PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 12674902

Limitations of electronic databases: a caution.

James E Svenson·Susan H Pollack·Mary E Fallat·Jeanne L Drapeau
View original source ↗Export metadata ↓

Abstract

OBJECTIVE: The purpose of this study was to assess the completeness and accuracy of information from two electronic datasets, one of which is voluntarily submitted, the other submitted by mandate. METHODS: Emergency department (ED) data have been voluntarily submitted by several hospitals to the Kentucky Emergency Medical Services Information System (KEMSIS). Similar information on all patients admitted to the hospital has been submitted by mandate to the state under the Uniform Billing Act (UB92). UB92 data for patients with at least one diagnosis code > or = 800 were available. The KEMSIS and UB 92 data for one hospital were compared to those manually abstracted from the ED log and medical records for completeness and accuracy. RESULTS: There were 316 patients listed on the ED log that were subsequently admitted to the hospital. The KEMSIS database contained 266 (84%) of these records, but only 91 (34%) were classified as having been admitted. Of those correctly classified as admitted, only 25 (27%, or 9% of the total 266) were correctly classified as to the hospital of admission (directly to hospital or transferred to another facility). Discharge diagnoses in the KEMSIS database and hospital records were concordant in 240 (90%) of the patients, even for those misclassified as to disposition. There were 37 patients listed in the ED log admitted during the study period with at least one discharge diagnosis field > or = 800. Eight patients were transferred to another institution, making the total population available for study period 29. Only eight (28%) of these patients were included in the UB92 database. The diagnosis codes were concordant between the UB 92 data and ED log in all cases. CONCLUSIONS: There is significant misclassification and/or omission in electronic databases. This is true regardless of whether data is reported voluntarily or by mandate. Electronic data must be independently validated before they are used for policy or research purposes.

Explore related subjects

Databases, FactualEmergency Medical ServicesEmergency Service, HospitalHospital Information SystemsHospital RecordsHospitalizationHumansKentuckyReproducibility of ResultsComparative StudyJournal ArticleResearch Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

James E Svenson, Susan H Pollack, Mary E Fallat, Jeanne L Drapeau. 2003. Limitations of electronic databases: a caution.. https://pubmed.ncbi.nlm.nih.gov/12674902/

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Performance of the FearID earprint identification system.

The Forensic Ear Identification (FearID) research project was started in order to study the strength of evidence of earprints found on crime scenes. For this purpose, a sample of earprints from 1229 donors over three countries was collected. From each donor three left and three right earprints were gathered. On the one hand, operators denoted contours of the earprints to facilitate segmentation of the images, on the other anthropological specialists denoted anatomically specific locations. On the basis of this, methods for automated classification were developed and used for training of a system that classifies pairs of prints as 'matching' or 'non-matching'. Comparing lab quality prints, the system has an equal error rate of 4%. Starting from a reference database containing two prints per ear, hitlist behaviour is such that in 90% of all query searches the best hit is in the top 0.1% of the list. The results become less favourable (equal error rate of 9%) for print/mark comparisons.

Databases, Factual↗
PubMed2007

Normal limits of ejection fraction and volumes determined by gated SPECT in clinically normal patients without cardiac events: a study based on the J-ACCESS database.

PURPOSE: Quantitative gated single-photon emission computed tomography (SPECT) is known to have high accuracy and precision for measurement of the principal cardiac functional parameters. We hypothesised that normal values for EF and LV volumes may differ among nationalities, and that optimal threshold values specific to the study population are required. METHODS: Among 4,670 consecutively registered patients for a J-ACCESS (Japanese investigation regarding prognosis based on gated SPECT) study from 117 hospitals, a total of 268 (149 women, 119 men) were selected who had no baseline cardiac diseases and had experienced no cardiac events during the preceding 3-year period. A gated SPECT study was performed with 99mTc-tetrofosmin and analysed with Cedars Sinai Medical Center's quantitative gated SPECT (QGS) software. The results in respect of ejection fraction (EF), end-diastolic volume (EDV), end-systolic volume (ESV) and stroke volume (SV), and EDV, ESV and SV normalised by body surface area (EDVI, ESVI and SVI), were calculated and summarised to obtain normal limits. RESULTS: EF for women and men was 74 +/- 9% and 63 +/- 7%, respectively (p < 0.0001). EDV, ESV and SV were significantly smaller in women than in men. Based on multiple regressions for linear models, the primary and secondary predictors of EF, EDVI, ESVI were gender and age. By stepwise multiple regression analysis, a statistically significant third predictor for EDV, ESV, SV and SVI was body weight. No colinearity was found between age and body weight. Important factors for the studied Japanese population included a high incidence of small hearts in women and the relatively advanced age of the population (the mean age +/-SD was 64.1 +/- 10.0 years for women and 60.9 +/- 11.7 years for men). CONCLUSION: EF and volumes determined by gated SPECT with QGS were significantly affected by gender and age, with body weight as a third predictor for volumes. Moreover, the normal limits were so specific for the population studied that standards appropriate for the study in question should be utilised.

Databases, Factual↗