Table 1. Description of model inputs.

Input parameter(s)	Detail	Data source
Risk factor
Physical activity	Minutes per week of moderate and vigorous physical activity (MVPA-METmins/week), weighted by MET value associated with activity. Heterogeneity by age, sex, and ethnicity.	New Zealand Health Survey 2011/12 Compendium of Physical Activities [17]
Distance travelled	Mode-specific total annual distance travelled (for pedestrians, cyclists, motorcyclists, and motor vehicles). Heterogeneity by age, sex, and ethnicity.	New Zealand Household Travel Survey 2003–2014 [18, 19]
Air pollution	Population-weighted annual fine particulate matter exposure (<2.5μm diameter). No heterogeneity.	Brauer et al [20]
Disease and injury parameters
Disease incidence, prevalence, case-fatality, and mortality rates	Each parameter was first estimated from linked health data, then simultaneously entered into DisMod II (an epidemiological calculator) to ensure coherence. Heterogeneity by age, sex, and ethnicity.	As per Cleghorn et al [21], with model inputs available at [22]
Injury incidence and mortality rates	Derived using GBD data on mode-specific incidence and mortality rates by age and sex, combined with Health Tracker data and NZBDS to estimate rates by ethnicity.	GBD Results Tool [23], NZBDS [24] and Health Tracker
Morbidity rates	In the main lifetable that simulated the QALYs, morbidity for each sex by ethnic by age group in BAU uses the years of life lived with disability (YLD) due to all causes from NZBDS, divided by the number of people in this strata to give a rate. This represents the average ‘background’ morbidity rate experienced. Disease-specific morbidity (or ‘disability’) rates are derived similarly, using disease specific YLDs from the NZBDS. Disease specific morbidity rates reflect the average disability experienced by someone with that specific disease. For example, CHD morbidity rate for 55–64 year old non-Māori males was calculated as follows. NZBDS pYLD estimate for 2006 (1,321) was scaled to account for demographic change to estimate pYLD value for 2011 (1,533). The scaled pYLD estimate was divided by the number of prevalent cases estimated from DisModII (17,326), to give a morbidity rate of 0.088.	GBD [25], NZBDS [24], as per methods described in detail in [21] and [26], with model inputs available at [22]
Healthcare costs (2011 NZ$)	The costs used represent excess annual health system costs for cases in first year of diagnosis, last year of life if dying of that disease, and otherwise prevalent years of diagnosis. Heterogeneity by age and sex, but not ethnicity.	As per Kvizhinadze et al [27], available in [16]

GBD: Global Burden of Disease Study

MET: Metabolic equivalent of task

MVPA: Moderate and vigorous physical activity

NZBDS: New Zealand Burden of Disease Study

YLD: Years lived with disability

See Technical Report [16] for further details on parameters, including uncertainty distributions.