Ninety-five Percent Limits Of Agreement Research Articles

Automatic measurement of the patellofemoral joint parameters in the Laurin view: a deep learning-based approach.

To explore the performance of a deep learning-based algorithm for automatic patellofemoral joint (PFJ) parameter measurements from the Laurin view. A total of 1431 consecutive Laurin views of the PFJ were retrospectively collected and divided into two parts: (1) the model development dataset (dataset 1, n = 1230) and (2) the hold-out test set (dataset 2, n = 201). Dataset 1 was used to develop the U-shaped fully convolutional network (U-Net) model to segment the landmarks of the PFJ. Based on the predicted landmarks, the PFJ parameters were calculated, including the sulcus angle (SA), congruence angle (CA), patellofemoral ratio (PFR), and lateral patellar tilt (LPT). Dataset 2 was used to assess the model performance. The mean of three radiologists who independently measured the PFJ parameters was defined as the reference standard. Model performance was assessed by the intraclass correlation coefficient (ICC), mean absolute difference (MAD), and root mean square (RMS) compared to the reference standard. Ninety-five percent limits of agreement (95% LoA) were calculated pairwise for each radiologist, reference standard, and model. Compared with the reference standard, U-Net showed good performance for predicting SA, CA, PFR, and LPT, with ICC = 0.85-0.97, MAD = 0.06-5.09, and RMS = 0.09-6.90 in the hold-out test set. Except for the PFR, the remaining parameters measured between the reference standard and the model were within the 95% LoA in the hold-out test dataset. The U-Net-based deep learning approach had a relatively high model performance in automatically measuring SA, CA, PFR, and LPT. • The U-Net model could be used to segment the landmarks of the PFJ and calculate the SA, CA, PFR, and LPT, which could be used to evaluate the patellar instability. • In the hold-out test, the automatic measurement model yielded comparable performance with reference standard. • The automatic measurement model could still accurately predict SA, CA, PFR, and LPT in patients with PI and/or PFOA.

Gait speed self-prediction accuracy for people with neurological conditions in inpatient rehabilitation

Introduction A component of engaging in functional motor tasks is the ability to assess one’s own motor performance in relationship to that task. Objective The primary purpose of this exploratory study is to investigate gait speed prediction accuracy for patients with neurological conditions participating in an Inpatient Rehabilitation (IPR) program. The secondary purpose was to evaluate what factors contributed to patient judgment error. Methods Nineteen subjects with a neurologic diagnosis who were admitted to an IPR unit were asked to predict their gait speed prior to execution of the 10-meter walk test. Data was analyzed to compare predicted versus actual gait speeds. Results Strong Pearson correlations (r) between actual and predicted gait speeds were found at time 1 (r = 0.90; 95%CI = 0.83, 0.96; p<.001), time 2 (r = 0.87; 95%CI = 0.74, 0.96; p<.001), and time 3 (r = 0.91; 95% CI = 0.82, 0.97; p<.001). Ninety-five percent limits of agreement (LOA) were calculated to be −0.07, 0.30 m/s at time 1; −0.21, 0.37 m/s at time 2; and −0.22, 0.34 at time 3. Conclusions Despite the strong correlations, there appears to be less certainty surrounding the relationship between actual gait speed and predicted gait speed when examining the limits of agreement. In general, subjects ambulated faster than they predicted, underestimating their actual performance. A mismatch between predicted gait speed and actual gait speed performance can have clinical implications.

Two-dimensional echocardiographic measures of left ventricular dimensions agree with M-mode measurements in dogs

IntroductionClinicians measure left ventricular dimensions in dogs from both M-mode and two-dimensional images. Little information currently exists as to whether these two methods provide measurements similar enough to be interchangeable. AnimalsThe animals included in this study are 206 client-owned dogs: 68 healthy, 105 with myxomatous mitral valve disease, 33 with other cardiac or extracardiac disease. Materials and methodsInvestigators measured left ventricular diastolic and systolic dimensions from archived M-mode and two-dimensional images obtained from the right parasternal short-axis view. Agreement between the pairs of measurements was examined using limits of agreement (Bland-Altman) plots. ResultsLeft ventricular diastolic dimensions showed no fixed or proportional bias but did show heteroscedasticity. Ninety-five percent limits of agreement for normalized differences approximated ±10%; 95% of the absolute differences for any pair of measurements were <3.9 mm regardless of bodyweight and <2.7 mm for dogs <15 kg. Left ventricular systolic dimensions showed slight proportional bias, with two-dimensional measurements being progressively larger than M-mode measurements as ventricular size increased. Ninety-five percent limits of agreement for normalized differences approximated ±20%; 95% of the absolute differences for any pair of measurements were <4.6 mm regardless of bodyweight and <3.5 mm for dogs <15 kg. Mitral valve disease did not appreciably affect these findings. ConclusionsLeft ventricular internal dimensions in dogs with and without cardiac disease measured from two-dimensional right parasternal short-axis images are interchangeable with those measured from M-mode images using the same view.

Inter-platform reproducibility of ultrasonic attenuation and backscatter coefficients in assessing NAFLD.

To assess inter-platform reproducibility of ultrasonic attenuation coefficient (AC) and backscatter coefficient (BSC) estimates in adults with known/suspected nonalcoholic fatty liver disease (NAFLD). This HIPAA-compliant prospective study was approved by an institutional review board; informed consent was obtained. Participants with known/suspected NAFLD were recruited and underwent same-day liver examinations with clinical ultrasound scanner platforms from two manufacturers. Each participant was scanned by the same trained sonographer who performed multiple data acquisitions in the right liver lobe using a lateral intercostal approach. Each data acquisition recorded a B-mode image and the underlying radio frequency (RF) data. AC and BSC were calculated using the reference phantom method. Inter-platform reproducibility was evaluated for AC and log-transformed BSC (logBSC = 10log10BSC) by intraclass correlation coefficient (ICC), Pearson's correlation, Bland-Altman analysis with computation of limits of agreement (LOAs), and within-subject coefficient of variation (wCV; applicable to AC). Sixty-four participants were enrolled. Mean AC values measured using the two platforms were 0.90 ± 0.13 and 0.94 ± 0.15dB/cm/MHz while mean logBSC values were - 30.6 ± 5.0 and - 27.9 ± 5.6dB, respectively. Inter-platform ICC was 0.77 for AC and 0.70 for log-transformed BSC in terms of absolute agreement. Pearson's correlation coefficient was 0.81 for AC and 0.80 for logBSC. Ninety-five percent LOAs were - 0.21 to 0.13dB/cm/MHz for AC, and - 9.48 to 3.98dB for logBSC. The wCV was 7% for AC. Hepatic AC and BSC are reproducible across two different ultrasound platforms in adults with known or suspected NAFLD. • Ultrasonic attenuation coefficient and backscatter coefficient are reproducible between two different ultrasound platforms in adults with NAFLD. • This inter-platform reproducibility may qualify quantitative ultrasound biomarkers for generalized clinical application in patients with suspected/known NAFLD.

Performance of a Deep-Learning Neural Network Model in Assessing Skeletal Maturity on Pediatric Hand Radiographs.

Purpose To compare the performance of a deep-learning bone age assessment model based on hand radiographs with that of expert radiologists and that of existing automated models. Materials and Methods The institutional review board approved the study. A total of 14 036 clinical hand radiographs and corresponding reports were obtained from two children's hospitals to train and validate the model. For the first test set, composed of 200 examinations, the mean of bone age estimates from the clinical report and three additional human reviewers was used as the reference standard. Overall model performance was assessed by comparing the root mean square (RMS) and mean absolute difference (MAD) between the model estimates and the reference standard bone ages. Ninety-five percent limits of agreement were calculated in a pairwise fashion for all reviewers and the model. The RMS of a second test set composed of 913 examinations from the publicly available Digital Hand Atlas was compared with published reports of an existing automated model. Results The mean difference between bone age estimates of the model and of the reviewers was 0 years, with a mean RMS and MAD of 0.63 and 0.50 years, respectively. The estimates of the model, the clinical report, and the three reviewers were within the 95% limits of agreement. RMS for the Digital Hand Atlas data set was 0.73 years, compared with 0.61 years of a previously reported model. Conclusion A deep-learning convolutional neural network model can estimate skeletal maturity with accuracy similar to that of an expert radiologist and to that of existing automated models. © RSNA, 2017 An earlier incorrect version of this article appeared online. This article was corrected on January 19, 2018.

Postural Change in Intraocular Pressure

To compare directly the results of 4 tonometers in measuring postural change in the intraocular pressure (IOP) between sitting and lying body positions. In 19 healthy subjects, the IOP was measured both eyes with a Goldmann applanation tonometer (GAT) while sitting, then using the same GAT with the subject in the left lateral decubitus position after lying for 15 minutes, using a novel system comprising a motorized bed and a modified slit-lamp table. On the next day, the sitting IOP was measured in 1 eye using GAT, and additionally in random order Tonopen XL and pneumatonometer. Then subjects were asked to lie down, and the IOP was measured with these tonometers and also with a hand-held applanation tonometer (HA-2) after lying for 15 minutes and for 45 minutes; measurements were made with subjects in the supine position except with GAT. There were 10 male and 9 female subjects, with a mean age of 33.0±12.4 years. On day 1, the sitting GAT IOP (mm Hg) was 13.7±3.0 right eye and 13.6±2.8 left eye (P=0.7) and the lying GAT IOP was 17.8±3.5 right eye and 18.1±3.2 left eye (P=0.3 for inter-eye postural IOP change). Postural change in the right and the left eyes was moderately correlated, with a coefficient of 0.453. On day 2, postural changes after lying for 15 minutes as measured with a GAT, a Tonopen, a pneumatonometer, and a HA-2 were 4.9±2.6, 1.6±1.8, 4.2±2.0, and 3.1±2.7, respectively. After lying for 45 minutes, it decreased to 3.3±2.1, 1.1±4.3, 3.1±2.5, and 2.6±3.3, respectively; this IOP decrease was statistically significant for the GAT and the pneumatonometer. The inter-subject variability was large, as in some subjects there was nearly no postural change in the IOP, whereas in some the IOP increased by 8 to 10 mm Hg. Ninety-five percent limits of agreement showed a poor agreement between GAT and the other 3 tonometers in the sitting and lying IOPs and postural IOP changes measured. The average postural IOP change was similar when measured with GAT and a pneumatonometer, smaller with a hand-held GAT, and much smaller with a Tonopen XL. This change decreased significantly when lying between the 15- and 45-minute time points. The inter-subject variability in postural IOP was remarkable, consistent with previous reports, suggesting the importance of this parameter in clinical glaucoma practice. Inter-device agreement was poor and similar to previous reports.

The reliability, minimal detectable change and concurrent validity of a gravity-based bubble inclinometer and iphone application for measuring standing lumbar lordosis

Purpose: To investigate the reliability, minimal detectable change (MDC90) and concurrent validity of a gravity-based bubble inclinometer (inclinometer) and iPhone® application for measuring standing lumbar lordosis. Methods: Two investigators used both an inclinometer and an iPhone® with an inclinometer application to measure lumbar lordosis of 30 asymptomatic participants. Results: ICC models 3,k and 2,k were used for the intrarater and interrater analysis, respectively. Good interrater and intrarater reliability was present for the inclinometer with Intraclass Correlation Coefficients (ICC) of 0.90 and 0.85, respectively and the iPhone® application with ICC values of 0.96 and 0.81. The minimal detectable change (MDC90) indicates that a change greater than or equal to 7° and 6° is needed to exceed the threshold of error using the iPhone® and inclinometer, respectively. The concurrent validity between the two instruments was good with a Pearson product-moment coefficient of correlation (r) of 0.86 for both raters. Ninety-five percent limits of agreement identified differences ranging from 9° greater in regards to the iPhone® to 8° less regarding the inclinometer. Conclusion: Both the inclinometer and iPhone® application possess good interrater reliability, intrarater reliability and concurrent validity for measuring standing lumbar lordosis. This investigation provides preliminary evidence to suggest that smart phone applications may offer clinical utility comparable to inclinometry for quantifying standing lumbar lordosis. Clinicians should recognize potential individual differences when using these devices interchangeably.

Reliability of the dial test using a handheld inclinometer

The primary purpose of this study was to investigate the intra-tester and inter-tester reliability of the dial test using a handheld digital inclinometer. Additionally, we examined the responsiveness of the test, and side-to-side differences for meaningful comparison. Twenty-four healthy subjects (22.5±2.8years) participated in the study. The dial test was performed on both knees at 30° and 90° of knee flexion with the subject supine. While maintaining a neutral position of the ankle, an inclinometer was positioned parallel to the medial border of the foot to quantify external rotation of the tibia. Two examiners performed the dial test in a blinded manner. The minimal detectable change across repeated measures and side-to-side difference was calculated. Intra-tester reliability for examiner 1 (E1) was 0.83 at 30° knee flexion and 0.89 at 90° knee flexion. Reliability values for examiner 2 (E2) were 0.86 at 30° and 0.87 at 90° knee flexion. Inter-tester reliability was 0.74 at 30° and 0.83 at 90°. The minimal detectable change (MDC) for E1 at 30° was ±9.4° and ±7.4° at 90°. For E2, the MDC value was ±9.1° at 30° and ±8.3° at 90°. Ninety-five percent limits of agreement for side-to-side difference was 16.1° and 11.3° ° for E1 at 30° and 90° and for E2 13.9° at 30° and 14.1° at 90°. This instrumented dial test using a handheld digital inclinometer to measure external rotation can be performed with acceptable reliability in the clinical setting. A difference of 10° between two measurements on the same knee suggests that a meaningful change has occurred. For right to left comparison, differences greater than 15° suggest clinical significance. III.

Prenatal evaluation of the position of the fetal conus medullaris

To determine the position of the fetal conus medullaris during pregnancy in relation to the last vertebral body and to examine its use in detecting skin-covered spinal dysraphism. This was a retrospective study involving 300 consecutive ultrasound examinations between 15 weeks of gestation and term. Two operators independently assessed images of the spine to determine whether the conus medullaris and the last vertebral body could be visualized in a single image in a midsagittal plane. The distance between these two landmarks (the conus distance) was measured twice by both operators who were not aware of any previous measurements. Intra- and interobserver variability was assessed by 95% limits of agreement. Linear regression analysis was used to determine the relevant contributors to the conus distance and a normal range was computed based on the best-fit model. The normal results were compared with five cases of prenatally detected skin-covered spinal dysraphism. In 84.7% of the 300 cases, both operators were able to visualize the conus medullaris and the last vertebral body. Ninety-five percent limits of agreement for the intraobserver variability in measurement of conus distance were ± 1.9 mm. For the interobserver variability, they were - 3.7 and 2.5 mm. We found a linear relationship between conus distance and gestational age, biparietal diameter and abdominal circumference. The strongest relationship was observed for femur length (conus distance = - 8.2 + femur length (mm)). In the five abnormal cases, conus distance was well below the 5(th) percentile. Determination of conus distance allows for an objective and feasible assessment of the conus medullaris position. This parameter promises to be useful in the prenatal detection of skin-covered spinal dysraphism.

Interchangeability of macular thickness measurements between different volumetric protocols of Spectralis optical coherence tomography in normal eyes

To evaluate in healthy eyes the interchangeability of retinal thickness measurements resulting from different protocols of Spectralis optical coherence tomography (OCT, Heidelberg Engineering, Germany). Three different central volumetric protocols were performed in one session by one single operator in a randomly selected eye of 31 healthy adults (range, 19-30 years): (1) area 6 × 6 mm, 25 scan pattern (SP), (2) area 6 × 6 mm, 49 SP, (3) area 6 × 3 mm, 97 SP. Nine ETDRS areas were analyzed, except outer areas for 97 SP. Interclass correlation coefficients (ICC) and limits of agreement (Bland-Altman) were calculated. Mean minimal foveal thickness was 224.0 ± 15.8 μm and central retinal thickness 280.8 ± 16.5 μm. ICC ranged from 0.973 to 0.993. Ninety-five percent limits of agreement between the protocols were <7 μm in all ETDRS areas. The difference between two measurements never exceeded 9 μm. Spectralis OCT retinal thickness measurements in healthy subjects showed good protocol interchangeability for all tested protocols. Larger differences are to be expected in eyes with macular pathology.

MR‐diffusion weighted imaging of healthy liver parenchyma: Repeatability and reproducibility of apparent diffusion coefficient measurement

To compare repeatability and reproducibility of four different methods of apparent diffusion coefficient (ADC) evaluation of liver parenchyma. In fact, repeatability and reproducibility assessment is mandatory in quantitative evaluations, however, these have not been accurately investigated in liver MR-diffusion-weighted studies. Diffusion-weighted sequences, b-value = 0-1000 s/mm(2), were acquired on 30 healthy volunteers by a 1.5T scanner whose reliability has been validated by a phantom study. Four sampling methods, evaluating various parenchyma percentages by different-sized region-of-interests (ROIs), were compared by two observers: 70% and 30% of the volume, 4%-one-ROI-per-segment, and 4%-one-ROI-per-slice in the right-lobe. Ninety-five percent limits of agreement and intraclass correlation coefficient (ICC) were calculated. Complete measurements on the left lobe could be obtained in less than half of patients. The 4%-one-ROI-per-slice and 4%-one-ROI-per-segment yielded lower mean values compared with 30-70% volume methods (1343-1373 versus 1463-1560.10(-6) mm(2)/s, respectively). Repeatability was acceptable (ICCs approximately 0.80) whereas reproducibility was low (ICCs <or= 0.45) for all methods. Averaging at least 3 measurements in middle-lower sections of the right lobe improved both repeatability (ICCs to >or=0.87) and reproducibility (ICCs to 0.82) for 30-70% V methods. ADC measurements were repeatable but not reproducible in our study. Reproducibility could be improved by taking averages on the right lobe with large ROI methods. Studies on procedures that standardize ADC measurements using more than two observers are needed.

Test–Retest Reliability of Health-Related Quality-of-Life Questionnaires in Adults with Strabismus

To report the test-retest variability of two health-related quality-of-life instruments: the new Adult Strabismus 20 (AS-20) and the National Eye Institute 25-item Visual Function Questionnaire (NEI VFQ-25), in adults with strabismus. Prospective case series. Fifty-five adult patients in a clinical practice with stable strabismus completed the AS-20 and the NEI VFQ-25 at 2 visits, without intervening treatment. Questionnaires were completed the second time either at a subsequent office visit, immediately before surgery, or by mail. Intraclass correlation coefficients were calculated. Ninety-five percent limits of agreement and 95% confidence intervals around the 95% limits of agreement also were calculated. There was excellent agreement of overall questionnaire scores for the AS-20 (intraclass correlation coefficient, 0.92) and NEI VFQ-25 (intraclass correlation coefficient, 0.94). The 95% limits of agreement for overall scores were 14.3 points (95% confidence interval, 10.9 to 17.7) for the AS-20 and 11.1 points (95% confidence interval, 8.5 to 13.8) for the NEI VFQ-25. The lower test-retest variability of the VFQ-25 seemed to be partly the result of ceiling effects with many scores at the normal end of the range. The new AS-20 and the NEI VFQ-25 show excellent test-retest reliability in adults with strabismus. Change exceeding 95% limits of agreement (14 points on the AS-20 and 11 points on the VFQ-25) is indicative of real change in an individual patient. The AS-20 may be more useful than the VFQ-25 because it is less prone to ceiling effects in adults with strabismus.

Same-Patient Reproducibility of State Entropy: A Comparison of Simultaneous Bilateral Measurements During General Anesthesia

State Entropy (SE) is an index of anesthetic depth similar to Bispectral Index (BIS). Both indices use a single-channel electroencephalogram, recorded from a unilaterally applied electrode on the forehead, as their input. Intrapatient reproducibility of BIS was questioned in a recent study in which simultaneous measurements from two electrodes applied to the same patient showed conflicting anesthetic depths. Our purpose was to determine whether SE results are similarly reproducible, even though their computation uses a different algorithm than BIS. In this study, we investigated the reproducibility of SE measurements simultaneously obtained from bilaterally applied electrodes in the same patient. Entropy electrodes were applied bilaterally on 21 patients under general inhaled anesthesia. Simultaneous SE measurements from both electrodes were recorded every 10 s from each patient. Data were analyzed with Bland-Altman statistics. We obtained 14,379 pairs of SE measurements. Four percent of the individual measurements suggested conflicting anesthetic depth along with a numeric difference more than 10 points. Bias was not clinically significant (-0.3). Ninety-five percent limits of agreement were -11.7 and +11.6. SE showed a clinically significant degree of disagreement when probes were applied on both sides of the forehead in the same patient. Bland-Altman statistics showed better same-patient reproducibility in SE than did a similar study on BIS. Nevertheless, 4% of the simultaneously measured pairs of SE suggested different anesthetic depths and differed by more than 10 points. Caution is advised when using SE as a clinical index of anesthetic depth.

Variation of Agreement in Polyp Size Measurement Between Computed Tomographic Colonography and Pathology Assessment: Clinical Implications

Clinical management of polyps discovered by computed tomographic (CT) colonography depends on polyp size. However, size measured by CT colonography is an estimate, and its agreement with other measures is not well characterized. We hypothesized that size measurement by CT colonography varies substantially compared with measurement by other methods. We performed a secondary data analysis of a multicenter study of CT colonography in comparison with colonoscopy. Polyp size was determined by CT colonography, at colonoscopy, and measurement prefixation with a ruler. Agreement was assessed using descriptive statistics and Bland-Altman methodology. Six hundred trial participants completed both tests. Ninety-five percent limits of agreement indicated that estimates of size by CT colonography were between 52% lower to 64% higher than prefixation polyp size estimates. Ninety-five percent limits of agreement stratified by categories of clinical importance indicated that estimates of size by CT colonography were between 44% lower to 84% higher for polyps 0.6 cm or smaller, 44% lower to 44% higher for polyps 0.6 to 0.9 cm, and 48% lower to 22% higher for polyps smaller than 0.6 cm, 44% lower to 44% higher for polyps 0.6 cm to 0.9 cm, and 48% lower to 22% higher for polyps larger than 0.9 cm compared with prefixation estimates. Analysis of participants with 1 identified polyp in the same colon segment showed that categorization based on CT colonography measurement (ie, <0.6 cm, 0.6-0.9 cm, or >0.9 cm) differed from prefixation measurement for 43% of participants. Polyp size estimation by CT colonography varies from prefixation and colonoscopic measures of size. Future studies should clarify whether size estimation by CT colonography is sufficiently reliable as a primary factor to guide clinical management.

How Reliable Is Estimation of Glomerular Filtration Rate at Diagnosis of Type 2 Diabetes?

The Cockcroft-Gault (CG) and Modification of Diet in Renal Disease (MDRD) equations previously have been recommended to estimate glomerular filtration rate (GFR). We compared both estimates with true GFR, measured by the isotopic (51)Cr-EDTA method, in newly diagnosed, treatment-naïve subjects with type 2 diabetes. A total of 292 mainly normoalbuminuric (241 of 292) subjects were recruited. Subjects were classified as having mild renal impairment (group 1, GFR <90 ml/min per 1.73 m(2)) or normal renal function (group 2, GFR >/=90 ml/min per 1.73 m(2)). Estimated GFR (eGFR) was calculated by the CG and MDRD equations. Blood samples drawn at 44, 120, 180, and 240 min after administration of 1 MBq of (51)Cr-EDTA were used to measure isotopic GFR (iGFR). For subjects in group 1, mean (+/-SD) iGFR was 83.8 +/- 4.3 ml/min per 1.73 m(2). eGFR was 78.0 +/- 16.5 or 73.7 +/- 12.0 ml/min per 1.73 m(2) using CG and MDRD equations, respectively. Ninety-five percent CIs for method bias were -11.1 to -0.6 using CG and -14.4 to -7.0 using MDRD. Ninety-five percent limits of agreement (mean bias +/- 2 SD) were -37.2 to 25.6 and -33.1 to 11.7, respectively. In group 2, iGFR was 119.4 +/- 20.3 ml/min per 1.73 m(2). eGFR was 104.4 +/- 26.3 or 92.3 +/- 18.7 ml/min per 1.73 m(2) using CG and MDRD equations, respectively. Ninety-five percent CIs for method bias were -17.4 to -12.5 using CG and -29.1 to -25.1 using MDRD. Ninety-five percent limits of agreement were -54.4 to 24.4 and -59.5 to 5.3, respectively. In newly diagnosed type 2 diabetic patients, particularly those with a GFR >/=90 ml/min per 1.73 m(2), both CG and MDRD equations significantly underestimate iGFR. This highlights a limitation in the use of eGFR in the majority of diabetic subjects outside the setting of chronic kidney disease.

Polyp measurement and size categorisation by CT colonography: effect of observer experience in a multi-centre setting

The extent measurement error on CT colonography influences polyp categorisation according to established management guidelines is studied using twenty-eight observers of varying experience to classify polyps seen at CT colonography as either 'medium' (maximal diameter 6-9 mm) or 'large' (maximal diameter 10 mm or larger). Comparison was then made with the reference diameter obtained in each patient via colonoscopy. The Bland-Altman method was used to assess agreement between observer measurements and colonoscopy, and differences in measurement and categorisation was assessed using Kruskal-Wallis and Chi-squared test statistics respectively. Observer measurements on average underestimated the diameter of polyps when compared to the reference value, by approximately 2-3 mm, irrespective of observer experience. Ninety-five percent limits of agreement were relatively wide for all observer groups, and had sufficient span to encompass different size categories for polyps. There were 167 polyp observations and 135 (81%) were correctly categorised. Of the 32 observations that were miscategorised, 5 (16%) were overestimations and 27 (84%) were underestimations (i.e. large polyps misclassified as medium). Caution should be exercised for polyps whose colonographic diameter is below but close to the 1-cm boundary threshold in order to avoid potential miscategorisation of advanced adenomas.

The Agreement and Repeatability of Tear Meniscus Height Measurement Methods

This study investigated the agreement and repeatability of five techniques of measuring tear meniscus height (TMH). Pairs of TMH measurements were obtained from one eye of 25 normal subjects with five techniques: optical coherence tomography in cross-section (OCT) and optical pachymetry in cross-section (PACH), without fluorescein; and 5 minutes after the instillation of fluorescein, optical pachymetry en-face (SL-F) and in cross-section (SL-X), and video capture en-face (VC). These 10 measurements were taken on different days within a period of 21 days. Additionally, immediately successive replicates were taken at the first visit when VC was used, allowing for the separation of technique repeatability from true day-to-day variation in TMH. Ninety-five percent confidence intervals for the mean TMH and 95% repeatability coefficients for repeated testing on separate occasions were: 0.25 to 0.29 mm, 0.14 mm--OCT, 0.30 to 0.36 mm, 0.13 mm--PACH, 0.27 to 0.34 mm, 0.092 mm--SL-F, 0.35 to 0.41 mm, 0.19 mm--SL-X, and 0.30 to 0.38 mm, 0.088 mm--VC. Reported repeatability coefficients are averages because, despite a relatively narrow experimental range, the measurements of low TMH were generally more repeatable than those of large TMH. Ninety-five percent limits of agreement were wide for all pairwise comparisons and included zero, indicating that the methods used in this study are interchangeable when taking single recordings, at least on subjects without dry eye. The average TMH of the five techniques was similar, although OCT and SL-X tended to record lower and higher values, respectively. However, their agreement for single recordings was poor, an inevitable consequence of the low intertest repeatability found for all methods. A much improved repeatability for immediate replicates indicated that this, in turn, was principally caused by a large day-to-day variation in TMH rather than measurement error.

Trends but Not Individual Values of Central Venous Oxygen Saturation Agree with Mixed Venous Oxygen Saturation during Varying Hemodynamic Conditions

Previous studies found contradictory results regarding the question whether mixed venous oxygen saturation (Svo2) and central venous oxygen saturation (Scvo2) are equivalent. The inconsistency of study results may result from different study designs and different, partly questionable, statistical approaches. The authors performed a prospective clinical trial comparing individual oxygen saturation values as well as the trend of values in blood from the superior vena cava (Scvo2), the right atrium (Srao2), and the pulmonary artery (Svo2) during varying hemodynamic situations. The subjects were 70 patients scheduled to undergo elective neurosurgical operations in the sitting position. Oxygen saturation was measured photospectrometrically in blood samples simultaneously taken at four different time points during supine and sitting positions. Statistical analysis was performed following the recommendations of Bland and Altman. Five hundred two comparative sets of measurements were obtained. Ninety-five percent limits of agreement ranging from +/-6.83 to +/-9.30% for single values were interpreted as clinically unacceptable. In contrast, correlations between changes of Svo2 and Scvo2 as well as of Svo2 and Srao2 were interpreted as clinically acceptable (R > or = 0.755, Pearson correlation coefficient; P < or = 0.0001). In this sample of patients, exact numerical values of Scvo2 and Srao2 are not equivalent to those of Svo2 in varying hemodynamic conditions. However, for clinical purposes, the trend of Scvo2 may be substituted for the trend of Svo2. In addition, previous studies investigating the agreement between Svo2 and Scvo2 were found to be lacking in their chosen statistical approaches.

Relationship between peripheral and central venous pressures in different patient positions, catheter sizes, and insertion sites

Objective: To investigate the relationship between peripheral and central venous pressures in different patient positions (supine, prone, lithotomy, Trendelenburg, and Fowler), different catheter diameters (18 G and 20 G), and catheterization sites (dorsal hand and forearm) during surgical procedures. Design: Prospective clinical study. Settings: University hospital. Participants: Five hundred adult patients Interventions: Peripheral over-the-needle intravenous catheters were placed in the dorsal hand or forearm. Central venous catheters were inserted via the internal jugular or subclavian vein after induction of anesthesia. Measurements and Main Results: Simultaneous measure-ments of central and peripheral venous pressures were made during stable conditions at random time points in surgery; 1953 paired measurements were performed. Mean central venous pressure was 11 ± 3.7 mmHg and peripheral venous pressure was 13 ± 4 mmHg ( p = 0.0001). The overall correlation between central venous and peripheral venous pressures was found to be statistically significant ( r = 0.89, r 2 = 0.8, p = 0.0001). Mean difference between peripheral and central venous pressure was 2 ± 1.8 mmHg. Ninety-five percent limits of agreement were 5.6 to −1.6 mmHg. Conclusion: It has been assumed that replacing central venous pressure by peripheral venous pressure would cause problems in clinical interpretation. If the validity of this data is confirmed by further studies, the authors suggest that central venous pressure could be estimated by using regression equations to compare the 2 methods.

Lung function measurement in general practice: a comparison of the Escort spirometer with the Micromed turbine spirometer and the mini-Wright peak flow meter

It is important that new types of spirometer for widespread clinical use are pragmatically evaluated in primary care. This study compared measurements taken by a new portable Fleisch pneumotachograph spirometer (known as the Escort) with those of the commonly used mini-Wright peak flow meter and the Micromed Pocket turbine spirometer. A pragmatic study was conducted in two phases during routine surgeries at Aldermoor Health Centre, Southampton. Phase 1 compared the new spirometer with the mini-Wright peak flow meter and Phase 2 compared the new spirometer and the turbine spirometer. One hundred patients aged 5–88 years (56 patients with a history of chronic respiratory complaints and 44 patients without) entered Phase 1, and 100 patients aged 6–82 years (62 patients with a history of chronic respiratory complaints and 38 patients without) entered Phase 2. Each patient contributed only once to each phase, but some entered both phases on separate occasions. Ninety-five percent limits of agreement (mean ± 2 sd) were wide for all comparisons. Graphical plots revealed trends towards higher Escort values as mean values rose compared with both mini-Wright and turbine readings for peak expiratory flow rate and forced expiratory volume in one second. Possible over-reading of peak expiratory flow rate with the mini-Wright meter at low mean values was also seen. Readings taken with these different types of meter cannot be interchanged with confidence in clinical practice. The clinical significance of the theoretically more accurate measures of lung function produced with the new meter, and indeed of spirometry itself, needs further investigation.