MODEL BASED ECONOMIC EVALUATION - CHEPScheps.or.id/wp-content/uploads/2017/04/Septiara-Putri_disemination... · Model based economic evaluation Economic evaluation has been defined

CENTER FOR HEALTH ECONOMICS AND POLICY

STUDIES (CHEPS)

UNIVERSITAS INDONESIA

2017

MODEL BASED

ECONOMIC EVALUATIONCost utility analysis (CUA) of Peritoneal

Dialysis versus Hemodialysis in End Stage

Renal Disease (ESRD) patients

Outline

Model based economic evaluation: basic concepts

Conceptual model development

Methodology

Result of Cost utility analysis, Incremental cost

effectiveness ratio (ICER)

Probabilistic sensitivity analysis CE Plane, CEAC

Discussion and Conclusion

Model based economic

evaluationEconomic evaluation has been defined as ‘‘the

comparative analysis of alternative courses of action interms of both their costs and their consequences’’

A cost-utility analysis a type of cost-effectiveanalysis that compares different procedures andoutcomes relative to a person's quality of life.

Decision analytical modeling compares the expectedcosts and consequences of decision options bysynthesising information from multiple sources andapplying mathematical techniques, usually withcomputer software.

“Essentially, all models are wrong, but

some are useful (George Box, British

Statistician)”

“A good decision is a logical decision – one

based on uncertainties, values, and

preferences of a decision‐maker.” (Ronald

Howard)

Conceptual model

Decision

problem :

CAPD vs HD

Markov Model

ICER

Sensitivity

Analysis and

Scenarios

Costs (IDR)

Transition probability

QALYs Utility Literature review

Epidemiological data

Discount rate

Model structure

Model design and assumptions

Adult cohort hypothetical, 50 years old patients.

Parameters were gathered from primary costs and utility data,Indonesia Renal Registry (IRR), and literature review.

Model was structured based on research question literaturereview with 2 modalities therapies as Renal ReplacementTherapy (RRT)

Initially, patients could receive CAPD or HD, and then move toanother states, (for instance: CAPD to HD, vice versa). Livingwith health risk alongside CAPD or hemodialysis.

50 years cycle (annual cycle), we performed “half-cyclecorrection”

Model design and assumptions

Transplant has very low probability due to the donor

transplant availability, suitability, data limitation and not as

first therapy for ESRD patients. This model excluded

transplant. The analysis focuses on Dialysis modalities.

Cost Utility Analysis (Adaptation from CEA), we estimated

ICER (Cost/Quality Adjusted Life Years) as final outcome

of this study

Probabilistic sensitivity analysis was applied.

Established several scenarios, deterministically.

Survival Data

-ln(1-tp(t))/interval

Exponential moving average

1-s(u)

ParametersParameter Nilai 95% CI Distribusi Referensi

Transition probabilities HDtoPD 0,111 -0,3-0,25 Beta Korevaar et al.,2003

PD to HD 0,198 0,11-0,28 Beta Jaar et al.,2009

Direct medical cost

CostDMHD 102.929.481 97.277.162-108.581.800 Gamma CHEPS CostDMPD 121.788.452 110.872.425-

132.704.478

Gamma CHEPS

CosComHD 23.252.295 8.890.352- 37.614.238 Gamma HTA Indonesia CostComPD 8.207.800 1.200.408-15.215.192 Gamma HTA Indonesia

Direct non medical cost

CostDnMHD 8.917.656 7.322.763- 10.512.550 Gamma CHEPS

CostDnMPD 4.506.421 1.932.270- 7.080.572 Gamma CHEPS

Indirect cost

CostIDHD 9.291.866 7.182.895-11.400.838 Gamma CHEPS CostIDPD 6.157.997 3.582.264-8.733.731 Gamma CHEPS

Utility

UHD 0,65 0,60-0,71 Beta CHEPS

UPD 0,81 0,73-0,88 Beta CHEPS

Discounting

Dcost 3% WHO Doutcome 3% WHO

Results

CAPD HD Incremental ICER

Cost 857,778,507 865,331,716 (7,553,209) (48,850,332)

QALY 3.66 3.51 0.15

Probabilistic Sensitivity

Analysis (CE plane)

-100000000.00

-80000000.00

-60000000.00

-40000000.00

-20000000.00

0.00

20000000.00

40000000.00

60000000.00

80000000.00

-0.40 -0.35 -0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55

Inc QALY

Inc c

osts

Cost Effectiveness Acceptability Curve

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

CEACThreshold (Rupiah)

Pro

ba

bili

ty to

be

co

st e

ffe

ctive

ScenarioParameters Total Costs

(IDR) QALY ICER

Scenario 1 839,788,559

3.56

(158,799,793)

Failure probability (actual)

Transition probability HD to PD (-5%) Utility PD (-5%)

Direct medical costs PD (-5%)

Cost of complication PD (+10%)

Scenario 2

841,467,003

3.46

(235,418,020)

Failure probability (actual) Transition probability HD to PD (-10%)

Utility PD (-10%)

Direct medical costs PD (-5%)

Cost of complication PD (+15%) Scenario 3

Probabilitas failure (EMA) 1,853,640,411 8.49 (145,726,883)

Transition probability HD to PD (-15%)

Utility PD (-5%)

Direct medical costs PD (-5%)

Cost of complication PD (+15%)

Model limitations Utilization of CAPD probability of switching between

modalities

Cost of complications should be improved, sample

limitation at primary data as well.

Limitation of survival data. Data only in patient who

received HD similar life years gained expert

consultation

Utility values EQ5D-3L, collected as cross

sectional requires more representative data

Playing with scenario considerable parameters

Conclusions Average direct medical costs CAPD is higher compared

to HD

CAPD potentially cost-effective compared to HD

Switching modalities and epidemiological data are

required

Future research needed: budget impact analysis, more

‘local parameters’ for mathematical modeling, sub

group analysis,

THANK YOU

Pusat Kajian Ekonomi Kebijakan dan Kesehatan (PKEKK) FKM UI

Gedung G Lantai 3 Ruang 311, FKM UI- Depok

Email : [email protected]

www.cheps.or.id

021 7875576