BACKGROUND: The Centor and McIsaac scores guide testing and treatment for group A streptococcal (GAS) pharyngitis in patients presenting with a sore throat, but they were derived on relatively small samples. We perform a national-scale validation of the prediction models on a large, geographically diverse population. METHODS: We analyzed data collected from 206,870 patients 3 years or older who presented with a painful throat to a United States national retail health chain from September 1, 2006, to December 1, 2008. Main outcome measures were the proportions of patients testing positive for GAS pharyngitis according to the Centor and McIsaac scores (both scales, 0-4). RESULTS: For patients 15 years or older, 23% (95% CI, 22%-23%) tested positive for GAS, including 7% (95% CI, 7%-8%) of those with a Centor score of 0; 12% (95% CI, 11%-12%) of those with a Centor score of 1; 21% (95% CI, 21%-22%) of those with a Centor score of 2; 38% (95% CI, 38%-39%) of those with a Centor score of 3; and 57% (95% CI, 56%-58%) of those with a Centor score of 4. For patients 3 years or older, 27% (95% CI, 27%-27%) tested positive for GAS, including 8% (95% CI, 8%-9%) of those testing positive with aMcIsaac score of 0; 14% (95% CI, 13%-14%) of those with a McIsaac score of 1; 23% (95% CI, 23%-23%) of those with a McIsaac score of 2; 37% (95% CI, 37%-37%) of those with a McIsaac score of 3; and 55% (95% CI, 55%-56%) of those with a McIsaac score of 4. The 95% CIs overlapped between our retail health chain-derived probabilities and the prior reports. CONCLUSION: Our study validates the Centor and McIsaac scores and more precisely classifies risk of GAS infection among patients presenting with a painful throat to a retail health chain.

STUDY OBJECTIVE: We analyze the risks and benefits of alternative treatment strategies for non-septic-appearing febrile patients with influenza-like illnesses and possible exposure to anthrax. METHODS: We used a decision analytic model to evaluate 6 testing and treatment strategies in an emergency department. Patients were non-septic-appearing and had influenza-like illnesses but low likelihood of exposure to anthrax. The following interventions were used: (1) no empiric antibiotics; (2) blood culture and treatment only if the result was positive; (3) rapid testing for influenza and, for those who tested negative, treatment with 60 days of ciprofloxacin; (4) a two-test strategy in which all patients were first tested for influenza; those who tested negative had a blood culture test and were treated empirically with ciprofloxacin for 3 days while waiting for blood culture results; (5) culture test for all patients and treatment with ciprofloxacin for up to 3 days while waiting for blood culture results; and (6) treatment of all patients with ciprofloxacin empirically for 60 days. Main outcome measures were deaths, complications from anthrax, adverse events from ciprofloxacin, and ciprofloxacin patient-days. RESULTS: For nonzero probabilities of anthrax, patient mortality was always lowest in the strategies in which all patients were treated empirically for anthrax either for 60 days or for 3 days pending blood culture results. These strategies, however, were associated with more morbidity (more ciprofloxacin patient-days and more antibiotic adverse events) than were strategies without empiric treatment. The numbers of adverse events and antibiotic patient-days were reduced substantially with the two-test strategy, in which patients with influenza were identified early and not treated. In general, for probabilities of anthrax equaling or exceeding 2%, treating all patients empirically for 60 days was best, but for probabilities between 0.1% and 2%, the sensitivity of blood culture for anthrax determined the optimal strategy: when the sensitivity exceeded 95%, a short course of empiric ciprofloxacin until blood culture results became available was best, but for sensitivities below 95%, more aggressive empiric antibiotics use was warranted. The proportion of patients with influenza in the community affected the choice of strategy, so that seasonal variation exists. CONCLUSION: During influenza season, our findings support rapid testing for influenza, followed by empiric treatment for anthrax pending blood culture results for those who test negative for influenza. Our results help to highlight the importance of developing rapid and sensitive tests for anthrax and of developing improved surveillance and methods to calculate the previous probability of attacks.

We present the approach taken in a Massachusetts-based national demonstration project to integrate the PING personally controlled health record (PCHR) with the MA-SHARE network, the state-wide inter-organizational data exchange. We describe how we have created a patient-controlled gateway to the network, and how PCHRs have become a first class data source in the network.

Syndromic surveillance systems are being deployed widely to monitor for signals of covert bioterrorist attacks. Regional systems are being established through the integration of local surveillance data across multiple facilities. We studied how different methods of data integration affect outbreak detection performance. We used a simulation relying on a semi-synthetic dataset, introducing simulated outbreaks of different sizes into historical visit data from two hospitals. In one simulation, we introduced the synthetic outbreak evenly into both hospital datasets (aggregate model). In the second, the outbreak was introduced into only one or the other of the hospital datasets (local model). We found that the aggregate model had a higher sensitivity for detecting outbreaks that were evenly distributed between the hospitals. However, for outbreaks that were localized to one facility, maintaining individual models for each location proved to be better. Given the complementary benefits offered by both approaches, the results suggest building a hybrid system that includes both individual models for each location, and an aggregate model that combines all the data. We also discuss options for multi-level signal integration hierarchies.

OBJECTIVE: To develop a novel diagnostic algorithm for Lyme disease among children with facial palsy by integrating public health surveillance data with traditional clinical predictors. DESIGN: Retrospective cohort study. SETTING: Children's Hospital Boston emergency department, 1995-2007. PATIENTS: Two hundred sixty-four children (aged 20 years) with peripheral facial palsy who were evaluated for Lyme disease. MAIN OUTCOME MEASURES: Multivariate regression was used to identify independent clinical and epidemiologic predictors of Lyme disease facial palsy. RESULTS: 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. CONCLUSIONS: 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.

Reams of data pertaining directly to the core health services research mission are accumulating in large-scale organizational and clinical information systems. Health services researchers who grasp the structure of information systems and databases and the function of software applications can use existing data more effectively, assist in establishing new databases, and develop new tools to survey populations and collect data. At the same time, informaticians are needed who can structure databases that serve the needs of health service research and who can design and evaluate applications that effectively improve health care delivery. As long as health services researchers and informaticians work in separate spheres, however, opportunities to use data from health care encounters to improve care, expand knowledge, and develop more effective policies will be missed. This paper provides a brief exploration of 1) existing successful collaborations between health services researchers and informaticians and 2) needs and opportunities for additional joint work in several core research areas.

BACKGROUND: Influenza and other winter respiratory viruses cause substantial morbidity among children. Previous estimates of the burden of illness of these viruses have neglected to include the emergency department, where a large number of patients seek acute care for respiratory illnesses. This study provides city- and statewide population estimates of the burden of illness attributable to respiratory viruses for children receiving emergency department-based care for respiratory infections during the winter months. METHODS: The number of patients or = 7 years of age presenting to the emergency department of an urban tertiary care pediatric hospital with acute respiratory infections was estimated by using a classifier based on presenting complaints. The rates of specific viral infections in this population were estimated by using the rates of positivity for respiratory syncytial virus, influenza virus, parainfluenza virus, adenovirus, and enterovirus. Local emergency department market share and US Census data enabled determination of the rates of emergency department visits in the Boston, Massachusetts, area and in Massachusetts. RESULTS: During the 11-year study period, the mean yearly number of patients or = 7 years of age presenting to the study emergency department during the winter season was 17397. On the basis of the respiratory classifier, the mean number of patients with an acute respiratory infection was 6923, or 398 per 1000 emergency department visits. In the city population, the mean number of emergency department visits for acute respiratory infections was 17906, which is equivalent to 113.9 per 1000 children residing in the city, and in the state population the mean number was 61529, or 94.5 per 1000 children residing in the state. At the state level, 23114 of the visits were for respiratory syncytial virus, 5650 for influenza, 1751 for parainfluenza virus, 2848 for adenovirus, and 798 for enterovirus. For patients 6 to 23 months of age in the state population, there were 19860 emergency department visits for acute respiratory infections, or 168 per 1000 children in this age group, with 6235 visits resulting from respiratory syncytial virus and 2112 resulting from influenza. CONCLUSION: There is a high incidence of emergency department visits for infectious respiratory illnesses among children. This important component of health care use should be included in estimates of the burden of illness attributable to influenza and other winter respiratory viruses.

OBJECTIVE: To determine whether women who frequently bring their neonates for problem-oriented primary care visits or emergency department visits are at elevated risk of having depressive symptoms. DESIGN: Analysis of 2 prospective cohort studies of mothers and their infants: (1) a telephone interview study of mothers and infants after birth at an urban teaching hospital (the hospital cohort) and (2) the 1988 National Maternal and Infant Health Survey, a nationally representative sample of women who had live births in 1988. PARTICIPANTS: A total of 1015 women in the hospital cohort surveyed at 3 and 8 weeks post partum and 6779 women with data from the national survey. MAIN OUTCOME MEASURE: Depressive symptoms above the Center for Epidemiologic Studies Depression Scale cutoff score of 15. RESULTS: After controlling for sociodemographic variables and parity, women exhibited high levels of depressive symptoms if their infants had more than 1 problem-oriented primary care visit (hospital cohort: odds ratio, 2.0 [95% confidence interval, 1.1-4.3]; national survey cohort: odds ratio, 2.0 [95% confidence interval, 1.5-3.0]). Women were more likely to have high levels of depressive symptoms if their infants had even 1 emergency department visit (hospital cohort: odds ratio, 3.2 [95% confidence interval, 1.5-6.9]). Frequent well-child visits were not associated with maternal depressive symptoms. CONCLUSIONS: Neonatal health care use patterns predict women at risk for postpartum depression. Recognition of these signature patterns of service use by pediatric health care providers may facilitate early diagnosis and treatment of postpartum depression and improve outcomes for women and their families.