DEVELOPMENT OF A MEASURE FOR CLINICAL DYSPNOEA: DYSPNOEA12

1
DEVELOPMENT OF A MEASURE FOR CLINICAL DYSPNOEA
DYSPNOEA-12 J Yorke (1), SH Moosavi (2), C
Shuldham (3), M Lau-Walker (2,3), C Haigh (1), PJ
Barnes (2), PW Jones (4). 1 University of
Salford, 2 Imperial College London, 3 Royal
Brompton and Harefield Trust, 4 St Georges
University of London.
Introduction
Results
Results

• Dyspnoea, like pain, consists of multiple
  dimensions including sensory quality and sensory
  affective.
• Dyspnoea is often measured using Borg scales or
  scales assessing its impact on activity and
  quality of life.
• No single measure is currently available to
  capture dyspnoea multidimensionalty and overall
  dyspnoea severity based upon its direct effect on
  the patient rather then through its impact on
  activity.

Figure 1 Process of Item Reduction

• Table 1 References for 81 items
• Papers describing the use of a unique list of
  descriptors for patient endorsement or where
  patients volunteered descriptors are listed.
• Bailey (2004)
• Caroci Lareau (2004)
• Edmonds et al. (2005)
• Elliot et al. (1991)
• Evans et al. (2002)
• Hardie et al. (2000)
• Heinzer, Bish Detwiler (2003)
• Killian et al. (2000)
• Mahler et al. (1996)
• Michaels Meek (2004)
• ODriscoll, Corner Bailey (1999)
• Parshall et al. (2001)
• Parshall (2002)
• Skevington et al. (1997)
• Simon et al. (1989) (1990)
• Tanaka et al. 2000

Table 2 Principal Components Analysis Items
loading gt 0.5 for each domain
Physical Emotional Domain
Domain Not in all the way 0.811
More work 0.715 0.333 Short of
breath 0.713 0.313 Difficulty catching
breath 0.696 0.320 Not enough air
0.688 0.398 Uncomfortable 0.653 0.456 Exhau
sting 0.578 0.479 Depressed 0.834 Mise
rable 0.821 Distressing 0.357 0.725 Agi
tated 0.488 0.658 Irritating 0.518 0.530

81 items
Hierarchical item reduction
47 items removed
Aim
34 items retained
Item Residual 6 items removed
To develop a valid scale of overall dyspnoea
magnitude using breathlessness descriptors.
Rasch Analysis
Differential Item Functioning Gender 7 items
removed
Methods

• 81 items were identified in the literature
  (Table 1).
• Patients responded none, mild, moderate or
  severe for each of the 81 items.
• Sample
• COPD 123 (age 698yr, 62 male, FEV1/FVC
  5512pred)
• Interstitial lung disease (ILD) 129 (age 5012
  yr, 47male, FEV1/FVC 7222pred)
• Chronic heart failure 106 (age 6811 yr, 72
  male, ejection fraction 3515).
• Hierarchical Item Reduction Items were excluded
  if 50 of patients rated them as none or if
  influenced by age (Plt0.01) (Figure 1).
• Rasch analysis informed decisions regarding
  further item removal (Figure 1) and to test
  overall fit to a unidimensional model - if data
  fit model expectations then all the items can be
  said to measure the same underlying construct,
  such as breathlessness, when tested together.
• Principal component analysis tested for the
  presence of different domains of the final list
  (Table 2).

Differential Item Functioning Diagnosis 9 items
removed
Items endorsed from a list Items volunteered
12 Items retained
Conclusions

• Rasch Analysis Reliability of the Dyspnoea-15
• The final 12 item list had a good Person
  Separation Index (0.89) showing that it could
  discriminate between patients with different
  degrees of breathlessness.
• The total item-trait statistic was
  non-significant (x2 95.5 P0.08) which shows
  good consistency between the patients and items.
• The 12 item list had a unidimensional structure
• Figure 2 illustrates targeting of the Dyspnoea-12
• Principal component analysis identified items
  that loaded gt0.5 on two domains physical and
  emotional.

• We have developed a novel instrument
  (Dyspnoea-12) for measuring dyspnoea that derives
  from the largest pool of respiratory descriptors
  collected from the literature reporting the
  patient experience.
• The Dyspnoea-12 provides an global score of
  breathlessness that incorporates both physical
  and emotional aspects of the experience.
• The Dyspnoea-12 is valid for use in obstructive
  and restrictive lung disorders and heart failure.
  

Future Directions
Figure 2 Suitability of the 12-item set for
study population Distribution of persons and
items based on Rasch logit. As the item severity
increases, patients must have worse dyspnea to
answer them positively.

• Further validation of the Dyspnoea-12 is now
  being carried out in patients with COPD, ILD and
  asthma (Funded by Action Medical Research).

This study was funded by Clinical Research
Committee Royal Brompton and Harefield Trust.