Impact of an ICU bed capacity optimisation method on the average length of stay and average cost of hospitalisation following implementation of Chinas open policy with respect to COVID-19: a difference-in-differences analysis based on information management system data from a tertiary hospital in southwest China

Zheng,Q.,Zeng,Z.,Tang,X.,Ma,L.

DOI: https://doi.org/10.1136/bmjopen-2023-078069

IF: 3.006

2024-04-21

BMJ Open

Abstract:Objectives Following the implementation of China's open policy with respect to COVID-19 on 7 December 2022, the influx of patients with infectious diseases has surged rapidly, necessitating hospitals to adopt temporary requisition and modification of ward beds to optimise hospital bed capacity and alleviate the burden of overcrowded patients. This study aims to investigate the effect of an intensive care unit (ICU) bed capacity optimisation method on the average length of stay (ALS) and average cost of hospitalisation (ACH) after the open policy of COVID-19 in China. Design and setting A difference-in-differences (DID) approach is employed to analyse and compare the ALS and ACH of patients in four modified ICUs and eight non-modified ICUs within a tertiary hospital located in southwest China. The analysis spans 2 months before and after the open policy, specifically from 5 October 2022 to 6 December 2022, and 7 December 2022 to 6 February 2023. Participants We used the daily data extracted from the hospital's information management system for a total of 5944 patients admitted by the outpatient and emergency access during the 2-month periods before and after the release of the open policy in China. Results The findings indicate that the ICU bed optimisation method implemented by the tertiary hospital led to a significant reduction in ALS (HR –0.6764, 95% CI –1.0328 to –0.3201, p=0.000) and ACH (HR –0.2336, 95% CI –0.4741 to –0.0068, p=0.057) among ICU patients after implementation of the open policy. These results were robust across various sensitivity analyses. However, the effect of the optimisation method exhibits heterogeneity among patients admitted through the outpatient and emergency channels. Conclusions This study corroborates a significant positive impact of ICU bed optimisation in mitigating the shortage of medical resources following an epidemic outbreak. The findings hold theoretical and practical implications for identifying effective emergency coordination strategies in managing hospital bed resources during sudden public health emergency events. These insights contribute to the advancement of resource management practices and the promotion of experiences in dealing with public health emergencies.

medicine, general & internal

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper aims to explore the impact of intensive care unit (ICU) bed optimization methods on the average length of stay (ALS) and the average cost of hospitalization (ACH) of hospitalized patients after the implementation of the COVID - 19 prevention and control relaxation policy (i.e., the "open - door policy") in China. Specifically, the research is set against the background that after the adjustment of China's epidemic prevention and control policy on December 7, 2022, hospitals faced a surge in infectious disease patients, leading to a shortage of medical resources, especially bed resources. To meet this challenge, hospitals took measures such as temporarily requisitioning and converting general wards into ICU beds to optimize bed resources.

### Main objectives of the research

1. **Evaluate the effectiveness of ICU bed optimization methods**: By comparing the data of the renovated ICUs and the unrenovated ICUs before and after the open - door policy, evaluate the specific impact of bed optimization methods on ALS and ACH.
2. **Provide theoretical and practical significance**: Verify the effectiveness of bed optimization methods and provide scientific basis and practical experience for medical resource management in future public health emergencies.

### Research design and methods

- **Research subjects**: ICU patients in a tertiary hospital in the southwest region.
- **Time range**: Data for two months before and after the open - door policy, that is, from October 5, 2022 to December 6, 2022 (before the policy), and from December 7, 2022 to February 6, 2023 (after the policy).
- **Analysis method**: The Difference - in - Differences (DID) model was used for analysis to control the influence of other potential confounding factors.

### Main findings

- **Significant reduction in average length of stay (ALS)**: The optimized ICU bed method after renovation reduced ALS by 0.6764 days (95% CI - 1.0328 to - 0.3201, p = 0.000).
- **Some reduction in average cost of hospitalization (ACH)**: Although not completely significant, ACH also showed a downward trend, decreasing by 0.2336 (95% CI - 0.4741 to - 0.0068, p = 0.057).
- **Heterogeneity of effects**: Among patients with different admission channels (outpatient and emergency), there are differences in the effectiveness of the optimization method.

### Conclusions

This study shows that the ICU bed optimization method has a significant positive role in alleviating the shortage of medical resources after the epidemic outbreak. These findings not only support theoretical research but also provide valuable references for practical operations, which are helpful for better coordinating and managing medical resources in future public health events.

### Formula representation

The DID model formulas used in the study are as follows:

\[
ALS_{it}=\alpha_1 + \beta_1Treat_i\times Post_t+X_{it}\gamma_1+\mu_i+\mu_t+\varepsilon_{it}
\]

\[
ACH_{it}=\alpha_2 + \beta_2Treat_i\times Post_t+X_{it}\gamma_2+\mu_i+\mu_t+\delta_{it}
\]

where:
- \(i\) represents the nursing unit;
- \(t\) represents the date;
- \(ALS_{it}\) and \(ACH_{it}\) represent the average length of stay and the average cost of hospitalization of the nursing unit \(i\) on the \(t\) - th day, respectively;
- \(Treat_i\) is a dummy variable indicating whether it is a renovated ICU (1 for renovated ICU, 0 for unrenovated ICU);
- \(Post_t\) is a dummy variable indicating whether it is within the time period after the implementation of the open - door policy (1 for after the policy, 0 for before the policy);
- \(X_{it}\) is a possible control variable;
- \(\mu_i\) and \(\mu_t\) are the fixed effects of the nursing unit and the date, respectively;
- \(\varepsilon_{it}\) and \(\delta_{it}\) are random disturbance terms.

Through these formulas, researchers can...