincombe A , Andrew Partington A , Alan Taylor A , Jodi Gray A , Alicia Murray C , Jennifer McInnes C , Cassandra Ryan C and Jonathan Karnon A
A
B
C
Abstract
This study aimed to compare clinical outcomes for patients admitted to Hospital in the Home (HITH) and traditional (bricks-and-mortar) hospitals and explore patient and carer experiences.
A mixed methods approach including triangulation of quantitative and qualitative data was used. Quantitative outcomes were compared using augmented inverse propensity weighting to adjust for differences in patient characteristics between groups. Qualitative data was collected by focus groups and interviews and analysed using reflexive thematic analysis. The study took place in metropolitan Adelaide and one adjacent regional health network in 2020–22. Participants were patients discharged from either hospital setting with 1 of 22 eligible diagnoses. Hospital administrative data informed a comparison of outcomes that included mortality, rate of emergency department re-presentations and re-admissions, length of stay and incidence of complications.
Patients treated in HITH were less unwell than traditional hospital patients. There were no safety or quality concerns identified in the clinical outcomes. Of 2095 HITH patients, the in-patient mortality rate was 0.2%, and 2.3% experienced a return to a bricks-and-mortar hospital during the HITH admission. For HITH patients, the mortality rate after 30 days was lower (−1.3%, 95% CI −2 to −0.5, P = 0.002), as were re-presentations in 28 days (−7.2%, 95% CI −9.5 to −5, P < 0.0001), re-admissions in 28 days (−4.9%, 95% CI −6.7 to −3.2, P < 0.001) and complications (−0.6%, 95% CI −0.8 to −0.5, P < 0.001). Interviews of 35 patients and six carers found that HITH was highly accepted and preferred by patients. HITH was perceived to free up resources for other, more acutely unwell patients.
HITH was preferred by patients and at least as effective in delivering quality health care as a traditional hospital, although the potential for unobserved confounding must be acknowledged.
Keywords: acceptability of health care, health services research, hospital, hospital-at-home, Hospital in the Home, model of care, quality and safety, routinely collected health data.
Introduction
Hospital in the Home (HITH) services provide ‘active treatment by healthcare professionals in the patient’s home for a condition that otherwise would require acute hospital inpatient care, and always for a limited time period’.1 The aim is to avoid admission to (i.e. admission avoidance, AA), or shorten the length of stay (LOS) in a traditional ‘bricks-and-mortar’ hospital (i.e. early supported discharge, ESD) through substitution of the treatment location.
Results from systematic reviews support the effectiveness and safety of HITH programs.2–5 There is a consensus that HITH programs are a viable substitute to a traditional hospital stay for patients and are as effective in improving patient outcomes.1–3,6–8
High levels of patient satisfaction are widely reported in HITH programs,4,6 which are generally preferred to traditional hospitals by patients and carers.9,10 In addition to providing more comfortable and patient-centred care, HITH programs can improve the continuity of care during and beyond the HITH hospitalisation by enhancing patient education and integrating acute care into long-term care management strategies.11 However, concerns for the safety of patients in the absence of 24/7 patient supervision is a recurring theme among stakeholders.11
HITH services are typically provided as an extension or outreach of a hospital’s own services. A South Australian HITH service, My Home Hospital (MyHH) commenced as a joint venture between private healthcare providers Calvary and Medibank in January 2021.12 Patients are referred into MyHH from a range of services (e.g. general practice, ambulance, emergency departments (EDs), hospital wards and residential aged care) in metropolitan Adelaide and some peri-urban areas. Doctors conduct daily telehealth visits, and nurses, clinical paramedics and allied health practitioners visit in-home to deliver care and support the use of telemonitoring. No face-to-face clinical care is provided after 10 pm. There is a 24/7 call centre that is staffed by a nurse, and an on-call medical officer is available for advice. For urgent care, patients are advised to call an ambulance or present to an ED. As a new stand-alone service operating with its own clinical governance and with a multitude of referring organisations, appropriate referrals into and out of MyHH are critical to maintain patient safety and continuity of care. However, we are not aware of any evaluations of stand-alone HITH services. We therefore conducted an evaluation of the new service to:
compare outcomes (mortality, rate of emergency department (ED) re-presentations and hospital re-admissions, LOS and incidence of hospital-acquired complications, HACs) for patients in HITH with patients in traditional hospital settings.
understand whether HITH is acceptable and/or preferable to patients compared to a traditional hospital.
Methods
A mixed methods approach including triangulation of quantitative and qualitative data was used. Integration of the quantitative and qualitative results was conducted using the ‘following a thread’ approach.13 We used the theoretical framework of acceptability (TFA) of healthcare interventions to support the interpretation of research findings related to the recipients’ acceptability of health interventions.14 The framework presents acceptability as a multifaceted construct represented by seven components: affective attitude, burden, perceived effectiveness, ethicality, intervention coherence, opportunity costs and self-efficacy.14
Quantitative phase
De-identified South Australia ED and inpatient activity data, linked to mortality data, for the period July 2019–July 2022 was used to compare outcomes for HITH patients and comparable patients using traditional hospital settings. These data were used to define patient journeys around admissions for all eligible Australian Refined Diagnosis Related Groups (AR-DRGs) including preceding ED presentations and admissions, related episodes of care (e.g. sub-acute episodes of care) and the clinical outcomes of interest.
The population of interest to this study was patients receiving acute admitted care under a medical officer over the period January 2020–July 2022 with a diagnosis (AR-DRG code from a list of 22) that was initially approved for care in the HITH program. The 22 AR-DRG codes are presented in Supplementary material Table S1.
The intervention evaluated is MyHH, an SA Health HITH service that provides hospital-level care for patients in their own place of residence. Between April 2021 and July 2022 there were 2095 admissions to HITH within the 22 AR-DRGs. Of these patients, 1505 were directly admitted to HITH (i.e. an AA), and 590 were transferred to continue a traditional admission (i.e. ESD).
There are two comparators: a historical control group (n = 13,589) comprising all inpatient admissions for the 22 eligible AR-DRGs in 2020, and a post-implementation control group (n = 14,975) comprising all non-HITH inpatient admissions for the 22 eligible AR-DRGs from April 2021 to July 2022.
Evaluation outcomes included: inpatient mortality and mortality at 30-day post discharge, 28-day ED re-presentations and 28-day re-admissions (all causes), LOS and percentage of patients with a HAC.
Adjustment variables included coded comorbidities, socio-demographic variables and health service utilisation (ED presentations and inpatient admissions in the 6 months prior to index admissions). These characteristics are summarised in Supplementary Table S1 and in the Results section.
Two sets of unadjusted and adjusted analyses were undertaken: (i) a pre–post comparison of outcomes before (2020) and after (April 2021–July 2022) implementation of HITH, and (ii) a comparison of HITH admissions with non-HITH admissions in the post period.
The unadjusted analyses were conducted using univariate methods, chi-square tests for categorical outcomes (e.g. re-admissions or re-presentations) and t-tests for continuous outcomes (e.g. LOS). The adjusted analyses were performed using augmented inverse propensity weighting (AIPW). AIPW is a doubly robust approach that combines regression adjustment and inverse propensity weights to estimate the differences in outcomes between groups after controlling for confounding.15,16 AIPW constructs propensity scores from a multivariable logistic model, which predicts the probability of being in the intervention group, given known confounders. The predicted outcomes are weighted for the exposed group as one minus the propensity score, and for the unexposed group as the propensity score. The predicted outcomes are deducted from the observed outcomes. Contrasts of the weighted averages estimate the average group differences with standard errors corrected for the three-step process.17,18 The AIPW analyses were conducted using the tefffects package in STATA v17.
Qualitative phase
Patients discharged from HITH (n = 1185) in the 6 months preceding October 2022 were invited to participate by letter. Interested participants consented over the telephone or completed a hardcopy consent form. Participants completed a short survey prior to the focus group/interview to collect information on their characteristics and HITH admission.
Focus groups and one-on-one interviews were conducted by a mix of three members of the research team (TS, CO, JG) with relevant expertise in qualitative research methods and health services research in late October to early November 2022 using an interview schedule (Supplementary material file S1).
Focus groups and interviews were transcribed verbatim and data were coded and developed into descriptive thematic categories using NVivo (QSR International). The categories were discussed by the three interviewers and synthesised across seven domains of acceptability of health interventions.14
Results
Quantitative phase
HITH admissions made up 2095 of 17,070 (12%) of post-period admissions. Patient characteristics and outcomes for the four different patient admission groups are presented in Table 1. Two of the five in-patient deaths were patients with respiratory infections and inflammations (minor complexity, E62B); others included heart failure and shock (minor complexity, F62B, n = 1), kidney and urinary tract infections (minor complexity, L63B, n = 1) and cellulitis (minor complexity, J64B, n = 1). Per AR-DRG, on average 9% (s.d. 10.4%) of post-period admissions were in HITH (range 0.2–35.3%) (Supplementary Table S1). Comorbidities were less common in HITH (Supplementary Table S1). HITH patients had less socio-economic disadvantage.
| Pre–post comparison | Post-implementation | ||||
|---|---|---|---|---|---|
| 2020 | 2021–22 | 2021–22 | |||
| Pre | Post | Non-HITH | HITH | ||
| n = 13,589 | n = 17,070 | n = 14,975 | n = 2095 | ||
| Patient characteristics | |||||
| Age (years) at admission (mean (s.d.)) | 64.9 (20.9) | 65.3 (20.7) | 64.9 (20.9) | 67.4 (19.3) | |
| Sex | |||||
| Male | 6638 (48.8%) | 8291 (48.6%) | 7220 (48.2%) | 1071 (51.1%) | |
| Female | 6951 (51.2%) | 8779 (51.4%) | 7755 (51.8%) | 1024 (48.9%) | |
| Marital status | |||||
| Married | 5897 (43.4%) | 7423 (43.5%) | 6519 (43.5%) | 904 (43.1%) | |
| Not married | 7692 (56.6%) | 9647 (56.5%) | 8456 (56.5%) | 1191 (56.8%) | |
| Insurance status | |||||
| Hospital insurance | 1124 (8.3%) | 1747 (10.2%) | 1386 (9.3%) | 361 (17.2%) | |
| No hospital insurance | 12,465 (91.7%) | 15,273 (89.8%) | 13,589 (90.7%) | 1734 (82.8%) | |
| Charlson comorbidity | |||||
| 0 | 6651 (48.9%) | 8594 (50.3%) | 7208 (48.1%) | 1386 (66.2%) | |
| 1 | 3798 (27.9%) | 4755 (27.9%) | 4335 (28.9%) | 420 (20%) | |
| 2 | 1388 (10.2%) | 1780 (10.4%) | 1606 (10.7%) | 174 (8.3%) | |
| 3 | 736 (5.4%) | 854 (5%) | 796 (5.3%) | 58 (2.8%) | |
| 4 | 1016 (7.5%) | 1087 (6.4%) | 1030 (6.9%) | 57 (2.7%) | |
| IRSAD category | |||||
| 0 (greatest disadvantage) | 4670 (34.4%) | 5625 (33.0%) | 5153 (34.4%) | 472 (22.5%) | |
| 1 | 4471 (32.9%) | 5644 (33.1%) | 5068 (33.8%) | 576 (27.5%) | |
| 2 (least disadvantage) | 4448 (32.7%) | 5801 (34.0%) | 4754 (31.7%) | 1047 (50%) | |
| Health service utilisation | |||||
| ED event prior to admission | 8427 (62.0%) | 10,361 (60.7%) | 9643 (64.4%) | 718 (34.3%) | |
| Inpatient event prior to admission | 873 (6.4%) | 1065 (6.2%) | 862 (5.8%) | 203 (9.7%) | |
| ED in previous 6 months | 7173 (52.8%) | 6641 (38.9%) | 6114 (40.8%) | 527 (25.2%) | |
| Inpatient in previous 6 months | 13,399 (98.6%) | 13,149 (77%) | 11,650 (77.8%) | 1499 (71.6%) | |
| Number of bed days in previous 6 months (mean (s.d.)) | 9.2 (14.6) | 6.6 (12.3) | 6.8 (12.8) | 5.5 (8.3) | |
| Unadjusted outcomesA | |||||
| Inpatient mortality | 411 (3.0%) | 554 (3.2%) | 549 (3.7%) | 5 (0.2%) | |
| Mortality in 30 days | 340 (2.5%) | 409 (2.4%) | 377 (2.5%) | 32 (1.5%) | |
| Re-presentations in 28 days | 2622 (19.3%) | 2788 (16.3%) | 2622 (17.5%) | 166 (7.9%) | |
| Re-admissions in 28 days | 1555 (11.4%) | 1706 (10%) | 1609 (10.7%) | 97 (4.6%) | |
| LOS (days) (mean (s.d.)) | 3.8 (4.6) | 3.7 (4.3) | 3.6 (4.5) | 4.3 (2.9) | |
| HACs | 106 (0.78%) | 126 (0.74%) | 117 (0.78%) | 9 (0.4%) | |
IRSAD (Index of Relative Social Advantage and Disadvantage,25 s.d. (standard deviation), LOS (length of stay), HACs (hospital acquired complications), HITH (Hospital in the Home)).
All values are numbers (percentages) unless otherwise indicated.
Inpatient mortality is defined as mortality occurring within any admission in a series of consecutive admissions for the same AR-DRG.
Mortality in 30 days is defined as that outcome occurring within 2 and 30 days after the admission.
Re-presentation and re-admission are defined as those events occurring within 2 and 28 days after the admission.
Length of stay (LOS) was calculated as the total number of days in hospital for the admission or the total number of days for the series of consecutive admissions for the same AR-DRG.
The HAC outcome was defined as the presence of any HAC that was recorded in an admission (or a series of consecutive admissions for the same AR-DRG).
Comparing pre- and post-periods, there was no difference between them in the rates of inpatient mortality and mortality at 30 days; however, re-presentation and re-admission rates fell slightly in the post-period, with falls of 1.6 percentage points (95% confidence interval (CI): −2.6%, −0.7%) and 0.8 percentage points (95% CI: −1.6%, −0.1) respectively (Table 2). LOS also fell on average over the period by around 0.5 days (95% CI: −0.6, −0.9). HACs were steady across both time periods (0.8%).
| Outcome | Pre–post comparison | Post-implementation | |||||||
|---|---|---|---|---|---|---|---|---|---|
| 2020 | 2021–22 | Difference | P-value | 2021–22 | Difference | P-value | |||
| Pre | Post | Non-HITH | HITH | ||||||
| (n = 13,589) | (n = 17,070) | (n = 14,975) | (n = 2095) | ||||||
| Inpatient mortality (%) | 3.8 (3.2, 4.3) | 3.3 (3.1, 3.6) | −0.4 (−1, 0.2) | 0.153 | 3.6 (3.3, 4) | 0.2 (0, 0.4) | −3.5 (−3.8, −3.1) | <0.001 | |
| Mortality in 30 days (%) | 2.6 (2.2, 3) | 2.5 (2.3, 2.7) | −0.1 (−0.6, 0.3) | 0.559 | 2.5 (2.2, 2.7) | 1.2 (0.5, 2) | −1.3 (−2, −0.5) | 0.002 | |
| Re-presentations 28 days (%) | 18.3 (17.5, 19) | 16.6 (16, 17.2) | −1.6 (−2.6, −0.7) | 0.001 | 17 (16.4, 17.6) | 9.7 (7.6, 11.9) | −7.2 (−9.5, −5.0) | <0.001 | |
| Re-admissions 28 days (%) | 11.0 (10.4, 11.6) | 10.2 (9.7, 10.6) | −0.8 (−1.6, −0.1) | 0.036 | 10.5 (10, 11) | 5.5 (3.9, 7.2) | −4.9 (−6.7, −3.2) | <0.001 | |
| LOS (days) | 4.2 (4.1, 4.3) | 3.7 (3.7, 3.8) | −0.5 (−0.6, −0.3) | <0.001 | 3.6 (3.6, 3.7) | 3.9 (3.6, 4.1) | 0.2 (0, 0.5) | 0.059 | |
| HACs (%) | 0.8 (0.6, 1) | 0.8 (0.6, 0.9) | −0.1 (−0.3, 0.2) | 0.609 | 0.8 (0.6, 0.9) | 0.1 (0, 0.3) | −0.6 (−0.8, −0.5) | <0.001 | |
LOS (length of stay), HACs (hospital acquired complications), HITH (Hospital in the Home), AR-DRGs (Australian Refined Diagnosis Related Groups).
Pre–post comparisons were made separately for each of the four most common HITH AR-DRGs (Table 3). Trends were similar across the AR-DRGs as overall, except for greater in-patient mortality in the post-period for ‘Kidney and urinary tract infections, minor complexity’ L63B (0.5, 95% CI: 0.1%, 1.0%) and longer LOS in the post-period for the same AR-DRG group (0.2 days, 95% CI: 0.1, 0.4).
| AR-DRG | Outcome | Pre–post comparison | Post-implementation | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2020 | 2021–22 | Difference | P-value | 2021–22 | Difference | P-value | ||||||||
| Pre | Post | Non-HITH | HITH | |||||||||||
| n (%) | mean (95% CI) | n (%) | mean (95% CI) | mean (95% CI) | n (%) | mean (95% CI) | n (%) | mean (95% CI) | mean (95% CI) | |||||
| E62B Respiratory infections and inflammations, minor complexity | Inpatient mortality (%) | 1673 (12.3) | 6.2 (5, 7.3) | 2174 (12.7) | 6.3 (5.3, 7.4) | 0.2 (−1.4, 1.7) | 0.808 | 1820 (12.2) | 7.7 (6.4, 9) | 354 (16.3) | 0.9 (−0.6, 2.5) | −6.8 (−8.8, −4.7) | <0.001 | |
| Mortality in 30 days (%) | 5.3 (4.2, 6.4) | 4.8 (3.9, 5.7) | −0.5 (−1.9, 0.9) | 0.494 | 5 (4, 6) | 1.7 (0.6, 2.8) | −3.3 (−4.8, −1.9) | <0.001 | ||||||
| Re-presentations in 28 days (%) | 16.2 (14.4, 18) | 12.8 (11.4, 14.2) | −3.4 (−5.6, −1.1) | 0.003 | 13.7 (12.1, 15.3) | 6.7 (3, 10.4) | −7 (−11, −3) | 0.001 | ||||||
| Re-admissions in 28 days (%) | 11.4 (9.9, 12.9) | 9.1 (7.9, 10.3) | −2.4 (−4.3, −0.4) | 0.018 | 9.8 (8.5, 11.2) | 6.1 (2.3, 9.9) | −3.7 (−7.8, 0.3) | 0.072 | ||||||
| LOS (days) | 3.3 (3.2, 3.5) | 3.3 (3.2, 3.4) | 0 (−0.2, 0.1) | 0.612 | 3.3 (3.1, 3.4) | 3.3 (3, 3.6) | 0 (−0.3, 0.3) | 0.812 | ||||||
| HACs (%) | 0.3 (0, 0.6) | 0.1 (0, 0.3) | −0.2 (−0.5, 0.1) | 0.301 | 0.1 (0, 0.3) | 0 | −0.1 (−0.3, 0) | 0.083 | ||||||
| J64A Cellulitis, major complexity | Inpatient mortality (%) | 997 (7.3) | 1.4 (0.7, 2.1) | 1191 (7.0) | 0.5 (0.1, 0.9) | −0.9 (−1.7, −0.1) | 0.034 | 1579 (10.5) | 0.6 (0.1, 1) | 863 (35.3) | 0 | −0.6 (−1, −0.1) | 0.014 | |
| Mortality in 30 days (%) | 1.7 (0.9, 2.5) | 2.4 (1.6, 3.3) | 0.7 (−0.5, 1.9) | 0.229 | 2.3 (1.3, 3.3) | 0.9 (0.1, 1.7) | −1.4 (−2.7, −0.1) | 0.038 | ||||||
| Re-presentations in 28 days (%) | 19.3 (16.8, 21.7) | 16 (14, 18.1) | −3.2 (−6.4, 0) | 0.05 | 17.4 (14.9, 19.9) | 6.2 (3.1, 9.4) | −11.2 (−15.2, −7.1) | <0.001 | ||||||
| Re-admissions in 28 days (%) | 11.6 (9.6, 13.6) | 9.8 (8.1, 11.5) | −1.8 (−4.4, 0.8) | 0.174 | 10.4 (8.4, 12.4) | 4.1 (1.7, 6.5) | −6.3 (−9.4, −3.2) | <0.001 | ||||||
| LOS (days) | 6.8 (6.3, 7.2) | 6.7 (6.3, 7.1) | −0.1 (−0.7, 0.5) | 0.75 | 6.9 (6.4, 7.3) | 4.8 (4.2, 5.5) | −2 (−2.8, −1.2) | <0.001 | ||||||
| HACs (%) | 2.2 (1.3, 3.1) | 2.4 (1.6, 3.3) | 0.2 (−1, 0.1) | 0.723 | 2.4 (1.3, 3.4) | 1.2 (0.1, 2.3) | −1.2 (−2.7, 0.4) | 0.145 | ||||||
| J64B Cellulitis, minor complexity | Inpatient mortality (%) | 1505 (11.1) | 0.1 (−0.1, 0.2) | 2442 (14.3) | 0.1 (0, 0.2) | 0 (−0.2, 0.2) | 0.861 | 915 (6.1) | 0.1 (0, 0.2) | 276 (16.5) | 0.1 (−0.1, 0.3) | 0 (−0.2, 0.3) | 0.684 | |
| Mortality in 30 days (%) | 0.5 (0.1, 0.8) | 0.5 (0.2, 0.8) | 0 (−0.4, 0.5) | 0.907 | 0.6 (0.2, 0.9) | 0.2 (0, 0.4) | −0.4 (−0.8, 0.1) | 0.107 | ||||||
| Re-presentations in 28 days (%) | 13.6 (11.9, 15.4) | 11.9 (10.6, 13.2) | −1.7 (−3.9, 0.5) | 0.121 | 15.2 (11.7, 18.6) | 7.7 (5.1, 10.2) | −7.5 (−11.8, −3.2) | 0.001 | ||||||
| Re-admissions in 28 days (%) | 5.4 (4.3, 6.6) | 5 (4.2, 5.9) | −0.4 (−1.8, 1) | 0.575 | 7.7 (4.2, 11.2) | 4.3 (1.9, 6.6) | −3.5 (−7.8, 0.8) | 0.111 | ||||||
| LOS (days) | 3.1 (2.9, 3.3) | 3.2 (3.1, 3.3) | 0.1 (−0.1, 0.3) | 0.448 | 2.6 (2.5, 2.8) | 4.2 (4, 4.4) | 1.5 (1.3, 1.7) | <0.001 | ||||||
| HACs (%) | 0 (0, 0) | 0.1 (0, 0.2) | 0.1 (0, 0.1) | 0.157 | 0.1 (0, 0.2) | 0 | −0.1 (−0.2, 0) | 0.158 | ||||||
| L63B Kidney and urinary tract infections, minor complexity | Inpatient mortality (%) | 10369 (7.6) | 0.1 (−0.1, 0.3) | 1410 (8.3) | 0.6 (0.2, 1.1) | 0.5 (0.1, 1) | 0.02 | 1178 (7.9) | 0.6 (0.2, 1) | 232 (16.5) | 0.3 (−0.3, 0.9) | −0.3 (−1, 0.4) | 0.378 | |
| Mortality in 30 days (%) | 1.4 (0.7, 2.2) | 1.7 (1, 2.4) | 0.3 (−0.7, 1.2) | 0.609 | 1.6 (0.8, 2.3) | 1.2 (0.1, 2.3) | −0.3 (−1.6, 1) | 0.617 | ||||||
| Re-presentations in 28 days (%) | 18.3 (15.9, 20.6) | 14.9 (13, 16.8) | −3.4 (−6.4, −0.4) | 0.026 | 15.5 (13.4, 17.5) | 8.4 (2.8, 13.9) | −7.1 (−13, −1.2) | 0.018 | ||||||
| Re-admissions in 28 days (%) | 12.9 (10.9, 14.9) | 11.1 (9.5, 12.8) | −1.8 (−4.4, 0.9) | 0.187 | 11.8 (9.9, 13.6) | 3.8 (0.8, 6.8) | −8 (−11.5, −4.5) | <0.001 | ||||||
| LOS (days) | 2.5 (2.4, 2.6) | 2.7 (2.6, 2.9) | 0.2 (0.1, 0.4) | 0.009 | 2.6 (2.4, 2.9) | 3.1 (2.7, 3.6) | 0.5 (0, 1.1) | 0.06 | ||||||
| HACs (%) | 0.1 (−0.1, 0.3) | 0.1 (−0.1, 0.2) | 0 (−0.3, 0.1) | 0.832 | 0.1 (−0.1, 0.2) | 0 | −0.1 (−0.2, 0.1) | 0.317 | ||||||
LOS (length of stay), HACs (hospital acquired complications), HITH (hospital-in-the-home), AR-DRGs (Australian Refined Diagnosis Related Groups).
Twenty-three percent (n = 49) of patients admitted to HITH experienced a return to hospital during their HITH admission. Most of these (40/49, 82%) were for the same AR-DRG (regardless of complexity) as for the HITH index admission, including cellulitis (J64, n = 20), respiratory infections and inflammations (E62, n = 9), kidney and urinary tract disorders (L63, n = 7), chronic obstructive airways disease (E65, n = 2) and heart failure (F62, n = 2). HITH was associated with lower rates for all outcomes after adjustment, except average LOS (0.2 days higher in HITH, 95% CI: 0, 0.5) (Table 2). In-patient mortality was 3.5% lower (95% CI: −3.8%, −3.1%) and mortality in 30 days was 1.3% lower (95% CI: −2.0%, −0.5%). Re-presentations were around 7 percentage points lower for HITH after adjustment (95% CI: −9.5%, −5.0%) and re-admissions around 5 percentage points lower (95% CI: −6.7%, −3.2%). HITH admissions were also associated with a reduction in HACs after adjustment (0.6 percentage points lower (95% CI: −0.8%, −0.5%)).
These findings were similar across the four most common AR-DRGs, with the exception of ‘Cellulitis, major complexity’ J64A, which demonstrated a 2 day shorter LOS (95% CI: −2.8, −1.2) and ‘Cellulitis, minor complexity’ J64B, which demonstrated no difference in inpatient mortality (95% CI: −0.2%, 0.3%) (Table 3).
Qualitative phase
Focus groups were conducted face-to-face (n = 6, 27 participants) and online (n = 1, 3 participants). Interviews were conducted on nine occasions (11 participants). A total of 35 consumers and six carers participated; five carers attended focus groups/interviews with a consumer, one carer was interviewed on their own. Demographics are presented in Table 4.
| Variable | n | (%) | ||
|---|---|---|---|---|
| Condition | Cellulitis | 11 | (35.5) | |
| Kidney and urinary tract infections | 6 | (19.4) | ||
| Orthopaedic post-operative care | 4 | (12.9) | ||
| Pneumonia | 3 | (9.7) | ||
| Deep vein thrombosis | 1 | (3.2) | ||
| Other | 6 | (19.4) | ||
| Age group (years) A | 18–39 | 1 | (2.6) | |
| 40–59 | 6 | (15.8) | ||
| 60–79 | 25 | (65.8) | ||
| 80+ | 6 | (15.8) | ||
| Referral path | Ambulance | 1 | (3.2) | |
| General practitioner | 1 | (3.2) | ||
| Hospital/local health network | 27 | (87.1) | ||
| Priority care centre | 1 | (3.2) | ||
| Respiratory specialist | 1 | (3.2) | ||
| Additional care | Aged care (home care package) | 5 | (16.1) | |
| Disability | 5 | (16.1) | ||
| No | 21 | (67.7) |
There was high acceptance of, and preference for, HITH. Table 5 maps the identified categories from the data against the components of acceptance developed by Sekhon. The predominant affective attitude to HITH was of acceptance, safety, familiarity, appreciation and gratitude, whereas distress and concerns about infection, including COVID-19, were expressed about traditional hospitals. Participants demonstrated high levels of understanding about how HITH works (intervention coherence) to free up capacity in traditional hospitals for acutely unwell patients or reduce stresses and risks of hospitalisation for immune-compromised patients. The perceived effectiveness of HITH was enhanced by personalised care delivered in a structured way by nurses and doctors but reduced by the difficulty of meeting all patients’ needs, especially those who required 12 hourly IV antibiotics. Participants’ self-efficacy was demonstrated through a preference for future hospitalisations to be delivered by HITH and undertaking usual activities including family duties and working from home. Self-efficacy could be enhanced for some participants through better preparation on admission to HITH (Table 5). Three components of acceptability (burden, ethicality and opportunity costs) were not identified in the qualitative data.
| Acceptability component (Sekhon) | Category | Description | Illustrations | |||
|---|---|---|---|---|---|---|
| Affective attitude | Highly acceptable to many patients | HITH is clearly highly acceptable to patients with many participants praising the program and commenting that they would highly recommend the program to others in a similar position. | I just can’t say enough good about it. It was just an amazing experience. I had nurses coming three times a day, the doctor ringing on cue, all of my bandages and medicines and everything were just arriving by courier; it was just an amazing experience. (FG_1_11am) | Yeah, godsend, it really is. It’s the future. (FG_2_1pm) | ||
| Affective attitude | Greater sense of familiarity and comfort from being at home | Participants described the benefits of being in their own home and close to their family members and pets. | I had my nice, clean shower I could shower in. Doors open and shut the way they were supposed to. Lights turn on and off the way they were supposed to. All that stuff. (FG_1_11am) | Having your own family around you and things like that…My husband still went to work, the kids still went to school. And I had no impact on anybody else because I just didn’t need to bother anybody. They didn’t have to come and visit me. They didn’t have to bring me anything. (FG_7_online) | I’ve got my pets, as well, animal therapy, you know. And more able to relax. And get to spend more time with my wife. (TS_4) | |
| Affective attitude | Feelings of safety, lower levels of stress and reassurance | Greater familiarity and comfort at home led to feelings of safety, lower levels of stress and reassurance. | I felt I had a good grasp on why they were there and what I had going on, and I felt, as I said, I felt quite safe, very reassured. (FG_7_online) | I couldn’t fault it … It was an absolute pleasure to have them come to the home and, you know, feel the way you feel without the additional stress of what goes on in the ED. (FG_7_online) | For me it was very reassuring and much easier being in your own home … So yeah, I thought it was brilliant. (FG_7_online). | |
| Affective attitude | Participants’ distressing experiences of the bricks and mortar hospital | Avoiding COVID-19 and other infections in hospital and disturbances to sleep were widely reported. In addition, many of the participants were very experienced as patients, and reported distressing incidents and general dislike for the hospital environment. | I’m very reluctant to go to hospital and had a couple of incidents when I’ve gone through the ED, which left me quite apprehensive … because of how anxious it was making me, my specialist actually organised for the home at home hospital to come. I thought the whole thing was brilliant. It was much more reassuring for me to be at home. (FG_7_online) | The guy across from me in the ward had his own guard because he was violent. The lady next to me, God bless her, she was passing away and it was – I never even saw her, I just heard the sounds through the curtains so that was quite distressing. (FG_1_11am) | I just couldn’t get any – I was trying to sleep, I couldn’t get any sleep and I was getting better because of the antibiotics but I was getting worse because I couldn’t get – and I was frightened to sleep because this guy would roam and streak. (FG_1_11am) | |
| Affective attitude | Gratitude for the service | Don’t hesitate to use it … All I can say is it’s, it’s just a wonderful experience and I would recommend it to anybody. And I’m just, you know, just so thankful and grateful that I’ve had the chance to receive it. (TS_2) | ||||
| Affective attitude | Appreciation of patients with caring roles | The HITH program supported patients to continue caring roles in their own home. | My husband has motor neuron disease and when they said I’d have to go to hospital, I thought I’d have to put my husband into the nursing home. When I found out I could stay at home, well it was the answer to my prayers. (TS_2) | |||
| Intervention coherence | Strength of support for HITH among vulnerable patients | Immuno-compromised patients and/or those living with a chronic respiratory disease were strikingly supportive of HITH. This group had experienced numerous ED visits and hospitalisations and were acutely aware of the risks of infection posed from a bricks-and-mortar hospital visit. | I think I’ve got a bit of posttraumatic stress disorder just from everything but definitely emergency and the way and the indignity and the being invisible and not having my legs [wheelchair]. To get quick attention, you’re going with an ambulance. But then I don’t have my legs. If you go in with a friend driving my modified car, then you get less attention. You just sit in the waiting room. (FG_7_online) | Because of some of the things that have gone on, I, I don’t need to be trying to breathe and deal with being frightened as well. Whereas at home I just could relax. It’s quite a different feeling … But honestly, if I got in that situation again, I just contact my specialist and say, can we organise this [MyHH] again? Because it’s such a reassuring way of going about things. (FG_7_online) | Safer for me because I probably hold off too long going to my doctor because I know he’s gonna say go to the ED. So if I could ring him knowing he’s gonna refer hospital at home, I’d probably do it two or three days earlier. So that gets the IV in earlier. I’ll probably wouldn’t get as sick and also wouldn’t have the two or three days of feeling like that and going through the ED. (FG_7_online) | |
| Intervention coherence | Role within the broader health setting | One of the perceived benefits of the program is that it frees up capacity in traditional hospitals for acutely unwell patients or shortens the length of stay. | I reckon if the government put more money into this program, they’d get more bang for buck than putting it straight into hospitals. … This is the greatest way to actually free up beds, and really the hospital should be all ICU, emergency stuff and everybody else should be on this program … It frees up nurses, frees up doctors, frees up a whole lot of things. But obviously that money has to go into the people that come to your house. (FG5_11am) | I preferred the MyHH to the [bricks-and-mortar] hospital, but I had to have the hospital in the beginning and then end up in the MyHH. Sometimes you have to go to the [bricks-and-mortar hospital] to get the MyHH you know? (TS_4) | I hope it’s not just a trial, I hope it’s ongoing. (FG_7_online) | |
| Perceived effectiveness | Nurses and doctors provided personalised care in a structured way with good communication | Communication and delivery of care was highly acceptable to patients, who described staff as being much more relaxed in the home environment, and non-clinical ‘in a good way, a less official way’ and that there was less rushing and more time for discussion. | The communication was brilliant and you always were given a time, a slot, that they should be here between there. I think maybe twice they were late, and they re-sent a message so you knew they were coming late. Once again, communicating with the doctor daily was great. Which you didn’t get in the hospital. (FG_4_1pm) | I found the MyHH program a lot better than what the hospital was. A lot more thorough and kinder. Staff were lot more gentle at home, I had a few bad experiences in hospital. (TS_4) | Cannula came out in the hospital; they didn’t do a thing about it. I had blood running down my arm. I just sat there waiting, looking around and the male nurse is sitting there looking at me the whole time and never even flinched. I thought, yep, it was this care factor in the hospital was zero. (FG_4_1pm) | |
| Perceived effectiveness | Difficulty of meeting all patients’ needs | Some patients were very aware of the difficulty of providing 12 hourly medications in a service that was nominally run from 08:00 to 20:00 hours. If the morning antibiotic was late, then the gap for the evening antibiotic would be less than 12 h. | … but I used to get quite stressed. I would be watching the clock and I didn’t care if they were an hour late, but when they started getting more than an hour the late …You know out of 12 days there must have been about 6 to 8 days when they were out by more than an hour. (TS_6) | I had trouble with my veins so a couple of times the nurses that were there couldn’t put the Jelco in – like a needle. So I had to go back to the hospital to get that done. (FG_3_11am) | The only problem that I actually had, the only thing negative I had was obviously the time restriction. Like [patient] needed antibiotics every six to eight hours three times a day, but because of the time … MyHH allows the nurses is between eight and eight pm which meant that he was receiving antibiotics sometimes within three hours or four, which is like really bad. (CO_2) | |
| Self-efficacy | A need to better prepare patients for the HITH admission | Some participants expressed some uncertainty or apprehension on entering HITH, which may be alleviated by better preparation of patients through education or better communication of information. Some patients were unclear about the restriction on not receiving medical care from other providers during their HITH admission (e.g. Medicare-funded care from their own GP). | A little bit more education. A little bit more information, not just say to you they’ll be delivered. (FG_1_11am) | … but after about 3 days, one nurse pulled out this booklet from the orange box and said ‘This is for you’. And I thought, well, every nurse that has been there prior to that had to be in that box to get things out. And they weren’t observant enough to see that I hadn’t actually been handed the pamphlets and the fact sheets? (TS_6) | I wasn’t as aware that … there is that issue that you ring up a [general] practice and do something like that, then they get into all sorts of trouble because you’ve got two Medicare claims being made … (FG_3_11am) | |
| Self-efficacy | Preferences for HITH over bricks-and-mortar hospital | Based on the high levels of acceptability to patients, their preferences for HITH over bricks-and-mortar hospital is unsurprising. A number of participants stated that they would request HITH for future care of a condition included in the HITH program. | If I had to go to the ______ Hospital again, depending on obviously why I’ve gotta be seen, I would be saying to them, am I a candidate for the My Home Hospital? (FG_7_online) | If I came out of the hospital and needed a service I would choose MyHH. They were very prompt and it was very good service and very efficient. (TS_3) | If I was given a choice, I would’ve preferred home anyway because of the flexibility and in a familiar environment. (FG_1_11am) | |
| Self-efficacy | Undertaking usual activities including family duties and work from home | In addition to enjoying the familiarity of home, HITH participants were able to undertake usual activities, which brought substantial advantages around not requiring leave from work. | It’s handy to be at home, because you can still potter around and do things yourself. Whereas if you’re sitting in a hospital, one is you get bored shitless. (FG_4_1pm) | They were flexible … I pushed them back or get them to come early or whatever it was, so I can attend my family function. (FG_1_11am) | I didn’t have to book any time off work. I didn’t have to use any holiday or sick leave. I just worked from home. (FG_7_online). | |
Discussion
Results across the quantitative and qualitative phases suggest positive outcomes for HITH patients and the health system, and importantly do not raise concerns about safety of patients in HITH. The incidence of ‘bricks-and-mortar’ admissions during a MyHH HITH admission (2.3%) is lower than estimates from Victorian HITH services (11.9 and 4.2%, respectively).18,19 However, patients in MyHH HITH were older, had less comorbidities and lower socio-economic disadvantage than non-HITH patients. These differences imply a selection bias, which, with the exception of age should result in (unadjusted) better outcomes for HITH patients, as seen in Table 1. Overall, this important finding indicates that HITH referral and acceptance pathways are working effectively to ensure that lower risk patients are being treated at home. The implications of these findings are also important to referrers, especially hospital-based referrers who may not otherwise be aware of patient outcomes from the HITH service.
The adjusted analysis found lower rates of mortality, re-presentations and re-admissions and HACs in the post-implementation HITH group compared to non-HITH. This suggests either the presence of an intervention effect or unobserved confounding. The magnitude of the HITH effect size (e.g. −3.5 and −1.3% for inpatient and 30 day mortality, respectively, and −0.6% for HACs) seems likely to be indicative of unobserved confounding in HITH patients. Confounding can arise when unobserved factors that influence treatment decisions (in this case, whether to admit a patient into HITH, such as fitness or frailty) are also independent determinants of health outcomes.19 While statistical approaches such as multivariable outcome models and propensity score methods can be used to control for observed confounders, the success of these methods are a function of the accuracy and completeness of the data. In studies dependent on administrative databases, many potentially important confounders are not collected (i.e. unobserved), poorly defined or measured with variable quality.19
If unobserved confounding explains the HITH/non-HITH comparison, then the results of the pre–post comparison (i.e. lower rates of re-presentations and re-admissions, and shorter LOS in the post-period compared to the pre-period) are likely indicative of a temporal change in these outcomes for the included AR-DRGs. However, the pre–post comparison is complicated by the potential impact of COVID-19 pandemic spikes in hospital admissions in South Australia in January and July 2022.20 Further investigation of temporal changes is warranted.
Systematic reviews of randomised controlled trials (RCTs) generally report lower (or at least equivalent) mortality in HITH programs compared with usual care in a ‘bricks-and-mortar’ hospital.1–3,6,7 Our estimates of mortality in HITH (0.2% for inpatients and 1.2% at 30 days) are low by international standards; for example, a recent meta-analysis of nine RCTs reported a mortality rate of 11.1% in HITH programs and 14.1% in usual care.2 This again suggests that referral patterns and inclusion of selected AR-DRGs in this HITH service are targeting lower risk patients, potentially due to its stand-alone nature. A more conservative approach to patient selection could readily result from a reluctance of either referrers to refer and/or the HITH service to accept more acutely unwell patients into the service. Reluctance of referrers could result from lack of familiarity with, or confidence in, the new service; whereas the HITH service could be seeking to reduce the risks of poorer patient outcomes. Further research is warranted.
Participants experienced acceptability, satisfaction and enhanced self-efficacy and relayed preferences for home care, similar to other qualitative studies of HITH.11 Support for HITH was very strong among vulnerable patients. Feeling safe and comforted at home is central to patient’s perceptions and satisfaction levels.21–23 While patients mostly perceived high levels of effectiveness and quality of care, gaps were identified in clinical skills of staff (e.g. Jelco insertion) and operating hours. Improving the competence of healthcare professionals and enhancing the design of the HITH service are two key strategies for HITH improvement.24 Nevertheless, in the majority of cases, patients seemed to rationalise that minor inconveniences, such as needing to attend a traditional hospital for Jelco insertion, were outweighed by the overall benefits of being at home. Under the TFA, the personal opportunity cost of forgoing a potentially higher quality of care within a hospital was acceptably ‘traded off’ by patients, due to the overwhelming gains they enjoyed in other aspects of their lives by being cared for at home. Additionally, HITH was perceived to help with releasing resources for more acutely unwell patients to receive care in traditional hospitals, implying an understanding of, and concern for, capacity constraints of traditional hospitals. This satisfied participants’ ethical or communal goals and contributed to their acceptance of inconveniences and other trade-offs.
Limitations
The use of routinely collected hospital administrative data limits observational study designs and the ability to control for sources of bias and confounding. Improvements to collections or de novo prospective data should be considered, but only where a higher standard of evidence warrants the collection costs. Approximately 3% of eligible patients self selected into the qualitative phase; it is possible that these research participants had unrepresentative experiences, either positive or negative, that they wished to convey to the research team
Conclusion
A stand-alone metropolitan-wide HITH program delivered safe care to patients. Further quantitative studies are required to investigate the effect of unobserved confounding on the magnitude of effectiveness. Patients and carers who directly interact with or receive HITH described their acceptability of the program and their preferences for home care. Although there were gaps in service provision, these were tolerated because of other personal benefits to being cared for at home and a sense of contributing to broader, communal goals for allocating scarce hospital resources.
Data availability
The data that support this study were obtained from SA Health by permission. Data will be shared upon reasonable request to the corresponding author with permission from SA Health. The authors had full access to all of the data included in this study.
Conflicts of interest
Researchers from Flinders University conducted this work as contract research, which was funded by SA Health. Three members of the team (AM, JM, CR) were employed by SA Health during the project.
Declaration of funding
The project was funded by South Australia Health, which defined the initial aims of the study and provided assistance in obtaining data and interpreting results. The researchers TS, CO, APi, APa, AT, JG and JK worked independently from the funder. AM, JM and CR were employed by the funder at the time of this work. All authors adhered to the Australian Code for the Responsible Conduct of Research, in particular items R23 (disseminate findings responsibly, accurately and broadly) and P7 (accountability and consideration of the consequences and outcomes of research prior to its communication).
Acknowledgements
The SA Health Enterprise Data and Information Branch provided de-identified patient data from the Emergency Department Data Collection and the Admitted Patient Care dataset. The contribution of 41 interviewees is gratefully acknowledged.
References
1 Shepperd S, Iliffe S, Doll HA, et al. Admission avoidance hospital at home. Cochrane Database Syst Rev 2016; 9: CD007491.
| Crossref | Google Scholar | PubMed |
2 Arsenault-Lapierre G, Henein M, Gaid D, et al. Hospital-at-Home Interventions vs In-Hospital Stay for Patients With Chronic Disease Who Present to the Emergency Department: A Systematic Review and Meta-analysis. JAMA Netw Open 2021; 4: e2111568.
| Crossref | Google Scholar | PubMed |
3 Curioni C, Silva AC, Damião J, et al. The Cost-Effectiveness of Homecare Services for Adults and Older Adults: A Systematic Review. Int J Environ Res Public Health 2023; 20(4): 3373.
| Crossref | Google Scholar | PubMed |
4 Leong MQ, Lim CW, Lai YF. Comparison of Hospital-at-Home models: a systematic review of reviews. BMJ Open 2021; 11: e043285.
| Crossref | Google Scholar | PubMed |
5 Sriskandarajah S, Hobbs J, Roughead E, et al. Safety and effectiveness of ‘hospital in the home’ and ‘outpatient parenteral antimicrobial therapy’ in different age groups: A systematic review of observational studies. Int J Clin Pract 2018; 72: e13216.
| Crossref | Google Scholar | PubMed |
6 Varney J, Weiland TJ, Jelinek G. Efficacy of hospital in the home services providing care for patients admitted from emergency departments: an integrative review. Int J Evid Based Healthc 2014; 12: 128-141.
| Crossref | Google Scholar |
7 Qaddoura A, Yazdan-Ashoori P, Kabali C, et al. Efficacy of Hospital at Home in Patients with Heart Failure: A Systematic Review and Meta-Analysis. PLoS One 2015; 10: e0129282.
| Crossref | Google Scholar | PubMed |
8 Gonçalves-Bradley DC, Iliffe S, Doll HA, et al. Early discharge hospital at home. Cochrane Database Syst Rev 2017; 6: CD000356.
| Crossref | Google Scholar | PubMed |
9 Jeppesen E, Brurberg KG, Vist GE, et al. Hospital at home for acute exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2012; CD003573.
| Crossref | Google Scholar | PubMed |
10 Ojoo JC, Moon T, McGlone S, et al. Patients’ and carers’ preferences in two models of care for acute exacerbations of COPD: results of a randomised controlled trial. Thorax 2002; 57: 167-169.
| Crossref | Google Scholar | PubMed |
11 Chua CMS, Ko SQ, Lai YF, et al. Perceptions of Hospital-at-Home Among Stakeholders: a Meta-synthesis. J Gen Intern Med 2022; 37: 637-650.
| Crossref | Google Scholar | PubMed |
12 SA Health. My Home Hospital. 2022. Available at https://www.sahealth.sa.gov.au/wps/wcm/connect/public+content/sa+health+internet/services/hospitals/my+home+hospital/my+home+hospital. [accessed 14 November 2022].
13 O’Cathain A, Murphy E, Nicholl J. Three techniques for integrating data in mixed methods studies. BMJ 2010; 341: c4587.
| Crossref | Google Scholar | PubMed |
14 Sekhon M, Cartwright M, Francis JJ. Acceptability of healthcare interventions: an overview of reviews and development of a theoretical framework. BMC Health Serv Res 2017; 17: 88.
| Crossref | Google Scholar | PubMed |
15 Rubin DB, van der Laan MJ. Empirical Efficiency Maximization: Improved Locally Efficient Covariate Adjustment in Randomized Experiments and Survival Analysis. Int J Biostat 2008; 4: 5.
| Crossref | Google Scholar | PubMed |
16 Tan Z. Bounded, efficient and doubly robust estimation with inverse weighting. Biometrika 2010; 97: 661-682.
| Crossref | Google Scholar |
17 Imbens GW. The Role of the Propensity Score in Estimating Dose-Response Functions. Biometrika 2000; 87: 706-710.
| Crossref | Google Scholar |
18 Robins JM, Hernán MÁ, Brumback B. Marginal Structural Models and Causal Inference in Epidemiology. Epidemiology 2000; 11: 550-560.
| Crossref | Google Scholar | PubMed |
19 Brookhart MA, Stürmer T, Glynn RJ, et al. Confounding control in healthcare database research: challenges and potential approaches. Med Care 2010; 48: S114-S120.
| Crossref | Google Scholar | PubMed |
20 Department of Health and Aged Care. COVID-19 reporting. 2023. Available at https://www.health.gov.au/topics/covid-19/reporting?language=und#covid19-impact-on-hospitals. [accessed 20 December 2023].
21 Leine M, Wahl AK, Borge CR, et al. Feeling safe and motivated to achieve better health: Experiences with a partnership-based nursing practice programme for in-home patients with chronic obstructive pulmonary disease. J Clin Nurs 2017; 26: 2755-2764.
| Crossref | Google Scholar | PubMed |
22 Schofield I, Knussen C, Tolson D. A mixed method study to compare use and experience of hospital care and a nurse-led acute respiratory assessment service offering home care to people with an acute exacerbation of chronic obstructive pulmonary disease. Int J Nurs Stud 2006; 43: 465-476.
| Crossref | Google Scholar | PubMed |
23 Wang Y, Haugen T, Steihaug S, et al. Patients with acute exacerbation of chronic obstructive pulmonary disease feel safe when treated at home: a qualitative study. BMC Pulm Med 2012; 12: 45.
| Crossref | Google Scholar | PubMed |
24 Wang X, Stewart C, Lee G. Patients’ and caregivers’ perceptions of the quality of hospital-at-home service: A scoping review. J Clin Nurs 2024; 33: 817-838.
| Crossref | Google Scholar |
