Table 2.

CVD microsimulation model parameters ^a

Parameter	Value	Reference
CVD natural history b
Probability of first CHD event type
Cardiac arrest	10%
MI	32% (males)/ 20% (females)	[23]
Angina	1 ‐ probability of other events	[24]
Acute (one‐year) mortality following a CVD event
Cardiac arrest	0.95	[25]
MI	0.05	[26]
Angina	0.045	[27]
Stroke	0.38	[28]
Annual mortality (post‐event)
MI	0.04	[29]
Angina	0.03	[29]
Stroke	0.05	[29]
Annual risk of new event in post‐event states
Repeated MI post‐MI	0.064	[30]
MI post‐angina	0.035	[31]
Repeated stroke post‐stroke	0.04	[32]
Intervention characteristics c
Annual screening coverage	20% of population	Assumption
Screening success	90%	[16]
NCD treatment uptake & long‐term medication management	50%	Assumption
NCD treatment effectiveness c , d
% reduction in CHD events with hypertension treatment among people		[18]
‐ With diabetes	88%
‐ Without diabetes	77%
% reduction in stroke events with hypertension treatment among people		[18]
‐ With diabetes	74%
‐ Without diabetes	74%
% reduction in CHD and stroke events with diabetes treatment	79%	[19]
Costs (2018 USD) c
Acute care for cardiac arrest	1,049.14	[33]
Acute care for MI	2,041.02	[33]
Acute care for angina	1,264.90	[33]
Acute care for stroke	1,916.26	[33]
Non‐acute care post‐CHD	331.15 per year	[34]
Non‐acute care post‐stroke	993.44 per year	[34]
Screening for HIV	21.50	[7]
Screening for hypertension and diabetes	1.22	[7]
HIV treatment (ART)	297.51 per year	[35]
Hypertension treatment	77.65 per year	[33]
Diabetes treatment	186.73 per year	[33]
Disability weights c
Angina	0.08	[36]
Cardiac arrest	0.08
MI	0.08 (first year) 0.072 (subsequent years)
Stroke	0.152
HIV	0.078 (on ART) 0.274 (off ART)

^aNon‐communicable diseases (NCD); Cardiovascular diseases (CVDs); Cardiovascular heart disease (CHD); Myocardial infarction (MI).

^bThe underlying structure of CVD natural history model and selected parameter values are based on Subramanian et al. (2019) [11].

^cParameters are varied within +/‐ 15% of their original values in sensitivity analysis.

^dNCD treatment effectiveness is estimated separately for hypertension and diabetes treatment, and the impact of combined treatments is modelled as independent and multiplicative (Data S1).