medical billing transcription supplier information

The New York City Department of Health and Mental Hygiene has established a syndromic surveillance system that monitors emergency department visits to detect disease outbreaks early. Routinely collected chief complaint information is transmitted electronically to the health department daily and analyzed for temporal and spatial aberrations. Respiratory, fever, diarrhea, and vomiting are the key syndromes analyzed. Statistically significant aberrations or "signals" are investigated to determine their public health importance. In the first year of operation (November 15, 2001, to November 14, 2002), 2.5 million visits were reported from 39 participating emergency departments, covering an estimated 75% of annual visits. Most signals for the respiratory and fever syndromes (64% and 95%, respectively) occurred during periods of peak influenza A and B activity. Eighty-three percent of the signals for diarrhea and 88% of the signals for vomiting occurred during periods of suspected norovirus and rotavirus transmission.

Two recent phenomena have contributed to widespread interest in monitoring nonspecific health indicator data to detect disease outbreaks early. The first is heightened concern about bioterrorism, particularly the ability of public health agencies to detect a large-scale bioterrorist attack in its early stages. The second is the proliferation of electronic databases in healthcare settings. Initially designed to facilitate billing, health information systems capture an increasingly rich array of clinical detail. Recent advances in information technology make extracting, transmitting, processing, and analyzing these data feasible for public health purposes. The emergency department surveillance system we describe is an early prototype of what may become a standard component of modern public health surveillance.

In New York City, emergency department chief complaint surveillance evolved out of the public health response to the September 11, 2001, World Trade Center attacks (1). When this labor-intensive effort ended, the New York City Department of Health and Mental Hygiene (DOHMH) began intensively recruiting hospitals capable of providing emergency department visit data in electronic formats. We describe the methods and chief results from the first 12 months of experience with this electronic system.

Materials and Methods

Data Transmission and Processing

Data files are transmitted to DOHMH 7 days per week, either as attachments to electronic mail messages or through direct file transfer protocol (FTP). Half of participating hospitals have automated the transmission process. Data processing and analysis are carried out on a laptop computer that can be operated either through the DOHMH local area network or through remote dial-up, which facilitates weekend and holiday analysis. Each morning, an analyst retrieves the files, inspects them for quality and completeness, and saves them for processing and analysis in SAS (version 8, SAS Institute Inc., Cary, NC). If a file is not received by 10:00 a.m., the analyst contacts hospitals to obtain missing data. The analysis is typically completed by 1 p.m.

Data files contain the following information for all emergency department patient visits logged during the previous midnight-to-midnight 24-hour period: date and time of visit, age in years, sex, home zip code, and free-text chief complaint. Additionally, some hospitals provide either a visit or medical record number. No other personal identifiers are included. Files arrive in several formats, most commonly as fixed-column or delimited ASCII text. Data are read and translated into a standard format, concatenated into a single SAS dataset, verified for completeness and accuracy, and appended to a master archive.

Syndrome Coding

Emergency department patient visits are categorized into exclusive syndromes based on the patient's chief complaint, a free-text field that captures the patient's own description of his/her illness. We developed a SAS algorithm that scans the chief complaint field for character strings assigned to a syndrome. The coding algorithm is designed to capture the wide variety of misspellings and abbreviations in the chief complaint field. If the chief complaint was blank or uninformative (e.g., "EVAL," "TRIAGE") the record was omitted. If it contained a word or phrase from a single category it was coded for that syndrome, i.e., "SHORTNESS OF BREATH" or "SOB" appearing alone would indicate the respiratory syndrome. If the chief complaint contained words or phrases from multiple categories, it was coded according to the following hierarchy: common cold > sepsis/dead on arrival > respiratory > diarrhea > fever > rash > asthma > vomiting > other visits. The hierarchy attempts to place each chief complaint into a single, specific syndrome (Table 1). Chief complaints containing text strings such as, "cold," "sneeze," "stuffy," or "nasal" are coded as cold and excluded to increase the specificity of the respiratory category for illnesses other than viral rhinitis. The two syndromes of particular interest for bioterrorism surveillance are the respiratory and fever syndromes in persons [greater than or equal to] 13 years of age. Children are excluded due to their high rates of febrile and respiratory illnesses and to limit the number of false signals generated. Respiratory and fever syndromes in children are examined by graphic and CUSUM analyses with SaTScan performed on an ad hoc basis. We monitor the diarrhea and vomiting syndromes in all ages in an effort to detect gastrointestinal outbreaks that may be due to contamination of food or water.