Skip to main content

Table 4 Level of discomfort, (lack of) comfort, pain, and respiratory distress in intervention phase versus pre-intervention phase for the control homes and the intervention homes together

From: Improving comfort in people with dementia and pneumonia: a cluster randomized trial

Level of suffering Unadjusted analyses Adjusted analysesa
Coefficient (95 % CI) ICC Coefficient (95 % CI) ICC*
Discomfort (DS-DAT) + 0.90* (0.82–0.98) 0.12 = 0.93 (0.85–1.01) 0.12
Lack of comfort (reverse EOLD-CAD) + 0.98* (0.96–0.99) 0.09 + 0.98* (0.97–1.00) 0.13
Pain (PAINAD) + 0.93* (0.88–0.99) 0.07 = 0.96 (0.91–1.01) 0.10
Respiratory distress (RDOS) + 0.91* (0.86–0.97) 0.16 + 0.92* (0.87–0.98) 0.18
  1. Discomfort (DS-DAT Discomfort Scale-Dementia of Alzheimer Type; range 0–27), (lack of) comfort (EOLD-CAD End Of Life in Dementia-Comfort Assessment in Dying; range 14–42), pain (PAINAD Pain Assessment in Advanced Dementia; range 0–10), respiratory distress (RDOS Respiratory Distress Observation Scale; range 0–16)
  2. Results from models without interaction terms of group and intervention phase with the level of observed sleepiness or death within 20 days
  3. Analyses were performed on log-transformed data after which coefficients and confidence intervals were back-transformed. Therefore, coefficients should be interpreted as a ratio, so that a coefficient of 1 means no difference
  4. CI confidence interval
  5. ICC intracluster correlation; *ICC for adjusted analyses determined with non-imputed data for covariates
  6. *p < 0.05
  7. + Significant association that is interpreted in terms of a positive attribute (favors intervention phase: decrease in discomfort, lack of comfort, pain and respiratory distress)
  8. = No significance
  9. aAdjusted for age, sex, season, severity of dementia (BANS-S score), type of dementia (Alzheimer’s or other), death within 20 days and baseline differences (Table 2 resident characteristics, and Additional file 1 Table S1 symptoms fo pneumonia). Adjusted analyses are performed on multiply imputed data for covariates