The second problem is that imprecise measures lead to distortions in statistical relations. Random variation introduces “noise” that can mask the relationship between an input measure and a result such as admission to intensive care. Random variation is erased, which could be a clear association . For example, in the United Kingdom, the current points system establishes an abrupt correlation between respiratory frequency and respiratory rate score between 20 and 25 (Figure 4). On the other hand, a machine-based decision tree produced a more gradual adaptation . With the decision tree, a rate of up to 18 points 0, 19 to 20 points 1, 21 to 25 points 2 and 25 or more is rated at 3, which gradually affects the measures of clinical respiratory frequency. Similarly, a logistic regression process can more easily classify less good results . In simple terms, the conversion of a continuous measure (for example. B respiratory rate) in categories (for example. B a score) using break values is statistically ineffective .
The records were analyzed with proprietary software (Spike2, version 5.19; CED, Cambridge, United Kingdom). Each breathing period was automatically identified and recorded using a threshold detection device in the display software to give a time sequence (accuracy >0.1 s) from the start of the recording (Figure 1a). For each patient, a total airflow per minute was calculated by deifying the total number of complete test cycles by the total duration of these cycles. This value represents the most accurate measure of respiratory rate for this patient (Figure 1b). The variability between measurements by health professionals in clinical practice and observers under research conditions highlights the imprecise methods that are applied. Such variability in clinical observations has been overlooked by research on physiological assessment and alert systems that do not take into account the inaccuracy of the underlying data. The first obvious problem is that the measure itself is imprecise. Given that recognition of this problem and attempts to reduce this problem were limited, we should consider the possibility that not only increased confidence in respiratory measurement for patient evaluation, but also that better information could improve management. Remember to evaluate the heart rate with an electrocardiogram compared to the palpation of the pulse! An important call to action, stemming from these results and applicable to all health care facilities, including low- and middle-income countries, is to return to basics and reinforce the importance for health care professionals of using one minute of time to accurately measure respiratory frequency. Reviewer #1: A well-written, sea-related study, but incredibly important from a clinical point of view. The only thing I think would reinforce the article are some suggestions in the debate, such as the evaluation and accuracy of the respiratory frequency count.
A total of 448 health professionals participated. The median measurements were slightly higher (1-3/min) than the actual respiratory rate, and 78.2% of the measurements were within 4 minutes of the actual rate. CCI was moderate (0.64, 95% CI 0.39-0.94). Comparing measured respiratory frequencies with categorical judgments, 14.5% were inconsistent.