Valerie B. Haley

Evaluation of spatial filters to create smoothed maps of health data

Spatial filters have been used as an easy and intuitive way to create smoothed disease maps. Birth weight data from New York State for 1994 and 1995 are used to compare the traditional filter type of fixed geographical size with a filter size of constant or nearly constant population size. The latter are more appropriate for mapping disease in geographic areas with widely varying population density, such as New York State. Issues such as the choice of population size for the filter, the scale of smoothing, the ability to detect true spatial variation and the ability to smooth over random spatial noise are evaluated and discussed.

Surveillance of the short-term impact of fine particle air pollution on cardiovascular disease hospitalizations in New York State.

BackgroundStudies have shown that the effects of particulate matter on health vary based on factors including the vulnerability of the population, health care practices, exposure factors, and the pollutant mix.MethodsWe used time-stratified case-crossover to estimate differences in the short-term impacts of PM2.5 on cardiovascular disease hospital admissions in New York State by geographic area, year, age, gender, co-morbid conditions, and area poverty rates.ResultsPM2.5 had a stronger impact on heart failure than other cardiovascular diagnoses, with 3.1% of heart failure admissions attributable to short-term PM2.5 exposure over background levels of 5 ug/m3. Older adults were significantly more susceptible to heart failure after short-term ambient PM2.5 exposure than younger adults.ConclusionThe short-term impact of PM2.5 on cardiovascular disease admissions, and modifications of that impact, are small and difficult to measure with precision. Multi-state collaborations will be necessary to attain more precision to describe spatiotemporal differences in health impacts.

Seasonality and trend in blood lead levels of New York State children

BackgroundEnvironmental exposure to lead remains a significant health problem for children. The costs of lead exposure in children are estimated to be considerably more than other childhood diseases of environmental origin. While long-term trends in blood lead levels (BLLs) among children are declining, seasonal variation persists. Cross-sectional studies have found a peak in summer months. Part of this variation may be due to increased exposure to lead paint on window sills and through increased contact with soils containing lead during the summer. The current study represents the largest published population-based study on seasonality and trends in the BLLs of children to date. In addition, the results offer a comparison of recent data on seasonality of BLLs in New York State children, to studies conducted over the past three decades.Methods262,687 New York State children born between 1994 and 1997 were screened for blood lead within 2 weeks of their first or second birthdays. Time series analyses of blood lead data from these children were conducted to study the seasonality and trends of BLLs.ResultsChildrens blood lead values showed a distinct seasonal cycle on top of a long-term decreasing trend. The geometric mean BLL declined by about 24% for children born between 1994 and 1997. The prevalence of elevated BLLs in two-year-olds was almost twice that in one-year-olds over the time period. Nearly twice as many children had elevated BLLs in the late summer compared to late winter/early spring. In this and previous cross-sectional studies, the amount of seasonality as a proportion of the mean ranged between 15% and 30%.ConclusionPediatricians should be aware of the seasonality of BLLs. For example, if a two-year-old receives a borderline result during the winter, it is possible that the levels would have been higher if he had been tested during the summer. However, physicians should continue to screen children at their normally scheduled well-child visits rather than delaying until summertime and possibly postponing the discovery of an elevated BLL. Age, season, and time trends still need to be considered in lead studies and result interpretation.

Geographic Analysis of Blood Lead Levels in New York State Children Born 1994–1997

We examined the geographic distribution of the blood lead levels (BLLs) of 677,112 children born between 1994 and 1997 in New York State and screened before 2 years of age. Five percent of the children screened had BLLs higher than the current Centers for Disease Control and Prevention action level of 10 μg/dL. Rates were higher in upstate cities than in the New York City area. We modeled the relationship between BLLs and housing and socioeconomic characteristics at the ZIP code level. Older housing stock, a lower proportion of high school graduates, and a higher percentage of births to African-American mothers were the community characteristics most associated with elevated BLLs. Although the prevalence of children with elevated BLLs declined 44% between those born in 1994 and those born in 1997, the rate of improvement may be slowing down. Lead remains an environmental health problem in inner-city neighborhoods, particularly in upstate New York. We identified areas having a high prevalence of children with elevated BLLs. These communities can be targeted for educational and remediation programs. The model locates areas with a higher or lower prevalence of elevated BLLs than expected. These communities can be studied further at the individual level to better characterize the factors that contribute to these differences.

Risk factors for coronary artery bypass graft chest surgical site infections in New York State, 2008

BACKGROUND All hospitals in New York State (NYS) are required to report surgical site infections (SSIs) occurring after coronary artery bypass graft surgery. This report describes the risk adjustment method used by NYS for reporting hospital SSI rates, and additional methods used to explore remaining differences in infection rates. METHODS All patients undergoing coronary artery bypass graft surgery in NYS in 2008 were monitored for chest SSI following the National Healthcare Safety Network protocol. The NYS Cardiac Surgery Reporting System and a survey of hospital infection prevention practices provided additional risk information. Models were developed to standardize hospital-specific infection rates and to assess additional risk factors and practices. RESULTS The National Healthcare Safety Network risk score based on duration of surgery, American Society of Anesthesiologists score, and wound class were not highly predictive of chest SSIs. The addition of diabetes, obesity, end-stage renal disease, sex, chronic obstructive pulmonary disease, and Medicaid payer to the model improved the discrimination between procedures that resulted in SSI and those that did not by 25%. Hospital-reported infection prevention practices were not significantly related to SSI rates. CONCLUSIONS Additional risk factors collected using a secondary database improved the prediction of SSIs, however, there remained unexplained variation in rates between hospitals.

Comparison of 2 Clostridium difficile surveillance methods: National Healthcare Safety Network's laboratory-identified event reporting module versus clinical infection surveillance.

OBJECTIVE To determine whether the Centers for Disease Control and Preventions National Healthcare Safety Network (NHSN) laboratory-identified (LabID) event reporting module for Clostridium difficile infection (CDI) is an adequate proxy measure of clinical CDI for public reporting purposes by comparing the 2 surveillance methods. DESIGN Validation study. SETTING Thirty New York State acute care hospitals. METHODS Six months of data were collected by 30 facilities using a clinical infection surveillance definition while also submitting the NHSN LabID event for CDI. The data sets were matched and compared to determine whether the assigned clinical case status matched the LabID case status. A subset of mismatches was evaluated further, and reasons for the mismatches were quantified. Infection rates determined using the 2 definitions were compared. RESULTS A total of 3,301 CDI cases were reported. Analysis of the original data yielded a 67.3% (2,223/3,301) overall case status match. After review and validation, there was 81.3% (2,683/3,301) agreement. The most common reason for disagreement (54.9%) occurred because the symptom onset was less than 48 hours after admission but the positive specimen was collected on hospital day 4 or later. The NHSN LabID hospital onset rate was 29% higher than the corresponding clinical rate and was generally consistent across all hospitals. CONCLUSIONS Use of the NHSN LabID event minimizes the burden of surveillance and standardizes the process. With a greater than 80% match between the NHSN LabID event data and the clinical infection surveillance data, the New York State Department of Health made the decision to use the NHSN LabID event CDI data for public reporting purposes.

Developing consistent data and methods to measure the public health impacts of ambient air quality for Environmental Public Health Tracking: progress to date and future directions.

Environmental Public Health Tracking (EPHT) staff at the state and national levels are developing nationally consistent data and methods to estimate the impact of ozone and fine particulate matter on hospitalizations for asthma and myocardial infarction. Pilot projects have demonstrated the feasibility of pooling state hospitalization data and linking these data to The United States Environmental Protection Agency (EPA) statistically based ambient air estimates for ozone and fine particulates. Tools were developed to perform case-crossover analyses to estimate concentration–response (C-R) functions. A weakness of analyzing one state at a time is that the effects are relatively small compared to their confidence intervals. The EPHT program will explore ways to statistically combine the results of peer-reviewed analyses from across the country to provide more robust C-R functions and health impact estimates at the local level. One challenge will be to routinely share data for these types of analyses at fine geographic and temporal scales without disclosing confidential information. Another challenge will be to develop C-R estimates which take into account time, space, or other relevant effect modifiers.

Surgical site infection risk factors identified for patients undergoing colon procedures, New York State 2009-2010.

BACKGROUND Since 2007, New York State (NYS) hospitals have been required to report surgical site infections (SSIs) following colon procedures to the NYS Department of Health, using the National Healthcare Safety Network (NHSN). The purpose of this study was to identify risk factors for the development of SSIs in patients undergoing colon procedures. METHODS NYS has been conducting validation studies at hospitals to assess the accuracy of the surveillance data reported by the participating hospitals. A sample of patients undergoing colon procedures in NYS hospitals were included in hospital-acquired infection program validation studies in 2009 and 2010. Medical chart reviews and on-site visits were performed to verify patient information reported and to evaluate additional risk factors for SSI. Bivariable and multivariable logistic regressions were performed. RESULTS A total of 2,656 colon procedures were included in this analysis, including 698 SSI cases. Multivariable analysis indicated that SSI following colon procedure was associated with body mass index greater than 30 (odds ratio [OR], 1.48 [95% confidence interval (CI), 1.21-1.80]), male sex (OR, 1.34 [95% CI, 1.10-1.64]), American Society of Anesthesiologists physical classification score greater than 3 (OR, 1.33 [95% CI, 1.08-1.64]), procedure duration, transfusion (OR, 1.32 [95% CI, 1.05-1.66]), left-side colon surgical procedures, other gastroenterologic procedures, irrigation, hospital bed size greater than 500, and medical school affiliation. CONCLUSIONS Male sex, obesity, transfusion, type of procedure, and prolonged duration were significant factors associated with overall infection risk after adjusting other factors. Additional factors not collected in the NHSN slightly improved prediction of SSIs.

Trends in validity of central line–associated bloodstream infection surveillance data, New York State, 2007-2010

BACKGROUND In 2007, New York State (NYS) hospitals began mandatory public reporting of central line-associated bloodstream infection (CLABSI) data associated with intensive care units (ICUs) into the National Healthcare Safety Network (NHSN). Facilities were required to use the NHSN device-associated CLABSI criteria to identify laboratory-confirmed bloodstream infections. METHODS Onsite audits were conducted in ICUs by NYS hospital-acquired infection program staff using a standardized database. Hospitals provided ICU patient medical records with a positive blood culture during a selected time frame. RESULTS Between 2007 and 2010, an average of 79% of all reporting hospitals were audited annually. Of the 5,697 patients audited, 3,104 (54%) had a central line in place, and 650 of the patients with a central line (21%) were identified as having a CLABSI by the hospital-acquired infection program reviewer. Between 2007 and 2010, the specificity increased from 90% to 99%, whereas the sensitivity remained stable at approximately 71%. As a result of the audit process, the NYS 2010 CLABSI rate increased by 5.6%. CONCLUSIONS A standardized audit process has helped improve the accuracy of CLABSI reporting. Data validation provides consistent data for measuring the progress of infection prevention strategies and allows for relevant comparison of ICU data.

Effect of 2013 National Healthcare Safety Network definition changes on central line bloodstream infection rates: Audit results from the New York State Department of Health

Surveillance criteria for central line-associated bloodstream infections (CLABSIs) are continually being refined to more accurately reflect infections related to central lines. An audit of 567 medical records from adult, pediatric, and neonatal intensive care unit patients with a central line and a positive blood culture showed a 16% decrease in CLABSI rates after the 2013 National Healthcare Safety Network definitions compared with the 2012 definitions.