Limits...
Inter-unit variability in two ParvoMedics TrueOne 2400 automated metabolic gas analysis systems.

Macfarlane DJ, Wu HL - Eur. J. Appl. Physiol. (2012)

Bottom Line: Absolute percentage errors (APE), coefficient of variations (CV), effect sizes and Bland-Altman analyses were undertaken on the metabolic data, including expired ventilation (V (E)), oxygen consumption (VO(2)) and carbon dioxide production (VCO(2)).The few statistically significant differences detected between the two duplicate systems were determined to have small or trivial effect sizes, and their magnitudes to be of little physiological importance.The two ParvoMedics TrueOne 2400 systems demonstrated excellent inter-unit agreement.

View Article: PubMed Central - PubMed

Affiliation: Institute of Human Performance, The University of Hong Kong, Pokfulam, Hong Kong. djmac@hku.hk

ABSTRACT
Knowing the inter-unit variability, especially the technological error, is important when using many physiological measurement systems, yet no such inter-unit analysis has been undertaken on duplicate automated gas analysis systems. This study investigated the inter-unit performance of two identical ParvoMedics TrueOne 2400 automated gas analysis systems during a range of submaximal steady-state exercises performed on an electromagnetic cycle ergometer. Fifteen adult males were tested on two separate days a rest, 30, 60, 90, and 120 Watts with the duplicate gas analysis units arranged (1) collaterally (2 min of steady-state expired gas was alternately passed through each system), and (2) simultaneously (identical steady-state expired gas was passed simultaneously through both systems). Total within-subject variation (biological + technological) was determined from the collateral tests, but the unique inter-unit variability (technological error between identical systems) was shown by the simultaneous tests. Absolute percentage errors (APE), coefficient of variations (CV), effect sizes and Bland-Altman analyses were undertaken on the metabolic data, including expired ventilation (V (E)), oxygen consumption (VO(2)) and carbon dioxide production (VCO(2)). The few statistically significant differences detected between the two duplicate systems were determined to have small or trivial effect sizes, and their magnitudes to be of little physiological importance. The total within-subject variations for VO(2), VCO(2) and V (E) each equated to a mean CV and mean APE value of ~4 and ~6 %, whilst the respective inter-unit technological errors equated to ~1.5 and ~2.1 %. The two ParvoMedics TrueOne 2400 systems demonstrated excellent inter-unit agreement.

Show MeSH
Bland–Altman plots from the collateral (a, c, e) and simultaneous (b, d, f) tests, showing the error scores for the two gas analysis systems (Test unit 1 − Test unit 2); data shown for VE (a, b), VO2 (c, d), and VCO2 (e, f). Solid horizontal line indicates the mean error, whilst the dashed horizontal lines indicate the 95 % LOA
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3569593&req=5

Fig2: Bland–Altman plots from the collateral (a, c, e) and simultaneous (b, d, f) tests, showing the error scores for the two gas analysis systems (Test unit 1 − Test unit 2); data shown for VE (a, b), VO2 (c, d), and VCO2 (e, f). Solid horizontal line indicates the mean error, whilst the dashed horizontal lines indicate the 95 % LOA

Mentions: Table 1 shows that the metabolic data from the two automated gas analysis units arranged collaterally were very similar across all workloads, with no statistically significant differences in any of the mean comparisons (after Holm–Bonferroni adjustment). Except for the Rest condition (when the relative error was expected to be higher), the APE and CV values tended to range between 2 and 4 %. All effect sizes were below 0.32, with 21 of the 30 (70 %) classified (Saunders 2004) as being trivial (<0.2) and the remaining 30 % as small (0.2–0.5). The Bland–Altman plots in Fig. 2a, c, e, show very minor systematic error (bias) between the two collateral systems, with almost no proportional random error, and small 95 % limits of agreement (LOA).Table 1


Inter-unit variability in two ParvoMedics TrueOne 2400 automated metabolic gas analysis systems.

Macfarlane DJ, Wu HL - Eur. J. Appl. Physiol. (2012)

Bland–Altman plots from the collateral (a, c, e) and simultaneous (b, d, f) tests, showing the error scores for the two gas analysis systems (Test unit 1 − Test unit 2); data shown for VE (a, b), VO2 (c, d), and VCO2 (e, f). Solid horizontal line indicates the mean error, whilst the dashed horizontal lines indicate the 95 % LOA
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3569593&req=5

Fig2: Bland–Altman plots from the collateral (a, c, e) and simultaneous (b, d, f) tests, showing the error scores for the two gas analysis systems (Test unit 1 − Test unit 2); data shown for VE (a, b), VO2 (c, d), and VCO2 (e, f). Solid horizontal line indicates the mean error, whilst the dashed horizontal lines indicate the 95 % LOA
Mentions: Table 1 shows that the metabolic data from the two automated gas analysis units arranged collaterally were very similar across all workloads, with no statistically significant differences in any of the mean comparisons (after Holm–Bonferroni adjustment). Except for the Rest condition (when the relative error was expected to be higher), the APE and CV values tended to range between 2 and 4 %. All effect sizes were below 0.32, with 21 of the 30 (70 %) classified (Saunders 2004) as being trivial (<0.2) and the remaining 30 % as small (0.2–0.5). The Bland–Altman plots in Fig. 2a, c, e, show very minor systematic error (bias) between the two collateral systems, with almost no proportional random error, and small 95 % limits of agreement (LOA).Table 1

Bottom Line: Absolute percentage errors (APE), coefficient of variations (CV), effect sizes and Bland-Altman analyses were undertaken on the metabolic data, including expired ventilation (V (E)), oxygen consumption (VO(2)) and carbon dioxide production (VCO(2)).The few statistically significant differences detected between the two duplicate systems were determined to have small or trivial effect sizes, and their magnitudes to be of little physiological importance.The two ParvoMedics TrueOne 2400 systems demonstrated excellent inter-unit agreement.

View Article: PubMed Central - PubMed

Affiliation: Institute of Human Performance, The University of Hong Kong, Pokfulam, Hong Kong. djmac@hku.hk

ABSTRACT
Knowing the inter-unit variability, especially the technological error, is important when using many physiological measurement systems, yet no such inter-unit analysis has been undertaken on duplicate automated gas analysis systems. This study investigated the inter-unit performance of two identical ParvoMedics TrueOne 2400 automated gas analysis systems during a range of submaximal steady-state exercises performed on an electromagnetic cycle ergometer. Fifteen adult males were tested on two separate days a rest, 30, 60, 90, and 120 Watts with the duplicate gas analysis units arranged (1) collaterally (2 min of steady-state expired gas was alternately passed through each system), and (2) simultaneously (identical steady-state expired gas was passed simultaneously through both systems). Total within-subject variation (biological + technological) was determined from the collateral tests, but the unique inter-unit variability (technological error between identical systems) was shown by the simultaneous tests. Absolute percentage errors (APE), coefficient of variations (CV), effect sizes and Bland-Altman analyses were undertaken on the metabolic data, including expired ventilation (V (E)), oxygen consumption (VO(2)) and carbon dioxide production (VCO(2)). The few statistically significant differences detected between the two duplicate systems were determined to have small or trivial effect sizes, and their magnitudes to be of little physiological importance. The total within-subject variations for VO(2), VCO(2) and V (E) each equated to a mean CV and mean APE value of ~4 and ~6 %, whilst the respective inter-unit technological errors equated to ~1.5 and ~2.1 %. The two ParvoMedics TrueOne 2400 systems demonstrated excellent inter-unit agreement.

Show MeSH