Differences in health outcomes for high-need high-cost patients across high-income countries

Home Heart Hospital Associated With Reduced Hospitalizations and Costs Among High-Cost Patients With Cardiovascular Disease

BACKGROUND

There is no widely accepted care model for managing high-need, high-cost (HNHC) patients. We hypothesized that a Home Heart Hospital (H3), which provides longitudinal, hospital-level at-home care, would improve care quality and reduce costs for HNHC patients with cardiovascular disease (CVD).

OBJECTIVE

To evaluate associations between enrollment in H3, which provides longitudinal, hospital-level at-home care, care quality, and costs for HNHC patients with CVD.

METHODS

This retrospective within-subject cohort study used insurance claims and electronic health records data to evaluate unadjusted and adjusted annualized hospitalization rates, total costs of care, part A costs, and mortality rates before, during, and following H3.

RESULTS

Ninety-four patients were enrolled in H3 between February 2019 and October 2021. Patients' mean age was 75 years and 50% were female. Common comorbidities included congestive heart failure (50%), atrial fibrillation (37%), coronary artery disease (44%). Relative to pre-enrollment, enrollment in H3 was associated with significant reductions in annualized hospitalization rates (absolute reduction (AR): 2.4 hospitalizations/year, 95% confidence interval [95% CI]: -0.8, -4.0; p < 0.001; total costs of care (AR: -$56 990, 95% CI: -$105 170, -$8810; p < 0.05; and part A costs (AR: -$78 210, 95% CI: -$114 770, -$41 640; p < 0.001). Annualized post-H3 total costs and part A costs were significantly lower than pre-enrollment costs (total costs of care: -$113 510, 95% CI: -$151 340, -$65 320; p < 0.001; part A costs: -$84 480, 95% CI: -$121 040, -$47 920; p < 0.001).

CONCLUSIONS

Longitudinal home-based care models hold promise for improving quality and reducing healthcare spending for HNHC patients with CVD.

Development and Internal Validation of an Interpretable Machine Learning Model to Predict Readmissions in a United States Healthcare System

(1) One in four hospital readmissions is potentially preventable. Machine learning (ML) models have been developed to predict hospital readmissions and risk-stratify patients, but thus far they have been limited in clinical applicability, timeliness, and generalizability. (2) Methods: Using deidentified clinical data from the University of California, San Francisco (UCSF) between January 2016 and November 2021, we developed and compared four supervised ML models (logistic regression, random forest, gradient boosting, and XGBoost) to predict 30-day readmissions for adults admitted to a UCSF hospital. (3) Results: Of 147,358 inpatient encounters, 20,747 (13.9%) patients were readmitted within 30 days of discharge. The final model selected was XGBoost, which had an area under the receiver operating characteristic curve of 0.783 and an area under the precision-recall curve of 0.434. The most important features by Shapley Additive Explanations were days since last admission, discharge department, and inpatient length of stay. (4) Conclusions: We developed and internally validated a supervised ML model to predict 30-day readmissions in a US-based healthcare system. This model has several advantages including state-of-the-art performance metrics, the use of clinical data, the use of features available within 24 h of discharge, and generalizability to multiple disease states.

Multiple hospital organisational factors are associated with adverse patient outcomes post-hip fracture in England and Wales: the REDUCE record-linkage cohort study

OBJECTIVES

Despite established standards and guidelines, substantial variation remains in the delivery of hip fracture care across the United Kingdom. We aimed to determine which hospital-level organisational factors predict adverse patient outcomes in the months following hip fracture.

METHODS

We examined a national record-linkage cohort of 178,757 patients aged ≥60 years who sustained a hip fracture in England and Wales in 2016-19. Patient-level hospital admissions datasets, National Hip Fracture Database and mortality data were linked to metrics from 18 hospital-level organisational-level audits and reports. Multilevel models identified organisational factors, independent of patient case-mix, associated with three patient outcomes: length of hospital stay (LOS), 30-day all-cause mortality and emergency 30-day readmission.

RESULTS

Across hospitals mean LOS ranged from 12 to 41.9 days, mean 30-day mortality from 3.7 to 10.4% and mean readmission rates from 3.7 to 30.3%, overall means were 21.4 days, 7.3% and 15.3%, respectively. In all, 22 organisational factors were independently associated with LOS; e.g. a hospital's ability to mobilise >90% of patients promptly after surgery predicted a 2-day shorter LOS (95% confidence interval [CI]: 1.2-2.6). Ten organisational factors were independently associated with 30-day mortality; e.g. discussion of patient experience feedback at clinical governance meetings and provision of prompt surgery to >80% of patients were each associated with 10% lower mortality (95%CI: 5-15%). Nine organisational factors were independently associated with readmissions; e.g. readmissions were 17% lower if hospitals reported how soon community therapy would start after discharge (95%CI: 9-24%).

CONCLUSIONS

Receipt of hip fracture care should be reliable and equitable across the country. We have identified multiple, potentially modifiable, organisational factors associated with important patient outcomes following hip fracture.

Indirect effect of the COVID-19 pandemic on hospital mortality in patients with hip fracture: a competing risk survival analysis using linked administrative data

BACKGROUND

Hip fracture is a leading cause of disability and mortality among older people. During the COVID-19 pandemic, orthopaedic care pathways in the National Health Service in England were restructured to manage pressures on hospital capacity. We examined the indirect consequences of the pandemic for hospital mortality among older patients with hip fracture, admitted from care homes or the community.

METHODS

Retrospective analysis of linked care home and hospital inpatient data for patients with hip fracture aged 65 years and over admitted to hospitals in England during the first year of the pandemic (1 March 2020 to 28 February 2021) or during the previous year. We performed survival analysis, adjusting for case mix and COVID-19 infection, and considered live discharge as a competing risk. We present cause-specific hazard ratios (HR_CS) for the effect of admission year on hospital mortality risk.

RESULTS

During the first year of the pandemic, there were 55 648 hip fracture admissions: a 5.2% decrease on the previous year. 9.5% of patients had confirmed or suspected COVID-19. Hospital stays were substantially shorter (p<0.05), and there was a higher daily chance of discharge (HR_CS 1.40, 95% CI 1.38 to 1.41). Overall hip fracture inpatient mortality increased (7.2% in 2020/2021 vs 6.4% in 2019/2020), but patients without concomitant COVID-19 infection had lower mortality rates compared with the year before (5.3%). Admission during the pandemic was associated with a 11% increase in the daily risk of hospital death for patients with hip fracture (HR_CS 1.11, 95% CI 1.05 to 1.16).

CONCLUSIONS

Although COVID-19 infections led to increases in hospital mortality, overall hospital mortality risk for older patients with hip fracture remained largely stable during the first year of the pandemic.

Developing a minimum common dataset for hip fracture audit to help countries set up national audits that can support international comparisons

AIMS

The aim of this study was to explore current use of the Global Fragility Fracture Network (FFN) Minimum Common Dataset (MCD) within established national hip fracture registries, and to propose a revised MCD to enable international benchmarking for hip fracture care.

METHODS

We compared all ten established national hip fracture registries: England, Wales, and Northern Ireland; Scotland; Australia and New Zealand; Republic of Ireland; Germany; the Netherlands; Sweden; Norway; Denmark; and Spain. We tabulated all questions included in each registry, and cross-referenced them against the 32 questions of the MCD dataset. Having identified those questions consistently used in the majority of national audits, and which additional fields were used less commonly, we then used consensus methods to establish a revised MCD.

RESULTS

A total of 215 unique questions were used across the ten registries. Only 72 (34%) were used in more than one national audit, and only 32 (15%) by more than half of audits. Only one registry used all 32 questions from the 2014 MCD, and five questions were only collected by a single registry. Only 21 of the 32 questions in the MCD were used in the majority of national audits. Only three fields (anaesthetic grade, operation, and date/time of surgery) were used by all ten established audits. We presented these findings at the Asia-Pacific FFN meeting, and used an online questionnaire to capture feedback from expert clinicians from different countries. A draft revision of the MCD was then presented to all 95 nations represented at the Global FFN conference in September 2021, with online feedback again used to finalize the revised MCD.

CONCLUSION

The revised MCD will help aspirant nations establish new registry programmes, facilitate the integration of novel analytic techniques and greater multinational collaboration, and serve as an internationally-accepted standard for monitoring and improving hip fracture services. Cite this article: Bone Joint J 2022;104-B(6):721-728.

Variation in revascularisation use and outcomes of patients in hospital with acute myocardial infarction across six high income countries: cross sectional cohort study

OBJECTIVES

To compare treatment and outcomes for patients admitted to hospital with a primary diagnosis of ST elevation or non-ST elevation myocardial infarction (STEMI or NSTEMI) in six high income countries with very different healthcare delivery systems.

DESIGN

Retrospective cross sectional cohort study.

SETTING

Patient level administrative data from the United States, Canada (Ontario and Manitoba), England, the Netherlands, Israel, and Taiwan.

PARTICIPANTS

Adults aged 66 years and older admitted to hospital with STEMI or NSTEMI between 1 January 2011 and 31 December 2017.

OUTCOMES MEASURES

The three categories of outcomes were coronary revascularisation (percutaneous coronary intervention or coronary artery bypass graft surgery), mortality, and efficiency (hospital length of stay and 30 day readmission). Rates were standardised to the age and sex distribution of the US acute myocardial infarction population in 2017. Outcomes were assessed separately for STEMI and NSTEMI. Performance was evaluated longitudinally (over time) and cross sectionally (between countries).

RESULTS

The total number of hospital admissions ranged from 19 043 in Israel to 1 064 099 in the US. Large differences were found between countries for all outcomes. For example, the proportion of patients admitted to hospital with STEMI who received percutaneous coronary intervention in hospital during 2017 ranged from 36.9% (England) to 78.6% (Canada; 71.8% in the US); use of percutaneous coronary intervention for STEMI increased in all countries between 2011 and 2017, with particularly large rises in Israel (48.4-65.9%) and Taiwan (49.4-70.2%). The proportion of patients with NSTEMI who underwent coronary artery bypass graft surgery within 90 days of admission during 2017 was lowest in the Netherlands (3.5%) and highest in the US (11.7%). Death within one year of admission for STEMI in 2017 ranged from 18.9% (Netherlands) to 27.8% (US) and 32.3% (Taiwan). Mean hospital length of stay in 2017 for STEMI was lowest in the Netherlands and the US (5.0 and 5.1 days) and highest in Taiwan (8.5 days); 30 day readmission for STEMI was lowest in Taiwan (11.7%) and the US (12.2%) and highest in England (23.1%).

CONCLUSIONS

In an analysis of myocardial infarction in six high income countries, all countries had areas of high performance, but no country excelled in all three domains. Our findings suggest that countries could learn from each other by using international comparisons of patient level nationally representative data.

High ward occupancy, bedspacing, and 60 day mortality for patients with myocardial infarction, stroke, and heart failure

Aims

To study the consequences of crowded wards among patients with cardiovascular disease.

Methods and results

This is a cohort study among 201 801 patients with 258 807 admissions who were acutely admitted for myocardial infarction (N = 107 895), stroke (N = 87 336), or heart failure (N = 63 576) to any Norwegian hospital between 2008 and 2016. The ward admitting most patients with the given clinical condition was considered a patient's home ward. We compared patients with the same condition admitted when home ward occupancy was different, at the same hospital and during comparable time periods. Occupancy was standardized such that a one‐unit difference corresponded to the interquartile range in occupancy in the given month. One interquartile increase in home ward occupancy was associated with 7% higher odds of admission to an alternate ward [odds ratio (OR) 1.07, 95% confidence interval (CI) 1.09 to 1.11], and length of stay was shorter (−0.10 days, 95% CI −0.18 to −0.09). Patients with heart failure had 15% higher odds of admission to alternate wards (OR 1.15, 95% CI 1.08 to 1.23) and increased mortality [hazard ratio (HR) 1.08, 95% CI 1.03 to 1.15]. We found no apparent effect on mortality for patients with myocardial infarction (HR 0.99, 95% CI 0.94 to 1.05) or stroke (HR 1.00, 95% CI 0.96 to 1.05).

Conclusions

Patients with heart failure had higher risk of admission to alternate wards when home ward occupancy was high. These patients may be negatively affected by full wards.

A methodology for identifying high-need, high-cost patient personas for international comparisons

OBJECTIVE

To establish a methodological approach to compare two high-need, high-cost (HNHC) patient personas internationally.

DATA SOURCES

Linked individual-level administrative data from the inpatient and outpatient sectors compiled by the International Collaborative on Costs, Outcomes, and Needs in Care (ICCONIC) across 11 countries: Australia, Canada, England, France, Germany, the Netherlands, New Zealand, Spain, Sweden, Switzerland, and the United States.

STUDY DESIGN

We outline a methodological approach to identify HNHC patient types for international comparisons that reflect complex, priority populations defined by the National Academy of Medicine. We define two patient profiles using accessible patient-level datasets linked across different domains of care-hospital care, primary care, outpatient specialty care, post-acute rehabilitative care, long-term care, home-health care, and outpatient drugs. The personas include a frail older adult with a hip fracture with subsequent hip replacement and an older person with complex multimorbidity, including heart failure and diabetes. We demonstrate their comparability by examining the characteristics and clinical diagnoses captured across countries.

DATA COLLECTION/EXTRACTION METHODS

Data collected by ICCONIC partners.

PRINCIPAL FINDINGS

Across 11 countries, the identification of HNHC patient personas was feasible to examine variations in healthcare utilization, spending, and patient outcomes. The ability of countries to examine linked, individual-level data varied, with the Netherlands, Canada, and Germany able to comprehensively examine care across all seven domains, whereas other countries such as England, Switzerland, and New Zealand were more limited. All countries were able to identify a hip fracture persona and a heart failure persona. Patient characteristics were reassuringly similar across countries.

CONCLUSION

Although there are cross-country differences in the availability and structure of data sources, countries had the ability to effectively identify comparable HNHC personas for international study. This work serves as the methodological paper for six accompanying papers examining differences in spending, utilization, and outcomes for these personas across countries.