Related Experiment Videos
Experience with two validation methods in a prevalence survey on nosocomial infections
P Gastmeier1, G Kampf, T Hauer
1Institute of Hygiene, Free University Berlin, Germany.
Infection Control and Hospital Epidemiology
|October 20, 1998
Summary
Despite validation efforts, a significant investigator effect was found in the first German nosocomial infection surveillance study (NIDEP). This highlights the need for robust data validation methods in healthcare epidemiology research.
Area of Science:
- Healthcare Epidemiology
- Infectious Disease Surveillance
- Medical Research Methodology
Background:
- The Nosokomiale Infektionen in Deutschland Erfassung und Prävention (NIDEP) study was the first German initiative to assess nosocomial infection prevalence.
- Ensuring data accuracy in large-scale surveillance studies is critical for reliable epidemiological insights.
Purpose of the Study:
- To evaluate the persistence of an investigator effect in the NIDEP study.
- To assess the efficacy of validation methods employed during the surveillance.
Main Methods:
- Two validation approaches were utilized: bedside validation and case study validation.
- Investigator results were compared against gold standard observers across 72 hospitals.
- Validation procedures were implemented at multiple points throughout the surveillance period involving 14,966 patients.
Main Results:
- Bedside validation yielded an overall sensitivity of 89.0% and specificity of 99.5%.
- Case study validation demonstrated an overall sensitivity of 95.6% and specificity of 92.8%.
- A notable investigator effect was identified by the end of the surveillance period.
Conclusions:
- An investigator effect was observed despite satisfactory validation results, underscoring its potential impact on surveillance data.
- The findings emphasize the critical importance of rigorous data validation protocols in epidemiological research.
- Recommendations include clarifying diagnostic criteria (e.g., Centers for Disease Control and Prevention criteria for pneumonia) and incorporating diagnostic test results to mitigate future investigator effects.