Respiratory rate:measurementofvariability over timeandaccuracyatdifferent counting periods

Therespiratory rates/minute of97children weremonitored every 10-15 minutes overone hour, byanobserver andbypneumogram, at which times two30second andone60second countswereobtained. Thechildren were under5yearsofagewithlowerrespiratory tractinfections (n=20), upperrespiratory tract infections (n=34), orcontrols without acuterespiratory infection (n=43). The difference betweenrespiratory ratecount determined simultaneously by observation andpneumogram inrelation totheir mean count wasanalysed forthe60second counting period, 30plus30second period, andthe30 second period doubled. Themeandifference forthe60second period was1-79, forthe30 plus30secondperiod 1-42, andforthe30 second period doubled 1-72. Thevariability between respiratory ratecounts determined byobservation andpneumogram wassignificantly lowerincounts obtained whenthe subject wassleeping andhigher whenagitated compared withobtaining acountwhenthe subject wasawakeandcalmorfeeding. The variability wasalsosignificantly lowerin subjects with lower respiratory tract infections compared withthose withupperrespiratory tract infections andcontrol subjects without respiratory symptoms. Inthesamepatient, overtheonehour,50%ofthe60second counts varied byupto14breaths/minute and 75%byupto21breaths/minute. Theleast variability wasseeninchidren withalower respiratory tractinfection, whotendedto maintain their rapid breathing incontrast to those with anupperrespiratory tract infection andcontrols without respiratory symptoms. About10%ofinitial 30second counts, 12%of 60second, and16%ofinitial andrepeat 30 secondattempts toobtain accurate counts failed. Failures occurred morefrequently in children <2months ofageandthose agitated. Thedatafromthisstudysuggest thatone minute's counting either atastretch orintwo blocks of30secondintervals isbetter than counting therespiratory ratefor30seconds, whenthechild iseither awakeandcalmor whenasleep.