Quantifying Tuberculosis y gBurden and Underrepresentation

in Malaysia 1990 2014in Malaysia, 1990-2014

Nurhuda Ismail, Awang M Bulgiba, Sanjay Rampal, Nicolaas JD Nagelkerke, Jiloris F. Dony, Omar AwangNicolaas JD Nagelkerke, Jiloris F. Dony, Omar Awang

22 October 201522 October 2015

Contents

1. Introduction

2. Methods

3 Results3. Results

4 Discussion4. Discussion

2

Question: 

Does TB incidence data in Malaysia reflectDoes TB incidence data in Malaysia reflect true burden?

125

103

17

Prevalence, incidence, and mortality rates per 100 000 population in countries and areas with an intermediate burden of TB in the Western Pacific Region 2006

4

Source: WHO Tuberculosis Control WPR Report, 2008

COMMUNICABLE DISEASES INCIDENCE RATE MORTALITY RATEAIDS 6.91 1.43HIV (all forms) 21.88 3.66

Incidence Rate and Mortality Rate of ( )

Chancroid 0.00 0.00Cholera 0.89 0.01Dengue fever 64.37 0.01Dengue Haemorrhagic fever 4.10 0.25

Mortality Rate of Communicable Diseases per 100 000 Population in Malaysia, 2006

Diphteria 0.00 0.00Dysentry 0.39 0.01Food Poisoning 26.04 0.01Gonoccocal Infection 1.90 0.00

y ,

•16,665 new TB cases Leprosy 0.89 0.00Malaria 19.87 0.08Measles 2.27 0.00Plague 0.00 0.00

,(mean ~ 46 cases /day)

• 1460 TB deathPolio Myelitis 0.00 0.00Rabies 0.00 0.00Relapsing Fever 0.00 0.00Syphillis 3.06 0.01

(mean ~ 4 death/day)

Tetanus Neonatorum 0.04 0.01Tetanus (Adult) 0.06 0.00

Tuberculosis 62.56 5.32Typhoid and Paratyphoid 0 77 0 02Typhoid and Paratyphoid 0.77 0.02Typhus 0.06 0.01Viral Enchepalitis 0.09 0.00Viral Hepatitis 9.37 0.18Whooping cough 0 02 0 00

Source: Ministry of Health, 2008Whooping cough 0.02 0.00Yellow fever 0.00 0.00Hand Foot and Mouth 19.30 0.02Ebola 0.00 0.00

Top 5 Communicable Disease,Top 5 Communicable Disease,Top 5 Communicable Disease, Malaysia (2010- 2013)

Top 5 Communicable Disease, Malaysia (2010- 2013)

Disease2011 2012 2013

IncidenceRate

Mortality Rate

IncidenceRate

Mortality  Rate

IncidenceRate

Mortality  RateRate  Rate   Rate  Rate Rate  Rate

Dengue Fever          63.75 0 72.20 0 143.27 0.21

Tuberculosis                71.35 5.68 77.41 4.82 81.0 5.4Hand, Food & Mouth Diseases 24.17 0 24.17 0 78.52 0.0Mouth Diseases        Food Poisoning           56.25 0.03 56.25 0.03 47.79 0.04

Malaria  18.32 0.06 16.11 0.05 1.30 0.01

Source: Ministry of Health, 2014Note:*Incidence Rate was per 100,000 population except for Malaria per 10,000 population*Mortality Rate was per 100 000 population except for Malaria per 10 000 population

6

Mortality Rate was per 100,000 population except for Malaria per 10,000 population

BCG vaccination coverage: 98% DOTS coverage: 98% XDR/MDRTB cases: 55-135 Success treatment rate: 85%

8182

82

86

77

81

74

78

82

opul

atio

n

6867 67

65 66 65 64

68

71

66

70

74

100,

000

po

65

61 6160

63 63

6059 60

6264

6562

59

64

60 6163 62 63 64

58

62

66

Rat

e pe

r

54

58

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

000

001

002

003

004

005

006

007

008

009

010

2011 012

013

014

Incidence rate of TB in Malaysia 1985 - 2014

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Year

Source: TBIS, Ministry of Health, 2015

TB Projection in Malaysia 1990-203035000

25000

30000

es

Projection of TB cases

Retrospective projection of TB cases

20000

f TB

cas

e Retrospective projection of TB cases

15000

umbe

r of

5000

10000Nu

0

5000

Time/year Source: Ismail N et al. 2012

Contents

1. Introduction

2. Methods

3 Results3. Results

4 Discussion4. Discussion

11

Study Profile

• Study design/approach: infectious disease mathematical modelling

• Study and target population: Malaysian population

• Main Data Sources:– Tuberculosis Information System (TBIS), Ministry of Health, Malaysia

(data from 1990-2010)– International Union Against Tuberculosis & Lung Disease (IUATLD)– WHO and World Bank

• Model system: deterministic, compartmental, SEIRode syste dete st c, co pa t e ta , S

• Numerical analysis system: ODE, 4th order Runge Kutta (RK4) iterative method

• Modelling Software: ModelMaker 4, Berkeley Madonna

12

Refinement of ModelRefinement of Model

Homogenous model

Homogenous modelmodelmodel

0 - 14 15 - 29 30 - 44˄ ˅

45 - 59 60 - 74 75 - 89

90+

Gender-structured model

Age-structured model

Age-structured model

90+

13

Contents

1. Introduction

2. Methods

3 Results3. Results

4 Discussion4. Discussion

14

Transmission of TB and Interventions

Non-infectedNon-infected •Early diagnosisNon-infectedPersons

Non-infectedPersons Infectious

TBC

Infectious TB

C

y g•Current treatment regimes

CasesCases

Infected Persons(LTBI)

Infected Persons(LTBI)

Hi h Ri kHigh RiskInfected Persons

•Identify with TST/IGRA•Intervention (IPT)•Identify with TST/IGRA•Intervention (IPT)

The Tuberculosis Transmission Dynamic Model for Malaysia

S R٨ µ

δ2kxS Rµ

Lλ rp

LE αdpµ

Id

(1 ‐ α)kt

LLdn

δ1kxµ

µ + µt

DIFFERENCE EQUATIONS DIFFERENTIAL EQUATIONS

St+1 = ٨tN + (1 - βIt - µt)St

LEt+1 = βItSt + δ2kxtRt+ (1 - µt - αdpt – ktt + αktt)LEt

LLt+1 = (1 ‐ α)kttLEt + (1 - δ1kxt - µt - dnt)LLt

It+1 = αdptLEt + δ1kxtLLt + dntLLt + ptRt + (1 - µtbt - µt - rt)It

Rt+1 = rtIt + (1 - pt - µt - δ2kxt)Rt

Various states of population in tuberculosis transmission dynamic of the model

State Descriptions Assumptions

S Number of susceptible individuals

S Number of susceptible individuals

LE

Number of early latent tuberculosis infected individuals

History of contacts within the first 5 years of exposure regardless of prior exposure status. Includes exogenous reinfection i e those reinfected following successfulreinfection i.e. those reinfected following successful treatment (recurrent exposure)

LL Number of late latent tuberculosis infected individuals

History of contacts more than 5 years of exposure individuals

I

Number of infectious tuberculosis cases

Equates annual incidence of tuberculosis at t time

R Number of those recovered from

tuberculosis following successful treatment

*N Total population N equals to total Malaysian population i.e. aggregation of all other states at time ‘t’, N(t) = S(t) + LE(t) + LL(t) + ( ) ( ) ( ) ( )I(t) + R(t)

 

18

Parameter Descriptions & Assumptions

Value/Unit

Reference

List of parameters used in the model (1)

٨

Recruitment rate. Equates only birth rate. BCG immunization does not play any role in this type of tuberculosis transmission dynamic

0.03/year

Malaysian Department of Statistic 2009; Malaysian

MOH, 2009

µ

Mortality rate due to other causes. The annual crude death rate minus µt

0.015/year

Malaysian Department of Statistic 2009; Malaysian

MOH, 2009

µt Mortality rate due to infectious tuberculosis. 0.1/year Malaysian MOH, 2009µt

yTakes into account the national TB death notification rate and does not include those with successful treatment

y

y ,

λ Force of infection, taking into account frequency dependent mechanism or mass

derivatives according to age

frequency dependent mechanism or mass

action with homogenous mixing: λ =

according to age and gender

Probability of effective tuberculosis transmission

0.35

Dye et al., 1998

c

Per capita contact rate

10/year

Malaysian MOH, 2009

α

Probability of those who develop primary infection within 5 years of exposure

0.07 Dye et al., 1998

kt

Rate of progression of those who have been infected yet to develop the disease into late latent group

0.2/year

Dye et al., 1998

kx Rate of exogenous reinfection. Exogenous reinfection rate k defined as k = aλ where

derivatives according to age

19

reinfection rate kx defined as kx = aλ where a is the co-efficient of exogenous reinfection and similar infection and reinfection rate, a = 1

according to age and gender

 

D i ti & A ti

V l /U it

R f

List of parameters used in the model (2)Parameter Descriptions & Assumptions Value/Unit Reference

δ1kx Rate of exogenous reinfection where

individuals at the state of late latent infection where δ1 is the coefficient of the first pathway of exogenous reinfection and δ k = δ k

0.75/year

Styblo K., 1986; Comstock G.W., 1982;

Verver S., 2005

δ1kx = δ2kx

δ2kx Rate of recurrent infection of tuberculosis following successful treatment (recurrent exposure). where δ2 is the coefficient of the second pathway of exogenous reinfection and δ1kx = δ2kx

0.75/year Styblo K., 1986; Comstock G.W., 1982;

Verver S., 2005

1 x 2 x

dp

Rate of developing infectious tuberculosis from early latent state (primary infection)

0.25/year

Vynnycky & Fine, 1997

dn Rate of developing infectious tuberculosis from late latent group via reactivation

0.00256/year Vynnycky & Fine, 1998

(endogenous route)

r

Rate of effective treatment of those with infectious tuberculosis

0.8/year Malaysian MOH, 2007

p Rate of relapse for tuberculosis. Relapse b k t i f ti t t dl f

0.025/year Malaysian MOH, 2007back to infectious state regardless of bacteriological status

Θ

Rate of effective IPT for early LTBI. Individuals in the early latent (LE) who do not progress into infectious state due to effective treatment of Isoniazid preventive

0.03/year Model output

effective treatment of Isoniazid preventive therapy

π Rate of effective IPT for late LTBI. Individuals in late latent (LL) who do not progress into infectious state due to effective treatment of Isoniazid preventive

0.007/year Model output

20

ptherapy

 

The optimized and best-fitted model showing predicted tuberculosis cases against observedtuberculosis cases in Malaysia from 1990 till 2010 indicated as blue line (Inftotal) and red dots(C N tifi ti ) ti l Th l d b d j ti f th d l l

21

(Case Notifications) respectively. The lower and upper border projections of the model are alsoshown as pink dot-dash line (minimum) and green dash line (maximum) respectively. Theoptimization statistics derived using Marquardt method yield r2 = 0.9315, p<0.0001.

TB Projection in Malaysia 1990-203035000

25000

30000

es

Projection of TB cases

Retrospective projection of TB cases

20000

f TB

cas

e Retrospective projection of TB cases

15000

umbe

r of

5000

10000Nu

0

5000

Time/year Source: Ismail N et al. 2012

TB Projection in Malaysia 1990-203035000

25000

30000

es

Observed number of TB casesProjection of TB casesR i j i f TB

20000

f TB

cas

e Retrospective projection of TB cases

15000

umbe

r of

5000

10000Nu

0

5000

Time/year Source: Ismail N et al. 2012

Inftotal Case Notifications

Estimates and Projection based on TB notification data in Malaysia from 1990‐2011

30000

35000

25000

ases

15000

20000

mbe

r of T

B c

10000

Num

0

5000

Year

Using data between 1990‐2010, the annual mean difference or underrepresentation is 13.11% (95%CI: 10.40;16.58)

Estimates and Projection based on TB notification data in Malaysia from 1990‐2014

Inftotal Case Notifications

30000

35000

25000

30000

20000

15000

5000

10000

0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

Using predicted data between 1990‐2010 and notification data between 1990‐2014, the annual mean difference or underrepresentation is 13.49% (95%CI: 

10 39 15 84)

1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

TB Underrepresentation  in Malaysia, 1990‐2014

30

20

25

ntation%

15

rrerep

resen

5

10

TB und

er

0

990

991

992

993

994

995

996

997

998

999

000

001

002

003

004

005

006

007

008

009

010

011

012

013

014

19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20

Year

Di ib i f l diff d i % b 1990 2014Distribution of annual mean difference  or underrepresentation % between 1990‐2014. The mean difference or underrepresentation range is between 13.11%‐13.49%

Contents

1. Introduction

2. Methods

3 Results3. Results

4 Discussion4. Discussion

27

• Underrepresentation is common

• Quantification provides evidence

• Knowledge for• Knowledge for– Evidence‐based planning and policy– Targeted/selective programme– Problem solvingg

Thank you for your attentionOther research supporting members:Dr Mohamed Paid Yusof, Disease Control Division, MOHDr Fuad Hashim, Institute of Public Health, MOHDr Fuad Hashim, Institute of Public Health, MOHDr Mariam bt Mohamad, UiTM Selayang CampusDr Badrul Hisham Abd Samad, Johor State Health Dept, MOHDr Mokhtar bin Pungut, Segamat District Health Dept, MOHg , g p ,Dr Mohd. Yusof bin Hashim, Seberang Prai District Health DeptDato’ Dr Haji Abdul Razak Muttalif, Inst. of Resp MedicineDatuk Dr Hjh Aziah bt Ahmad Mahayiddin, Inst. Resp MedicineAssoc. Prof. Dr Pang Yong Kek, UMMCDr Karina Razali, HART Consultancy

Acknowledgement:Director General of Health, Ministry of HealthDisease Control Division Ministry of HealthDisease Control Division, Ministry of HealthInstitute of Respiratory Medicine, Ministry of Health

The Tuberculosis Transmission Dynamic Model for Malaysia

S R٨ µ

δ2kxS Rµ

Lλ rp

θ LE αdpµ

θ

Id

(1 ‐ α)kt

LLdn

δ1kxµ

ϕ

µ + µt

• Framework (scope, epidemiology, heterogeneity)• Data collection and analysis of parameters (year 1990-2010)

Model development

• Model baseline output• Model validation (optimization, fitting & comparability)

• Model used as a tool• Estimate and projection of tuberculosis burden in Malaysia 1990-2030

Model application

Estimate and projection of tuberculosis burden in Malaysia 1990 2030• Effectiveness of potential intervention strategies• Quantification of effectiveness of IPT (coverage) as a combined strategy• Selection of LTBI for optimal coverage• Overall potential impact of IPT combined strategy

Formulation of policy

• Translation of model outputs to ‘reality’• Recommendation of evidence for implementation• Integration into current management protocol & health system• Return on investment (ROI)policy

recommend-ation

• Informed decision making

31

TB SituationTB SituationTB Burden Ranking TB among HCW

Disease

Number of casesIncidence rate (IR)

2003 2004 2005 2006 2007 2007Dengue Infection

14,76161 65

12.75551 50

15,86264 53

17,14768 47

23,31090 65

46,517177 92Infection 61.65 51.50 64.53 68.47 90.65 177.92

TB 15,58263.29

15,42960.30

15,99161.20

16,66562.56

16,91862.26

17,50663.10

FP 6,62426.45

5,95723.40

4,64117.76

6,93826.04

14,45553.19

17,32262.47

HFMD 1,265 378 6,325 5,141 12,558 15,564, , , 12,55846.21

15,56456.13

HIV 6,75631.27

6,42729.61

6,12028.09

5,83028.79

4,57721.01

4,57716.71

TB among Immigrant TB‐HIV Co‐infection

12

14

Peratus kes TB di kalangan wagra asing, 2003 ‐ 2008

TB among Immigrant TB‐HIV Co‐infection

4

6

8

10

%

32

0

2

2003 2004 2005 2006 2007 2008

TB SituationTB SituationTB in Malaysia vs ASEAN Country TB Cases and IR 1985‐2008

Disease

Number of casesIncidence rate (IR)

2003 2004 2005 2006 2007 2007Dengue Infection

14,76161 65

12.75551 50

15,86264 53

17,14768 47

23,31090 65

46,517177 92

Country No. of cases

IR EstimatedNo. of cases

EstimatedIR

%case

detectionBrunei 207 53 230 59 90

S 64

66

68

70

14000160001800020000

0s

TB Cases and Incidence Rate 1985-2008

Infection 61.65 51.50 64.53 68.47 90.65 177.92TB 15,582

63.2915,42960.30

15,99161.20

16,66562.56

16,91862.26

17,50663.10

FP 6,62426.45

5,95723.40

4,64117.76

6,93826.04

14,45553.19

17,32262.47

HFMD 1,265 378 6,325 5,141 12,558 15,564

Singapore 1359 31 1176 27 115.5

Malaysia 16129 61 27439 103 58.8

Thailand 54793 86 90878 142 60.3 52

54

56

58

60

62

64

02000400060008000

1000012000

IR/1

00,0

00

No.

of c

ases

,46.21

,56.13

HIV 6,75631.27

6,42729.61

6,12028.09

5,83028.79

4,57721.01

4,57716.71

TB among Children Death Associated with TB

Indonesia 275193 119 528063 228 52.1 Year

TB among Children Death Associated with TB

45

ge 3456

r 100

,000

la

tion

0123

2000 2001 2002 2003 2004 2005 2006 2007 2008

Perc

enta

g

012

2000 2001 2002 2003 2004 2005 2006 2007 2008

Year

Rat

e pe

rpo

pu

Year Death Rate (per 100,000 population) Target: <3 per 100,000 population

TBHIV in TB : 1990‐2012

12,025000

8,0

10,020000

6,0

10000

15000

% T

BH

IV in

TB

No

of C

ases

2,0

4,0

5000

%

0,00

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

YearNo.TB cases Number Of New TB Cases With HIV Positive %TBHIV inTB

The optimized and best-fitted model for predicting tuberculosis in Malaysia.

35

The optimization statistics derived using Marquardt method yield r2 = 0.9315, p<0.0001.