Skip to main content
. 2018 Mar 9;15(3):e1002526. doi: 10.1371/journal.pmed.1002526

Table 3. PACE-UP and PACE-Lift studies: Imputation analyses for step counts at 3 years (PACE-UP) and 4 years (PACE-Lift).

Imputation models PACE-UP study PACE-Lift study
Postal versus Control Nurse versus Control Intervention versus Control
  N Effect (95% CI)
p-value
Effect (95% CI) p-value N Effect (95% CI) p-value
All participants with follow-up data 681 627 (198–1,056)
0.004
670 (237–1,102) 0.002 225 407 (−177–992) 0.17
Imputed using treatment group, baseline steps, gender, age, practice, month baseline accelerometry 1,023 597 (174–1,020)
0.006
679 (268–1,089) 0.001 298 429 (−152–1,010) 0.15
Imputed using treatment group, baseline steps, gender, age, practice, month baseline accelerometry, baseline deprivation, baseline self-reported pain, and baseline body fat mass 996 634 (211–1,057)
0.003
735 (293–1,178) 0.001 292 437 (−154–1,028) 0.15
Imputed using treatment group, baseline steps, gender, age, practice, month baseline accelerometry, and 12-month steps†† 965 625 (217–1,033)
0.003
683 (270–1,095) 0.001 280 367 (−181–916) 0.19

Multiple imputations were carried out using the Stata commands mi impute followed by mi estimate in Stata V12, with three different sets of covariates, as listed in the table. The results shown are the change in each intervention group relative to the change in their control group. The effect estimates, 95% confidence intervals, and p-values were obtained from the model output.

Imputed values were only available for 996 (PACE-UP) and 292 (PACE-Lift) when including baseline deprivation in the model (PACE-UP NS-SEC and PACE-Lift Index of Multiple Deprivation).

††Imputed values only available for 965 (PACE-UP) and 280 (PACE-Lift) when including 12-month steps in the model.

Abbreviations: NS-SEC, National Statistics Socio-economic classification; PACE-Lift, Pedometer Accelerometer Consultation Evaluation-Lift; PACE-UP, Pedometer And Consultation Evaluation-UP.