Inflammatory biomarkers in infective endocarditis: machine learning to predict mortality

Infective endocarditis (IE) is the cardiac disease with the highest rates of mortality. New biomarkers that are able to identify patients at risk for death are required to improve patient management and outcome. This study aims to investigate if cytokines, chemokines and growth factors measured at IE diagnosis can predict mortality. Patients with definite IE, according to the Duke’s modified criteria, were included. Using high-performance Luminex assay, 27 different cytokines, chemokines and growth factors were analyzed. Machine learning techniques were used for the prediction of death and subsequently creating a decision tree, in which the cytokines, chemokines and growth factors were analyzed together with C-reactive protein (CRP). Sixty-nine patients were included, 41 (59%) male, median age 54 [interquartile range (IQR) = 41–65 years] and median time between onset of the symptoms and diagnosis was 12 days (IQR = 5–30 days). The in-hospital mortality was 26% (n = 18). Proinflammatory cytokines interkeukin (IL)-15 and C-C motif chemokine ligand (CCL4) were found to predict death, adding value to CRP levels. The decision tree predicted correctly the outcome of 91% of the patients at hospital admission. The high-risk group, defined as CRP ≥ 72 mg/dL, IL-15 ≥ 5·6 fg/ml and CCL4 ≥ 6·35 fg/ml had an 88% in-hospital mortality rate, whereas the patients classified as low-risk had a mortality rate of 8% (P = < 0·001). Cytokines IL-15 and CCL4 were predictors of mortality in IE, adding prognostic value beyond that provided by CRP levels. Assessment of cytokines has potential value for clinical risk stratification and monitoring in IE patients.     

Infective endocarditis in Greece: a changing profile. Epidemiological, microbiological and therapeutic data

The epidemiology, and clinical and microbiological spectrum, of infective endocarditis (IE) in Greece was analysed in a prospective 4-year study in a tertiary hospital and a heart surgery centre in Athens. In total, 101 cases of IE (71 men, 30 women, aged 54.4 +/- 17.1 years) were studied, with a follow-up period of 3 months. Seventy-seven cases were definite and 24 possible; 59 involved native valves (native valve endocarditis; NVE), 31 prosthetic valves (prosthetic valve endocarditis; PVE), of which nine were early and 22 late, and 11 permanent pacemakers (pacemaker endocarditis; PME). There was a predominant involvement of aortic (48/101) and mitral (40/101) valves. Seven patients had rheumatic valvular disease, two had mitral valve prolapse, and eight had a previous history of IE. Thirteen and six patients had undergone dental and endoscopic procedures, respectively. In 13 patients, intravenous catheters were used within the 3 months before diagnosis of IE. There were three intravenous drug users among the patients. Staphylococcus aureus was the most important pathogen, isolated in 22% of cases, followed by viridans streptococci (19%) and coagulase-negative staphylococci (16%). Enterococcus spp. were responsible for 3%, HACEK group for 2%, and fungi for 6% of cases. Viridans streptococci were the leading cause of NVE (29%), Staphylococcus epidermidis of PVE (16%), and S. aureus of PME (54.5%). Six of 22 S. aureus and ten of 16 S. epidermidis isolates were methicillin-resistant. Surgical intervention, including total pacemaker removal, was performed in 51.5% of patients. Overall mortality was 16%, but was 29% with PVE, and was significantly higher with medical than with combined surgical and medical therapy (24.5% vs. 8%). Compared with previous studies, there were changing trends in the epidemiology, microbiology, treatment and prognosis of IE in Greece.     

Chronic antiplatelet therapy and mortality among patients with infective endocarditis

Whether antiplatelet therapy is associated with better outcomes among patients with infective endocarditis (IE) remains controversial. A retrospective study was conducted concerning all patients with IE, treated in a tertiary-care centre of Canada between 1991 and 2006, who satisfied the modified Duke criteria for a definite or possible IE. The primary outcome was all-cause mortality within 90 days of diagnosis. A secondary outcome was the development of major systemic embolism. In total, 241 patients satisfied the inclusion criteria, 75 of whom had been on chronic antiplatelet therapy prior to developing endocarditis. Seventy-one (29.5%) patients died. According to multivariate analysis, age, a high Charlson score, aortic valve involvement, myocardial infarction and presence of a perivalvular abscess were strongly associated with mortality. Undergoing valvular replacement (adjusted OR (AOR) 0.28, 95% CI  0.09–0.84) and chronic antiplatelet therapy before IE (AOR  0.27, 95% CI  0.11–0.64) correlated with lower mortality. There was a trend for lower mortality among patients started on antiplatelet drugs after admission (AOR  0.29, 95% CI  0.08–1.13). The effect of aspirin on mortality was much the same in patients who received 325 or 80 mg daily. Chronic antiplatelet therapy was not associated with a significantly lower risk of major embolism. In conclusion, chronic antiplatelet therapy was associated with lower mortality among patients with IE, independently of any effect on major embolism. Whether or not a beneficial effect could be replicated by initiating antiplatelet therapy at the time of diagnosis remains unproven.     

Complement activation in infective endocarditis: correlation with extracardiac manifestations and prognosis

In an infectious process complement activation is necessary for a proper immune and inflammatory response, but when exacerbated may cause tissue injuries. In infective endocarditis (IE) patients tend to develop high titres of circulating immune complexes (CIC) that activate complement. The aim of this study was to evaluate for the first time complement activation in IE for possible correlation with extracardiac manifestations and clinical prognosis. Twenty patients with IE, 14 healthy controls and 15 patients presenting mitral and aortic valve lesions (with no signs of either infection or other associated diseases), were studied. Plasma levels of C3adesArg, SC5b-9, C1rs-C1Inh and C3b(Bb)P were determined by ELISA and C3d by double decker immunoelectrophoresis. C3 and C4 levels were assayed by turbidimetry and CIC by ELISA. Elevation of plasma levels of all complement activation products, with the exception of C3b(Bb)P, indicated a significant classical pathway activation in IE patients when compared to controls (C3d: P < 0.00004; C3adesArg: P < 0.03, SC5b-9: P < 0.01, C1rs-C1Inh: P < 0.00007). CIC levels were significantly increased (P < 0.005) and C3 reduced in IE patients (P < 0.05). Elevated C3d (P < 0.02) and C3adesArg (P < 0.03) levels were associated with pulmonary manifestations. In addition, C3d was significantly elevated in the patients who died when compared to those who had a good recovery (P < 0.02). Our data demonstrate the activation of the complement classical pathway, most probably mediated by CIC, in IE and suggests C3d and C3adesArg as possible markers for extracardiac lesion and severity of the disease.     

Case clustering in infective endocarditis: the role of availability bias

Limited data exist regarding the impact of variations in clinical practice and physicians' cognitive bias on the diagnosis of infective endocarditis (IE). As an illustration of these effects, unexpected clustering of IE diagnosis was encountered in a prospectively studied cohort. Transoesophageal echocardiography examinations for suspected IE were performed more frequently following a diagnosis of IE, and were associated with a subsequent cluster of IE cases. The cognitive bias of physicians resulting from a recent case of IE can lead to a transient increase in diagnosing additional cases of IE.     

A pragmatic approach for mortality prediction after surgery in infective endocarditis: optimizing and refining EuroSCORE

ObjectiveTo simplify and optimize the ability of EuroSCORE I and II to predict early mortality after surgery for infective endocarditis (IE).MethodsMulticentre retrospective study (n = 775). Simplified scores, eliminating irrelevant variables, and new specific scores, adding specific IE variables, were created. The performance of the original, recalibrated and specific EuroSCOREs was assessed by Brier score, C-statistic and calibration plot in bootstrap samples. The Net Reclassification Index was quantified.ResultsRecalibrated scores including age, previous cardiac surgery, critical preoperative state, New York Heart Association >I, and emergent surgery (EuroSCORE I and II); renal failure and pulmonary hypertension (EuroSCORE I); and urgent surgery (EuroSCORE II) performed better than the original EuroSCOREs (Brier original and recalibrated: EuroSCORE I: 0.1770 and 0.1667; EuroSCORE II: 0.2307 and 0.1680). Performance improved with the addition of fistula, staphylococci and mitral location (EuroSCORE I and II) (Brier specific: EuroSCORE I 0.1587, EuroSCORE II 0.1592). Discrimination improved in specific models (C-statistic original, recalibrated and specific: EuroSCORE I: 0.7340, 0.7471 and 0.7728; EuroSCORE II: 0.7442, 0.7423 and 0.7700). Calibration improved in both EuroSCORE I models (intercept 0.295, slope 0.829 (original); intercept –0.094, slope 0.888 (recalibrated); intercept –0.059, slope 0.925 (specific)) but only in specific EuroSCORE II model (intercept 2.554, slope 1.114 (original); intercept –0.260, slope 0.703 (recalibrated); intercept –0.053, slope 0.930 (specific)). Net Reclassification Index was 5.1% and 20.3% for the specific EuroSCORE I and II.ConclusionsThe use of simplified EuroSCORE I and EuroSCORE II models in IE with the addition of specific variables may lead to simpler and more accurate models.     

Clinical and molecular epidemiology of community-acquired, healthcare-associated and nosocomial methicillin-resistant Staphylococus aureus in Spain

A prospective cohort study including all new cases of methicillin-resistant Staphylococcus aureus (MRSA) colonization or infection in 64 Spanish hospitals during June 2003 was performed to investigate the epidemiology of MRSA in Spain. Only patients who yielded clinical MRSA-positive samples were included. Epidemiological and clinical data for a total of 370 cases were collected. Genotyping was performed using pulsed-field gel electrophoresis and multilocus sequence typing. Panton–Valentine leukocidin genes and the staphylococcal chromosomal cassette mec (SCC mec ) were identified in representative isolates. MRSA was considered to be nosocomially acquired in 202 cases (55%), healthcare-associated (HCA) in 139 cases (38%), community-acquired (CA) in three cases, and of uncertain mode of acquisition in 26 (7%) cases. The pooled population-based rate was 2.31 cases/100 000 population/month, and the pooled nosocomial rate was 0.21 cases/1000 hospital stays (20.2% of S. aureus ). Peripheral vascular disease, respiratory tract infections, catheter infections, bloodstream infections and crude mortality were more frequent among HCA cases, whereas neoplasia and urinary tract infections were more frequent among nosocomially acquired cases. Two clones related to the paediatric clone ST5-IV accounted for 71% of the isolates; EMRSA-16 has emerged in two different geographical areas. Only one isolate belonged to the formerly predominant Iberian clone. The three CA isolates were related to the USA300 clone. SCC mec type IV was the most frequent type in nosocomial and HCA isolates. The epidemiology of MRSA has changed in Spain; outpatients with previous healthcare contact represent a very important reservoir of MRSA, and community isolates are emerging.     

Changes in the treatment of Enterococcus faecalis infective endocarditis in Spain in the last 15 years: from ampicillin plus gentamicin to ampicillin plus ceftriaxone

The aim of this study was to assess changes in antibiotic resistance, epidemiology and outcome among patients with Enterococcus faecalis infective endocarditis (EFIE) and to compare the efficacy and safety of the combination of ampicillin and gentamicin (A+G) with that of ampicillin plus ceftriaxone (A+C). The study was a retrospective analysis of a prospective cohort of EFIE patients treated in our centre from 1997 to 2011. Thirty patients were initially treated with A+G (ampicillin 2 g/4 h and gentamicin 3 mg/kg/day) and 39 with A+C (ampicillin 2 g/4 h and ceftriaxone 2 g/12 h) for 4–6 weeks. Increased rates of high-level aminoglycoside resistance (HLAR; gentamicin MIC ≥512 mg/L, streptomycin MIC ≥1024 mg/L or both) were observed in recent years (24% in 1997–2006 and 49% in 2007–2011; p 0.03). The use of A+C increased over time: 1997–2001, 4/18 (22%); 2002–2006, 5/16 (31%); 2007–2011, 30/35 (86%) (p <0.001). Renal failure developed in 65% of the A+G group and in 34% of the A+C group (p 0.014). Thirteen patients (43%) in the A+G group had to discontinue treatment, whereas only one patient (3%) treated with A+C had to discontinue treatment (p <0.001). Only development of heart failure and previous chronic renal failure were independently associated with 1-year mortality, while the individual antibiotic regimen (A+C vs. A+G) did not affect outcome (OR, 0.7; 95% CI, 0.2–2.2; p 0.549). Our study shows that the prevalence of HLAR EFIE has increased significantly in recent years and that alternative treatment with A+C is safer than A+G, with similar clinical outcomes, although the sample size is too small to draw firm conclusions. Randomized controlled studies are needed to confirm these results.     

Time to blood culture positivity: An independent predictor of infective endocarditis and mortality in patients with Staphylococcus aureus bacteraemia

Objectives

Time to blood culture positivity (TTP), a routinely available parameter in automated blood culture systems, may be a proxy for infectious burden in patients with bloodstream infections. We aimed to study the association between TTP and infective endocarditis (IE), or death, in patients with Staphylococcus aureus bacteraemia.

Prognostic models for mortality after cardiac surgery in patients with infective endocarditis: a systematic review and aggregation of prediction models

BackgroundThere are several prognostic models to estimate the risk of mortality after surgery for active infective endocarditis (IE). However, these models incorporate different predictors and their performance is uncertain.ObjectiveWe systematically reviewed and critically appraised all available prediction models of postoperative mortality in patients undergoing surgery for IE, and aggregated them into a meta-model.Data sourcesWe searched Medline and EMBASE databases from inception to June 2020.Study eligibility criteriaWe included studies that developed or updated a prognostic model of postoperative mortality in patient with IE.MethodsWe assessed the risk of bias of the models using PROBAST (Prediction model Risk Of Bias ASsessment Tool) and we aggregated them into an aggregate meta-model based on stacked regressions and optimized it for a nationwide registry of IE patients. The meta-model performance was assessed using bootstrap validation methods and adjusted for optimism.ResultsWe identified 11 prognostic models for postoperative mortality. Eight models had a high risk of bias. The meta-model included weighted predictors from the remaining three models (EndoSCORE, specific ES-I and specific ES-II), which were not rated as high risk of bias and provided full model equations. Additionally, two variables (age and infectious agent) that had been modelled differently across studies, were estimated based on the nationwide registry. The performance of the meta-model was better than the original three models, with the corresponding performance measures: C-statistics 0.79 (95% CI 0.76–0.82), calibration slope 0.98 (95% CI 0.86–1.13) and calibration-in-the-large –0.05 (95% CI –0.20 to 0.11).ConclusionsThe meta-model outperformed published models and showed a robust predictive capacity for predicting the individualized risk of postoperative mortality in patients with IE.Protocol registrationPROSPERO (registration number CRD42020192602).     

Clinical and echocardiographic predictors of embolism in infective endocarditis: systematic review and meta-analysis

Objectives

Prediction of embolic events (EEs) in infective endocarditis (IE) could inform clinical decisions, such as surgical timing. We conducted a systematic review to more precisely define associations between risk factors and EEs.

Course and outcome of Staphylococcus aureus bacteraemia: a retrospective analysis of 308 episodes in a Swiss tertiary-care centre

Staphylococcus aureus bacteraemia (SAB) is associated with substantial morbidity and mortality worldwide. The charts of adult patients with SAB who were hospitalised in a Swiss tertiary-care centre between 1998 and 2002 were studied retrospectively. In total, 308 episodes of SAB were included: 2% were caused by methicillin-resistant strains; 49% were community-acquired; and 51% were nosocomial. Bacteraemia without focus was the most common type of community-acquired SAB (52%), whereas intravenous catheter-related infection predominated (61%) among nosocomial episodes of SAB. An infectious diseases (ID) specialist was consulted in 82% of all cases; 83% received appropriate antibiotic treatment within 24 h of obtaining blood cultures. Overall hospital-associated mortality was 20%. Community-acquired SAB was associated independently with a higher mortality rate than nosocomial SAB (26% vs. 13%; p 0.009). Independent risk-factors for a fatal outcome were age (p < 0.001), immunosuppression (p 0.007), alcoholism (p < 0.001), haemodialysis (p 0.03), acute renal failure (p < 0.001) and septic shock (p < 0.001). Consultation with an ID specialist was associated with a better outcome in univariate analysis (p < 0.001). Compared with a previous retrospective analysis performed at the same institution between 1980 and 1986, there was a 140% increase in community-acquired SAB, a 60% increase in catheter-related SAB, and a 14% reduction in mortality. In conclusion, mortality in patients with SAB remained high, despite effective antibiotic therapy. Patients with community-acquired SAB were twice as likely to die as patients with nosocomial SAB. Consultation with an ID specialist may reduce mortality in patients with SAB.     

Predictors of clinical and microbiological treatment failure in patients with methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia: a retrospective cohort study in a region with low MRSA prevalence

Invasive infections with methicillin-resistant Staphylococcus aureus (MRSA) have been associated with increased morbidity and mortality. The aim of the present study was to identify independent predictors of early mortality and treatment failure in patients with MRSA bacteraemia. A total of 132 adult patients who developed MRSA bacteraemia during hospitalization in the University Hospital of Vienna between 2000 and 2011 were screened and 124 were included in a retrospective cohort study. Patient demographics, source of bacteraemia, antimicrobial treatment and microbiological characteristics were evaluated. The 28-day crude mortality was 30.6%. Predictors of early mortality identified in multivariate Cox regression analysis included higher patientage (adjusted hazard ratio (aHR) 1.03, 95% CI 1.01–1.06, p 0.006), pneumonia (aHR 3.86, 95% CI 1.83–8.12, p <0.001) and failure to use MRSA active treatment (aHR 8.77, 95% CI 3.50–21.93, p <0.001). Ninety-one (73.4%) patients received glycopeptides as specific MRSA treatment. Of 63 patients treated with vancomycin, only 14 (22.6%) patients had aimed trough levels of 15–20 mg/L. Vancomycin MIC ≥ 2 mg/L was detected in 28.2% and was associated with glycopeptide pretreatment (p 0.001). All MRSA isolates were susceptible to linezolid and tigecycline. Persistent bacteraemia ≥ 7 days was documented in 25 (20.2%) patients. Independent determinants for microbiological eradication failure in patients with MRSA bacteraemia included endocarditis (p <0.001) and vancomycin trough levels (p 0.014), but not vancomycin MIC. Failure of clinical and microbiological eradication of MRSA among patients with MRSA bacteraemia was associated with clinical entity rather than with bacterial traits. Pharmacokinetic parameters seem to be decisive on microbiological and clinical success.     

Methicillin-resistant Staphylococcus aureus: risk assessment and infection control policies.

The endemic state of methicillin-resistant Staphylococcus aureus (MRSA) occurs through a constant influx of MRSA into the healthcare setting from newly admitted MRSA-positive patients, followed by cross-transmission among inpatients and an efflux of MRSA from the hospital with discharged patients. To date, most MRSA prevention strategies have targeted cross-transmission among hospitalised patients. Intensive concerted interventions that include isolation can reduce the MRSA incidence substantially. However, debate continues about the cost-effectiveness of infection control policies, including screening protocols, to control the influx of MRSA into hospitals. The rationale and cost-effectiveness of wide screening, as compared to targeted screening, should be further studied using appropriate statistical approaches and economic modelling.     

Serum IL-27 predicts the severity and prognosis in patients with community-acquired pneumonia: a prospective cohort study

Background: The previous studies have revealed that IL-27 was involved in the pathophysiology of pulmonary inflammatory diseases. However, the role of IL-27 in community-acquired pneumonia (CAP) was unclear. The goal of this research was to explore the associations of serum IL-27 with the severity and prognosis among CAP patients through a prospective cohort study.Methods: The whole of 239 healthy population and 239 CAP patients were enrolled. Fasting blood samples were collected. Inflammatory cytokines were detected using enzyme linked immunosorbent assay (ELISA). Demographic characteristics and clinical information were analyzed.Results: Serum IL-27 on admission was significantly risen in CAP patients compared with control subjects. Besides, serum IL-27 was gradually increased in line with CAP severity scores. Correlative analysis suggested that serum IL-27 was associated with blood routine indices, renal function, liver function, myocardial function and inflammatory cytokines. Linear and logistic regression analyses revealed that serum IL-27 was positively correlated with CAP severity scores. Logistic regression analysis demonstrated that serum higher IL-27 on admission elevated the risks of vasoactive agent usage and longer hospital stay during hospitalization among CAP patients.Conclusions: Serum IL-27 is markedly and positively associated with the severity and poor prognosis among CAP patients, indicating that IL-27 may involve in the pathophysiological process of CAP. Serum IL-27 may be used as a biomarker for diagnosis and prognosis in CAP patients.     

Time-trends for Gram-negative and multidrug-resistant Gram-positive bacteria associated with nosocomial infections in German intensive care units between 2000 and 2005.

This study analysed the time-trends for bacteria associated with nosocomial lower respiratory tract infections (LRTIs), bloodstream infections (BSIs) and urinary tract infections (UTIs) that were reported to the German Nosocomial Infection Surveillance System for intensive care units (ICUs). Data concerning 19 822 nosocomial infections were submitted by 139 ICUs between 2000 and 2005. There was a significant increase in the proportion of Gram-negative bacteria causing LRTIs (from 63.9% to 68.4%) and UTIs (from 65.3% to 68.6%). The proportion of BSIs caused by Gram-negative bacteria declined significantly, from 36.4% to 22.7%. The frequency of methicillin-resistant Staphylococcus aureus among all S. aureus isolates increased from 19.8% to 37.2%.     

Detection of Panton–Valentine leukocidin gene in Staphylococcus aureus by LightCycler PCR: clinical and epidemiological aspects

The prevalence of the Panton-Valentine leukocidin (PVL) gene in Staphylococcus aureus was investigated with a simple, reproducible and rapid real-time LightCycler SYBR Green I PCR assay. The PVL gene was detected in one isolate from 65 patients with S. aureus bacteraemia, in four isolates from 55 patients with respiratory tract infections, and in two isolates from 91 patients with cutaneous infections. In contrast, 15 of 25 cutaneous isolates of methicillin-resistant S. aureus (MRSA) were positive. All PVL-positive cutaneous MRSA isolates were community-acquired and comprised three different clones as determined by pulsed-field gel electrophoresis. The PVL gene was detected in isolates from patients with recurrent primary skin infections and S. aureus bacteraemia, but PVL did not seem to be an important virulence factor in the pathogenesis of staphylococcal bacteraemia.     

Detection of antibiotic resistant E. coli and Enterococcus spp. in stool of healthy growing children in Portugal

From stool specimens of 118 healthy children's (1–14 years) in Portugal 92 E. coli and 101 Enterococcu s spp. strains have been isolated. Almost half (40.2%) of the E. coli isolates were resistant to ampicillin, 25.0% were resistant to tetracycline and 26.1% were resistant to streptomycin. Resistance genes detected by specific PCR included bla_TEM and/or bla_SHV and/or bla_CTX‐M (33 of 37 ampicillin and/or cefotaxime resistant isolates), tet (A) and/or tet (B) (16 of 23 tetracycline‐resistant isolates), aad A (19 of 24 streptomycin‐resistant isolates), cml A (in the two chloramphenicol‐resistant isolates), aac (3)‐II with/without aac (3)‐IV (in the four gentamicin‐resistant isolates), sul 1 and/or sul 2 and/or sul 3 (in all trimethoprim/sulfamethoxazole resistant isolates). The majority of the resistant E. coli isolates (69.1%) belonged to phylogenetic group B2. Of the enterococci isolates E. faecium (n = 53), E. faecalis (n = 41), E. hirae (n = 4) and E. durans (n = 3) more than one‐fourth (28.7%) of the isolates were resistant to tetracycline; 21.8% were resistant to erythromycin and 8.9% were resistant to kanamycin. Resistance genes detected by PCR in enterococci included aph (3)′‐IIIa (in all kanamycin‐resistant isolates), aac (6′) (in all gentamicin‐resistant isolates), tet (M) and/or tet (L) (26 of 29 tetracycline‐resistant isolates), erm (B) (17 of 22 erythromycin‐resistant isolates). This survey showed that faecal bacteria such as E. coli and enterococci of healthy growing children's could be a reservoir of antimicrobial resistance genes. (© 2009 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Caregiver supervision and child-injury risk: I. Issues in defining and measuring supervision; II. Findings and directions for future research

OBJECTIVE: To discuss the role of caregiver supervision in child-injury risk, with attention given to definitional and methodological issues and outlining important questions to be addressed in future research. METHODS: Analysis, synthesis, and critique of existing literature. RESULTS: Comparisons across studies are difficult because of insufficient specificity regarding what constitutes supervision. Hence, a multi-dimensional definition of supervision is developed based on the literature. Numerous issues arise when attempting to measure supervision and these are extensively discussed, along with reporting on the recent development of two questionnaire measures of supervision (Beliefs About Supervision Questionnaire and Parent Supervision Attributes Profile Questionnaire) that have shown good validity and hold promise for addressing the problem of measuring caregiver supervision in reliable and valid ways. A review of the findings on relations between supervision and child-injury risk reveals that many substantive questions remain unanswered. A number of recommendations for future research are given and a conceptual model is presented that focuses attention on the need for research that examines how factors interact to influence child-injury risk. This model has relevance not only for research but also for prevention and serves to emphasize the complementary nature of environment-oriented and person-oriented approaches to child-injury prevention. CONCLUSION: Direct evidence linking supervision to child-injury risk is scarce and many important questions remain unanswered. Based on the conceptual model presented, in future research it is important to examine how supervision interacts with other key factors to influence children's risk of injury.