Surveillance of selected post-caesarean infections based on electronic registries: validation study including post-discharge infections

https://doi.org/10.1016/j.jhin.2009.11.018Get rights and content

Summary

The importance of surveillance of post-discharge infections has increased as a consequence of shorter hospital stay after surgical procedures. This study examined the ability of a computer-based surveillance system to identify urinary tract infections (UTIs) and postoperative wound infections (PWIs) within 30 days after caesarean section. We assessed the use of data from various electronic registries to identify patients with post-caesarean UTI and PWI classified according to a reference standard. The standard was based on information from medical records and self-reported data (questionnaire) using modified Centers for Disease Control and Prevention definitions. The sensitivity of the computer system in detecting UTI diagnosed during hospital stay, readmission or at visits to hospital outpatient clinics was 80.0%; the specificity was 99.9%. For post-discharge UTIs diagnosed outside the hospital, sensitivity and specificity were 76.3% and 99.9%, respectively. For PWIs diagnosed in hospital and post-discharge outside hospital, sensitivities were 77.1% and 68.9%, and the specificities 99.5% and 98.2%. We conclude that a computer-based surveillance system may identify in-hospital infections and post-discharge infections with a relatively high sensitivity and excellent specificity.

Introduction

Surveillance of hospital-acquired infections (HAIs) with feedback to the clinical staff has in previous studies been shown to be associated with reduced rates of wound infections of 25–55%.1, 2, 3 Several countries have implemented nosocomial infection surveillance systems, often as regularly conducted prevalence studies.4, 5, 6 Prevalence studies are, however, time-consuming and may, despite clear definitions of infection, be subjective with some interobserver variability.7, 8 The increasing amount of medical data stored in electronic databases may be used as data source in surveillance systems. A number of studies have validated the use of such electronically stored information in different patient groups or types of infections.9, 10, 11, 12, 13 In none of these studies were post-discharge infections included, except for readmitted patients.

Women having undergone caesarean section (CS) normally have a short hospital stay after delivery, which means that post-caesarean infections often occur after discharge.14, 15 The prevalence of post-discharge infections after delivery has been shown to range from 36% to 94%.16, 17, 18, 19

The aim of this study was to validate monitoring of postpartum infections in a study population of women having undergone CS. We compared a computer registration of infections based on data from various electronic hospital registries with a reference standard based on all available clinical and para-clinical data from medical records together with information about infections obtained from patient-administered questionnaires, confirmed by the general practitioner (GP).

Section snippets

Methods

We conducted this prospectively designed cohort study in the former County of Aarhus, Denmark from 1 February 2007 to 31 August 2008. In Denmark, health service is free of charge with free access for all residents to GPs and public hospitals for examinations and treatments. During the study period virtually all births in the county took place at three public hospitals.

Since 1968, all Danish residents have been assigned a unique civil registration number that enables unambiguous electronic

Results

During the 19 month period, 2492 women underwent CS of whom 2453 women were invited to participate in the questionnaire survey. A total of 1513 (61.7%) of these women answered the questionnaire. These 1513 women became our study group, of which 833 (55%) underwent emergency CS and 680 (45%) elective CS.

We found that 3.2% (48/1513) of the women had a UTI within 30 days postpartum; 38/48 (79%) were diagnosed and treated post-discharge outside hospital. Forty-nine of 1513 women had their urine

Discussion

We included more than 1500 women who underwent CS in validating the ability of an electronic surveillance system to identify post-caesarean infections. Post-caesarean UTI and PWI within 30 days was seen in 10.4% of the women. Nearly two-thirds of these infections were diagnosed and treated by the GP. The HAIR system was able to identify about 80% of in-hospital PWIs and UTIs. Post-discharge PWIs and UTIs diagnosed outside hospital were identified in 69% and 76% of the cases, respectively.

The

References (25)

  • B.M. Andersen et al.

    Hospital-acquired infections before and after healthcare reorganization in a tertiary university hospital in Norway

    J Public Health

    (2009)
  • A. Troelstra

    Hospital infection control in the Netherlands

    J Hosp Infect

    (2007)
  • Cited by (14)

    • Impacts of structuring the electronic health record: Results of a systematic literature review from the perspective of secondary use of patient data

      2017, International Journal of Medical Informatics
      Citation Excerpt :

      The quality of structured patient data was evaluated usually as the consistency and reliability of data. The most common structuring method evaluated was the international classification of diseases, ICD in nine articles [11,62,66,71,81,85,86,87,89], various other classifications and terminologies in four articles [64,79,82,97], and different documentation standards and standardized forms in 15 articles [13,17,20,23,24,27,30,34,46,47,57,59,72,78,86]. In addition, numeric data was used in one study as a source [60].

    • Data elements and validation methods used for electronic surveillance of health care-associated infections: A systematic review

      2015, American Journal of Infection Control
      Citation Excerpt :

      ICD-9,4,9-11,15-19 Systematized Nomenclature of Medicine—Clinical Terms,20-22 and free text from notes were used to determine hospital billing diagnosis and procedures.10,11,17,21 Microbiology results were formatted in institution-specific codes,10,11,13,14,17 textual results,3-5,9,12,16,18,22-30 and Logical Observation Identifiers Names and Codes.22 Validation of the numerator was performed most often (80%; 24 out of 30 studies).

    • Methodology of the Norwegian Surveillance System for Healthcare-Associated Infections: The value of a mandatory system, automated data collection, and active postdischarge surveillance

      2013, American Journal of Infection Control
      Citation Excerpt :

      There is a limited, but growing, amount of literature on computerized surveillance systems for SSIs. Most of the literature focuses on the actual case finding (detecting infections) through computer-based algorithms in electronic medical records or administrative data sets.18-21 We have found that over 80% of the infections are detected after discharge, and most of these will not be detected through hospital systems unless patients return to hospital for follow-up.

    View all citing articles on Scopus
    View full text