Agreement between methods of measurement with multiple observations per individual.

Department of Health Sciences, University of York, York, UK.
Journal of Biopharmaceutical Statistics (Impact Factor: 0.72). 02/2007; 17(4):571-82. DOI: 10.1080/10543400701329422
Source: PubMed

ABSTRACT Limits of agreement provide a straightforward and intuitive approach to agreement between different methods for measuring the same quantity. When pairs of observations using the two methods are independent, i.e., on different subjects, the calculations are very simple and straightforward. Some authors collect repeated data, either as repeated pairs of measurements on the same subject, whose true value of the measured quantity may be changing, or more than one measurement by one or both methods of an unchanging underlying quantity. In this paper we describe methods for analysing such clustered observations, both when the underlying quantity is assumed to be changing and when it is not.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Cardiac output (CO) plays a crucial role in the hemodynamic management of critically ill patients treated in the intensive care unit and in surgical patients undergoing major surgery. In the field of cardiovascular dynamics, innovative techniques for CO determination are increasingly available. Therefore, the number of studies comparing these techniques with a reference, such as pulmonary artery thermodilution, is rapidly growing. There are mainly two outcomes of such method comparison studies: (1) the accuracy of agreement and (2) the precision of agreement. The precision of agreement depends on the precision of each method, i.e., the precision that the studied and the reference technique are able to achieve. We call this "precision of method". A decomposition of variance shows that method agreement does not only depend on the precision of method but also on another important source of variability, i.e., the method's general variability about the true values. Ignorance of that fact leads to falsified conclusions about the precision of method of the studied technique. In CO studies, serial measurements are frequently confused with repeated measurements. But as the actual CO of a subject changes from assessment to assessment, there is no real repetition of a measurement. This situation equals a scenario in which single measurements are given for multiple true values per subject. In such a case it is not possible to assess the precision of method.
    International Journal of Clinical Monitoring and Computing 05/2015; DOI:10.1007/s10877-015-9711-x · 1.45 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Transpulmonary thermodilution (TPTD) derived parameters are used to direct fluid management in ICU-patients. Extravascular lung water EVLW and its ratio to pulmonary blood volume (pulmonary vascular permeability index PVPI) have been associated with mortality. In single indicator TPTD pulmonary blood volume (PBV) is estimated to be 25% of global end-diastolic volume (GEDV). A recent study demonstrated marked overestimation of GEDV indexed to body-surface area (BSA; GEDVI) when using a femoral central venous catheter (CVC) for indicator injection due to the additional volume measured in the vena cava inferior. Therefore, a correction formula derived from femoral TPTD and biometric data has been suggested. Consequence, one of the commercially available TPTD-devices (PiCCO; Pulsion Medical Systems, Germany) requires information about CVC site. Correction of GEDVI for femoral CVC can be assumed. However, there is no data if correction also pertains to unindexed GEDV, which is used for calculation of PBV and PVPI. Therefore, we investigated, if also GEDV, PBV and PVPI are corrected by the new PiCCO-algorithm. In this prospective study 110 triplicate TPTDs were performed within 30 hours in 11 adult ICU-patients with PiCCO-monitoring and femoral CVC. We analyzed if the femoral TPTD correction formula for GEDVI was also applied to correct GEDV. Furthermore, we compared PVPIdisplayed to PVPIcalculated which was calculated as EVLWdisplayed/(0.25*GEDVdisplayed). Multiplication of GEDVIdisplayed by BSA resulted in GEDVcalculated which was not significantly different to GEDVdisplayed (1459 ± 365 mL vs. 1459 ± 366 mL) suggesting that correction for femoral indicator injection also pertains to GEDVdisplayed. However, PVPIdisplayed was significantly lower than PVPIcalculated (1.64 ± 0.57 vs. 2.27 ± 0.72; p < 0.001). In addition to a bias of -0.64 ± 0.22 there was a percentage error of 22%. Application of the correction formula suggested for GEDVI to PVPIdisplayed reduced the bias of PVPIdisplayed compared to EVLW/PBV from -0.64 ± 0.22 to -0.10 ± 0.05 and the percentage error from 22% to 4%. Correction for femoral CVC in the PiCCO-device pertains to both GEDVIdisplayed and GEDVdisplayed, but not to PVPIdisplayed. To provide consistent information, PVPI should be calculated based on GEDVcorrected in case of femoral CVC.
    BMC Anesthesiology 09/2014; 14:81. DOI:10.1186/1471-2253-14-81 · 1.33 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this research was to validate a new procedure (SkanLab) for the three-dimensional estimation of total arm volume. SkanLab is based on a single structured-light Kinect sensor (Microsoft, Redmond, WA, USA) and on Skanect (Occipital, San Francisco, CA, USA) and MeshLab (Visual Computing Lab, Pisa, Italy) software. The volume of twelve plastic cylinders was measured using geometry, as the reference, water displacement and SkanLab techniques (two raters and repetitions). The right total arm volume of thirty adults was measured by water displacement (reference) and SkanLab (two raters and repetitions). The bias and limits of agreement (LOA) between techniques were determined using the Bland–Altman method. Intra-and inter-rater reliability was assessed using the intraclass correlation coefficient (ICC) and the standard error of measurement. The bias of SkanLab in measuring the cylinders volume was −21.9 mL (−5.7%) (LOA: −62.0 to 18.2 mL; −18.1% to 6.7%) and in measuring the volume of arms' was −9.9 mL (−0.6%) (LOA: −49.6 to 29.8 mL; −2.6% to 1.4%). SkanLab's intra-and inter-rater reliabilities were very high (ICC >0.99). In conclusion, SkanLab is a fast, safe and low-cost method for assessing total arm volume, with high levels of accuracy and reliability. SkanLab represents a promising tool in clinical applications.
    Sensors 05/2015; 15:12342-12357. DOI:10.3390/s150612342 · 2.05 Impact Factor


Available from