Skip to main content

Advertisement

Table 1 Model input parameters

From: The cost-effectiveness of a theory-based online health behaviour intervention for new university students: an economic evaluation

Parameter Distribution Parameters* Source
Individual characteristics
Age (years) Individual-level data Mean = 18.90 U@Uni RCT
SD = 2.49
Gender Individual-level data 42% Male
58% Female
Fruit and vegetables (portions per day) Individual-level data Mean = 6.52
SD = 4.94
Alcohol (units per week) Individual-level data Mean = 11.51
SD = 18.62
Physical activity (minutes per week) Individual-level data Mean = 163.35
SD = 121.74
Smoking status Individual-level data 12% Smoker
88% Non-smoker
Costs
U@Uni cost (full development) Lognormal Mean ln(cost) = 12.2763 Costing analysis
SD ln(cost) = 0.0513
U@Uni cost (roll-out) Lognormal Mean ln(cost) = 10.3349 Costing analysis
SD ln(cost) = 0.0663
Intervention effect regression coefficients
β0 Fruit and vegetables: constant Multivariate normal (see section C of Additional file 1 for covariance matrix) Mean = 3.0095 U@Uni RCT
β1 Fruit and vegetables: baseline behaviour coefficient Mean = 0.2482
β2 Fruit and vegetables: age coefficient Mean = 0.0586
β3 Fruit and vegetables: gender coefficient (1 = male; 0 = female) Mean = -0.0252
β4 Fruit and vegetables: intervention coefficient (i.e. mean effect of U@Uni on portions of fruit and vegetables per day compared to control) Mean = -0.1116
ϵ Fruit and vegetables: residual Normal Mean = -3.18 × 10-09 U@Uni RCT
SD = 4.8973
β0 Alcohol: constant Multivariate normal (see section C of Additional file 1 for covariance matrix) Mean = 18.9219 U@Uni RCT
β1 Alcohol: baseline behaviour coefficient Mean = 0.4834
β2 Alcohol: age coefficient Mean = -0.5772
β3 Alcohol: gender coefficient (1 = male; 0 = female) Mean = -0.4252
β4 Alcohol: intervention coefficient (i.e. mean effect of U@Uni on units of alcohol per week compared to control) Mean = -0.4275
ϵ Alcohol: residual Normal Mean = 1.85 × 10-08 U@Uni RCT
SD = 19.9055
β0 Physical activity: constant Multivariate normal (see section C of Additional file 1 for covariance matrix) Mean = 161.3326 U@Uni RCT
β1Physical activity: baseline behaviour coefficient Mean = 0.2339
β2 Physical activity: age coefficient Mean = -2.7763
β3 Physical activity: gender coefficient (1 = male; 0 = female) Mean = -2.0189
β4 Physical activity: intervention coefficient (i.e. mean effect of U@Uni on minutes of physical activity per week compared to control) Mean = 4.0141
ϵ Physical activity: individual residual Normal Mean = -1.38 × 10-07 U@Uni RCT
SD = 106.0016
Probability smokers quit smoking (U@Uni) Beta α = 27 U@Uni RCT
β = 33
Probability non-smokers start smoking (U@Uni) Beta α = 14 U@Uni RCT
β = 466
Probability smokers quit smoking (do nothing) Beta α = 19 U@Uni RCT
β = 45
Probability non-smokers start smoking (do nothing) Beta α = 27 U@Uni RCT
β = 462
Lag effects (years until full effect of behaviour change on mortality risk)
Fruit and vegetables lag Lognormal Mean = 2.7438 Expert elicitation
SD = 0.1247
Alcohol lag Gamma α = 1.3541 Expert elicitation
β = 0.6537
Physical activity lag Normal Mean = 5.5000 Expert elicitation
SD = 1.4642
Smoking lag Normal Mean = 5.5000 Expert elicitation
SD = 1.1110
Distribution of individual-level duration of U@Uni behavioural effect (years)**
Mean duration Beta α = 1.8179 Expert elicitation
β = 0.1304
Scale = 4.5000
Standard deviation of duration Beta α = 2.9109 Expert elicitation
β = 0.2691
Scale = 3.3800
Hazard ratios for effect of health behaviours on mortality risk
Fruit and vegetable consumption Lognormal Mean = 0.0953 Kvaavik et al. (2010) [11]
SD = 0.0673
Alcohol consumption Lognormal Mean = 0.1655 Kvaavik et al. (2010) [11]
SD = 0.0840
Physical activity Lognormal Mean = 0.3577 Kvaavik et al. (2010) [11]
SD = 0.0641
Smoking status Lognormal Mean = 0.3577 Kvaavik et al. (2010) [11]
SD = 0.0873
Utility ordinary least squares regression model coefficients
β0 Constant Multivariate normal (see section F of Additional file 1 for covariance matrix) Mean = 0.9490 Analysis of Health Survey for England 2008 [1]
β1 Age coefficient Mean = -0.0038
β2 Gender (1 = male; 0 = female) coefficient Mean = 0.0142
β3 Fruit and vegetables (portions per day) coefficient Mean = 0.0207
β4 Alcohol (units per week) coefficient Mean = 0.0016
β5 Smoke (smoker = 1; non-smoker = 2) coefficient Mean = -0.0541
β6 Physical activity (minutes per week) coefficient Mean = 0.0002
β7 Age2 Mean = -4.31 × 10-06
β8 Fruit and vegetables2 Mean = -0.0033
β9 Fruit and vegetables3 Mean = 0.0001
β10 Alcohol2 Mean = -2.77 × 10-05
β11 Alcohol3 Mean = 6.45 × 10-08
β12 Physical activity2 Mean = -2.59 × 10-07
β13 Physical activity3 Mean = 4.88 × 10-11
β14 Age*Fruit and vegetables β0 interaction Mean = 4.94 × 10-05
β15 Age*Alcohol interaction Mean = 1.61 × 10-05
β16 Age*Physical activity Mean = 2.47 × 10-06
  1. *SD standard deviation.
  2. **The sampled mean and standard deviation from the beta distributions are then converted to log (mean) and log (standard deviation) and used as parameters for the lognormal distribution for individual-level durations of response.