This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Interactive Journal of Medical Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.i-jmr.org/, as well as this copyright and license information must be included.
The LACE index was designed to predict early death or unplanned readmission after discharge from hospital to the community. However, implementing the LACE tool in real time in a teaching hospital required practical unavoidable modifications.
The purpose of this study was to validate the implementation of a modified LACE index (LACE-rt) and test its ability to predict readmission risk using data in a hospital setting.
Data from the Canadian Institute for Health Information’s Discharge Abstract Database (DAD), the National Ambulatory Care Reporting System (NACRS), and the hospital electronic medical record for one large community hospital in Toronto, Canada, were used in this study. A total of 3855 admissions from September 2013 to July 2014 were analyzed (N=3855) using descriptive statistics, regression analysis, and receiver operating characteristic analysis. Prospectively collected data from DAD and NACRS were linked to inpatient data.
The LACE-rt index was a fair test to predict readmission risk (C statistic=.632). A LACE-rt score of 10 is a good threshold to differentiate between patients with low and high readmission risk; the high-risk patients are 2.648 times more likely to be readmitted than those at low risk. The introduction of LACE-rt had no significant impact on readmission reduction.
The LACE-rt is a fair tool for identifying those at risk of readmission. A collaborative cross-sectoral effort that includes those in charge of providing community-based care is needed to reduce readmission rates. An eHealth solution could play a major role in streamlining this collaboration.
Unplanned hospital readmission has been a major challenge in health care worldwide [
Other tools addressing hospital readmission, such as the UK Nuffield Trust model [
While van Walraven et al developed LACE based on a secondary analysis of a multicenter, prospective cohort study of patients in 11 hospitals, our study examined the use of a modified LACE index (LACE-rt) tailored for use in real time in an active setting in the general medicine unit at a large community teaching hospital in Toronto. In order to use the LACE tool in real time to help identify those discharged patients who are at higher risk of readmission, some practical unavoidable modifications had to be made to the LACE index. Accordingly, the purpose of this study was to implement a modified LACE index in a real-time setting (hence the name LACE-rt) and examine its reliability as well as its ability to discriminate between high- and low-risk patients.
The hospital is a community teaching hospital with 426 acute care beds. Secondary data covering the period September 2013 to July 2014 were obtained from the hospital. A total of 3 datasets were provided:
Inpatient information extracted from the Canadian Institute for Health Information’s Discharge Abstract Database (DAD); it included patient identifier, encounter identifier, admission and discharge dates, location of admission, and basic demographic information such as age and sex.
Emergency department visit data extracted from the National Ambulatory Care Reporting System (NACRS).
“LACE-rt” related information extracted from the hospital electronic medical record.
The data included 7676 admissions from 6332 patients. Among these admissions, we selected those who were admitted to 1 of the 4 medicine units that implemented LACE-rt (Stroke, Acute Geriatrics, Cardiology, and Respirology and Gastrointestinal) and were assessed by a nurse using the LACE-rt tool before being discharged to home, another hospital, or a long-term care facility. The total number of admissions analyzed in our study was 3855 (N=3855).
The “L” value is calculated differently in LACE-rt than in the original LACE index. When managers at the hospital decided to implement the LACE index, they faced the practical challenge of needing to start preparing for discharge as soon as the patient is admitted; waiting until the discharge day to compute the “L” score would delay discharge planning, making the original LACE approach untenable from a practical standpoint. In LACE-rt the managers therefore decided to compute “L” based on the patient’s length of stay during the previous (instead of current) acute care admission within the last 30 days.
The attributes L, A, C, and E are computed in the same way in LACE-rt and the original LACE; their corresponding values and points are provided in
The Charlson comorbidity score (C) is calculated as follows: 1 point for history of myocardial infarction, peripheral vascular disease, cerebrovascular disease, or diabetes without complications; 2 points for congestive heart failure, chronic obstructive pulmonary disease, mild liver disease or cancer, diabetes with end-organ damage, and any tumor (including lymphoma or leukemia); 3 points for dementia or connective tissue disease; 4 points for moderate to severe liver disease or human immunodeficiency virus infection; and 6 points for metastatic cancer.
Both the original LACE index and the LACE-rt index scores range from 0 to 19, where a higher score indicates an increased chance of readmission or early death (
The LACE and LACE-rt index attributes and the corresponding values and points.
Attribute | Value | Points |
Length of staya, days | <1 | 0 |
1 | 1 | |
2 | 2 | |
3 | 3 | |
4-6 | 4 | |
7-13 | 5 | |
≥14 | 7 | |
Acute (emergent) admission | Yes | 3 |
No | 0 | |
Comorbidity (Charlson comorbidity index score) | 0 | 0 |
1 | 1 | |
2 | 2 | |
3 | 3 | |
≥4 | 5 | |
Emergency department visit (within the last 6 months) | 0 | 0 |
1 | 1 | |
2 | 2 | |
3 | 3 | |
≥4 | 4 |
aLACE: during the current admission (van Walraven et al); LACE-rt: during the last 30 days.
In this study, nurses checked the hospital’s electronic patient chart to estimate the values for “L,” “A,” “C,” and “E,” then entered those values into a software interface that computes the patient’s LACE-rt score. However, discussion with staff suggested that the extraction and recording of the “L,” “E,” and “C” values are often done quickly.
To check the data entry accuracy for the “L” and “E” components in our dataset, we computed “L” and “E” using the DAD and NACRS data, respectively, and compared the calculations from the administrative data with those values entered manually by the nurses. Even though we had a rationale for investigating the accuracy of “C,” this was not feasible as it would have required a complex time-consuming clinical assessment.
According to Statistics Canada, “non-elective return to an acute care hospital for any cause is counted as a readmission if it occurs within 30 days of the index episode of inpatient care” [
Readmission rate formula.
Statistical analyses were performed using IBM SPSS Statistics 22 (IBM Corporation). Descriptive statistical analysis was carried out describing the population’s basic demographic characteristics. On the basis of previous literature, patients with LACE-rt score of 10 or higher were defined as a high-risk group and those with a score lower than 10 were defined as the low-risk group [
This project obtained ethical approval from the hospital Research Ethics Board and all researchers obtained the “Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans Course on Research Ethics” certificate (TCPS 2: CORE).
Descriptive statistics (
Descriptive statistics: patients’ sex and age groups.
Hospital admission characteristics (N=3855) | Value | Count (%) |
Sex | Male | 1859 (48.22) |
Female | 1996 (51.78) | |
Age, years | Mean age | 74.29 |
<20 | 10 (0.26) | |
20-29 | 82 (2.13) | |
30-39 | 99 (2.57) | |
40-49 | 182 (4.72) | |
50-59 | 346 (8.98) | |
60-69 | 484 (12.56) | |
70-79 | 763 (19.79) | |
≥80 | 1889 (49.00) |
Descriptive statistics: LACE-rt elements and their corresponding frequencies.
LACE-rt elements (N=3855) | Value | Count (%) | |
Length of stay in the last 30 days | Less than 1 day | 2559 (66.38) | |
1 day | 648 (16.81) | ||
2 days | 148 (3.84) | ||
3 days | 94 (2.44) | ||
4-6 days | 179 (4.64) | ||
7-13 days | 134 (3.48) | ||
≥14 days | 93 (2.41) | ||
Acute (emergent) admission | Yes | 163 (4.23) | |
No | 3692 (95.77) | ||
Comorbidity (Charlson comorbidity index score) | 0 | 1159 (30.06) | |
1 | 979 (25.40) | ||
2 | 625 (16.21) | ||
3 | 559 (14.50) | ||
≥4 | 533 (13.83) | ||
Emergency department visit (within the last 6 months) | 0 visits | 1776 (46.07) | |
1 visit | 1025 (26.59) | ||
2 visits | 541 (14.03) | ||
3 visits | 246 (6.38) | ||
≥4 visits | 267 (6.93) |
Differences between the high- and low-risk groups were compared in a cross-tabulation. The readmission rate for the low-risk group was 10.6% compared with 23.9% for the high-risk group. The chi-square analysis indicated that there is a statistically significant difference between the 2 groups (χ21=65.5, N=3855,
Readmission rates by LACE-rt scores.
The binary logistic regression analysis revealed an odds ratio Exp(B)=2.648,
Of the readmissions, 11.9% were for female patients and 12.4% were for male patients. An analysis of the readmission rates by sex indicated that there is no significant difference between the 2 groups (χ21=.3,
An analysis of the readmission rates by age groups indicated that the readmission rates were 10%, 7.3%, 5.1%, 11.0%, 7.2%, 11.2%, 11.4%, and 14.3% for the age groups <20 years, 20s, 30s, 40s, 50s, 60s, 70, and ≥80 years, respectively (χ2=23.6,
We used the nonparametric Mann-Whitney
To assess the accuracy of the LACE-rt index in predicting readmission, we conducted an ROC curve analysis. The ROC analysis was statistically significant (
Our results suggest that the LACE-rt index can predict readmission with a reasonable degree of accuracy and that a threshold of 10 is useful for differentiating between patients who are at high versus low risk of readmission. Our results further showed that the readmission rates at the hospital are 10.6% and 23.9% for the low-risk and high-risk groups, respectively. These results are consistent with Gruneir and colleagues [
Current discussion of readmissions in the literature often focuses on demographic and socioeconomic status (SES) factors that explain readmission in specified populations (eg, patients with congestive heart failure). However, demographic and SES predictors are not routinely collected by hospitals; moreover, hospitals would benefit more from tools that work across multiple conditions rather than tools that are specific to certain health conditions. Van Walraven and colleagues recently improved the predictive power of LACE by incorporating age and sex into LACE+ [
The original LACE index required a modification in order to implement it in a hospital setting. As mentioned above, the “Length of stay” could not be implemented in the manner originally designed and had to be modified to measure patients’ length of stay in the last 30 days instead of during the current admission. However, our results suggest that the LACE-rt index remains useful for identifying patients at high risk of readmission. In our sample, higher LACE-rt scores were associated with higher readmission rates. Moreover, the chi-square analysis indicated that patients with a LACE-rt score of ≥10 were significantly more likely to be readmitted than those with a LACE-rt score of <10. This is particularly interesting given no demographic or SES factors were used in these analyses—although most admitted patients we studied were elderly, the LACE-rt tool was still able to distinguish between the high- and low-risk groups.
The ROC analysis showed a C statistic that is lower than the one found in the population studied by van Walraven et al (C statistic .684, 95% CI 0.679-0.691) [
Our analysis showed that the LACE-rt implementation itself had no effect on readmission rates. Although hospitals can use the LACE-rt tool to identify patients at high risk of readmission, it is unlikely that use of this type of tool
Our study was not able to take early death into account. Patients who died would appear as patients with no readmission in our dataset. It is therefore likely that our data underestimate actual readmission rates. The fact that we were only able to examine readmissions to the same hospital further contributes to underestimation of our readmission rates.
These limitations should not detract from the purpose of this study, which was to examine utility of the LACE-rt index as a tool for quality improvement. Indeed, methodological concerns related to the measurement of readmission rates have led to suggestions that readmission data are better suited to quality improvement than accountability purposes [
Our main research aim was to examine the extent to which the LACE-rt index could be used as a predictor of readmission in real time in a large community hospital setting. Our results suggest the LACE-rt index can be practically applied and is a good predictor of readmission. We suggest exploring ways to incorporate basic demographic and socioeconomic data into the tool. We already know that geography has an impact on patient’s health [
Discharge Abstract Database
National Ambulatory Care Reporting System
receiver operating characteristic
socioeconomic status
Scottish Patients at Risk of Readmission
None declared.