Article Text

Download PDFPDF

Primary care
HOSPITAL Score, LACE Index and LACE+ Index as predictors of 30-day readmission in patients with heart failure
Free
  1. Abdisamad M Ibrahim1,
  2. Cameron Koester1,
  3. Mohammad Al-Akchar2,
  4. Nitin Tandan1,
  5. Manjari Regmi1,
  6. Mukul Bhattarai2,
  7. Basma Al-Bast1,
  8. Abhishek Kulkarni2,
  9. Robert Robinson1
  1. 1 Internal Medicine, SIU School of Medicine, Springfield, Illinois, USA
  2. 2 Division of Cardiology, SIU School of Medicine, Springfield, Illinois, USA
  1. Correspondence to Dr Abdisamad M Ibrahim, Internal Medicine, SIU School of Medicine, Springfield, IL 62781, USA; aibrahim0912{at}gmail.com

Abstract

This study aimed to evaluate the accuracy of the HOSPITAL Score (Haemoglobin level at discharge, Oncology at discharge, Sodium level at discharge, Procedure during hospitalization, Index admission, number of hospital admissions, Length of stay) LACE index (Length of stay, Acute/emergent admission, Charlson comorbidy index score, Emerency department visits in previous 6 months) and LACE+ index in predicting 30-day readmission in patients with diastolic dysfunction. Heart failure remains one of the most common hospital readmissions in adults, leading to significant morbidity and mortality. Different models have been used to predict 30-day hospital readmissions. All adult medical patients discharged from the SIU School of Medicine Hospitalist service from 12 June 2016 to 12 June 2018 with an International Classification of Disease, 10th Revision, Clinical Modification diagnosis of diastolic heart failure were studied retrospectively to evaluate the performance of the HOSPITAL Score, LACE index and LACE+ index readmission risk prediction tools in this patient population. Of the 730 patient discharges with a diagnosis of heart failure with preserved ejection fraction (HFpEF), 692 discharges met the inclusion criteria. Of these discharges, 189 (27%) were readmitted to the same hospital within 30 days. A receiver operating characteristic evaluation showed C-statistic values to be 0.595 (95% CI 0.549 to 0.641) for the HOSPITAL Score, 0.551 (95% CI 0.503 to 0.598) for the LACE index and 0.568 (95% CI 0.522 to 0.615) for the LACE+ index, indicating poor specificity in predicting 30-day readmission. The result of this study demonstrates that the HOSPITAL Score, LACE index and LACE+ index are not effective predictors of 30-day readmission for patients with HFpEF. Further analysis and development of new prediction models are needed to better estimate the 30-day readmission rates in this patient population.

  • heart failure readmission
  • readmission
  • HOSPITAL Score
  • LACE Index
  • LACE+ Index

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Key messages

What is already known about this subject?

  • The HOSPITAL Score, LACE index and LACE+ index are all validated tools for 30-day hospital readmissions in the general population.

  • Hospital readmission is one of the leading causes of healthcare cost in low-income countries.

  • Heart failure remains one of the leading causes of hospital readmissions, and there are no good prediction models to predict readmissions.

What are the new findings?

  • Our study demonstrates that the HOSPITAL Score, LACE index and LACE+ index did not predict 30-day readmission for patients with heart failure with preserved ejection fraction (HFpEF).

How might it impact on clinical practice in the foreseeable future?

  • We believe our findings in this article are of clinical significance. Heart failure remains one of the leading causes of hospital readmissions in the USA. There are numerous protocols and prediction models that were developed in order to identify patients who are at risk of readmission. The Hospital Score, LACE index and LACE+ index are predictions models that are used globally for hospital readmission. Our study could not replicate this for patients with HFpEF. Further analysis and prediction models are needed to better estimate the 30-day readmission rates in patients with HFpEF to facilitate targeted interventions to improve hospital outcomes.

Background

Heart failure is one of the leading causes of hospital readmission and accounted for over 1 million hospitalisations each year from 2000 to 2010. This rate of heart failure readmissions leads to an extraordinary expense of an estimated over 30 billion US dollars in the year 2012, according to one report.1 More importantly, heart failure has an enormous impact on mortality as the same report documents an approximately 50% 5-year mortality in patients diagnosed with heart failure.1 Thirty-day hospital readmission rates have been used as a metric to evaluate the quality of care provided by hospitals across the country. After the implementation of the Affordable Care Act, this metric has subsequently been used to determine reimbursement for hospitals and hospital systems.2 As a result of the shift from a volume-defined reimbursement model to a value-based reimbursement system with the Centers for Medicare and Medicaid Services, hospitals have developed many interventions to help achieve these core measures and to reduce readmission rates.3

One such intervention that has gained traction over recent years has been predictive models to identify patients with a higher risk of hospital readmission. It has become increasingly more common for hospitals to leverage a robust data set that can be analysed within an electronic medical record to aid in the identification of patients in which a specific intervention may prevent readmission rates. Multiple scoring systems have been proposed, analysed and validated to help predict patients at high risk of readmission. Of these, the most cited readmission risk assessment tools are the HOSPITAL Score (Haemoglobin level at discharge, Oncology at discharge, Sodium level at discharge, Procedure during hospitalization, Index admission, number of hospital admissions, Length of stay),4 LACE index5 (Length of stay, Acute/emergent admission, Charlson comorbidy index score, Emerency department visits in previous 6 months) and LACE+6 index. The HOSPITAL Score is a simple tool that uses seven routinely gathered data points to predict readmission rates and has been validated using a large population internationally.7 Likewise, the LACE index and LACE+ index use similar, readily available data points and were validated in a Canadian population in a study by van Walraven et al.5 6 A national study by Garrison et al showed that these two prediction tools performed similarly,8 whereas two studies done outside the USA actually showed that the HOSPITAL Score might outperform the LACE index.9 10

It has been estimated that heart failure with preserved ejection fraction (HFpEF) is present in up to 50% of patients, and as awareness of this specific disease entity grows, the incidence will likely increase. As this diagnosis becomes more prevalent in the population, it becomes imperative to recognise that the rates of hospitalisation and death in patients with HFpEF approach that of patients with heart failure with reduced ejection fraction.11 For patients with systolic dysfunction who are at high risk of readmission, there are well known and thoroughly studied interventions that have been proven to improve survival and decrease readmission rates. Unfortunately, the same cannot be said about HFpEF. These patients are known to account for a significant percentage of readmissions; however, it is difficult to predict which patients are at higher risk and what interventions may reduce readmission rates. Our study aimed to assess the accuracy of the HOSPITAL Score, LACE index and LACE+ index in predicting 30-day readmission in patients with HFpEF.

Methods

All medical patients 18 years of age or older, discharged from Memorial Medical Centre by the SIU School of Medicine (SOM) Academic Hospitalist service from 12 June 2016 to 12 June 2018 with a primary or secondary International Classification of Disease, 10th Revision, Clinical Modification diagnosis of diastolic heart failure and a documented left ventricular ejection fraction of ≥50% were studied retrospectively to evaluate the effectiveness and accuracy of several readmission prediction tools, namely, the HOSPITAL Score, LACE index and LACE+ index. Exclusion criteria were leaving the hospital against medical advice, transfer to another acute care hospital, discharging with hospice or death. The evaluated outcome was all-cause readmission within 30 days of discharge from the index hospitalisation.

Patient information from other local and regional facilities and practices within the same facility were not available for analysis.

Data on age, gender, International Classification of Diseases (ICD) diagnosis codes, emergency department visits in the previous 6 months, length of index hospitalisation, hospital readmission within 30 days, discharge medications, left ventricular ejection fraction, as well as additional variables included in the HOSPITAL Score, LACE index and LACE+ index were obtained from the electronic health record in a deidentified manner for analysis. Because of the deidentified nature of the data, exported information contains a binary variable indicating if the patient was readmitted within 30 days. The cause for this new admission is not linked to the prior admission.

There is no distinct oncology admitting service at the study hospital, as patients hospitalised for oncological diagnoses are admitted through the SIU SOM Hospitalist service with the support of oncology consulting services. This is standard practice at the study hospital for a hospitalist to admit patients and subsequently consult the haematology/oncology service. To address the increased risk of readmission in patients admitted for primarily oncological diagnoses seen in the HOSPITAL Score, this study classified patients with oncology-related diagnosis ICD codes as discharged from an oncology service.

Statistical analysis

The HOSPITAL Score, LACE index and LACE+ index performance for predicting the risk of all-cause hospital readmission within 30 days of discharge were evaluated. Qualitative variables were compared using Pearson χ2 test or Fisher's exact test and were reported as frequency (%). Quantitative variables were compared using the non-parametric Mann-Whitney U test and were reported as mean±SD.

The HOSPITAL Score, LACE index and LACE+ index scores were calculated for each admission. HOSPITAL Scores of 0–4 points were classified as low risk of readmission (5%); 5–6 points were classified as intermediate risk (10%); and 7 or more points were classified as high risk (20%) based on the initial validation study of the HOSPITAL Score.7 LACE indexes ranged from 0 to 19, with an expected probability for readmission of 0%–43.7% based on the initial validation study of the LACE index score.5 The LACE+ index score ranged from −10 to >109 and is an improved version of the LACE index score, which incorporates population data regarding the reason for hospital readmission.6 Brier Scores were calculated for each readmission risk predictor.

Most statistical analyses were performed using SPSS V.25. The Brier Score was calculated with R V.3.4.2 (R Foundation for Statistical Computing, Vienna, Austria). Two-sided p values of <0.05 were considered significant.

Results

During the study period, 730 discharges with a diagnosis of HFpEF were recorded for the SIU SOM Hospitalist service. The analysis includes data for the 692 discharges that met the inclusion criteria (figure 1). Of these discharges, 189 (27%) were readmitted to the same hospital within 30 days.

Figure 1

Study flow diagram. HFpEF, heart failure with preserved ejection fraction.

The baseline characteristics of age, gender, length of stay, left ventricular ejection fraction and emergency department visits in the last 6 months were not statistically significant between the patients who were readmitted and those who were not readmitted (table 1). Medical comorbidities were similar between the groups, except that paralysis was found to be less common in the group that was not readmitted. Low haemoglobin and high serum creatinine were found at higher rates in the group readmitted to the hospital within 30 days of discharge. Patients who were readmitted had higher HOSPITAL Score, LACE index and LACE+ index values. These differences were statistically significant.

Table 1

Baseline characteristics of the study population by 30-day readmission status

A multivariate logistic regression of potential risk factors for hospital readmission within 30 days of discharge showed the HOSPITAL Score (OR 1.24 (95% CI 1.08 to 1.42), p=0.002) and LACE+ index score (OR 1.01 (95% CI 1.00 to 1.03), p=0.091) to be weak but statistically significant predictors of hospital readmission within 30 days of discharge. The LACE index was not found to be a significant independent predictor of hospital readmission in regression analysis (table 2).

Table 2

Multivariate logistic regression of potential risk factors for hospital readmission within 30 days of discharge

A receiver operating characteristic evaluation of the HOSPITAL Score showed a C-statistic of 0.595 (95% CI 0.549 to 0.641); that of LACE index was 0.551 (95% CI 0.503 to 0.598); and that of LACE+ index was 0.568 (95% CI 0.522 to 0.615), indicating poor specificity in predicting 30-day readmission (figure 2 and table 3). Brier scores for the HOSPITAL Score (0.245), LACE index (0.215) and LACE+ index (0.201) readmission tools showed marginal overall performance.

Table 3

Receiver operating characteristic data on readmission risk prediction tools

Figure 2

Receiver operating characteristic curves for readmission risk assessment scores for patients with HFpEF. HFpEF, heart failure with preserved ejection fraction. HOSPITAL(Haemoglobin level at discharge, Oncology at discharge, Sodium level at discharge, Procedure during hospitalization, Index admission, Number of hospital admissions, Length of stay). LACE (Length of stay, Acute/emergent admission, Charlson Comorbidity Index score, Emergency department visits in previous 6 months).

Discussion

In this single-institution study, we found that general-purpose readmission risk prediction tools, including the HOSPITAL Score, LACE index and LACE+ index, were ineffective at predicting the 30-day all-cause readmission risk of patients with HFpEF. This is in stark contrast to the performance of these validated risk prediction tools in medically diverse populations.4–6 Our study questions the utility of using general-purpose readmission risk prediction scores for patients with HFpEF and highlights the importance of developing HFpEF-specific readmission risk prediction tools.

The readmission rate in this study is high (27%) but is comparable with previous results at this centre (12%–27%).12–15 Local factors that may contribute to this higher than expected rate of readmission are a lack of a multidisciplinary heart failure management clinic and a high proportion of patients with negative social determinants of health, such as poverty, poor healthcare access and rural residence.

Hospital readmission is one of the leading problems facing the US healthcare system. The Centers for Medicare and Medicaid Services and other payers increasingly emphasise reduction of readmission rates, improvement of quality of care and reduction of cost.12 The Agency for Healthcare Research and Quality finds that nearly one in five of all hospital patients covered by Medicare are readmitted within 30 days, costing $15 billion a year.12 With heart failure being one of the top reasons for admission of adults aged more than 65 years with associated cost, morbidity and mortality, it is vital to address and attempt to predict readmission in order to reduce cost, morbidity and mortality.16

While many of readmission prediction scores, including HOSPITAL Score, LACE index and LACE+ index, have been validated in predicting readmission of patients in general, there have been only limited studies addressing their utility in certain patient populations such as patients with heart failure. Yazdan-Ashoori et al looked at 30-day readmission in patients with heart failure using the LACE index and found that higher LACE index score predicted readmission when threshold was increased to 13.17 An Australian group of researchers developed a readmission predictive model and found it to be effective in predicting 30-day readmission in patients with heart failure, and further stratified them into HFpEF and heart failure with reduced ejection fraction.18 The model incorporated standard clinical and administrative data, sociodemographic and socioeconomic status, and mental health.18 This multidisciplinary approach to readmission prediction has not been incorporated in HOSPITAL Score, LACE index and LACE + index.

The performance of the HOSPITAL Score in this population was considerably worse (0.595 vs 0.72) than what was reported in the multicentre international validation study of the HOSPITAL Score and the initial derivation study. Previous work in the same practice setting showed the performance of the HOSPITAL Score in a medically diverse population to be similar to the validation study (0.75 vs 0.71).13 The prevalence of congestive heart failure of any type was 22% in the derivation study for the HOSPITAL Score, which differs considerably from the population in this study. This difference alone may be responsible for the lower performance of the HOSPITAL Score for patients with HFpEF.

The LACE index was developed by van Walraven et al and has been validated in multiple practice settings worldwide.5 Performance of the LACE index in this study was poor (C-statistic of 0.551 vs 0684) when compared with the derivation study by van Walraven and colleagues. The use of the LACE index for patients with any type of congestive heart failure showed a C-statistic of 0.56, which is comparable with the results in this study. This performance is only slightly better than chance in both studies.

The LACE+ index was developed from a large, population-based sample in Ontario, Canada, and was highly discriminate (C-statistic 0.771, 95% CI 0.767 to 0.775) and performed better than the LACE index.6 Studies using the LACE+ index in specific populations, such as patients admitted with brain tumours, did not predict 30-day readmissions. In a study that looked at 30-day readmission in a brain tumour population, the LACE+ index had poor specificity in predicting 30-day readmission (C-statistic=0.58).19 In our patients with HFpEF, we have achieved similar results as those of the brain tumour population. This shows that the LACE+ index might not be able to accurately predict the 30-day readmission for specific populations.

Limitations

There are several limitations to consider when analysing the results of this study. First, this is a retrospective, single-centre study with a small sample size. Second, the LACE+ index was studied and validated in Ontario, Canada, a country with different metrics and readmission rates than the USA. Furthermore, this study looked at all-cause of readmission and did not specifically evaluate if readmissions were related to HFpEF. Third, adherence to medication and diet was not assessed in this study, a factor that may play a role in heart failure readmission. More studies are needed to evaluate readmission scores in patients with HFpEF.

Conclusion

Our study demonstrates that the HOSPITAL Score, LACE index and LACE+ index did not predict 30-day readmission for patients with HFpEF. Further analysis and prediction models are needed to better estimate the 30-day readmission rates in this patient population. Using a model that will accurately predict the 30-day readmission risk for patients with HFpEF will facilitate targeted interventions such as education, chronic disease management and self-care management with the goal of improving the quality of healthcare for patients with HFpEF.

References

Footnotes

  • Twitter @aibrahim0912

  • Contributors All authors contributed to the preparation of the manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval Institutional review board review for this study was obtained from the Springfield Committee for Research Involving Human Subjects. This study was determined not to meet the criteria for research involving human subjects according to 45 CFR 46.101 and 45 CFR 46.102.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.