Impact of the Implementation of a Short Stay Observation Unit Associated With Admitter‐Rounder Model Onto Other Internal Medicine Units in a French University Hospital

ABSTRACT

Introduction

Few data on the impact of specific interventions against Emergency Rooms ‘or Hospitals overcrowding are available in France.

Methods

In the present report, we retrospectively investigated the impact of the implementation of a short‐stay observation unit associated with the admitter‐rounder model, especially onto the other in‐patient internal medicine units in a French University Hospital.

Results

During the first 100 days, 242 patients were admitted into the short‐stay observation unit. The median length of stay (LOS) was 5 days, but it was lower during the first and third parts with the admitter‐rounder model than during the second part without: 4 versus 6 days (p = 0.007).

Internal medicine bed‐spaced patients accounted for 19.5% of the bed‐spaced patients during the study period versus 28.1% during the same period the previous year (p = 0.04). The median LOS increased significantly in two units of internal medicine: 10 versus 8 days (p = 0.005) and 10 versus 8 days (p = 0.01) during the study period versus the same period the previous year, respectively.

Discussion

The reduced number of Internal Medicine bed‐spaced patients, the reduced LOS of patients in short‐stay observation unit when associated with the admitter‐rounder model and the increase of LOS among some of the in‐patient internal medicine units observed in this study should be evaluated elsewhere.

1. Introduction

Emergency rooms (ER) or Hospitals overcrowding is associated with increased mortality [1]. During the last decade, multiple interventions against these latter have been evaluated in North‐American Hospitals: admitter‐rounder model [2, 3], active bed management [4, 5], observation units [6], boarder services [7] and bed‐spacing [8].

ER overcrowding has been studied in France [9, 10, 11], where the the efficiency of the universalist healthcare system has been criticized [12]. But, few data on the impact of the aforementioned interventions are available [13, 14].

In the present study, we specifically investigated in a newly created internal medicine SSU (i) the impact of an associated admitter‐rounder model onto the SSU patients outcome and Lenght of stay (LOS) (ii) the impact of the implementation of this SSU plus admitter‐rounder model onto the proportion of bed‐spaced patients and the LOS in other in‐patient internal medicine units.

2. Methods

A newly created Internal Medicine 12‐bed SSU named: ‘Post Emergency room medical unit’ (in French UPUM) was opened in Reims University Hospital, an 1100‐bed teaching hospital in North‐eastern France (with four in‐patient internal medicine units, a previously implemented active bed management team and bed‐spacing procedures), at the location of the previous in‐patient dermatology ward on 6 November 2023. The target patients of UPUM should (i) be aged less than 75 years, (ii) come from the ER and (iii) have a presumed LOS lower than 5–7 days, regardless of the reason for admission.

The study period corresponded to the first 100 days of UPUM (6 November 2023—14 February 2024). During the first and the third parts of the study period (6 November 2023—9 December 2023 and 13 January 2024–14 February 2024), some of the admissions were made by a senior consultant (external admitter: Y.N.G.) during office hours whereas the UPUM medical staff (rounders: M.H. and A.B.) was the same during the whole study period (Appendix A).

All the patients hospitalised in UPUM during the study period were enroled. Letters informing that data would be collected from medical records for research purpose were sent to the patients who were still alive, on 12 March 2024, in accordance with French legislation. Three patients refused the collection of data from their medical records. Among previously deceased patients, none had previously objected to the further use of their medical data during their hospitalisation before they died. Thus, the data (age, point of origin, LOS, and outcome [discharge, transfer to another medical or surgical acute care unit or death]) from 242 patients were secondarily extracted from medical records on 12 April and data confidentiality were preserved throughout this internal study (Reims University Hospital GDPR register number MR00416102023), in accordance with the principles of the Declaration of Helsinki.

The data concerning the ER or the bed‐spaced patients and concerning the LOS of other in‐patient internal medicine units were provided by the bed management team (L.L.) and by the medical information department (C.V.) respectively.

Quantitative variables usually expressed as median +interquartile range were compared using the Mann–Whitney U test and qualitative variables expressed as percentages were compared using Fischer's Exact Test or Pearson's test if applicable. A p value < 0.05 was considered as significant. Statistical analyses were performed using EpiInfo 7.2.2.6 (CDC, Atlanta, Georgia, United States) and MedCalc version 22.007 (MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org; 2023) softwares.

3. Results

3.1. UPUM Patients and Length of Stay

During the first 100 days, 242 patients were admitted to UPUM (124 male [51.2%], median age 67 [18;47;76;100]). The median LOS in UPUM was 5 days [1;3;7;26] and the correlation between age and LOS in UPUM is shown in Figure 1A (r = 0.26). One hundred and eighty‐nine out of the 242 (78.1%) patients were admitted directly from the ER while 86, 67 and 89 patients had been admitted respectively during each part of the study period (first part 6 November 2023–9 December 2023, second part 10 December 2023—12 January 2024 and third part 13 January 2024–14 February 2024). Sixty‐one out of the 89 patients (68.5%) admitted during the third part of the study period, came from ER versus 128 out of the 153 patients (83.6%) admitted during the first and the second parts (p = 0.0006). One hundred and eighty‐three (75.6%), and 57 (23.6%) out of the 242 patients were discharged and transferred to another medical or surgical acute care unit, respectively. Two patients (0.8%) died while hospitalised in UPUM.

Figure 1.

(A) Correlation between the age of the admitted patients (years) and the length of stay in post‐emergency room medical unit (UPUM LOS) (days) by linear regression (MedCalc version 22.007) (r = 0.26). (B) Left. Number of discharged patients among all admitted patients during the first and third part of the study period with admitter‐rounder model and the second part without (p = 0.26). Right. Length of stay in post‐emergency room medical unit (UPUM LOS) (days) during the first and third part of the study period with admitter‐rounder model and the second part without. * p < 0.05. (C) Number of Internal Medicine bed‐spaced patients among all bed‐spaced patients during the study period (November 2023 to February 2024), October 2023 and the same period the previous year (November 2022 to February 2023). * p < 0.05. (D) Length of Stay (LOS) (days) in each of the four Internal medicine units (namely units 11, 41, 82 and 83) during the study period (November 2023 to February 2024) and the same period the previous year (November 2022 to February 2023). * p < 0.05.

The number of discharged patients was not significantly higher during the first and third parts of the study period (with admitter‐rounder model) than during the second part (without admitter rounder model: 129 out of 175 (73.7%) vs. 54 out of 67 (80.5%) (p = 0.26). However, the median LOS was significantly lower during the first and third parts of the study period than during the second part: 4 [1;3;6.75;26] versus 6 [1;4;8;26] days (p = 0.007) (Figure 1B).

3.2. The Impact of UPUM Onto Bed‐Spaced Patients and Length of Stay in Other In‐Patient Internal Medicine Units

The data concerning the bed‐spaced patients in Reims University Hospital are shown in Figure 1C. The internal medicine bed‐spaced patients accounted for 50 of the 256 bed‐spaced patients (19.5%) during the period November 2023 to February 2024 versus 45 of the 160 bed‐spaced patients (28.1%) during the same period the previous year before the implementation of UPUM, that is, November 2022 to February 2023 (p = 0.04). The number of ER in and out‐patients was grossly the same during these two periods (Appendix B).

The LOS in the four other in‐patient internal medicine units (namely units 11, 41, 82 and 83) are shown in Figure 1D. The median LOS were 10 versus 8 days (p = 0.005) for unit 82 and 10 versus 8 days (p = 0.01) for unit 83 during the period November 2023 to February 2024 versus the same period the previous year, respectively. There were no significant differences for the two other in‐patient internal medicine units 11 versus 11 days (p = 0.28) for unit 11 and 9 versus 8 days (p = 0.45) for unit 41 during the same periods, respectively.

4. Discussion

In the present report, we first observed that the implementation of a new SSU plus admitter‐rounder model was associated with (i) a statistically significant reduction of the number of Internal Medicine bed‐spaced patients (ii) a statistically significant increase of LOS in some in‐patient internal medicine units.

The statistically significant reduction of the number of Internal Medicine bed‐spaced patients after the implementation of UPUM suggested that Internal Medicine patients were especially those for whom beds were lacking before. The previously ‘Internal Medicine bed‐spaced patients’ have been admitted in the new SSU, where a bed was available. The absence of a reduction of the number of all bed‐spaced patients could be explained by an increase of other medical specialties bed‐spaced patients such as those of the previously closed in‐patient dermatology ward, as described elsewhere [15].

Moreover, the patients, who should have been admitted previously in in‐patient internal medicine wards for a few days, were also preferably admitted in the new SSU instead of internal medicine units, whose beds were therefore filled by other patients with higher LOS.

Taken together, this suggested that the implementation of a new SSU, even associated with an admitter rounder model, did not create new patients. The shift of patients described above should be taken in consideration when evaluating the efficiency of a new SSU at the whole hospital level. (Supporting Information S1: Appendix C)

Because this study was observational, we could not exclude the existence of unappreciated confounding bias affecting the link we would like to establish between the implementation of UPUM and the reduction of Internal Medicine bed‐spaced patients as well as the increase of LOS in other internal medicine units. However, we did not observe or remember any significant variation between the study period and the same period the previous year (same medical and paramedical staff in internal medicine units and ER, same patients [Appendix B], same season without new pandemic due to previously unknown emerging virus…).

Second, if we assumed that the admitted patients did not fully correspond to the target patients, the presence of a senior consultant making some of the admissions (admitter) and available during office hours was associated here with a statistically significant reduced LOS in the SSU (Figure 1B). To the best of our knowledge, we did not find other references about an association between reduced LOS in the SSU and admitter‐rounder model, because previous manuscripts mainly focused on the Hospital LOS [2, 3]. Interestingly, Potts and colleagues described a reduced LOS in the paediatrics SSU when the patients were admitted by an attending physician but also managed and discharged by three advanced practice providers available each day [16].

Combined interventions such as the implementation of SSU and the admitter‐rounder model should be globally evaluated elsewhere in universalist healthcare systems such as the one in France [17].

Ethics Statement

This study was approved by Reims University Hospital (approval number GDPR register number MR00416102023).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Discussion and supplementary references.

Appendix A.

See Figure A1.

Figure A1.

Potential admissions in the Short Stay Unit (UPUM) during the different parts of the study period. Supplementary sources of admissions were possible during the first and third parts of the study period, where an admitter rounder model was available. The ‘rounders’ were M.H. and A.B. during the whole study period. Y.N.G. is a senior consultant previously working in internal medicine (units 82 and 83) and now working mainly with surgery in‐patients and Emergency Room outpatients. He was the ‘admitter’ during the first and the third parts of the study period and made admissions in UPUM of patients with a presumed Length of Stay (LOS) lower than 5–7 days. When a bed‐spaced patient was secondarily transferred to UPUM, this patient was nevertheless counted as a bed‐spaced patient.

Appendix B.

See Table B1.

Table B1.

Number and demographic data of the emergency room's (ER) in and out‐patients during the study period and during the same time period the previous year, before the implementation of the Short Stay Unit UPUM.

Time Periods	Nov 06 2022–Feb 14 2023	Nov 06 2023–Feb 14 2024
ER external Consultations (n=)	14,812	14,655
All hospitalisations from ER (n=)	3302	3331
Mean age of patients (years)	49.8	50.0
Male sex (n=)	7655	7657

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.