Multivariate time series analysis in nosocomial infection surveillance: a case study

Area of Science:

• Epidemiology
• Hospital Infection Control
• Time Series Analysis

Background:

• Nosocomial infections pose a significant challenge in healthcare settings.
• Evaluating the impact of hospital operations on infection rates is crucial for effective management.
• Work-related factors, including personnel management and external events, may influence infection incidence.

Purpose of the Study:

• To investigate the relationship between hospital operational factors and the incidence of nosocomial infections.
• To quantify the impact of interventions like personnel training and surveillance systems on infection rates.
• To model and statistically evaluate the influence of strikes and personnel changes on infection trends.

Main Methods:

• Utilized time series analysis on monthly nosocomial infection data from March 1982 to December 1990.
• Employed curve fitting, Autoregressive Integrated Moving Average (ARIMA) modeling (Box-Jenkins), and intervention analysis.
• Applied dynamic regression analysis to assess the impact of specific events on infection incidence.

Main Results:

• Personnel control and training by the surveillance system decreased infection incidence by 3.63%.
• Medical strikes were associated with a 4.34% increase in infection incidence, with lasting effects.
• Each new nursing contract correlated with a 0.18% increase in infection incidence.
• Evidence suggested a link between vacation periods and increased infection rates.

Conclusions:

• Strict control of hospital personnel activities is recommended to mitigate infection risks.
• Implementing preventative measures during vacation periods is essential to reduce nosocomial infections.
• Time series analysis provides a valuable tool for evaluating the effectiveness of infection control strategies.