A comparison of competing methods for the detection of surgical-site infections in patients undergoing total arthroplasty of the knee, partial and total arthroplasty of hip and femoral or similar vascular bypass

doi:10.1016/j.jhin.2004.03.020

Journal of Hospital Infection

Volume 57, Issue 3, July 2004, Pages 189-193

https://doi.org/10.1016/j.jhin.2004.03.020 Get rights and content

Abstract

Recent research suggests that the retrospective review of the International Classification of Disease (ICD-9-CM) codes assigned to a patient episode will identify a similar number of healthcare-acquired surgical-site infections as compared with prospective surveillance by infection control practitioners (ICP). We tested this finding by replicating the methods for 380 surgical procedures. The sensitivity and specificity of the ICP undertaking prospective surveillance was 80% and 100%, and the sensitivity and specificity of the review of ICD-10-AM codes was 60% and 98.9%. Based on these results we do not support retrospective review of ICD-10-AM codes in preference prospective surveillance for SSI.

Introduction

Cadwallader et al.¹ recently compared two methods for identifying surgical-site infections (SSI) in patients undergoing orthopaedic surgery. The study was undertaken in a 560-bed tertiary teaching hospital in Western Australia and compared routine prospective surveillance undertaken by the infection control practitioner (ICP) with retrospective identification of cases using ICD-9-CM² coding from the patient administration system. They concluded that careful analysis of the ICD-9-CM² codes, assigned to an episode by clinical coders, would identify a similar number of SSIs to that identified using prospective surveillance by the ICP. They suggest the analysis of the ICD-9-CM codes is preferable, because prospective surveillance incurs greater cost.

This hypothesis was tested by repeating the study using different groups of patients admitted to a tertiary referral hospital in Queensland. The aim of the present study was to compare the numbers of SSIs identified using prospective surveillance by ICPs with the number identified by review of ICD-10-AM codes.3., 4.

Section snippets

Materials and methods

All cases of total arthroplasty of the knee, partial and total arthroplasty of the hip and femoral or similar vascular bypass grafting carried out at the Princess Alexandra Hospital (PAH), a 720 bed tertiary referral teaching hospital affiliated with the University of Queensland, between 1 February 2001 and 1 November 2002 were included. Although post-discharge surveillance is undertaken by the ICPs at PAH for all targeted procedures, clinical coding would only identify these infections if the

Results

The only difference in the coding of SSI between the ICD-9-CM and the second and third editions of ICD-10-AM was one new code, ‘T81.41—wound infection following a procedure’. Three hundred and eighty procedures were identified: 33 aorta-iliac-femoral bypasses using synthetic graft, 104 femoro-popliteal bypasses, 12 femoro-tibial bypasses, 109 total hip replacements and 122 total knee replacements. Twenty SSIs were identified by the ICPs using eICAT (5.26%) and 19 infections were identified from

Discussion

There were differences between the prospective surveillance described by Cadwallader et al.¹ and prospective surveillance conducted at the PAH. The former involved the identification of the surveillance population by the nursing staff of the operating theatre, post-procedure monitoring was conducted by the clinical nurse consultants for the relevant wards, ward staff, allied health staff, the surgical team and staff in extended-care units, and only if an infection was notified did ICPs review

Acknowledgements

We thank Christine Carroll, Michelle Somlyay, Maria O'Neil, Kristen Tarvitt, Rebecca Tarvitt, Gary Waller and Kristen McKenzie for all their help. This research was supported by the Quality Improvement and Enhancement Program (QIEP) administered through Queensland Health.

References (9)

H.L. Cadwallader et al.
A comparison of two methods for identifying surgical site infections following orthopaedic surgery
J Hosp Infect
(2001)
The Australian Version of the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM)
(1996)
International Classification of Diseases and Related Health Problems, 10th revision, Australian Modification (ICD-10-AM)
(2000)
International Classification of Diseases and Related health Problems, 10th revision, Australian Modification (ICD-10-AM)
(2002)

There are more references available in the full text version of this article.

Cited by (19)

Predicting Infections After Total Joint Arthroplasty Using a Prescription Based Comorbidity Measure
2015, Journal of Arthroplasty
Citation Excerpt :
Additionally, an ICD-10 coding algorithm and prescription of antibiotics were used to ascertain infections in our study. The ICD-10 diagnostic codes used have been evaluated with acceptable sensitivity and specificity in Australian cohorts [22,23], but the procedure codes for lavage of knee and hip and antibiotic medication were not previously evaluated as infection indicators. We included them, after consultation with a clinical expert, in an attempt to capture all surgical site infections and this may have over reported infections.
This study evaluated the association and predictive ability of co-morbidities measured by RxRisk-V, Elixhauser and Charlson measures and post-total hip (THA) and total knee arthroplasties (TKA) infection. THAs and TKAs (2001–2012) were identified using the Australian Department of Veterans’ Affairs data. Infections within 90 days post-surgery were the study endpoint. Co-morbidities were identified using pharmacy (RxRisk-V) and hospitalization history (Elixhauser, Charlson). Of the 11,848 THAs, 3.1% (N = 364) had infections and out of 18,972 TKAs 3.4% (N = 648). Comorbidity burden and specific conditions were associated with infection likelihood. RxRisk-V performed better than other measures, but none had high predictive ability and differences were small. The best performing infection prediction models resulted when a combination of conditions identified by all measures was used.
Use of the French medico-administrative database (PMSI) to detect nosocomial infections in the University hospital of Lyon
2011, Revue d'Epidemiologie et de Sante Publique
Surveillance is an effective element in the fight against nosocomial infections, but the monitoring methods are often cumbersome and time consuming. The detection of infection in computerized databases is a means to alleviate the workload of health care teams. The objective of this study was to evaluate the performance of using discharge summaries in medico-administrative databases (PMSI) for the identification of nosocomial infections in surgery, intensive care and obstetrics.
The retrospective assessment study included patients who were hospitalized in general surgery, intensive care and obstetrics at different periods of time in 2006 and 2007 depending on the wards. Patients were monitored according to standard protocols which are coordinated at the regional level by the Southeast coordinating centre (CCLIN). The performance of identifying cases of nosocomial infection from discharge diagnoses coded by using the International Classification of Diseases (tenth revision) was evaluated by a study of sensitivity, specificity, positive and negative predictive values with their 95% confidence intervals.
Using a limited number of diagnostic codes, the sensitivity and specificity were, respectively, 26.3% (95% CI 13.2–42.1) and 99.5% (95% 98.8–100.0) for the identification of surgical site infections. By expanding the number of diagnostic codes, the sensitivity and specificity were 78.9% (95% CI 65.8–92.1) and 65.7% (95% CI 61.0–70.3). The sensitivity and specificity for case identification of nosocomial infections in intensive care were 48.8% (95% CI 42.6–55.0) and 78.4% (95% CI 76.1–80.1), and were 42.9% (95% CI 25.0–60.7) and 87.3% (95% CI 85.2–89.3) for identification of postpartum infections.
The PMSI is not a sufficiently efficient method in terms of sensitivity to be used in surveillance of nosocomial infections. A reassessment of the PMSI must be considered, with changes in coding of comorbidity that occurred in 2009.
La surveillance est un élément efficace de la lutte contre les infections nosocomiales mais les méthodes de surveillance sont souvent lourdes et consommatrices de temps. Le repérage des cas d’infection dans des bases de données informatisées est un moyen pour alléger la charge de travail des équipes de soins. L’objectif de cette étude était d’évaluer les performances de l’utilisation des diagnostics des résumés de sortie du Programme médicalisé du système d’information (PMSI) pour le repérage des infections nosocomiales en chirurgie, réanimation et obstétrique.
L’étude d’évaluation rétrospective incluait des patients ayant été hospitalisés en chirurgie générale, réanimation et obstétrique, à différentes périodes de 2006 et 2007 selon les services. Les patients étaient surveillés selon des protocoles standardisés et coordonnés régionalement par le CClin Sud-Est. Les performances du repérage des cas d’infection nosocomiale à partir des codes de la Classification internationale des maladies (dixième révision) des diagnostics de sortie ont été évaluées par une étude de sensibilité, spécificité, valeurs prédictives positives et négatives, avec leurs intervalles de confiance à 95 %.
La sensibilité et la spécificité étaient, respectivement, de 26,3 % (IC à 95 % 13,2–42,1) et 99,5 % (IC à 95 % 98,8–100,0) pour le repérage des infections du site opératoire, avec une restriction à l’utilisation de codes spécifiques et de 78,9 % (IC à 95 % 65,8–92,1) et 65,7 % (IC à 95 % 61,0–70,3) avec l’élargissement à des codes moins spécifiques, de 48,8 % (IC à 95 % 42,6–55,0) et 78,4 % (IC à 95 % 76,1–80,1) pour le repérage des cas d’infections nosocomiales en réanimation et de 42,9 % (IC 95 % 25,0–60,7) et 87,3 % (IC à 95 % 85,2–89,3) pour le repérage des cas d’infections nosocomiales du post-partum en obstétrique.
Le PMSI ne représente pas une méthode suffisamment performante en termes de sensibilité pour être utilisé dans la surveillance des infections nosocomiales. Une nouvelle évaluation du PMSI devra être envisagée, avec les modifications de codage de la comorbidité survenues en 2009.
Economic rationale for infection control in Australian hospitals
2009, Healthcare Infection
The objective of the present study was to predict the economic consequences of healthcare-acquired infections arising among admissions to Australian acute care hospitals. A quantitative algorithm informed by epidemiological and economic data was developed. All acute care hospitals in Australia were included in the study and the participants included all admissions to general medical and general surgical specialties. The main outcome measures were the numbers of cases of healthcare-acquired infection and bed days lost annually. It was estimated that there are 175 153 (95% credible interval 155 911 : 195 168) cases of healthcare-acquired infection among admissions to Australian hospitals annually, and the extra stay in hospital to treat symptoms accounts for 854 289 bed days (95% credible interval 645 091 : 1 096 244). If rates were reduced by 1%, then 150 158 bed days would be released for alternative uses. This would allow ~38 500 new admissions. Healthcare-acquired infections in patients cause bed blocks in Australian hospitals. The cost-effectiveness of hospital services might be improved by allocating more resources to infection control, releasing beds and allowing new admissions. There exists an opportunity to improve the efficiency of the Australian health care system.
Surveillance of operative site infections in an orthopedic and traumatology surgery department: An example of methodology
2007, Revue de Chirurgie Orthopedique et Reparatrice de l'Appareil Moteur
Le but de cette étude a été de décrire une méthode de recensement des infections nosocomiales de site opératoire (ISO) et rapporter les résultats observés dans un service de chirurgie orthopédique et traumatologie.
Le recensement de ces infections repose sur le signalement par un praticien du service de bactériologie de tous prélèvements de site opératoire ayant une culture positive. Un collège de 4 médecins spécialistes décide du caractère infectant du germe et de l’origine nosocomiale de l’infection. Un dossier informatique est alors créé pour chaque patient ayant une infection de site opératoire regroupant des données cliniques, biologiques et microbiologiques. Le fichier informatique permet alors un suivi prospectif et une analyse des cas ainsi répertoriés.
Entre 2000 et 2002, 9397 interventions ont été réalisées et nous avons relevé 86 infections de site opératoire considérées comme nosocomiales. Les patients avaient un âge moyen de 58 ans, un index de masse corporelle moyen à 25,7 et 72 % avaient un score ASA supérieur ou égal à II. L’infection concernait une arthroplastie dans 23 cas, un acte de traumatologie dans 21 cas, le traitement d’une tumeur dans 24 cas. Elle a été diagnostiquée dans les 30 jours suivant l’intervention dans 75 % des cas, et après la sortie de l’hôpital dans 65,4 % des cas. Les infections monomicrobiennes étaient les plus fréquentes (n = 59) et un staphylococcus aureus a été isolé dans 80,23 % des infections. Mais en chirurgie tumorale, les infections plurimicrobiennes étaient statistiquement plus fréquentes associant des staphylocoques à coagulase négatif et des bacilles Gram négatif. Nous déplorons 6 décès directement imputables à une infection de site opératoire.
A program for the prevention of nosocomial infections, including operative site infections (OSI) is a legal obligation in France. According to the CDC, in orthopedic surgery, nosocomial infection is defined as any infection occurring within 30 days of operation, or within one year in the event of material implantation. No surveillance system has been validated and the rate of OSI is unknown in orthopedic surgery. We report the number of OSI observed during a three year period in our unit and describe the characteristic features.
Data were collected from the bacteriology reports on operative site samples with a positive culture. A group of specialists determined the infective nature of the germ and the nosocomial nature of the OSI. Clinical and bacteriological data were noted on a standard datasheet used for prospective follow-up of the number of cases and data processing. During a three-year period (2000, 2001, 2002), among 9397 orthopedic and traumatology operations performed, 86 OSI were identified. Mean patient age was 58 years and mean body mass index was 25.7. The ASA score was ≥ II for 72% of patients.
The OSI involved an arthroplasty in 23 cases, a traumatology procedure in 21, and tumor treatment in 24. The diagnosis was established within 30 days of operation for 75% and after discharge from hospital in 65.4%. Single-germ infections predominated (n = 59). Staphylococcus aureus was isolated in 80.23% of infections. For tumor surgery, the statistically more frequent multiple-germ infections associated coagulase negative Staphylococcus and Gram-negative bacilli. There were six OSI-related deaths.
Two criticisms can be formulated concerning our surveillance system. First, infections with no identified germ could be missed. The frequency of such infections has been estimated at 2.8 to 19% by different authors. Although patients are automatically recalled for consultation, we were unable to determine the number of patients lost to follow-up at one year. It was thus not possible to determine a precise rate of OSI. Data in the literature have not demonstrated any system providing an exhaustive surveillance, particularly because of the long postoperative period after material implantation. Excepting tumor surgery, Staphylococcus aureus infections predominated. Factors of risk of OSI include the patient's general status, particularly for arthroplasty. We had a mortality rate of 7% for our OSI, corroborating earlier studies and illustrating the severity of such infections.
Surveillance of OSI in orthopedic surgery requires the development of a system responding to the problem of a long observation period. It would be important to know the precise number of OSI and their characteristic features in order to develop comparison tools.
An enhanced benchmark for prosthetic joint replacement infection rates
2006, American Journal of Infection Control
Citation Excerpt :
Forty-four percent of the total SSIs were identified during a readmission, and 4% were identified during the index or original admission. Fifty-two percent is on the upper end of the 20% to 70% range, which is reported by a number of sources as the observed percentage of total SSIs detected during the postdischarge period.1-5 In addition, excluding the infections identified during the postdischarge period, the IHCS rates for both hips and knees are statistically significantly lower than NNIS rates (See Tables 1 and 2).
The National Nosocomial Infection Surveillance System (NNIS) has historically provided the infection control community with the most accurate benchmark for healthcare-associated infections. However, NNIS does not require postdischarge surveillance. For medical centers where comprehensive postdischarge surveillance is possible, the efficiency of surgical site infection (SSI) detection is enhanced and rates may be higher than those provided by NNIS.
From 1999 to 2004, a large integrated healthcare system (IHCS) used a standard surveillance methodology inclusive of the postdischarge period. This article compares IHCS and NNIS SSI data.
IHCS infection rates, stratified and weighted average (hip, 1.7; knee, 2.1) for the study period are higher than the corresponding NNIS rates (hip, 1.4; knee, 1.2) (hip, P = .006; knee, P = .012) when infections detected by the IHCS during the postdischarge period are included.
The data from the study period show that when comprehensive postdischarge surveillance is used by the IHCS, SSI rates are higher than those reflected in the NNIS database.
Evaluation of postdischarge surveillance of surgical site infections after total hip and knee arthroplasty
2005, American Journal of Infection Control
Citation Excerpt :
Moreover, there are still attempts to devise more effective approaches of postdischarge surveillance of SSI in surgery, but a perfect strategy has not yet been established. The collection and processing of health care data from different sources, eg, pharmacy activity and antibiotic administration, might be useful approaches.19,27,28 However, it is uncertain whether a universally valid improvement of existing surveillance systems can be found and implemented for different health care systems.
Artificial joint replacement of hip (HPRO) and knee (KPRO) are 2 of about 20 categories of operative procedures of the surveillance of surgical site infection (SSI) as stated by nosocomial infections surveillance systems in the United States and in Germany. Periprosthetic SSI can manifest itself after a long period.
Seven hundred fifty-six orthopedic patients from 2 centers were evaluated after HPRO (n = 508) or KPRO (n = 248). SSI was recorded during hospitalization and for 12 month postdischarge. The surveillance regimen was extended by also sending patients a questionnaire after 12 months postdischarge. All complaints were followed up by contacting the patients and any clinicians and general practitioners (GPs) involved. Stratified infection rates and standardized infection ratio (SIR) were calculated and compared with reference data of the national surveillance system.
The total response rate to the postal questionnaire survey was 85.2%. SSI was recorded in 16 patients (3.15%) after HPRO; 12 were detected by predischarge surveillance, and the 4 cases found postdischarge were all organ/space SSI. In total, only 1 SSI was detected after KPRO before discharge and none after discharge (SSI rate 0.40%). Time between discharge and detection of SSI cases ranged from 8 days to 8 months. SIR of HPRO was 1.25 and SIR of KPRO was 0.36.
Because 25% of SSIs after HPRO occurred after discharge and all were organ/space SSI, highlights the importance of postdischarge surveillance of nosocomial infections (NIs). Because all SSIs were reported already by current surveillance, the extended postdischarge surveillance appears to be unnecessary. The pursuit of shorter hospital stay after surgery may challenge the methods of surveillance systems in future.

View all citing articles on Scopus

View full text

A comparison of competing methods for the detection of surgical-site infections in patients undergoing total arthroplasty of the knee, partial and total arthroplasty of hip and femoral or similar vascular bypass

Abstract

Introduction

Section snippets

Materials and methods

Results

Discussion

Acknowledgements

J Hosp Infect

The Australian Version of the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM)

International Classification of Diseases and Related Health Problems, 10th revision, Australian Modification (ICD-10-AM)

International Classification of Diseases and Related health Problems, 10th revision, Australian Modification (ICD-10-AM)