Skip to main content

Table 4 Showing the results of the linear mixed model analyses

From: Ambient bright light treatment improved proxy-rated sleep but not sleep measured by actigraphy in nursing home patients with dementia: a placebo-controlled randomised trial

Variables n Nmb obs Baseline to week 8 Baseline to week 16 Baseline to week 24
Regression coefficient Std.error P CI Regression coefficient Std.error P CI Regression coefficient Std.error P CI
SEa 65 201 −0.32 3.18 .920 −6.56-5.92 2.97 3.36 .377 −3.60-9.55 −3.65 3.26 .264 −10.04-2.73
TST 24 hb,c 65 194 −52.18 30.86 .093 −112.66-8.03 −36.11 33.95 .290 −102.64-30.43 −12.46 31.41 .692 −74.03-49.10
TST dayc 65 201 6.65 14.03 .636 −20.84-34.14 24.83 14.79 .096 −4.16-53.81 21.44 14.35 .138 -6.69-49.57
WASOd 65 201 14.42 10.43 .169 −6.03-34.86 2.59 10.98 .814 −18.92-24.11 20.92 10.68 .052 0.01–41.85
SDI transe 66 213 −0.04 .02 .094 −0.09-0.01 −0.06 .03 .020 −0.11--0.01 − 0.05 0.02 .028 −0.10--0.01
  1. Note: The R package lme4 assumes an unstructured variance-covariance structure
  2. CI confidence interval, Nmb obs Number of observations, P p-value, SDI Sleep Disorder Inventory, SE (variable) sleep efficiency, std.error standard error, trans transformed, TST total sleep time, WASO wake after sleep onset
  3. Note: The regression coefficients reflect the change from baseline in the intervention group compared to the control group
  4. aEye disease included as covariate
  5. bDrop-out included as covariate
  6. cTotal Functional Assessment Staging score in the relevant week included as covariate
  7. dAge included as covariate
  8. eAverage number of psychotropic medications included as covariate