Table 5, Results of studies of cough frequency assessment instruments - Assessment and Management of Chronic Cough

Table 5Results of studies of cough frequency assessment instruments

Study	Index Test	Comparator(s)	Validity Assessed?	Reliability Assessed?	Responsiveness Assessed?	Results
Studies in Adults and Adolescents
Barnabe, 1995³⁴	VAS	NA	No	No	Yes	Mean VAS scores of cough frequency decreased over the 2 days of treatment.
Barry, 2006³⁵	Hull Automatic Cough Counter	Cough count (by a human)	No	Yes	No	SN: 0.80 SP: 0.96 Reproducibility: 100%
Birring, 2008³⁶	Leicester Cough Algorithm	Cough count (by a human)	No	Yes	No	Correlation: 0.9 (p<0.001) (stage 1); 0.93 (p<0.001) (stage 2) SN: 0.91 (stage1); 0.86 (stage 2) SP: 0.99 (stage 1); 0.99 (stage 2) Reliability: (ICC): 0.9
Birring, 2006³⁷	Leicester Cough Monitor	Electronic cough recorder (sound or pressure)	No	Yes	No	Reliability (ICC): 0.8
		Leicester Cough Questionnaire	Yes	No	No	Correlation Total score: -0.6 (p=0.03) Physical: -0.6 (p=0.03) Psychological: -0.5 (p=0.08) Social: -0.7 (p=0.01)
		Capsaicin tussigenic challenge (C5)	Yes	No	No	Correlation: 0.9 (p=0.008)
		Capsaicin tussigenic challenge (C2)	Yes	No	No	Correlation: 0.8 (p≤0.05)
Coyle, 2005⁴⁵	LifeShirt system	Video cough recorder	No	Yes	No	Overall (95% CI) SN: 78.1 (76.7 to 79.4) SP:99.6 (99.5 to 99.6) PPV: 84.6 (83.3 to 85.8) NPV: 99.4 (99.0 to 99.1) Comparison of nighttime with daytime: SN, SP, PPV, NPV=p<0.0001
Crawford, 2008⁴⁶	Electronic cough recorder	CASA-Q Cough symptoms	Yes	Yes	Yes	The cough symptom and sputum symptom domains did not correlate with cough recordings.
Decalmer, 2007⁴⁹	Electronic cough recorder	VAS Cough challenge	Yes	Yes	No	“The repeatability of cough monitoring was excellent,” both for day and night recordings. There was a significant inverse correlation between log10 daytime cough rates and log C5 (Pearson's r=20.45, p=0.001). Subjective cough measures tended to correlate more strongly with objective cough rates for overnight than for daytime recordings and the VAS correlated more strongly than cough scores.
Faruqi, 2011⁵⁵	Hull Automated Cough Counter	NA	No	No	Yes	Median cough counts were significantly lower 8 weeks after baseline assessment.
		Leicester Cough Questionnaire	Yes	No	No	Correlation: -0.62 (p<0.001)
		SAS	Yes	No	No	Correlation: 0.54 (p<0.001)
		VAS	Yes	No	No	Correlation: 0.38 (p=0.007)
		C2	Yes	No	No	Correlation: -0.45 (p=0.001)
		C5	Yes	No	No	Correlation: -0.55 (p<0.001)
		Reproducibility	No	Yes	No	r=0.91, (p<0.001)
Fisman, 2001⁵⁷	Cough frequency score	Cough severity score	No	No	Yes	Cough severity and cough frequency scores decreased, respectively, from 2.6 ± 1.1 to 0.7 ± 1.0 (p<0.001) and from 6.9 ± 2.2 to 2.1 ± 2.4 (p<0.001).
Freestone, 1997⁵⁹	Electronic cough recorder	NA	Yes	No	Yes	Correlation: r=0.524 (95% CI, 0.32 to 0.68), p<0.0001 between cough frequency and CSPLs Responsiveness: Decrease in all 3 measures of cough (p<0.001)
		Score	Yes	No	No	Correlation: NS
		VAS	Yes	No	No	Correlation: NS
Hsu, 1994⁶⁵	24-hour ambulatory recorder and EMG	Cough count	No	Yes	No	Correlation: 0.99, p<0.005
Hsu, 1994⁶⁵	24-hour ambulatory recorder and EMG	Score	Yes	No	No	Correlation: 0.96, p<0.005
Kelsall, 2011⁷²	24-hour cough recording	Cough counts VAS	Yes	Yes	Yes	Intraclass correlation: r=0.98 (p<0.001) Change in cough frequency did not correlate with the change in cough VAS scores or cough scores during the day ( r=0.26; p=0.052, and r=0.23; p=0.08, respectively) and correlated only with cough VAS scores at night ( r=0.45; p=0.001, and r=0.23; p=0.10, respectively).
Kelsall, 2009⁷⁰	24-hour ambulatory recorder	Cough challenge	Yes	No	Yes	Log 2 and log 24-hr cough frequency and log overnight cough frequency correlated with cough reflex sensitivity to citric acid (logC5).
Kelsall, 2008⁷¹	24-hour ambulatory recorder Parameters measured: cough epochs, explosive phases, cough seconds	Cough count (by a human)	No	Yes	No	Correlation: Cough epochs vs. cough seconds (r²=0.84) Correlation: Cough epochs vs. explosive phases (r²=0.80) Correlation: Cough seconds vs. explosive phases (r²=0.98)
		Daytime VAS	Yes	No	No	Correlation: with explosive phases: r=0.45 (p≤0.001)
			Yes	No	No	Correlation: with cough seconds: r=0.44 (p≤0.001)
			Yes	No	No	Correlation: with cough epochs: r=0.40 (p=0.001)
		Nighttime VAS	Yes	No	No	Correlation: with explosive phases: r=0.67 (p≤0.001)
			Yes	No	No	Correlation: with cough seconds: r=0.64 (p≤0.001)
			Yes	No	No	Correlation: with cough epochs: r=0.60 (p≤0.001)
		Leicester Cough Questionnaire	Yes	No	No	Correlation: with explosive phases: r=-0.53 (p≤0.001)
			Yes	No	No	Correlation: with cough seconds: r=-0.53 (p≤0.001)
			Yes	No	No	Correlation: with cough epochs: r=-0.46 (p≤0.001)
Key, 2010⁷³	Electronic cough recorder	Cough count (by a human)	No	Yes	No	Correlation: Mean difference 0.9 (±1.7)
		VAS	Yes	No	No	Correlation: Day r=0.80 (p<0.001) Night r=0.71 (p=0.001)
		Leicester Cough Questionnaire	Yes	No	No	Correlation Total score: -0.80 (p<0.001) Physical: -0.76 (p<0.001) Psychological: -0.76 (p<0.001) Social: 0.74 (p<0.001)
Krahnke, 2004⁷⁴	Audio cough recorder	NA	No	No	Yes	Cough counts decreased over 3 days of treatment for seasonal allergic rhinitis.
		Daytime score	Yes	No	No	Correlation: 0.54 (p<0.0001)
		Nighttime score	Yes	No	No	Correlation: 0.39 (p=0.0006)
		Daytime plus nighttime score	Yes	No	No	Correlation: 0.51 (p<0.001)
Krajnik, 2010⁷⁵	Portable automatic cough analyzer	Daytime total cough incidents, score	Yes	No	No	Correlation: 0.63 (p=0.22)
		Daytime cough incidents per hour, score	Yes	No	No	Correlation: 0.60 (p=0.29)
		Time spent coughing per hour during the day, score	Yes	No	No	Correlation: 0.48 (p=0.1)
		Nighttime cough incidents per hour, score	Yes	No	No	Correlation: 0.29 (p=0.34)
		Time spent coughing per hour during the night, score	Yes	No	No	Correlation: 0.26 (p=0.4)
		Total time spent coughing per hour	Yes	No	No	Correlation: 0.37 (p=0.21)
		Total cough incidents per hour	Yes	No	No	Correlation: 0.52 (p=0.066)
Leconte, 2011¹⁰⁵	LR102 recording device	Video recorder	No	Yes	No	The cough meter was well tolerated by all but one patient who complained of itching at the electrode sites. The two recording methods produced cough frequencies that were closely correlated (r=0.87 for number of cough episodes per hour; r=0.89 for number of single coughs per hour). There was no systematic difference between the two measures across the spectrum of cough frequency. ICCs were also good (ICC=0.86 for episode [95% CI, 0.75 to 0.92] and 0.88 for single cough [95% CI, 0.78 to 0.93]). The number of coughs per hour measured by the cough meter was significantly higher than that measured by counting coughs on the video recording (number of cough episodes per hour 22.57 vs. 18.77, respectively; number of single coughs per hour 65.22 vs. 52.67, respectively). The mean difference between the two methods was 3.8 for cough episodes per hour (p=0.04) and 12.5 for single coughs per hour (p<0.01).
Marsden, 2008⁷⁷	Electronic cough recorder	Tussigenic challenge (C5)	Yes	No	No	Correlation: 0.08 (p=0.65)
		Tussigenic challenge (C2)	Yes	No	No	Correlation: 0.39 (p=-0.03)
		Score	Yes	No	No	Correlation: 0.32 (p=0.45)
		Leicester Cough Questionnaire	Yes	No	No	Correlation Total score: 0.54 (p<0.001) Physical: -0.45 (p=0.001) Psychological: -0.48 (p≤0.001) Social: -0.55 (p<0.001)
Matos, 2007⁷⁸	Leicester Cough Monitor	Cough count (by a human)	No	Yes	No	SN: 97.8 (IQR, 88.1 to 99.1)
Ribeiro, 2007⁹³	Diary VAS	NA	No	No	Yes	There was a significant difference in the cough diaries and VAS scores before and after beclomethasone treatment compared with before and after placebo use (difference of differences, 1.0; 95% CI, 0.4 to 1.5; p=0.002 for diaries; and difference of differences, 1.1; 95% CI, 0.6 to 1.8; p=0.01 for the VAS).
Shaheen, 2011⁹⁴	Fisman cough frequency score	NA	No	No	Yes	Cough frequency score decreased from 6.2 (SD 1.8) to 3.0 (SD 1.8) in the PPI group and form 6.8 (SD 2.0) to 4.5 (SD 2.5), p=0.3.
Smith, 2006⁹⁸	Video recorder	Cough count (by a human)	No	Yes	No	Manual counting of cough sounds from digital audio recordings has excellent agreement with simultaneous video recordings in laboratory conditions.
Smith, 2006⁹⁷	Cough recorder	Cough score VAS	Yes	No	Yes	Correlation with day cough frequency: Change in cough score r=0.34, p=0.204 Change in VAS r=0.47, p=0.070 Correlation with night cough frequency: Change in cough score r=0.19, p=0.510 Change in VAS r=0.81, p=0.001
Smith, 2006⁹⁶	Cough recorder	Cough challenge Cough score VAS	Yes	No	Yes	There were no significant correlations between the change in time spent coughing and the change in cough threshold (r=20.27; p=0.30), change in cough score (day r=20.07, p=0.78; night r=0.17, p=0.48), or change in VAS (day r=20.07, p=0.79; night r=0.30, p=0.24).
Smith, 2006⁹⁵	Cough recorder	Tussigenic challenge	Yes	No	No	Subjective measures of cough and cough reflex sensitivity are statistically related to time spent coughing in patients with COPD, but with low-to-moderate levels of correlation.
		Electronic recorder	Yes	No	No
		Score	Yes	No	No
		Cough-specific quality of life	Yes	No	No
Thomas, 1978¹⁰²	DATA recording system	Cough count (by a human)	No	Yes	No	Reliability (ICC): Coefficient of variation: 1.8% Reproducibility: Able to detect effect of codeine with a probability of <0.01 Accuracy: Group 1 correlation coefficient : 0.99; coefficient of determination=0.98 Group 2 correlation coefficient: -0.94; coefficient of determination=0.88
Woodcock, 2010¹⁰⁰	Lifeshirt cough recorder	NA	No	No	Yes	The intervention drug (SCH486757) was associated with reduced cough counts as measured by the Lifeshirt. On day 5, SCH486757 reduced cough counts by a median of 7.0 coughs/hour, codeine reduced counts by 13.8 coughs/hour, and placebo reduced counts by 7.8 coughs/hour over the first 4 hours.
Woolf 1964¹⁰³	Human cough count Hospital room equipped with a recording system	Human cough count	No	Yes	Yes	There was nearly perfect correlation of cough counts between two human observers. Cough curves were constructed for each of the 4 treatment periods for this single-subject study. For the most part, the cough curves did not cross over the 24-hour study period.
Studies in Adults, Adolescents, and Children
Hamutcu, 2002⁶²	Logan Research (LR 100) cough monitor	Electronic cough recorder	No	Yes	No	Correlation: 0.96, p<0.0001
Paul, 2006⁹⁰	Accelerometer	Video recorder	No	Yes	No	Correlation: 0.997 (p<0.001)
Studies in Children Only
Archer, 1985³¹	Cough recorder	Diary	Yes	No	No	There was no correlation between night cough counts and diary card scores for night-time wheeze, daytime wheeze, daytime activity or for 24 hour diary scores calculated for the 24 hour period beginning with, and the period ending with, the night recording.
Chang, 2003⁴²	Ambulatory cough recording	Parent cough score Child cough score Cough challenge	Yes	No	No	Log cough frequency correlated with parent and child recorded log cough score (r(s)=0.42, p=0.02) and r(s)=0.44, p=0.01, respectively). Cough scores did not correlate with cough challenge test.
Chang, 1998⁴⁰	Electronic cough monitor	Parent-completed diary Child-competed diary	Yes	No	No	The agreement adjusted for chance (Cohen's Kappa) between the subjective and objective presence of daytime cough for the subjects was 100%, and that for controls was 0.61 for child-completed cards and 0.44 for parent-completed cards. When both groups were considered together, the agreement was 0.67 and 0.47, respectively. For night-time cough, the agreement between the subjective and objective presence of cough was poor both in subjects and controls. In the subjects, there was no difference between the parent and child scores.
Chang, 1997⁴¹	Holter monitor cough meter	Electronic cough recorder (sound or pressure)	No	Yes	No	Mean difference of -0.3 coughs x h(-1), 95% CI (-0.7, 0.2); limits of agreement -2.2 to 1.7 coughs x h(-1).
Corrigan, 2003⁴⁴	Logan Research (LR 100) cough monitor	Video cough recorder	No	Yes	No	SN: 81.00% PPV: 0.8
Dales, 1997⁴⁷	Electronic cough recorder	Cough count (by a human)	No	Yes	No	SN: 80 SP: 95 Accuracy: 90
Dales, 1997⁴⁷	Electronic cough recorder	Diary	Yes	No	No	Kappa: ≤0.10
Falconer, 1993⁵³	Electronic cough recorder	Diary	Yes	No	No	Mean Kappa between reported and recorded cough was 3.0 (range: -0.17 to 1.0), representing poor agreement beyond chance.
Fuller, 1998⁶¹	Video recording	Parent questionnaire Parent diary	Yes	No	No	On the second night of recording the between-subject correlation coefficient for individual coughs and percentage of the night awake was 0.25, p=0.13 (n=37) and for individual coughs and percentage of the night awake and restless was 0.36, p=0.03 (n=37). The within-subject correlation coefficient log % awake time on log number of coughs was 0.26 (p=0.01; r=0.13, SE 0.036) with a residual coefficient of variation of 74%. The within-subject correlation coefficient (log % awake time + % restless) was 0.17 (p=0.02; r=0.016, SE 0.0071) with a residual coefficient of variation of 17%.
Hoskyns, 1991⁶⁴	Cough recorder	Parental night cough diary	Yes	No	Yes	Correlation: (r=0.588, p<0.02) during medication week; no correlation during placebo week
Newcombe, 2010⁸⁶	Digital voice recorder	PC-QOL psychological	Yes	No	No	Correlation, Time 1: -0.10 (p=0.521) Correlation, Time 2: -0.28 (p=0.089)
		PC-QOL physical	Yes	No	No	Correlation, Time 1: -0.21 (p=0.188) Correlation, Time 2: -0.46 (p=0.003)
		PC-QOL social	Yes	No	No	Correlation, Time 1: -0.11 (p=0.487) Correlation, Time 2: -0.51 (p=0.001)
		PC-QOL physical	Yes	No	No	Correlation, Time 1: -0.15 (p=0.329) Correlation, Time 2: -0.42 (p=0.008)
	Cough count/hour	NA	No	No	Yes	Effect size for responsiveness to change over time: 0.19
Zihlif, 2005¹⁰¹	LR100 cough recorder	Parent cough scores	Yes	No	No	There was significant correlation between parental scoring of day and night cough compared with actual number of recorded coughs (r=0.930, p<0.0001 for daytime cough; r=0.711 for nighttime cough, p=0.002).

: CASA-Q = Cough and Sputum Assessment Questionnaire; CI = confidence interval; COPD = chronic obstructive pulmonary disease; DATA = Discriminator and Accumulator of Tussive Activity; EMG = electromyogram; ICC = intraclass correlation coefficient; IQR = interquartile range; NA = not applicable; NPV = negative predictive value; NS = not statistically significant; PC-QOL = Parent Cough-specific Quality-of-Life questionnaire; PPV = positive predictive value; SE = standard error; SN = sensitivity; SP = specificity; VAS = visual analog scale

From: Results

Cover of Assessment and Management of Chronic Cough

Assessment and Management of Chronic Cough [Internet].

Comparative Effectiveness Reviews, No. 100.

McCrory DC, Coeytaux RR, Yancy WS Jr, et al.

Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Jan.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.