Use of population health data to refine diagnostic decision-making for pertussis.

Division of Emergency Medicine, Children's Hospital Boston and Department of Pediatrics, Harvard Medical School, Boston, Massachusetts 02115, USA.
Journal of the American Medical Informatics Association (Impact Factor: 3.93). 01/2010; 17(1):85-90. DOI: 10.1197/jamia.M3061
Source: DBLP

ABSTRACT To improve identification of pertussis cases by developing a decision model that incorporates recent, local, population-level disease incidence.
Retrospective cohort analysis of 443 infants tested for pertussis (2003-7).
Three models (based on clinical data only, local disease incidence only, and a combination of clinical data and local disease incidence) to predict pertussis positivity were created with demographic, historical, physical exam, and state-wide pertussis data. Models were compared using sensitivity, specificity, area under the receiver-operating characteristics (ROC) curve (AUC), and related metrics.
The model using only clinical data included cyanosis, cough for 1 week, and absence of fever, and was 89% sensitive (95% CI 79 to 99), 27% specific (95% CI 22 to 32) with an area under the ROC curve of 0.80. The model using only local incidence data performed best when the proportion positive of pertussis cultures in the region exceeded 10% in the 8-14 days prior to the infant's associated visit, achieving 13% sensitivity, 53% specificity, and AUC 0.65. The combined model, built with patient-derived variables and local incidence data, included cyanosis, cough for 1 week, and the variable indicating that the proportion positive of pertussis cultures in the region exceeded 10% 8-14 days prior to the infant's associated visit. This model was 100% sensitive (p<0.04, 95% CI 92 to 100), 38% specific (p<0.001, 95% CI 33 to 43), with AUC 0.82.
Incorporating recent, local population-level disease incidence improved the ability of a decision model to correctly identify infants with pertussis. Our findings support fostering bidirectional exchange between public health and clinical practice, and validate a method for integrating large-scale public health datasets with rich clinical data to improve decision-making and public health.

Download full-text


Available from: Lise E Nigrovic, Jul 04, 2015
  • [Show abstract] [Hide abstract]
    ABSTRACT: To develop a novel diagnostic algorithm for Lyme disease among children with facial palsy by integrating public health surveillance data with traditional clinical predictors. Retrospective cohort study. Children's Hospital Boston emergency department, 1995-2007. Two hundred sixty-four children (aged <20 years) with peripheral facial palsy who were evaluated for Lyme disease. Multivariate regression was used to identify independent clinical and epidemiologic predictors of Lyme disease facial palsy. Lyme diagnosis was positive in 65% of children from high-risk counties in Massachusetts during Lyme disease season compared with 5% of those without both geographic and seasonal risk factors. Among patients with both seasonal and geographic risk factors, 80% with 1 clinical risk factor (fever or headache) and 100% with 2 clinical factors had Lyme disease. Factors independently associated with Lyme disease facial palsy were development from June to November (odds ratio, 25.4; 95% confidence interval, 8.3-113.4), residence in a county where the most recent 3-year average Lyme disease incidence exceeded 4 cases per 100,000 (18.4; 6.5-68.5), fever (3.9; 1.5-11.0), and headache (2.7; 1.3-5.8). Clinical experts correctly treated 68 of 94 patients (72%) with Lyme disease facial palsy, but a tool incorporating geographic and seasonal risk identified all 94 cases. Most physicians intuitively integrate geographic information into Lyme disease management, but we demonstrate quantitatively how formal use of geographically based incidence in a clinical algorithm improves diagnostic accuracy. These findings demonstrate potential for improved outcomes from investments in health information technology that foster bidirectional communication between public health and clinical settings.
    JAMA Pediatrics 01/2011; 165(1):61-7. DOI:10.1001/archpediatrics.2010.250
  • [Show abstract] [Hide abstract]
    ABSTRACT: Clinical prediction rules do not incorporate real-time incidence data to adjust estimates of disease risk in symptomatic patients. To measure the value of integrating local incidence data into a clinical decision rule for diagnosing group A streptococcal (GAS) pharyngitis in patients aged 15 years or older. Retrospective analysis of clinical and biosurveillance predictors of GAS pharyngitis. Large U.S.-based retail health chain. 82 062 patient visits for pharyngitis. Accuracy of the Centor score was compared with that of a biosurveillance-responsive score, which was essentially an adjusted Centor score based on real-time GAS pharyngitis information from the 14 days before a patient's visit: the recent local proportion positive (RLPP). Increased RLPP correlated with the likelihood of GAS pharyngitis (r(2) = 0.79; P < 0.001). Local incidence data enhanced diagnostic models. For example, when the RLPP was greater than 0.30, managing patients with Centor scores of 1 as if the scores were 2 would identify 62, 537 previously missed patients annually while misclassifying 18, 446 patients without GAS pharyngitis. Decreasing the score of patients with Centor values of 3 by 1 point for an RLPP less than 0.20 would spare unnecessary antibiotics for 166, 616 patients while missing 18, 812 true-positive cases. Analyses were conducted retrospectively. Real-time regional data on GAS pharyngitis are generally not yet available to clinicians. Incorporating live biosurveillance data into clinical guidelines for GAS pharyngitis and other communicable diseases should be considered for reducing missed cases when the contemporaneous incidence is elevated and for sparing unnecessary antibiotics when the contemporaneous incidence is low. Delivering epidemiologic data to the point of care will enable the use of real-time pretest probabilities in medical decision making.
    Annals of internal medicine 09/2011; 155(6):345-52. DOI:10.1059/0003-4819-155-6-201109200-00002
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Mycoplasma pneumoniae (M. pneumoniae) is a significant cause of community-acquired pneumonia in children and adolescents. Treatment with macrolide antibiotics is recommended. However, M. pneumoniae is difficult to diagnose based on clinical symptoms and signs. Diagnostic uncertainty can lead to inappropriate antibiotic prescribing, which may worsen clinical prognosis and increase antibiotic resistance. The objectives of this review are (i) to assess the diagnostic accuracy of symptoms and signs in the clinical recognition of M. pneumoniae in children and adolescents with community-acquired pneumonia; and (ii) to assess the influence of potential sources of heterogeneity on the diagnostic accuracy of symptoms and signs in the clinical recognition of M. pneumoniae. We searched MEDLINE (January 1950 to 26 June 2012) and EMBASE (January 1980 to 26 June 2012). We identified additional references by handsearching the reference lists of included articles and snowballing. We searched the reference lists of relevant systematic reviews identified by searching the Medion database, Database of Reviews of Effects 2012, Issue 6 (25 June 2012) and the Cochrane Register of Diagnostic Test Accuracy studies (2 July 2012). Experts in the field reviewed our list of included studies for any obvious omissions. We included peer-reviewed published studies which prospectively and consecutively recruited children with community-acquired pneumonia from any healthcare setting, confirmed the presence of M. pneumoniae using serology with or without other laboratory methods and reported data on clinical symptoms and signs in sufficient detail to construct 2 x 2 tables. One review author scanned titles to exclude obviously irrelevant articles. Two review authors independently scanned the remaining titles and abstracts, reviewed full-text versions of potentially relevant articles, assessed the quality of included articles and extracted data on study characteristics and the following clinical features: cough, wheeze, coryza, crepitations, fever, rhonchi, shortness of breath, chest pain, diarrhea, myalgia and headache.We calculated study-specific values for sensitivity, specificity and positive and negative likelihood ratios with 95% confidence intervals (CIs). We estimated the post-test probability of M. pneumoniae based on the absence or presence of symptoms and signs.We calculated pooled sensitivities, specificities, positive and negative likelihood ratios with 95% CIs for symptoms and signs where data were reported by at least four included studies by fitting a bivariate normal model for the logit transforms of sensitivity and specificity. We explored potential sources of heterogeneity by fitting bivariate models with covariates using multi-level mixed-effects logistic regression. We performed sensitivity analyses excluding data from studies for which we were concerned about the representativeness of the study population and/or the acceptability of the reference standard. Our search identified 8299 articles (excluding duplicates). We examined the titles and abstracts of 1125 articles and the full-text versions of 97 articles. We included seven studies in our review, which reported data from 1491 children; all were conducted in hospital settings. Overall, study quality was moderate. In two studies the presence of chest pain more than doubled the probability of M. pneumoniae. Wheeze was 12% more likely to be absent in children with M. pneumoniae (pooled positive likelihood ratio (LR+) 0.76, 95% CI 0.60 to 0.97; pooled negative likelihood ratio (LR-) 1.12, 95% CI 1.02 to 1.23). Our sensitivity analysis showed that the presence of crepitations was associated with M. pneumoniae, but this finding was of borderline statistical significance (pooled LR+ 1.10, 95% CI 0.99 to 1.23; pooled LR- 0.66, 95% CI 0.46 to 0.96). M. pneumoniae cannot be reliably diagnosed in children and adolescents with community-acquired pneumonia based on clinical symptoms and signs. Although the absence of wheeze is a statistically significant diagnostic indicator, it does not have sufficient diagnostic value to guide empirical macrolide treatment. Data from two studies suggest that the presence of chest pain more than doubles the probability of M. pneumoniae. However, further research is needed to substantiate this finding. More high quality large-scale studies in primary care settings are needed to help develop prediction rules based on epidemiological data as well as clinical and baseline patient characteristics.
    Cochrane database of systematic reviews (Online) 01/2012; 10(10):CD009175. DOI:10.1002/14651858.CD009175.pub2