PubMed Health⌕ Search

Biomedical subjects

Genevieve B Melton

Publications and source records attributed to Genevieve B Melton.

14 recordsLinked to original sources
PubMed2006

Using discordance to improve classification in narrative clinical databases: an application to community-acquired pneumonia.

Data mining in electronic medical records may facilitate clinical research, but much of the structured data may be miscoded, incomplete, or non-specific. The exploitation of narrative data using natural language processing may help, although nesting, varying granularity, and repetition remain challenges. In a study of community-acquired pneumonia using electronic records, these issues led to poor classification. Limiting queries to accurate, complete records led to vastly reduced, possibly biased samples. We exploited knowledge latent in the electronic records to improve classification. A similarity metric was used to cluster cases. We defined discordance as the degree to which cases within a cluster give different answers for some query that addresses a classification task of interest. Cases with higher discordance are more likely to be incorrectly classified, and can be reviewed manually to adjust the classification, improve the query, or estimate the likely accuracy of the query. In a study of pneumonia--in which the ICD9-CM coding was found to be very poor--the discordance measure was statistically significantly correlated with classification correctness (.45; 95% CI .15-.62).

Adult↗
PubMed2006

Inter-patient distance metrics using SNOMED CT defining relationships.

BACKGROUND: Patient-based similarity metrics are important case-based reasoning tools which may assist with research and patient care applications. Ontology and information content principles may be potentially helpful tools for similarity metric development. METHODS: Patient cases from 1989 through 2003 from the Columbia University Medical Center data repository were converted to SNOMED CT concepts. Five metrics were implemented: (1) percent disagreement with data as an unstructured "bag of findings," (2) average links between concepts, (3) links weighted by information content with descendants, (4) links weighted by information content with term prevalence, and (5) path distance using descendants weighted by information content with descendants. Three physicians served as gold standard for 30 cases. RESULTS: Expert inter-rater reliability was 0.91, with rank correlations between 0.61 and 0.81, representing upper-bound performance. Expert performance compared to metrics resulted in correlations of 0.27, 0.29, 0.30, 0.30, and 0.30, respectively. Using SNOMED axis Clinical Findings alone increased correlation to 0.37. CONCLUSION: Ontology principles and information content provide useful information for similarity metrics but currently fall short of expert performance.

Algorithms↗
PubMed2006

Sacral resection for recurrent rectal cancer: analysis of morbidity and treatment results.

PURPOSE: Composite sacropelvic resection for locally advanced recurrent rectal cancer is a high-risk procedure that benefits select patients. We reviewed our recent institutional experience to evaluate case selection, morbidity, and outcomes. METHODS: Between 1987 and 2004, 29 patients underwent composite resection for recurrent locoregional rectal cancer (17 females; median age, 60 years). Clinicopathologic indicators were evaluated as indicators of survival by log-rank test and Cox proportional hazards model. RESULTS: Of 29 total patients, 27 (93 percent) received radiotherapy with their previous surgery (n = 10; 34 percent) or before sacrectomy (n = 17; 59 percent), and 12 (41 percent) received intraoperative therapy. Sacral resections were performed at S2/S3 (55 percent) or S4/S5 (45 percent) using anterior (41 percent) or combined anterior-posterior approach (59 percent), with adherence to (62 percent) or cortical invasion in (38 percent) the sacrum. A majority of those who had undergone previous abdominoperineal resection had total exenteration (9/13), whereas most patients who had undergone a previous sphincter-preserving procedure had abdominoperineal resection (12/16) and none had exenteration. Pedicle flaps (omental, 11; abdominal rectus, 7) often were used. A median of five (range, 1-33) units of blood was given intraoperatively. Transfusions were associated with previous abdominoperineal resection (P < 0.03), correlating strongly with postoperative morbidity (P < 0.02). There were 33 complications in 17 (59 percent) patients, most commonly perineal wound breakdown (9 (31 percent)) and pelvic abscess (5 (17 percent)). Median hospital stay was 18 (range, 7-56) days, significantly longer in patients with previous abdominoperineal resection (P < 0.02) or postoperative morbidity (P < 0.03). The only postoperative death was from pelvic sepsis. Resection was complete (R0) in 18 patients (62 percent), with microscopically positive margins (R1) in 10 (34 percent) and grossly positive margins (R2) in 1 (3 percent). Two-year and five-year recurrence rates were 47 and 85 percent, respectively; disease-specific survival was 63 and 20 percent, respectively. Less transfusion (P = 0.03), R0 resection (P = 0.005), lack of anterior organ involvement (P = 0.02), and absence of cortical bone invasion (P < 0.001) were associated with better survival on univariate analysis; original colorectal cancer stage was not. CONCLUSIONS: Sacrectomy for rectal cancer is a high-risk procedure that can achieve clear resection margins with low mortality in select patients. This procedure has a low cure rate but may provide local disease control with acceptable morbidity.

Adenocarcinoma↗
PubMed2005

A temporal constraint structure for extracting temporal information from clinical narrative.

INTRODUCTION: Time is an essential element in medical data and knowledge which is intrinsically connected with medical reasoning tasks. Many temporal reasoning mechanisms use constraint-based approaches. Our previous research demonstrates that electronic discharge summaries can be modeled as a simple temporal problem (STP). OBJECTIVE: To categorize temporal expressions in clinical narrative text and to propose and evaluate a temporal constraint structure designed to model this temporal information and to support the implementation of higher-level temporal reasoning. METHODS: A corpus of 200 random discharge summaries across 18 years was applied in a grounded approach to construct a representation structure. Then, a subset of 100 discharge summaries was used to tally the frequency of each identified time category and the percentage of temporal expressions modeled by the structure. Fifty random expressions were used to assess inter-coder agreement. RESULTS: Six main categories of temporal expressions were identified. The constructed temporal constraint structure models time over which an event occurs by constraining its starting time and ending time. It includes a set of fields for the endpoint(s) of an event, anchor information, qualitative and metric temporal relations, and vagueness. In 100 discharge summaries, 1961 of 2022 (97%) identified temporal expressions were effectively modeled using the temporal constraint structure. Inter-coder evaluation of 50 expressions yielded exact match in 90%, partial match with trivial differences in 8%, partial match with large differences in 2%, and total mismatch in 0%. CONCLUSION: The proposed temporal constraint structure embodies a sufficient and successful implementation method to encode the diversity of temporal information in discharge summaries. Placing data within the structure provides a foundational representation upon which further reasoning, including the addition of domain knowledge and other post-processing to implement an STP, can be accomplished.

Database Management Systems↗
PubMed2005

Automated detection of adverse events using natural language processing of discharge summaries.

OBJECTIVE: To determine whether natural language processing (NLP) can effectively detect adverse events defined in the New York Patient Occurrence Reporting and Tracking System (NYPORTS) using discharge summaries. DESIGN: An adverse event detection system for discharge summaries using the NLP system MedLEE was constructed to identify 45 NYPORTS event types. The system was first applied to a random sample of 1,000 manually reviewed charts. The system then processed all inpatient cases with electronic discharge summaries for two years. All system-identified events were reviewed, and performance was compared with traditional reporting. MEASUREMENTS: System sensitivity, specificity, and predictive value, with manual review serving as the gold standard. RESULTS: The system correctly identified 16 of 65 events in 1,000 charts. Of 57,452 total electronic discharge summaries, the system identified 1,590 events in 1,461 cases, and manual review verified 704 events in 652 cases, resulting in an overall sensitivity of 0.28 (95% confidence interval [CI]: 0.17-0.42), specificity of 0.985 (CI: 0.984-0.986), and positive predictive value of 0.45 (CI: 0.42-0.47) for detecting cases with events and an average specificity of 0.9996 (CI: 0.9996-0.9997) per event type. Traditional event reporting detected 322 events during the period (sensitivity 0.09), of which the system identified 110 as well as 594 additional events missed by traditional methods. CONCLUSION: NLP is an effective technique for detecting a broad range of adverse events in text documents and outperformed traditional and previous automated adverse event detection methods.

Academic Medical Centers↗
PubMed2005

Interpretation of 99mTc sestamibi parathyroid SPECT scan is improved when read by the surgeon and nuclear medicine physician together.

OBJECTIVE: Parathyroid gland localization and lateralization are important before surgery, particularly for minimally invasive parathyroidectomy (MIP) and recurrent hyperparathyroidism. We hypothesized that readings of Tc sestamibi scans with single photon emission computed tomography (SPECT) by a surgeon and nuclear medicine physician together (NMP+S) compared to a nuclear medicine physician alone (NMP alone) might affect scan interpretation accuracy. METHODS: Between May 1999 and December 2002, 127 hyperparathyroid patients had preoperative localization with sestamibi SPECT. Scans were prospectively interpreted by an endocrine surgeon and nuclear medicine physician attending together (NMP+S) and a nuclear medicine physician attending alone (NMP alone). These readings were compared to intra-operative findings, which served as the 'gold standard'. RESULTS: There were 120 patients with primary hyperparathyroidism (55 underwent MIP) and seven with secondary or tertiary hyperparathyroidism; seven patients had recurrent hyperparathyroidism. Of 127 patients, 83 had single adenomas; 27, double adenomas; 15, hyperplasia; one, MENIIA; and one, parathyroid cancer. Sensitivity and positive predictive values were 58.6% and 67.4% for NMP alone compared to 81.9% and 70.0% for NMP+S. The overall accuracy of correct localization was 45.7% vs. 60.6% (P<0.01) and of correct lateralization was 69.3% vs. 80.3% (P<0.01) for NMP alone versus NMP+S respectively. The most common finding interpreted incorrectly by NMP alone and correctly by NMP+S was an ectopic superior parathyroid adenoma in the inferior position. Ninety-eight per cent of patients were cured of their hyperparathyroidism. CONCLUSIONS: Parathyroid sestamibi SPECT scan interpretation by an endocrine surgeon reading with a nuclear medicine attending resulted in improved accuracy of gland localization and lateralization compared to a nuclear medicine attending reading alone. This improvement may be due to increased awareness of clinical factors and head-and-neck anatomy.

Adolescent↗
PubMed2005

Surgical management of bile duct injuries sustained during laparoscopic cholecystectomy: perioperative results in 200 patients.

OBJECTIVE: A single institution retrospective analysis of 200 patients with major bile duct injuries was completed. Three patients died without surgery due to uncontrolled sepsis. One hundred seventy-five patients underwent surgical repair, with a 1.7% postoperative mortality and a complication rate of 42.9%. SUMMARY BACKGROUND DATA: The widespread application of laparoscopic cholecystectomy (LC) has led to a rise in the incidence of major bile duct injuries (BDI). Despite the frequency of these injuries and their complex management, the published literature contains few substantial reports regarding the perioperative management of BDI. METHODS: From January 1990 to April 2003, a prospective database of all patients with a BDI following LC was maintained. Patients' charts were retrospectively reviewed to analyze perioperative surgical management. RESULTS: Over 13 years, 200 patients were treated for a major BDI following LC. Patient demographics were notable for 150 women (75%) with a mean age of 45.5 years (median 44 years). One hundred eighty-eight sustained their BDI at an outside hospital. The mean interval from the time of BDI to referral was 29.1 weeks (median 3 weeks). One hundred nine patients (58%) were referred within 1 month of their injury for acute complications including bile leak, biloma, or jaundice. Twenty-five patients did not undergo a surgical repair at our institution. Three patients (1.5%) died after delayed referral before an attempt at repair due to uncontrolled sepsis. Twenty-two patients, having intact biliary-enteric continuity, underwent successful balloon dilatation of an anastomotic stricture. A total of 175 patients underwent definitive biliary reconstruction, including 172 hepaticojejunostomies (98%) and 3 end-to-end repairs. There were 3 deaths in the postoperative period (1.7%). Seventy-five patients (42.9%) sustained at least 1 postoperative complication. The most common complications were wound infection (8%), cholangitis (5.7%), and intraabdominal abscess/biloma (2.9%). Minor biliary stent complications occurred in 5.7% of patients. Early postoperative cholangiography revealed an anastomotic leak in 4.6% of patients and extravasation at the liver dome-stent exit site in 10.3% of patients. Postoperative interventions included percutaneous abscess drainage in 9 patients (5.1%) and new percutaneous transhepatic cholangiography and stent placement in 4 patients (2.3%). No patient required reoperation in the postoperative period. The mean postoperative length of stay was 9.5 days (median 9 days). The timing of operation (early, intermediate, delayed), presenting symptoms, and history of prior repair did not affect the incidence of the most common perioperative complications or length of postoperative hospital stay. CONCLUSIONS: This series represents the largest single institution experience reporting the perioperative management of BDI following LC. Although perioperative complications are frequent, nearly all can be managed nonoperatively. Early referral to a tertiary care center with experienced hepatobiliary surgeons and skilled interventional radiologists would appear to be necessary to assure optimal results.

Adult↗
PubMed2005

Mining a clinical data warehouse to discover disease-finding associations using co-occurrence statistics.

This paper applies co-occurrence statistics to discover disease-finding associations in a clinical data warehouse. We used two methods, chi2 statistics and the proportion confidence interval (PCI) method, to measure the dependence of pairs of diseases and findings, and then used heuristic cutoff values for association selection. An intrinsic evaluation showed that 94 percent of disease-finding associations obtained by chi2 statistics and 76.8 percent obtained by the PCI method were true associations. The selected associations were used to construct knowledge bases of disease-finding relations (KB-chi2, KB-PCI). An extrinsic evaluation showed that both KB-chi2 and KB-PCI could assist in eliminating clinically non-informative and redundant findings from problem lists generated by our automated problem list summarization system.

Chi-Square Distribution↗
PubMed2005

Document ontology: supporting narrative documents in electronic health records.

Electronic health records (EHRs) are beginning to manage an increasing volume of narrative data, such as clinical notes pertaining to admission, patient progress, shift change, follow-up, consultation, procedures, etc. These documents fall into a wide variety of classes, based on who is writing them, for what purpose, and in which location, suggesting the need for a document ontology (DO) to model our knowledge of health care documents and their properties. This paper focuses on one aspect of the Health Level 7 (HL7)/ Logical Observation Identifiers, Names, and Codes (LOINC) DO, the Subject Matter Domain (SMD). We created a new polyhierarchical structure for the SMD that combines the current value lists from the LOINC database with another value list from the American Board of Medical Specialties (ABMS). We refined and evaluated the new structure through expert review of the ontology, a survey of medical specialty boards, and specification of SMDs for a corpus of clinical notes.

Attitude of Health Personnel↗
PubMed2005

Clinicians' perceptions of usability of eNote.

The purpose of this evaluation was to assess perceptions of usability of a new semi-structured electronic clinical note. Two focus groups were held, one with attending physicians and one with residents. Physicians described their experiences with eNote and their perceptions about the system. Transcripts of the focus groups underwent content analysis, and four major themes emerged. These were "time", "hardware-system issues", "eNote application issues", and "patients' perceptions."

Attitude of Health Personnel↗
PubMed2004

Percutaneous management of bile duct strictures and injuries associated with laparoscopic cholecystectomy: a decade of experience.

BACKGROUND: The 1990s were associated with a dramatic increase in bile duct injuries with the widespread use of laparoscopic cholecystectomy (LC). Interventional radiology has an integral role in diagnosing and managing these injuries. Definitive percutaneous management with balloon dilatation might be possible in select patients with intact biliary-enteric continuity, but longterm data are limited. STUDY DESIGN: Data were collected prospectively on 51 consecutive patients with major bile duct stricture or injury associated with LC, treated with percutaneous management, January 1, 1990, to December 31, 1999. Percutaneous transhepatic cholangiography and biliary catheter placement were followed by balloon dilatation and stenting. Outcomes were assessed with direct patient contact or hospital records. RESULTS: All patients completed treatment, and 50 (98%) were stent free at mean followup of 76 months. The success rate of percutaneous management was 58.8%, without need for subsequent intervention. Presenting symptoms, level of injury, and number of stents or dilatations did not predict outcomes. Percutaneous treatment was more likely to fail in patients stented for less than 4 months (p < 0.001). Operative repair at Hopkins before percutaneous management was predictive of a successful outcome (p < 0.05). Including subsequent operations or percutaneous management, successful outcomes were achieved in 98% of patients. CONCLUSIONS: Major bile duct injuries after LC remain a clinical challenge. Although surgical reconstruction is the treatment cornerstone, selected patients with biliary-enteric continuity can achieve successful long-term results with definitive percutaneous management. The combination of percutaneous management and surgical reconstruction results in successful outcomes in virtually all patients.

Adult↗
PubMed2004

Effect of calcium channel blockers on the sensitivity of preoperative 99mTc-MIBI SPECT for hyperparathyroidism.

BACKGROUND: Technetium 99m ( 99m Tc)-methoxyisobutylisonitrile (MIBI) single photon emission computed tomography (SPECT) is frequently used in the evaluation of patients with hyperparathyroidism. Calcium channel blockers (CACBs) may affect 99m Tc-MIBI uptake by parathyroid cells. This study examines the effect of CACB therapy on the sensitivity of 99m Tc-MIBI SPECT localization for hyperparathyroidism. METHODS: Two hundred fifty-three operated patients with hyperparathyroidism were retrospectively reviewed. The potential effect of CACB therapy on 99m Tc-MIBI scan sensitivity was examined by using logistic regression analysis. Possible confounding factors were considered. RESULTS: Among 235 patients, those with multiple endocrine neoplasia, type I (MEN-I), MEN-IIA, 4-gland hyperplasia, secondary hyperparathyroidism, and tertiary hyperparathyroidism exhibited no difference associated with CACB use. Of the remaining 198 patients with primary hyperparathyroidism, 7/30 (23%) with negative 99m Tc-MIBI SPECT scans compared to 24/168 (14%) with positive scans used CACBs. After correcting for age, gender and gland weight, the odds ratio (OR) for a negative study in patients taking CACBs was 2.88 (95% CI, 1.03-8.10; P = .045). Atherosclerosis, hypertension, diabetes mellitus, preoperative calcium and parathyroid hormone levels, and thyroid hormone use were not confounding factors. CONCLUSIONS: CACB therapy reduces the sensitivity of 99m Tc-MIBI parathyroid SPECT in patients with primary hyperparathyroidism. Further studies are required to determine the potential reversibility of this effect with termination of CACB therapy.

Calcium Channel Blockers↗
PubMed2002

Major bile duct injuries associated with laparoscopic cholecystectomy: effect of surgical repair on quality of life.

OBJECTIVE: To assess the quality of life (QOL) of patients after surgical reconstruction of a major bile duct injury from laparoscopic cholecystectomy (LC). SUMMARY BACKGROUND DATA: The incidence of bile duct injuries has increased dramatically since the introduction and widespread use of LC. Previous reports show that at long-term follow-up, most patients surgically repaired will have a successful outcome as measured by standard clinical parameters. However, there is a general impression that these patients have an impaired QOL. Data addressing QOL of these patients are limited. METHODS: A standard QOL questionnaire was sent to 89 patients after successful surgical repair of a major bile duct injury from a LC treated at the Johns Hopkins Hospital between 1990 and 2000. The instrument consisted of 30 items on a visual analog scale categorized into physical (15 items), psychological (10 items), and social (5 items) domains. The same questionnaire was sent to age- and sex-matched healthy controls (n = 100) and to patients who underwent uncomplicated LC (n = 100). An additional portion of the questionnaire inquired about outcome measures and legal action undertaken by patients. RESULTS: Overall QOL scores for bile duct injury patients in the three domains (physical, psychological, and social) were 76%, 77%, and 75%, respectively. QOL scores were comparable to those of patients undergoing uncomplicated LC and healthy controls in the physical and social domains but were significantly different in the psychological domain. Presenting symptoms, prior repair, level of injury, number of stents, length of postoperative stenting, and length of follow-up did not influence QOL scores. Repaired patients reported similar rates of abdominal pain, change in bowel habits, use of pain medications, and recent symptoms of fever or chills as LC controls. Thirty-one percent of responding bile duct injury patients reported having sought legal recourse for their injury. All QOL domain scores were significantly lower in the patients who pursued a lawsuit versus those who did not. CONCLUSIONS: This study provides formal data evaluating QOL after surgical repair of major bile duct injuries from LC. Although there was a significant difference in the QOL as evaluated from a psychological dimension, bile duct injury patients reported QOL scores in the physical and social domains comparable to those of control patients. The decreased QOL assessment in the psychological dimension may be attributable to the prolonged, complicated, and unexpected nature of these injuries. The presence of a lawsuit appears to be associated with a poorer QOL assessment.

Adult↗