Accuracy and reliability of forensic handwriting comparisons

Forensic handwriting examination involves the comparison of writing samples by forensic document examiners (FDEs) to determine whether or not they were written by the same person. Here we report the results of a large-scale study conducted to assess the accuracy and reliability of handwriting comparison conclusions. Eighty-six practicing FDEs each conducted up to 100 handwriting comparisons, resulting in 7,196 conclusions on 180 distinct comparison sets, using a five-level conclusion scale. Erroneous “written by” conclusions (false positives) were reached in 3.1% of the nonmated comparisons, while 1.1% of the mated comparisons yielded erroneous “not written by” conclusions (false negatives). False positive rates were markedly higher for nonmated samples written by twins (8.7%) compared to nontwins (2.5%). Notable associations between training and performance were observed: FDEs with less than 2 y of formal training generally had higher error rates, but they also had higher true positive and true negative rates because they tended to provide more definitive conclusions; FDEs with at least 2 y of formal training were less likely to make definitive conclusions, but those definitive conclusions they made were more likely to be correct (higher positive predictive and negative predictive values). We did not observe any association between writing style (cursive vs. printing) and rates of errors or incorrect conclusions. This report also provides details on the repeatability and reproducibility of conclusions, and reports how conclusions are affected by the quantity of writing and the similarity of content.

Forensic science is under scrutiny, particularly for pattern-based disciplines in which source conclusions are reported. The National Research Council report Strengthening Forensic Science in the United States: A Path Forward (1) stated that “The scientific basis for handwriting comparisons needs to be strengthened” and noted that “there has been only limited research to quantify the reliability and replicability of the practices used by trained document examiners.” The President’s Council of Advisors on Science and Technology (PCAST) report Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods (2) expressed concerns regarding the validity and reliability of conclusions made by forensic examiners, and called for empirical testing: “The only way to establish the scientific validity and degree of reliability of a subjective forensic feature-comparison method—that is, one involving significant human judgment—is to test it empirically by seeing how often examiners actually get the right answer. Such an empirical test of a subjective forensic feature-comparison method is referred to as a ‘black-box test.’” The National Commission on Forensic Science also called for such testing (3). Although the accuracy and reliability of conclusions made by forensic document examiners (FDEs) have been the focus of multiple studies over the years (4–10), the designs of those studies are notably different from this study (and from PCAST’s recommendations), and therefore the resulting rates are not directly comparable (in particular, when comparing open-set to closed-set studies, comparing studies based on one-to-one vs. one-to-many examinations, and comparing studies that use notably different conclusion scales; see SI Appendix, Appendix B for a summary).This study was conducted to provide data that can be used to assess the scientific validity of handwriting comparisons, for use by policy makers, laboratory managers, the legal community, and FDEs. This study follows the approach used in the previous FBI Laboratory–Noblis latent print black box study (11) and later recommended by the PCAST report. The design utilizes open-set, one-to-one document comparisons to evaluate the conclusions reached by practicing FDEs when comparing writing samples selected to be broadly comparable to casework. The primary purposes of the study are to measure the accuracy of conclusions by FDEs when comparing handwriting samples and to assess reliability by measuring the reproducibility (interexaminer variability) and repeatability (intraexaminer variability) of those conclusions. Secondary purposes include reporting any associations between the accuracy of the decisions in this study, factors related to the participants (such as training or experience), and factors related to the samples (such as quantity of writing, comparability of content, limitations, or style of writing).     

Renal failure in sick hypertensive premature infants receiving captopril therapy

A retrospective study of nine sick premature infants with chronic lung disease who received captopril for control of systemic hypertension (systolic blood pressure (BP) greater than 113 mm Hg) was carried out to determine efficacy of therapy and associated complications. All nine infants had markedly elevated peripheral renin values, 134.3 +/- 128.1 ng/mL/hr (mean +/- SD). Five infants had abnormal renal sonographic and perfusion scans with evidence of renal artery thrombosis, parenchymal disease, or both. Captopril therapy (0.3 mg/kg) was instituted at a postnatal age of 123 +/- 108 days. After the initial dose, the systolic BP decreased significantly in all infants, the decrease ranging from 21% to 58% of the pretreatment value. Dosage was subsequently halved in all infants. Seventeen episodes of unpredictable decreases in BP more than 40% from baseline occurred during the reduced maintenance therapy. Four infants had a total of seven episodes during which the BP decreased by 57 +/- 10% from baseline; this decrease persisted for 17 +/- 6 hours and was unresponsive to volume reexpansion and inotropic therapy. All seven episodes were accompanied by oliguria (urine output less than 1 mL/kg/hr) that persisted for 18 +/- 12 hours. These episodes were accompanied by neurologic signs (subtle seizures, lethargy, and/or apnea) within 18 +/- 6 hours after the onset of oliguria. The remaining five infants had a total of 13 episodes of decreased BP of 50 +/- 8% of baseline, which were of significantly shorter duration and responded to volume reexpanders, inotropic therapy, or both and were unaccompanied by oliguria. These data suggest the need for close observation of BP in infants receiving maintenance captopril therapy.     

Cytokine expression in response to bacterial antigens in preterm and term infant cord blood monocytes

BACKGROUND: Neonatal susceptibility to bacterial infection is associated with an immature immune system, but the role of different bacterial antigens in specific responses is largely unknown. OBJECTIVE: To evaluate differences in intracellular cytokine response to physiologically relevant bacterial antigens in term and preterm infants as compared with adults. METHODS: Cord blood samples from preterm and term neonates and adult peripheral blood samples were cultured ex vivo with and without whole heat-killed bacteria. Intracellular leukocyte production of interleukin (IL)-6, IL-10, IL-12, and IL-8 responses was assessed by flow cytometry. RESULTS: Monocytes were the primary producers of all mediators. Escherichia coli was the most potent stimulant. Lactobacillus plantarum 299v activated fewer monocytes as compared with E. coli for all responses (p < 0.05), except for IL-12 in term neonates. IL-6 response to Staphylococcus epidermidis was lower in both groups of neonates as compared with adults (p = 0.023 and p = 0.001). IL-8 response to S. epidermidis was lower in term as compared with preterm neonates and adults (p = 0.003). IL-10 response to group B streptococci was lower in term neonates as compared with adults and higher in preterm as compared with term neonates (p = 0.015). CONCLUSIONS: Monocytes from term neonates compared to preterm neonates show a downregulated anti-inflammatory response to specific bacteria. High neonatal response to pathogenic E. coli in the preterm infant could cause uncontrolled inflammatory response, while lower IL-6 response to S. epidermidis in neonates may indicate a basis for vulnerability to S. epidermidis infection.     

Interleukin-3, GM-CSF, and TPA induce distinct phosphorylation events in an interleukin 3-dependent multipotential cell line

The mechanism of action of the hemopoietic growth factor, murine interleukin-3 (mIL-3), was investigated using an mIL-3-dependent multipotential hematopoietic cell line, B6SUtA1. Murine granulocyte- macrophage colony-stimulating factor (mGM-CSF) was as potent as mIL-3 in stimulating these cells. In addition, sodium orthovanadate, an inhibitor of phosphotyrosine phosphatase, and 12-O-tetradecanoyl- phorbol-13-acetate (TPA), a known activator of protein kinase C, also stimulated DNA synthesis in these cells, suggesting that protein phosphorylation might be involved in the mechanism of action of mIL-3 and mGM-CSF. To assess this possibility, intact B6SUtA1 cells exposed for brief periods to mIL-3, mGM-CSF, and TPA were analyzed for changes in phosphorylation patterns using metabolic 32P-labeling and antibodies to phosphotyrosine. Both mIL-3 and mGM-CSF induced the serine-specific phosphorylation of a 68-Kd cytosolic protein, whereas all three agents stimulated the serine-specific phosphorylation of a 68-Kd membrane protein. Furthermore, mIL-3 stimulated tyrosine phosphorylation of the 68-Kd membrane protein, as well as of 140-, 90-, 55, and 40-Kd proteins. The 90-Kd protein was also tyrosine phosphorylated in response to mGM-CSF. These phosphotyrosine containing proteins were not detected in TPA-treated cells. These results indicate that protein phosphorylations on tyrosine and serine residues occur in B6SUtA1 cells following short-term incubation with mIL-3 or mGM-CSF and that most of these phosphorylation events are mediated by kinases other than protein kinase C (PkC).     

How do family practitioners perceive surgery for the morbidly obese?

BACKGROUND: Little is known about the level of knowledge and comfort with bariatric surgery among family practice physicians. METHODS: Surveys were sent to all family practitioners in Connecticut querying the practice type and knowledge of bariatric surgery. The results were analyzed for the prevalence of opinion. RESULTS: Of 620 surveys sent out, 129 (21%) were completed. Of the 129 respondents, 73% were men, aged 31-79 years, and 92% were board certified, with an average of 19 years' experience. The average body mass index of respondents was 26 kg/m2 (range 16-40). Only 4% of respondents had a body mass index >30 kg/m2. Physicians reported a patient obesity rate of 43%. Of the 129 respondents, 88% believed obesity was difficult to control with diet and exercise alone. Only 6% thought obesity was best controlled surgically. Also, 85% of respondents had referred a patient for gastric bypass, although only 57% were comfortable explaining the procedure. The most common reason for refusal to refer was fear of complications and death. Additionally, 55% correctly listed a body mass index of 40 kg/m2 as qualifying for bariatric surgery without comorbidities; 48% identified the mortality rate of surgery as <1%, with 4% of respondents reporting >10%; and 84% were familiar with gastric bypass, 66% with LapBand, 33% with vertical banded gastroplasty, and 5% with duodenal switch. The respondents believed that nausea was the most common side effect, followed by anemia and fatigue. Finally, 53% believed bowel obstruction was common. CONCLUSIONS: The results of our study have shown that misconceptions about bariatric surgery exist in the family practice community despite the increasing frequency of these procedures. Educational programs need to be designed to assist family practitioners in treating and referring obese patients.     

Inter-Rater Reliability of the PIPES Tool: Validation of a Surgical Capacity Index for Use in Resource-Limited Settings

Introduction

In response to the need for simple, rapid means of quantifying surgical capacity in low resource settings, Surgeons OverSeas (SOS) developed the personnel, infrastructure, procedures, equipment and supplies (PIPES) tool. The present investigation assessed the inter-rater reliability of the PIPES tool.

Methods

As part of a government assessment of surgical services in Santa Cruz, Bolivia, the PIPES tool was translated into Spanish and applied in interviews with physicians at 31 public hospitals. An additional interview was conducted with nurses at a convenience sample of 25 of these hospitals. Physician and nurse responses were then compared to generate an estimate of reliability. For dichotomous survey items, inter-rater reliability between physicians and nurses was assessed using the Cohen’s kappa statistic and percent agreement. The Pearson correlation coefficient was used to assess agreement for continuous items.

Results

Cohen’s kappa was 0.46 for infrastructure, 0.43 for procedures, 0.26 for equipment, and 0 for supplies sections. The median correlation coefficient was 0.91 for continuous items. Correlation was 0.79 for the PIPES index, and ranged from 0.32 to 0.98 for continuous response items.

Conclusions

Reliability of the PIPES tool was moderate for the infrastructure and procedures sections, fair for the equipment section, and poor for supplies section when comparing surgeons’ responses to nurses’ responses—an extremely rigorous test of reliability. These results indicate that the PIPES tool is an effective measure of surgical capacity but that the equipment and supplies sections may need to be revised.     

MR imaging of facial nerve enhancement in Bell palsy or after temporal bone surgery

The authors evaluated magnetic resonance (MR) images obtained with intravenously administered gadolinium in ten patients who had facial paralysis and no facial nerve tumor. In patients with either Bell palsy (four patients) or facial paralysis after temporal bone surgery (six patients), intratemporal facial nerve enhancement was seen. Facial nerve enhancement on MR images proved to be a nonspecific finding.