Initialization

This section provides additional details about hypertension and hypercholesterolemia treatment and how medications are changed and/or increased over time. For blood pressure, this is also referred to as the “SBP ramp”.

Links to documentation for relevant risk pages

Todo

add tobacco risk exposure

Todo

add tobacco risk effect

Todo

add bmi risk effect

Todo

add fpg risk effect

Initialization parameters

Key parameters for initialization
Parameter	Reference	Data Source for Simulation	Notes
Outpatient visit rate	GBD outpatient envelope	outpatient_visits=HealthcareEntity (name=’outpatient_visits’, kind=’healthcare_entity’, gbd_id=me_id(19797), utilization=me_id(19797),)	Outpatient utilization envelope from GBD; will want to update to use NHANES data in future
Follow-up visit rate for cardiometabolic risk management	AHA/ACC recommendations	uniform distribution from 3 to 6 months
SBP measurement error	Br J Gen Pract 2011; DOI: 10.3399/bjgp11X593884	Normal distribution, mean=0, SD=2.9	85% measurements within +/- 3 mm Hg; 15% within +/- 4-9 mm Hg
SBP therapeutic inertia	Hypertension. J Hypertens 39:1238–1245 DOI:10.1097/HJH.0000000000002783; https://doi. org/10.1371/journal.pone.0182807	0.4176	48% uncontrolled htn (NHANES); 87% of the time this is due to therapeutic inertia
SBP prescription initiation rate	Assumption for current run; will reevaluate in future	100 %
SBP adherence rate	Medical Expenditure Panel Survey, 2014	/share/scratch/projects/cvd_gbd/cvd_re/simulation_science/pdc_meps_2014.csv
SBP treatment efficacy	BMJ 2009 May 19;338:b1665. doi: 10.1136/bmj.b1665.	/share/scratch/projects/cvd_gbd/cvd_re/simulation_science/drug_efficacy_sbp.csv
SBP baseline coverage rate for each ramp position	Egan et al. Hypertension. 2012;59:1124- 1131.	/share/scratch/projects/cvd_gbd/cvd_re/simulation_science/tx_percent_initialize.csv
Proportion of Group 2 from SBP ramp algorithm receiving combination therapy	Byrd et al Am Heart J 2011;162:340-6.	45%	Represents non-compliance with guidelines
SBP drug combinations	Medical Expenditure Panel Survey, 2014
LDL-C measurement error	BMJ 2020;368:m149 doi: 10.1136/bmj.m149	normal distribution from 2 to 5%; mean and standard deviation
LDL-C therapeutic inertia	https://pesquisa.bvsalud.org/portal/resource/fr/ibc-171028	0.194
LDL-C prescription initiation rate	Assumption; will revisit later	100%
LDL-C adherence rate	Medical Expenditure Panel Survey
LDL-C treatment efficacy
LDL-C baseline coverage rate
Medication outreach effectiveness on medication adherence	Circulation. 2005;111(10):1298-1304. doi:10.1161/01.CIR.0000157734.97351.B2	OR 2.3 (95% CI 1.39-3.88)
Medication outreach baseline coverage	Assumption	0%
Polypill effectiveness on medication adherence
Polypill baseline coverage rate
Lifestyle Modification Education effectiveness on BMI, FPG, and Tobacco Initiation/Cessation
Lifestyle Modification Education baseline coverage rate

Visit types

Health care visit types

Baseline Coverage Data

Baseline coverage of treatment for elevated SBP and elevated LDL-c is substantial and expected to vary by age, sex, and time. Bask To initialize simulants, the research team has fit a multinomial regression to NHANES data.

\(\ln(\frac{P(tx=SBPonly)}{P(tx=none)}) = b_{10} + b_{11}(SBP_{level}) + b_{12}(LDL_{level}) + b_{13}age_{(yrs)} + b_{14}sex\) \(\ln(\frac{P(tx=LDLonly)}{P(tx=none)}) = b_{20} + b_{21}(SBP_{level}) + b_{22}(LDL_{level}) + b_{23}age_{(yrs)} + b_{24}sex\) \(\ln(\frac{P(tx=Both)}{P(tx=none)}) = b_{30} + b_{31}(SBP_{level}) + b_{32}(LDL_{level}) + b_{33}age_{(yrs)} + b_{34}sex\)

###### Setup ######
rm(list=ls())

suppressMessages(library(data.table))
library(ggplot2)
library(nnet)

###### Files and paths ######
file_path <- "/share/scratch/projects/cvd_gbd/cvd_re/simulation_science/nhanes/"

###### Read in file ######
load(paste0(file_path, "nhanes_microdata.rdata"))

# Recode treatment variables to account for skip pattern
data[,sbptx:=ifelse(highbp==0 & is.na(bpmeds), 0, bpmeds)]
data[,choltx:=ifelse(highchol==0 & is.na(cholmeds), 0, cholmeds)]
data[,tx:=ifelse(sbptx==0 & choltx==0, "none", ifelse(sbptx==1 & choltx==0, "bponly",
                ifelse(sbptx==0 & choltx==1, "cholonly", ifelse(sbptx==1 & choltx==1, "both", NA))))]
data[,tx2:=factor(tx, levels=c("none", "bponly", "cholonly", "both"))]

meds <- multinom(tx2 ~ bpsys + lbdldl + sex_id + age_year, data=data)

# weights:  24 (15 variable)
initial  value 21425.179351
iter  10 value 16793.908492
iter  20 value 14903.770849
final  value 14903.720511
converged

summary(meds)
Call: multinom(formula = tx2 ~ bpsys + lbdldl + sex_id + age_year,
  data = data)

Coefficients:
         (Intercept)        bpsys       lbdldl     sex_id   age_year
bponly     -6.746432  0.024905946 -0.004474287  0.1578084 0.05006270
cholonly   -4.234380 -0.002564668 -0.005063271 -0.1900133 0.06173726
both       -6.262507  0.018470096 -0.013548739  0.1326292 0.06909707

Std. Errors:
         (Intercept)       bpsys       lbdldl     sex_id    age_year
bponly     0.1863489 0.001265926 0.0006439986 0.04686429 0.001632670
cholonly   0.2665387 0.001872484 0.0009045871 0.06485975 0.002270549
both       0.2067298 0.001371421 0.0007557389 0.05139671 0.001875866

Residual Deviance: 29807.44
AIC: 29837.44

[[Should this also predict which simulants are non-adherent to treatment?]]

This initialization scheme will also allow initialization of “untreated LDL-C” and “untreated SBP” attributes, which refer to what a simulants risk exposure would be, if they were not receiving treatment. Individuals who are initialized to be receive treatment will also need to be initialized to have a follow-up visit date somehow.

Baseline coverage of polypill, medication outreach, and lifestyle modification education are all low, and we will assume that they are 0%. (This means that we will can initialize the untreated BMI, FPG, and smoking risk exposures to be equal to the actual BMI, FPG, and smoking exposures.)

Weighted means of treatment (not specific to drug class) by age, sex, and SBP category (in 10 mm Hg groups) are here: /share/scratch/projects/cvd_gbd/cvd_re/simulation_science/nhanes_sbp_tx_info.csv

Baseline coverage data
Location	Subpopulation	Coverage parameter	Value	Note
USA	General Population	Hypertension Treatment	Distribution from NHANES
USA	General Population	Lipid lowering therapy	Distribution from NHANES	empirical calibration needed
USA	General Population	Polypill	0.0%	assumption
USA	General Population	Medication outreach	0.0%	assumption
USA	General Population	Lifestyle modification education	0.0%	assumption