Automating a severity score guideline for community-acquired pneumonia employing medical language processing of discharge summaries

Obtaining encoded variables is often a key obstacle to automating clinical guidelines. Frequently the pertinent information occurs as text in patient reports, but text is inadequate for the task. This paper describes a retrospective study that automates determination of severity classes for patients with community-acquired pneumonia (i.e. classifies patients into risk classes 1-5), a common and costly clinical problem. Most of the variables for the automated application were obtained by writing queries based on output generated by MedLEE1, a natural language processor that encodes clinical information in text. Comorbidities, vital signs, and symptoms from discharge summaries as well as information from chest x-ray reports were used. The results were very good because when compared with a reference standard obtained manually by an independent expert, the automated application demonstrated an accuracy, sensitivity, and specificity of 93%, 92%, and 93% respectively for processing discharge summaries, and 96%, 87%, and 98% respectively for chest x-rays. The accuracy for vital sign values was 85%, and the accuracy for determining the exact risk class was 80%. The remaining 20% that did not match exactly differed by only one class.

A reliability study for evaluating information extraction from radiology reports

GOAL

To assess the reliability of a reference standard for an information extraction task.

SETTING

Twenty-four physician raters from two sites and two specialties judged whether clinical conditions were present based on reading chest radiograph reports.

METHODS

Variance components, generalizability (reliability) coefficients, and the number of expert raters needed to generate a reliable reference standard were estimated.

RESULTS

Per-rater reliability averaged across conditions was 0.80 (95% CI, 0.79-0.81). Reliability for the nine individual conditions varied from 0.67 to 0.97, with central line presence and pneumothorax the most reliable, and pleural effusion (excluding CHF) and pneumonia the least reliable. One to two raters were needed to achieve a reliability of 0.70, and six raters, on average, were required to achieve a reliability of 0.95. This was far more reliable than a previously published per-rater reliability of 0.19 for a more complex task. Differences between sites were attributable to changes to the condition definitions.

CONCLUSION

In these evaluations, physician raters were able to judge very reliably the presence of clinical conditions based on text reports. Once the reliability of a specific rater is confirmed, it would be possible for that rater to create a reference standard reliable enough to assess aggregate measures on a system. Six raters would be needed to create a reference standard sufficient to assess a system on a case-by-case basis. These results should help evaluators design future information extraction studies for natural language processors and other knowledge-based systems.

Representing genomic knowledge in the UMLS semantic network

Genomics research has a significant impact on the understanding and treatment of human hereditary diseases, and biomedical literature concerning the genome project is becoming more and more important for clinicians. The Unified Medical Language System (UMLS) is designed to facilitate the retrieval and integration of information from multiple-readable biomedical information resources. This paper describes our efforts to integrate concepts important to genomics research with the UMLS semantic network. We found that the UMLS contains over 30 semantic types and most of the semantic relations that are essential for representing the underlying genomic knowledge. In addition, we observed that the organization of the network was appropriate for representing the hierarchical organization of the concepts. Because some of the concepts critical to the genomic domain were found to be missing, we propose to extend the network by adding six new semantic types and sixteen new semantic relations.

Large multi-chromosomal duplications encompass many members of the olfactory receptor gene family in the human genome

The human genome contains thousands of genes that encode a diverse repertoire of odorant receptors (ORs). We report here on the identification and chromosomal localization of 74 OR-containing genomic clones. Using fluorescence in situ hybridization (FISH), we demonstrate a striking homology among a set of approximately 20 OR locations, illustrating a history of duplications that have distributed OR sequences across the genome. Half of the OR-containing BACs cloned from total genomic DNA and 86% of cosmids derived from chromosome 3 cross-hybridize to a subset of these locations, many to 17 of them. These paralogous regions are distributed on 13 chromosomes, and eight lie in terminal bands. By analyzing clones from an approximately 250 kb clone-walk across one of these sites (3p13), we show that the homology among these sites is extensive (>150 kb) and encompasses both OR genes and intergenic genomic sequences. The FISH signals appear significantly larger at some sites than at the native location, indicating that portions of some duplicons have undergone local amplification/attrition. More restricted duplications involving pairs of other genomic locations are detected with 12% of the OR-BACs. Only a small subset of OR locations is sufficiently diverged from the others that clones derived from them behave as single-copy FISH probes. We estimate that duplications encompassing members of the OR gene family account for >0.1% of the human genome. A comparison of FISH signals at orthologous locations in other primates indicates that a portion of this OR 'subgenome' has been in flux during the divergence of primates, possibly as a mechanism for evolving the repertoire of olfactory receptors.

Evaluating natural language processors in the clinical domain

Evaluating natural language processing (NLP) systems in the clinical domain is a difficult task which is important for advancement of the field. A number of NLP systems have been reported that extract information from free-text clinical reports, but not many of the systems have been evaluated. Those that were evaluated noted good performance measures but the results were often weakened by ineffective evaluation methods. In this paper we describe a set of criteria aimed at improving the quality of NLP evaluation studies. We present an overview of NLP evaluations in the clinical domain and also discuss the Message Understanding Conferences (MUC) [1-4]. Although these conferences constitute a series of NLP evaluation studies performed outside of the clinical domain, some of the results are relevant within medicine. In addition, we discuss a number of factors which contribute to the complexity that is inherent in the task of evaluating natural language systems.

A gene recently inactivated in human defines a new olfactory receptor family in mammals

The olfactory receptor (OR) gene family constitutes one of the largest multigene families and is distributed among many chromosomal sites in the human genome. Four OR families have been defined in mammals. We previously demonstrated that a high fraction of human OR sequences have incurred deleterious mutations, thus reducing the repertoire of functional OR genes. In this study, we have characterized a new OR gene, 912-93, in primates. This gene is unique and it defines a new OR family. It localizes to human chromosome 11q11-12 and at syntenical sites in other hominoids. The sequence marks a previously unrecognized rearrangement of pericentromeric material from chromosome 11 to the centromeric region of gibbon chromosome 5. The human gene contains a nonsense point mutation in the region corresponding to the extracellular N-terminus of the receptor. This mutation is present in humans of various ethnic groups, but is absent in apes, suggesting that it probably appeared during the divergence of humans from other apes, <4 000 000-5 000 000 years ago. A second mutation, a frameshift at a different location, has occurred in the gorilla copy of this gene. These observations suggest that OR 912-93 has been recently silenced in human and gorilla, adding to a pool of OR pseudogenes whose growth may parallel a reduction in the sense of smell in primates.

Cloning, characterization, and the complete 56.8-kilobase DNA sequence of the human NOTCH4 gene

The first complete mammalian genomic sequence reported thus far in the Notch gene family, including a putative promoter region and 30 exons of the human NOTCH4 gene spanning 56.8 kb of DNA, were sequenced. The NOTCH4 locus contains a TATA-less promoter with two putative transcription initiation sites (Inr), three RBP-Jkappa sites, and two GATA recognition sites. Two cDNA isoforms, NOTCH4(L) and NOTCH4(S),were identified. Whereas the NOTCH4(S) isoform contains the entire coding sequence, the NOTCH4(L) isoform has two unspliced intronic sequences between exons 11 and 12 and exons 20 and 21 and a misspliced exon 6. Consistent with these results, two alternatively spliced isoforms of transcripts of approximately 9.3 and 6.7 kb were detected by Northern blot analysis. The predicted amino acid sequence of the NOTCH4 protein based on the NOTCH4(S) cDNA sequence contains 2003 amino acids and includes the predominant motifs of the Notch family: 29 epidermal growth factor (EGF)-like repeats, 3 Notch/lin-12 repeats, a transmembrane region, 6 cdc10/Ankyrin repeats, and a PEST domain.

A survey of infection control professional staffing patterns at University HealthSystem Consortium institutions

BACKGROUND

Proper staffing of infection control departments has long been a topic of interest. The most complete report on the subject, the Study on the Efficacy of Nosocomial Infection Control, was published in 1985. To provide current benchmarking comparison data for expected staff reductions at the University of Michigan Health System, a survey of University HealthSystem Consortium members was performed.

METHODS

A survey tool was developed to obtain general demographic, staffing, and case-mix information. An infection control professional at each institution was contacted to obtain most of the information. Additional information was obtained from standard references.

RESULTS

Responses were obtained from 45 University HealthSystem Consortium members (67%). Full-time equivalent ratios were based on the following parameters and compared for the institutions: number of occupied beds (according to occupancy rate, median 137 occupied beds/full-time equivalent), number of intensive care unit beds (median 28 beds/full-time equivalent), number of admissions or discharges (median 6686 admissions/full-time equivalent), number of ambulatory care visits (median 104,426 visits/full-time equivalent), and case-mix index (median 1.75).

CONCLUSIONS

Many institutions are using benchmarking comparison data to make decisions regarding staff reductions. This survey provides preliminary data for determining the "best practice" in staffing for infection control departments. More information may be needed to evaluate other factors that affect infection control professionals' workload.

Single-dose oral granisetron has equivalent antiemetic efficacy to intravenous ondansetron for highly emetogenic cisplatin-based chemotherapy

PURPOSE

To compare the antiemetic efficacy of a single dose of an oral antiemetic (granisetron 2 mg) with a single dose of an intravenous (i.v.) antiemetic (ondansetron 32 mg) given before cisplatin-based chemotherapy.

PATIENTS AND METHODS

This was a multicenter, randomized, double-blind, parallel-group study. Patients (N = 1,054) scheduled to receive cisplatin (> or = 60 mg/m2)-based chemotherapy were randomized to receive either 2 mg of oral granisetron tablets 1 hour before chemotherapy (n = 534) or i.v. ondansetron (32 mg) 30 minutes before chemotherapy (n = 520). The primary efficacy end point was total control (no emesis, no nausea, and no use of antiemetic rescue medication) over the initial 24 hours after the start of chemotherapy. Dexamethasone or methylprednisolone were permitted, but not required, as concomitant prophylactic antiemetics.

RESULTS

Total control was equivalent 24 hours after cisplatin chemotherapy for single-dose oral granisetron (54.7%) and i.v. ondansetron (58.3%) (95% confidence interval [CI], -9.6 to 2.4). Similar proportions of patients remained nausea-free in the granisetron group (55.4%) and the ondansetron group (59%) (95% CI, -9.6 to 2.4). The rate of complete control of emesis was 61.2% in the granisetron group and 67.1% in the ondansetron group (95% CI, -11.7 to -0.1). Both treatment regimens were well tolerated, with similar patterns of adverse reactions, generally of a mild degree. The most common side effects included constipation (14%), headache (15%), and diarrhea (10%).

CONCLUSION

Oral granisetron, administered as a single 2-mg dose, provided equivalent total antiemetic control when compared with i.v. ondansetron (32 mg) in patients who received highly emetogenic, cisplatin-based chemotherapy.

Requirements for infrastructure and essential activities of infection control and epidemiology in hospitals: a consensus panel report. Society for Healthcare Epidemiology of America

The scientific basis for claims of efficacy of nosocomial infection surveillance and control programs was established by the Study on the Efficacy of Nosocomial Infection Control project. Subsequent analyses have demonstrated nosocomial infection prevention and control programs to be not only clinically effective but also cost-effective. Although governmental and professional organizations have developed a wide variety of useful recommendations and guidelines for infection control, and apart from general guidance provided by the Joint Commission on Accreditation of Healthcare Organizations, there are surprisingly few recommendations on infrastructure and essential activities for infection control and epidemiology programs. In April 1996, the Society for Healthcare Epidemiology of America established a consensus panel to develop recommendations for optimal infrastructure and essential activities of infection control and epidemiology programs in hospitals. The following report represents the consensus panel's best assessment of needs for a healthy and effective hospital-based infection control and epidemiology program. The recommendations fall into eight categories: managing critical data and information; setting and recommending policies and procedures; compliance with regulations, guidelines, and accreditation requirements; employee health; direct intervention to prevent transmission of infectious diseases; education and training of healthcare workers; personnel resources; and nonpersonnel resources. The consensus panel used an evidence-based approach and categorized recommendations according to modifications of the scheme developed by the Clinical Affairs Committee of the Infectious Diseases Society of America and the Centers for Disease Control and Prevention's Hospital Infection Control Practices Advisory Committee.

Requirements for infrastructure and essential activities of infection control and epidemiology in hospitals: A consensus panel report. Society for Healthcare Epidemiology of America

The scientific basis for claims of efficacy of nosocomial infection surveillance and control programs was established by the Study on the Efficacy of Nosocomial Infection Control project. Subsequent analyses have demonstrated nosocomial infection prevention and control programs to be not only clinically effective but also cost-effective. Although governmental and professional organizations have developed a wide variety of useful recommendations and guidelines for infection control, and apart from general guidance provided by the Joint Commission on Accreditation of Healthcare Organizations, there are surprisingly few recommendations on infrastructure and essential activities for infection control and epidemiology programs. In April 1996, the Society for Healthcare Epidemiology of America established a consensus panel to develop recommendations for optimal infrastructure and essential activities of infection control and epidemiology programs in hospitals. The following report represents the consensus panel's best assessment of needs for a healthy and effective hospital-based infection control and epidemiology program. The recommendations fall into eight categories: managing critical data and information; setting and recommending policies and procedures; compliance with regulations, guidelines, and accreditation requirements; employee health; direct intervention to prevent transmission of infectious diseases; education and training of healthcare workers; personnel resources; and nonpersonnel resources. The consensus panel used an evidence-based approach and categorized recommendations according to modifications of the scheme developed by the Clinical Affairs Committee of the Infectious Diseases Society of America and the Centers for Disease Control and Prevention's Hospital Infection Control Practices Advisory Committee.

Respiratory isolation of tuberculosis patients using clinical guidelines and an automated clinical decision support system

OBJECTIVE

To evaluate a clinical guideline and an automated computer protocol for detection and respiratory isolation of tuberculosis (TB) patients.

DESIGN

An automated computer protocol was tested on a retrospective cohort of adult culture-positive TB patients admitted from 1992 to 1993 to Columbia-Presbyterian Medical Center and evaluated prospectively from July 1995 until July 1996.

SETTING

A large teaching hospital in New York City.

PATIENTS

171 adult patients admitted from 1992 to 1993 and 43 patients admitted between July 1995 and July 1996.

INTERVENTIONS

The 1990 Centers for Disease Control and Prevention guidelines for preventing transmission of TB were adapted to formulate clinical guidelines to ensure early isolation of TB patients at Columbia-Presbyterian Medical Center.

RESULTS

Implementation of a clinical respiratory isolation protocol resulted in a significant improvement in TB patient isolation rates, from 45 (51%) of 88 in 1992 to 62 (75%) of 83 in 1993 (P<.001). In testing automated protocols, the theoretical improvement would have identified an additional 27 patients not isolated by clinicians, making the overall isolation rate 134 (78%) of 171. For the prospective evaluation, 30 (70%) of 43 TB patients were isolated by clinicians adhering to the clinical protocol. Four additional patients were identified by the automated TB protocol, making the combined isolation rate 34 (79%) of 43.

CONCLUSIONS

A clinical policy to isolate TB patients and suspected human immunodeficiency virus-infected patients with cough, fever, or radiographic abnormalities improved isolation of culture-documented TB patients from 1992 to 1993. Automated computer protocols were successful in identifying additional potentially infectious patients that clinicians failed to place on respiratory isolation. Clinical and automated protocols combined resulted in better isolation rates than a clinical protocol alone.

Comparison of single-dose oral granisetron versus intravenous ondansetron in the prevention of nausea and vomiting induced by moderately emetogenic chemotherapy: a multicenter, double-blind, randomized parallel study

PURPOSE

The antiemetic effectiveness and safety of single-dose oral granisetron were compared with intravenous (I.V.) ondansetron in chemotherapy-naive patients who received moderately emetogenic chemotherapy.

PATIENTS AND METHODS

In this double-blind, parallel-group study, patients naive to emetogenic chemotherapy (N = 1,085) who were scheduled to receive cyclophosphamide- (500 to 1,200 mg/m2) or carboplatin (> or = 300 mg/m2) based chemotherapy, were randomized to receive either oral granisetron (n = 542) or I.V. ondansetron (n = 543). Efficacy assessments included the proportion of patients in each treatment group with total control over the 24 and 48 hours following chemotherapy initiation, as well as incidence and severity of nausea and emesis and use of antiemetic rescue medication. Prophylactic corticosteroids were allowed. Safety assessment was based on patients' reports of adverse experiences.

RESULTS

Approximately 80% of patients received prophylactic corticosteroids. Single-dose oral granisetron (2 mg) and I.V. ondansetron (32 mg) resulted in equivalent levels of total emetic control during the first 48 hours after chemotherapy. The proportion of nausea- and emesis-free patients at 24 and 48 hours were also approximately equivalent. The most commonly reported adverse experiences were headache, asthenia, and constipation. More patients who received ondonsetron than granisetron reported dizziness (9.6% v 5.4%, respectively; P = .011) and abnormal vision (4.2% v 0.6%, respectively; P < .001).

CONCLUSION

A single oral dose of granisetron (2 mg) resulted in equivalent levels of antiemetic protection as I.V. ondansetron (32 mg). Both agents were well tolerated, although more dizziness and abnormal vision were reported with ondansetron. Because the two antiemetic regimens exhibited equivalent efficacies, additional factors such as convenience and cost of therapy should be considered.

Comparable safety and antiemetic efficacy of a brief (30-second bolus) intravenous granisetron infusion and a standard (15-minute) intravenous ondansetron infusion in breast cancer patients receiving moderately emetogenic chemotherapy

PURPOSE

For patients receiving chemotherapy, optimization of antiemetic therapy in terms of safety of administration, efficacy, cost, and convenience remains a subject of intense clinical research. In this study, we evaluated and compared the safety and antiemetic efficacy of a single 30-second intravenous bolus infusion of granisetron which those of a standard 15-minute intravenous infusion of ondansetron in chemotherapy-naive breast cancer patients receiving moderately emetogenic chemotherapy.

PATIENTS AND MATERIALS

This was a randomized, double-blind, double-dummy, multicenter crossover study of 623 chemotherapy-naive patients (two male, 621 female) receiving moderately emetogenic chemotherapy (cyclophosphamide plus doxorubicin, with or without 5-fluorouracil) for breast cancer. Patients were assigned randomly to receive either granisetron or ondansetron in cycle 1 and the other agent in cycle 2. Granisetron (10 micrograms/kg) was administered as a 30-second intravenous bolus infusion within 5 minutes before the start of chemotherapy, and ondansetron (32 mg) was administered as a 15-minute intravenous infusion beginning 30 minutes before the start of chemotherapy. A total of 573 patients received the two planned chemotherapy cycles. Safety assessment was based on the type and frequency of adverse experiences reported by the patient at 24 and 48 hours after chemotherapy began. Efficacy assessments included the occurrence of nausea or emesis and the proportion of patients who achieved total emetic control at 24 and 48 hours after chemotherapy.

RESULTS

Similar proportions of patients in both treatment groups remained free of emesis at the 24-hour assessment period (58.6% and 62.7% of granisetron- and ondansetron-treated patients, respectively) and at 48 hours (42.2% and 45.0% for granisetron vs ondansetron) in both cycles. The proportions of granisetron- and ondansetron-treated patients who remained nausea-free for the first 24 hours after treatment in both cycles were 44.0% and 48.5%, respectively, and at 48 hours were 26.7% and 31.0%, respectively. Statistical analysis demonstrated no significant treatment-by-cycle interaction. The 30-second granisetron infusion and the 15-minute ondansetron infusion were well tolerated. However, administration of ondansetron as a 15-minute intravenous infusion produced abnormal vision (6.28%) in significantly more patients than granisetron (0.35%).

DISCUSSION

These two intravenous 5-hydroxytryptamine3-receptor antagonist antiemetics were similarly effective in controlling acute nausea and emesis during two cycles of moderately emetogenic chemotherapy. Granisetron (30 seconds) and the longer infusion of ondansetron (15 minutes) were well tolerated; however, ondansetron was associated with a greater proportion of patients reporting abnormal vision. Granisetron administered as a 30-second bolus infusion allows a considerably shorter waiting time between the end of the antiemetic infusion and the initiation of chemotherapy. This shorter administration time may enhance patient convenience and provider efficiency.

The human homolog of rat Jagged1 expressed by marrow stroma inhibits differentiation of 32D cells through interaction with Notch1

A cDNA clone encoding the human homolog of rat Jagged1 was isolated from normal human marrow. Analyses of human stromal cell lines indicate that this gene, designated hJagged1, is expressed by marrow stromal cells typified by the cell line HS-27a, which supports the long-term maintenance of hematopoietic progenitor cells. G-CSF-induced differentiation of 32D cells expressing Notch1 was inhibited by coculturing with HS-27a. A peptide corresponding to the Delta/Serrate/LAG-2 domain of hJagged1 and supernatants from COS cells expressing a soluble form of the extracellular portion of hJagged1 were able to mimic this effect. These observations suggest that hJagged1 may function as a ligand for Notch1 and play a role in mediating cell fate decisions during hematopoiesis.

An evaluation of natural language processing methodologies

Medical language processing (MLP) systems that codify information in textual patient reports have been developed to help solve the data entry problem. Some systems have been evaluated in order to assess performance, but there has been little evaluation of the underlying technology. Various methodologies are used by the different MLP systems but a comparison of the methods has not been performed although evaluations of MLP methodologies would be extremely beneficial to the field. This paper describes a study that evaluates different techniques. To accomplish this task an existing MLP system MedLEE was modified and results from a previous study were used. Based on confidence intervals and differences in sensitivity and specificity between each technique and all the others combined, the results showed that the two methods based on obtaining the largest well-formed segment within a sentence had significantly higher sensitivity than the others by 5% and 6%. The method based on recognizing a complete sentence had a significantly worse sensitivity than the others by 7% and a better specificity by .2%. None of the methods had significantly worse specificity.

Members of the olfactory receptor gene family are contained in large blocks of DNA duplicated polymorphically near the ends of human chromosomes

We have identified three new members of the olfactory receptor (OR) gene family within a large segment of DNA that is duplicated with high similarity near many human telomeres. This segment is present at 3q, 15q, and 19p in each of 45 unrelated humans sampled from various populations. Additional copies are present polymorphically at 11 other subtelomeric locations. The frequency with which the block is present at some locations varies among populations. While humans carry seven to 11 copies of the OR-containing block, it is located in chimpanzee and gorilla predominantly at a single site, which is not orthologous to any of the locations in the human genome. The observation that sequences flanking the OR-containing segment are duplicated on larger and different sets of chromosomes than the OR block itself demonstrates that the segment is part of a much larger, complex patchwork of subtelomeric duplications. The population analyses and structural results suggest the types of processes that have shaped these regions during evolution. From its sequence, one of the OR genes in this duplicated block appears to be potentially functional. Our findings raise the possibility that functional diversity in the OR family is generated in part through duplications and inter-chromosomal rearrangements of the DNA near human telomeres.

Extracting findings from narrative reports: software transferability and sources of physician disagreement

While natural language processing systems are beginning to see clinical use, it remains unclear whether they can be disseminated effectively through the health care community. MedLEE, a general-purpose natural language processor developed for Columbia-Presbyterian Medical Center, was compared to physicians' ability to detect seven clinical conditions in 200 Brigham and Women's Hospital chest radiograph reports. Using the system on the new institution's reports resulted in a small but measurable drop in performance (it was distinguishable from physicians at p = 0.011). By making adjustments to the interpretation of the processor's coded output (without changing the processor itself), local behavior was better accommodated, and performance improved so that it was indistinguishable from the physicians. Pairs of physicians disagreed on at least one condition for 22% of reports; the source of disagreement appeared to be interpretation of findings, gauging likelihood and degree of disease, and coding errors.

MRIT, a novel death-effector domain-containing protein, interacts with caspases and BclXL and initiates cell death

Activation of the cascade of proteolytic caspases has been identified as the final common pathway of apoptosis in diverse biological systems. We have isolated a gene, termed MRIT, that possesses overall sequence homology to FLICE (MACH), a large prodomain caspase that links the aggregated complex of the death domain receptors of the tumor necrosis factor receptor family to downstream caspases. However, unlike FLICE, the C-terminal domain of MRIT lacks the caspase catalytic consensus sequence QAC(R/Q)G. Nonetheless MRIT activates caspase-dependent death. Using yeast two-hybrid assays, we demonstrate that MRIT associates with caspases possessing large and small prodomains (FLICE, and CPP32/YAMA), as well as with the adaptor molecule FADD. In addition, MRIT simultaneously and independently interacts with BclXL and FLICE in mammalian cells. Thus, MRIT is a mammalian protein that interacts simultaneously with both caspases and a Bcl-2 family member.

Prevalence and correlates of skin damage on the hands of nurses

OBJECTIVE

To describe the prevalence and correlates of skin damage on nurses' hands.

DESIGN

Prevalence survey using self-report questionnaire of hand care regimens, problems, and skin condition, and visual examination of the hands at 30X magnification by trained investigators to evaluate degree of skin scaling.

SETTING

Four hospitals: two in the Mid-Atlantic and two in the northern United States.

PARTICIPANTS

Convenience sample of 410 nurses working 30 hours or more per week in acute care units.

OUTCOME MEASURES

Damage to skin of the hands.

RESULTS

Approximately one fourth of subjects (n = 106) met the criteria for currently damaged hands; 85.6% (n = 351) reported ever having skin problems. Damage was not correlated with age (p = 0.43), sex (p = 0.14), or skin type (p = 0.25), type of soap used at home (p = 0.58), reported duration of handwashing (p = 0.12), or glove brand (p = 0.90), but was significantly correlated with the type of soap used at work (p = 0.01), number of hand washes per shift (p = 0.0003), number of times gloves were worn (p = 0.008), and study site (p = 0.009). Variables significantly predictive of skin damage in a logistic regression analysis were type of soap used at work and number of times gloves were worn (p = 0.04). Geographic location was not a factor, because both the highest and lowest prevalence of skin damage occurred in the northern study institutions.

CONCLUSIONS

Damage to skin of the hands is a common and potentially serious problem among nurses, and is associated with gloving and handwashing practices rather than with geographic or demographic factors. Efforts to improve skin condition must focus on improving products and identifying any interactive effects between hand care products and glove materials and brands.

Analysis of the variation in chromosome size among diverse human populations by bivariate flow karyotyping

Chromosomes sampled from seven human populations were analyzed by flow cytometry to survey normal variation in chromosome size. The populations include two African Pygmy groups, two Amerindian tribes, Druze, Khmer Cambodians, and Melanesians. Mitotic chromosomes were isolated from cultured cells and stained with Hoechst 33,258 and chromomycin A3. The relative DNA content and base-pair composition of each homolog was quantified by bivariate flow karyotyping. Significant variation in DNA content, ranging from 10-40%, was observed for chromosomes 1, 13-16, 19, 21, 22, and Y. The measurements for each population appeared to be a random sampling of the total set of 33 individuals for the majority of chromosomes. A few significant differences in the distributions of chromosomal DNA content were observed among the populations, however. The data, when combined with an earlier study of 33 unrelated individuals of unknown ethnic origin, provide a good representation of the variation in chromosome size among humans.

The human SHIP gene is differentially expressed in cell lineages of the bone marrow and blood

The macrophage colony-stimulating factor receptor and several other hematopoietic growth factor receptors induce the tyrosine phosphorylation of a 145- to 150-kD protein in murine cells. We have previously cloned a cDNA for the murine 150-kD protein, SHIP, and found that it encodes a unique signaling intermediate that binds the SHC PTB domain through at least one tyrosine phosphorylated (NPXY) site in the carboxyl-terminal region. SHIP also contains several potential SH3 domain-binding sites, an SH2 domain for binding other tyrosine phosphorylated proteins, and an enzymatic activity that removes the phosphate from the 5 position of phosphatidylinositol 3,4,5-phosphate or from inositol 1,3,4,5-phosphate. SHIP has a negative effect on cell growth and therefore loss or modification may have profound effects on hematopoietic cell development. In this study, we have cloned a cDNA for human SHIP and examined mRNA and protein expression of SHIP and related species in bone marrow and blood cells. Flow cytometry indicates that at least 74% of immature CD34+ cells express SHIP cross-reacting protein species, whereas within the more mature population of CD33+ cells, only 10% of cells have similar expression. The majority of T cells react positively with the anti-SHIP antibodies, but significantly fewer B cells are positive. Immunoblotting detects up to seven different cross-reacting SHIP species, with peripheral blood mononuclear cells exhibiting primarily a 100-kD protein and a CD34+ acute myeloblastic leukemia expressing mainly 130-kD and 145-kD forms of SHIP. Overall, these results indicate that there is an enormous diversity in the size of SHIP or SHIP-related mRNA and protein species. Furthermore, the expression of these protein species changes according to both the developmental stage and differentiated lineage of the mature blood cell.

Identification of findings suspicious for breast cancer based on natural language processing of mammogram reports

There is need for encoded data for computerized clinical decision support, but most such data are unavailable as they are in free-text reports. Natural language processing offers one alternative for encoding such data. MedLEE is a natural language processing system which is in routine use for encoding chest radiograph and mammogram reports. In this paper, we study MedLEE's ability to identify mammogram findings suspicious for breast cancer by comparing MedLEE's encoding with a logbook of all suspicious findings maintained by the mammography center. While MedLEE was able to identify all the suspicious findings, it varied in the level of granularity, particularly about the location of the suspicious finding. Thus, natural language processing is a useful technique for encoding mammogram reports in order to detect suspicious findings.

Towards a comprehensive medical language processing system: methods and issues

Natural language processing (NLP) systems can help solve the data entry problem by providing coded data from textual reports for clinical applications. A number of NLP systems have shown promise, but have not yet achieved wide-spread use for practical applications. In order to achieve such use, a system must have broad coverage of the clinical domain and not be restricted to limited applications. In addition, an NLP system must perform satisfactorily for real-world applications. This paper describes methods and issues associated with an ongoing extension of MedLEE, an operational NLP system, from a limited domain to a domain that encompasses comprehensive clinical information.