Table 1.
Key model variables.
| Parameters | Mean Values and 95% UI | Data Source and Assumptions |
|---|---|---|
| Intervention effect estimate | ||
| Mean minutes per day watching TV, by age and SEIFA IRSD quintile | See Supplementary Information 3 | Sampled from a normal distribution, from Government sources [44]. Adjusted for time spent using TV screens for other uses [50]. |
| Number of advertisements per hour for HFSS foods during children’s peak viewing times | 3.4 (95% UI 1.9–5.2) | Sampled from a pert distribution, minimum 1.5 maximum 6.5 from a 2012 Australian review of outcomes for studies that reported non-core TV advertising during children’s peak viewing times (based on television audience patterns, generally weekday evenings and weekend mornings) [15]. Most likely 3.1 taken from Australian study 2017 [13]. |
| TV advertisement length (seconds) | 29.9 (95% UI 19.2–40.9) | Sampled from pert distribution, minimum 15, most likely 30, maximum 45. Based on logical reasoning and published estimates [20]. |
| Reduction factor for application of experimental effect to real-world setting | 0.50 (95% UI 0.16–0.85) | Sampled from a pert distribution, minimum 0.00, most likely 0.50, maximum 1.00. Based on assumption. |
| Mealtime compensation effect for snacking | 0.37 (95% UI 0.22–0.61) | Sampled from a pert distribution, minimum 0.20, most likely 0.30, maximum 0.80 compensation index [35]. |
| Kcal effect per minute of TV ad exposure per day | 38 (95% UI 15.5–60.6) | Sampled from a normal distribution (mean 37.94, 95% UI 15.6–60.3), see Supplementary Information 2. After base-case reduction factor for application of experimental effect to real-world setting and mealtime compensation are applied, the kcal effect per minute of TV ad exposure per day is estimated as 12 (95% UI 3–27). |
| Intervention cost estimate | ||
| Cost of legislation (including RIS process) | AUD1,089,650 (95% UI AUD940,351–1,240,624) | Sampled from a gamma distribution [37]. |
| Weekly wage of personnel for legislation administration | AUD1242 (95% UI AUD1127–1358) | Sampled from a gamma distribution (mean 1240.90, se 58.90) Administrative and Support Services, fulltime adult [38]. |
| Labour on-costs, 14% salary cost | AUD174 (95% UI AUD155–195) | Sampled from a pert distribution (+/−10%), from Government sources [39]. |
| Annual leave loading, 17.5% weekly salary cost, 4 weeks per annum | AUD870 (95% UI AUD773–975) | Sampled from a pert distribution (+/−10%), from Government sources [40]. |
| Sensitivity analysis, worst case analysis | ||
| Assumed loss of network revenue, year one of intervention | 2.5% (95% UI 0.4–5.1) | Sampled from a pert distribution (minimum 0, most likely 2%, maximum 7%), based on 2010 network advertising revenue of AUD3.9B [51,52,53]. |
| Kcal effect per minute of TV ad exposure per day | 27.6 (95% UI 19.3–35.8) | Sampled from a normal distribution (mean 27.6, 95% UI 19.5–35.7), see Supplementary Information 2. |
| Reduction factor for application of experimental effect to real-world setting | 0.67 (95% UI 0.30–0.95) | Sampled from a pert distribution, minimum 0.00, most likely 0.75, maximum 1.00. Based on assumption. |
| Proportion of time spent watching paid or streamed TV services (assumed no advertisements) | 0.22 (95% UI 0.20–0.24) | Sampled from a pert distribution, minimum 0.2, most likely 0.22, maximum 0.24 (+/−10%) from published estimate [54]. |
95% UI = 95% uncertainty interval; ABS = Australian Bureau of Statistics; AUD = Australian dollars; B = billion; BMI = body mass index; Kcal = kilocalories; RIS = regulatory impact statement; se = standard error; SEIFA = Socioeconomic Indexes for Areas Index of Relative Socioeconomic Disadvantage; TV = television.