Every paramedic will have been taught how to take a respiratory rate at the very outset of their career, and will recognise it as one of the vital signs that should be recorded for (almost) every patient encounter. It is a relatively easy procedure to undertake in most circumstances and it provides essential information that can help identify the deteriorating patient. Given the relative ease of the procedure, it may seem excessive to dedicate an entire article to the respiratory rate; however, most of this article will focus on the why rather than the how.
It will address how to take a respiratory rate, but will also consider questions such as why we should take a respiratory rate and why it is essential for us to be accurate. It will also address some of the issues around what is considered ‘normal’ as a consensus on this is necessary before what is ‘abnormal’ can be judged. Although we have only limited understanding of how paramedics perform in this area, we do have evidence from other healthcare professions that may cause us to reflect on our own practice. Some of those findings will be outlined later in this article.
Respiratory rate
Respiratory rate (RR) is the number of times that a person breathes in and out each minute and is the single best predictor of severe illness. Changes in RR appear particularly sensitive to deterioration and provide a strong predictor of morbidity and mortality in adults and children (Fieselmann et al, 1993; Hodgetts et al, 2002; Cretikos et al, 2008; Churpek et al, 2012; Mochizuki et al, 2016; Daw et al, 2020). As well as being a valuable indicator of our patients' condition, respiratory deterioration is also one of the major causes of critical illness in the UK. Given this evidence, it is reasonable to argue that the accurate assessment and management of respiratory function is paramount in patient care and should never be estimated.
What is a ‘normal’ respiratory rate?
There is some disagreement over what constitutes a ‘normal’ rate in adults, although an accepted range is typically between 12–20 breaths per minute (Chourpiliadis and Bhardwaj, 2022). Respiratory rates vary with age (higher in children and older adults) and this will need to be considered when assessing the patient.
Respiratory norms for children are provided by the Joint Royal Colleges Ambulance Liaison Committee (JRCALC) (Association of Ambulance Chief Executives (AACE), 2019) on their ‘page for age’ in the iCPG app and are used widely across UK ambulance services. However, it is not clear where the figures came from, nor how accurately those figures represent the child population. A second set of norms are provided by the Advanced Life Support Group (ALSG) (2016) and though there is overlap with JRCALC, there are also clear differences. Fleming et al (2011) undertook a systematic review of observational studies of norms for respiratory and heart rates in children and provided centile charts based upon 20 studies involving 3881 children. Their centile charts can be found at https://pmc.ncbi.nlm.nih.gov/articles/PMC3789232/figure/F2/ (open access) and provide a useful benchmark for the guideline figures generally used in clinical practice. The JRCALC figures do not always include the median respiratory rate from the data provided from Fleming's systematic review, which could lead to higher-than-necessary proportions of children being classed as ‘severe’ or ‘high risk’. For example, at 3 months of age, the median respiratory rate in Fleming's group was around 42 breaths per minute, but JRCALC gives a maximum of 40 for this age group. This means that infants with a median average respiratory rate have a rate that is above the maximum range postulated by JRCALC. Further studies such as Bonafide et al (2013) and O'Leary et al (2015) add to the confusion as their data did not mirror Fleming's.
Table 1. Comparison of norms for respiratory rates in children
| JRCALC | APLS | |
|---|---|---|
| Age | Respiratory rate (breaths/minute) | Respiratory rate (breaths/minute) 5th–95th centile |
| Birth | 40–60 | 25–50 |
| 1 month | ||
| 3 months | 30–40 | 25–45 |
| 6 months | 20–40 | |
| 9 months | ||
| 12 months | 25–35 | |
| 18 months | 20–35 | |
| 2 years | 25–30 | 20–30 |
| 3 years | ||
| 4 years | ||
| 5 years | 20–25 | |
| 6 years | ||
| 7 years | ||
| 8 years | 15–24 | |
| 9 years | ||
| 10 years | ||
| 11 years | ||
| 12 years | As adult values | 12–24 |
| 14 years |
Likewise, the Advanced Paediatric Life Support (APLS) guidelines (ALSG, 2016) have been criticised for providing norms that are too low (Brennan et al, 2023). Brennan assessed over 192 000 records from admissions to hospitals in the North-west of England and compared them with the APLS figures – their conclusion was that the findings in their study suggested that the heart and respiratory rates used in the current APLS guidelines are too low, especially among younger children.
This information is challenging as it shows disparity between centiles and guidelines but does not provide us with a solution. It would be reasonable to suggest that those writing the guidelines should review their figures in line with the growing data on respiratory norms; however, making a specific recommendation for how an individual clinician should respond is beyond the remit of this article.
Older adults
Ageing is a statistically significant factor in determining the normal respiratory rate in older patients (>59 years to <100 years) (Takayama et al, 2019), with an expectation that respiratory rate increases with age, independent of underlying disease. What is less clear is whether the increase in rate progresses linearly with age. It is reasonable to suggest that healthcare professionals recognise the influence of ageing on the respiratory rate in order to optimise vital sign assessment. However, it is not yet possible to quantify the impact of ageing on respiratory rates.
Measuring respiratory rate
Do we do it accurately?
Measuring the respiratory rate accurately is essential but evidence suggests that it is often the most poorly recorded vital sign (Cretikos et al, 2008; Parkes, 2011; Ansell et al, 2014; Hosking et al, 2014; Badawy et al, 2017; Flenady et al, 2017). These studies tend to focus on nurses rather than paramedics, but it would be unwise to suggest that we are better than nurses unless we can provide the evidence. The study by Badawy et al (2017) is particularly interesting. It used data from 28 511 patients representing 220 665 unique hospital days and found that respiratory rates were not normally distributed. There was little variation in respiratory rates (maximum respiratory rate equalled 18 or 20 in 75% of all hospital days), even in those with cardiopulmonary compromise or immediately prior to transfer to intensive care. This contrasts with the more objectively measured heart rate, which was more normally distributed.
Why is RR not accurately measured?
There is limited understanding as to why clinicians often fail to accurately record regular respiratory rates but the evidence that is available suggests issues such as workload, time pressures, interruptions to workflow, a lack of specialised training, and little appreciation of the value of this vital sign with regard to clinical deterioration (Elliot, 2016; Flenady et al, 2017). Flenady et al (2017) used grounded theory to help explain why emergency department nurses failed to properly assess a respiratory rate and found the following:
- Nurses believed that counting respiratory rates at each round of observations is superfluous and wastes valuable time
- This cohort of nurses believe they are enhancing patients' outcomes by performing tasks other than counting respiratory rate
- Recording an accurate respiratory rate was not a priority unless the patient was exhibiting signs of respiratory distress, was a paediatric patient, or had a history of respiratory-related illness.
The importance of accurate assessment and recording cannot be over-emphasised as illustrated by two large studies. The first study (Bleyer et al, 2011) involved data from over 1 million adult patients and found that an increase in respiratory rate to 24–28 breaths per minute represented an increase to the risk of mortality by 5%. The second study (Ljunggren et al, 2016) found that when the respiratory rate fell to 8 or less, a patient has up to 18.1 times higher odds of death within 24 hours of presentation compared with a patient who has a normal respiratory rate. The respiratory rate matters and should not be seen as a tick-box exercise.
How should we measure respiratory rate?
- The respiratory rate should be counted for a full minute. It may be tempting to count for 15 or 30 seconds and then multiply the finding by 4 or 2 respectively but this should be avoided. Using these shortcuts results in an estimate of the actual respiratory rate and studies have shown these methods to be inaccurate when compared with counting for the full minute (Kallioinen et al, 2019; Rimbi et al, 2019). The following conclusion was made by Rimbi et al (2019; 145) in their study:‘Rates observed over 1 min that scored 3 National Early Warning Score points were not identified by half the rates derived from 15 s and a quarter of the rates derived from 30 s.’
- The rate is probably best counted by direct visualisation of the chest. This also allows for an appreciation of depth of breathing and an assessment of the inspiratory/expiratory (I:E) ratio. The I:E ratio should normally be between 1:1.5 and 1:2, with expiration lasting between 1.5 and two times as long as inspiration (Wang et al, 2013). However, it must be noted that this may change in diseased states.
Important note
It is not possible to count the patient's respirations when they are talking. This may sound obvious but it is something to consider if your colleague is undertaking the assessment when you decide to ask a question.
Does it matter if the patient knows we are counting their respirations?
The evidence shows that patient awareness of assessment of their breathing causes a decrease in their RR irrespective of how the RR is measured (Kallioinen et al, 2021). This means that the reduction is directly attributable to their awareness of the assessment.
Tip: take the patient's pulse and then immediately count their respirations while still holding the pulse
What else contributes to RR errors?
Value bias
A number of studies have looked into suggestions that particular respiratory rates are over-represented in manually measured RR because they represent an estimate rather than a count. Kallioinen et al (2021) helpfully summarised their findings and made the comment that those studies (especially those that employed criterion-measured comparisons) provide clear support for the suggestion that there is a tendency to bias RR data by recording values that are thought to be common or ‘normal’. It is not common to see odd numbers recorded as these cannot be derived by counting for 15 or 30 seconds. One study also highlighted another peculiarity where an RR of 15 did occur frequently, along with an RR of 20 – these coincided with the lower and higher thresholds for scoring one on the hospital's track-and-trigger system.
Recording
Many studies reported by Kallioinen et al (2021) showed large numbers of recording errors where the RR was missed off altogether. They cite a study by Ramgopal et al (2018) who assessed the recording of vital signs in emergency medical services staff and found an omission rate of 1.85% for RR. That is good compared with the other studies; however, we still do not know whether the figures written down represented the actual RR or an estimated/guessed RR.
Conclusion
Counting RR is not difficult but there is evidence to suggest that healthcare professionals do not measure RR as accurately as other vital signs. We do not know how well we are doing in paramedic practice, but the lack of evidence of wrongdoing on our part does not mean that we get it right all of the time. A few small changes to practice and an appreciation of the value and significance of the RR could be beneficial to our patients. It is perhaps worth considering that the movement from a ‘normal’ to ‘abnormal’ RR is relatively small in numerical terms, especially compared with the heart rate. An RR change from 12 to 24 is the same percentage increase as a HR change from 60 to 120 beats per minute.