The Health Economics and Epidemiology Research Office (HE ... · health economics and epidemiology to help understand the effectiveness, benefits, and costs of interventions. Established

The Health Economics and

Epidemiology Research Office (HE2RO):

An overview of research methods and

results

Gesine Meyer-Rath

Lawrence Long

Kathryn Schnippel

Health Economics and Epidemiology Research Office

Wits Health Consortium University of the Witwatersrand

HE RO2

Introducing HE2RO

HE2RO conducts applied, policy- and program-relevant research

and evaluation on issues of public health of importance in South

Africa. Focusing on HIV and TB, it works at the intersection of

health economics and epidemiology to help understand the

effectiveness, benefits, and costs of interventions.

Established in 2004

Funded by Partnered with

Collaboration between Wits and BU

INROADS

HIV care and treatment:

linkage, retention,

and outcomes

TB/HIV: case-finding,

diagnosis, and

treatment

Service delivery

models and treatment guidelines

Cost and budget

modeling

Sexual and Reproductive

Health

Capacity Development

NACM

NTCM



HE RO2

Epi analyses: Methods

• All epidemiological analyses are based on primary data

• Most analyses are based on data from the Themba Lethu

Clinic cohort (Helen Joseph Hospital)

– currently 38,000 patients on ART

– collected in TherapyEdge real-time database

– part of IeDEA network

• New large cohort planned including 9 clinics across

Gauteng



HE RO2

Epi analyses:

Summary of studies – Treatment outcomes

Impact of

• d4T dosing

• tenofovir

• older age

• down-referral

• switching to second-line

– Loss to initiation and treatment

• Meta-analysis of loss between testing

and treatment

• Update of mortality estimates with vital

registry data

– Eligibility at 200 vs 350 CD4

cells/microl

– 7-year outcomes



HE RO2

Cost analyses: Methods

• All cost analyses are based on primary data

• Resource use from a sample of 150-500 patient files

– Consecutive sample

– All resources used (visits, drugs, lab tests, other diagnostics) collected

– Fixed cost per visit calculated based on number of all visits to clinic

(includes cost of space, equipment, utilities, non-patient facing staff)

• Unit costs for all resources based on financial and HR records

at clinic/ hospital or public records (drugs tenders, NHLS prices)



HE RO2

Cost analyses: Results NB: Please refer to the published papers for full details before using these estimates, on

www.heroza.org

Item Year Sample size Unit cost per

Adult ART (outpatient, op)

- Nurse managed ART 2011 712 $492 pt yr

- Doctor managed ART 2011 2136 $551 pt yr

- Second-line ART 2010 293 $1037 pt yr

Adult ART (inpatient, ip) 2010 448 $131 pt yr

Adult HIV (ip) 2013 3906 $72 (pre-ART), $117 (ART) pt yr

Paediatric ART (op) 2012 288 $678 - 830 (yr 1), $717 - 782 (yr 2) pt yr

Paediatric HIV (ip)

(first year of life only)

2010 516 $346 (early tx), $1237 (deferred),

$2523 (routine)

pt yr

Adult - MDR-TB (ip) 2013 121 $17,164 pt

Diagnostic tests

- Xpert MTB/Rif 2012 $18 (R166) test

- CD4 POC (Pima) 2012 $24 test



HE RO2

Future cost analyses

• HIV:

– NIMART (op)

– adult HIV (ip, by diagnosis)

• TB diagnosis:

– active case finding

– Xpert for diagnosis of extra-pulmonary TB

• TB treatment:

– drug-sensitive TB (ip + op)

– decentralised MDR-TB (op)

• Cervical cancer diagnosis:

– Pap smear/ visual inspection with acetic acid/ HPV DNA



HE RO2

Budget models: Methods

• National ART Cost Model (NACM): ART, PMTCT, HIV

testing; adults and children

• National TB Cost Model (NTCM): TB diagnosis and

treatment (drug sensitive/ resistant TB; PTB, adults only)

• Population models with health-state transition kernel

• Fixed time period (NACM: 2009/10 to 2016/17; NTCM:

2011/12 to 2016/17)

• Allow for change of guidelines and slow or fast scale-up of

new interventions/ coverage

National ART Cost Model (NACM)

• Number of patients initiating ART

from NSP model (Leigh Johnson, UCT)

• Transition probabilites and rates of

mortality, loss to follow-up, and first-

line treatment failure based on 2 large

Johannesburg cohorts:

• Themba Lethu Clinic Cohort

(n= 9,502)

• Harriet Shezi Children’s Clinic

(n= 3,748)

• Transition probabilities and rates

depend on CD4 cell count/ percentage

and on time on treatment

• Cost from bottom-up cost analyses

in the same clinics, regulalry updated

to current tender prices etc

NACM: Example output The cost of the old guidelines vs. 2010 guidelines vs. full WHO guidelines

0

500,000

1,000,000

1,500,000

2,000,000

2,500,000

3,000,000

3,500,000

4,000,000

-

20,000

40,000

60,000

80,000

100,000

120,000

140,000

Old Guidelines New Guidelines Full WHO Guidelines

To

tal p

atie

nts

To

tal c

os

t [m

illi

on

s Z

AR

]

Total cost (Full cost)

Total cost (Reduced cost)

Total patients initiated on ART

-33%

-25%

-16%

With the new drug purchasing mechanism and task-shifting,

the total cost of the ART programme under the New (2010) Guidelines will be 25% less

than under the Old Guidelines, while reaching 15% more patients;

implementing the Full WHO Guidelines will be 16% cheaper than continuing the Old

Guidelines, while reaching 24% more patients.



HE RO2

NACM: Uses and results

• New ART guidelines 2010 and 2011, task-shifting, new drug tenders

• Doubling of conditional grant (CG) budget between 2009/20 and

2011/12

• Aids2031

• Global Fund proposals for round 10

• NSP 2012-2016; PSPs

• CG budget/ provincial business plans 2009/10 to 2012/13

• Currently working on CG budget for 2013/14



HE RO2

NACM: Why use a moving

average cost?

2012/2013 2013/2014 2014/2015 2015/2016 2016/2017

Year starting 1 April of 2012 2013 2014 2015 2016

ART (adults)

Total patients on ART at end of year 2,167,938 2,683,206 2,989,228 3,190,005 3,365,005

Total cost [million 2011 ZAR] 10,600 13,291 14,986 16,192 17,276

Cost per patient 4,889 4,953 5,013 5,076 5,134

ART (children)

Total patients on ART at end of year 122,677 144,625 169,353 197,725 227,472

Total cost [million 2011 ZAR] 847 1,001 1,193 1,411 1,649

Cost per patient 6,902 6,918 7,043 7,135 7,251

ART (all patients)

Total patients on ART at end of year 2,290,615 2,827,832 3,158,581 3,387,731 3,592,477

Total cost [million 2011 ZAR] 11,447 14,291 16,178 17,603 18,925

Cost per patient 4,997 5,054 5,122 5,196 5,268

National TB Cost Model (NTCM)

• Represents diagnostic pathway

from TB suspect to TB

diagnosis/ treatment

• Old algorithm (sputum/

smear)

• Xpert MTB/Rif algorithm

(scaled up by end 2014)

• Population data:

• national-level electronic TB

register (ETR) (n=57,688)

• 10% sample of NHLS TB

database (n=83,977)

• Xpert implementation studies

• Cost data:

• ingredients costing of Xpert

• literature



HE RO2

NTCM: Example output The cost and yield of the old vs. the Xpert diagnostic algorithm

Results by 2014 (full Xpert scale) Smear +

culture

Xpert

scenario

% change

Number of suspects 2.6 million

TB cases diagnosed

(% of suspects)

338,407

(13%)

426,588

(17%) +26%

Resistant TB cases diagnosed

(% of all PTB)

5,032

(1.5%)

12,628

(3.0%) +151%

% of patients diagnosed by second visit 46% 87% +89%

Treatment initiated

(% of diagnosed)

258,770

(66%)

359,271

(85%) +39%

Cost per TB suspect tested $48 $60 +24%

Cost per TB patient initiated on treatment $468 $456 -3%

Total annual cost of first-line TB diagnosis

and treatment $283 million $348 million +23%



HE RO2

NTCM: Uses and results

• Planning of and budgeting

for Xpert MTB/Rif roll-out

by NHLS

• NSP 2012-2016; PSPs

• Analysis of placement of

Xpert in labs vs. clinics

• Analysis of alternative

options for diagnosing TB

in Xpert negative patients



HE RO2

Future model analyses

• HIV:

– Test and Treat/ Treatment as Prevention

– Cost-benefit analysis of national ART programme

– Update NACM

• Add provincial populations from Spectrum (March 2013)

• Extend until 2020/21

• More detail to second line and treatment failure populations

• Scale-up of NIMART

• TB treatment:

– Centralised vs. decentralised MDR-TB



HE RO2

HE2RO resources

• Publications

• Policy briefs

• Model manuals

• NACM/ NTCM training for NDoH/ NT/ SANAC/

anyone else

www.heroza.org



HE RO2

The Health Economics and Epidemiology Research Office (HE ... · health economics and epidemiology to help understand the effectiveness, benefits, and costs of interventions. Established

Documents