Challenges of service integration: the TB model Linda-Gail Bekker The Desmond Tutu HIV Centre, Faculty of Health Sciences, University of Cape Town, South.

Challenges of service integration: the TB modelLinda-Gail Bekker

The Desmond Tutu HIV Centre, Faculty of Health Sciences, University of Cape Town, South Africa

AIDS ConferenceMelbourneJuly 2014

Global TB in 2012

• 33 million HIV infected • 1/3 have TB• 8.6 million new cases of TB globally• 13% co-infected with HIV.• 1.3 million deaths (320 000 deaths in HIV/TB)• SA: >300 cases notified annually (3rd largest number)• 933/100 000 Largest number TB/HIV co-infected

(65%)

TB incidence rates : 2012

FoI 5-10%

FoI <1%

Generalized

Individualized

HIV prevalence/ new TB cases : 2012

Number of HIV-infected persons receiving antiretroviral treatment (ART) and percentage of persons receiving

concomitant tuberculosis treatment in Africa 2002-2007

HIV in a sea of TB/TB in a sea of HIV

• In our TB clinics: reason for HIV testing and diagnosis of HIV– Similar symptoms– Now an indication for referral for ART

• In our HIV Clinics: TB not always symptomatic but must be actively investigated for.– Indication to start ART regardless of CD4.

• In our primary health clinics : both conditions often present simultaneously

Annual TB notifications

0

1000

2000

3000

4000

5000

0-4 5-910-1

415-1

920-2

425-2

930-3

435-3

940-4

445-4

950-5

455-5

960-6

465-6

970-7

475+

Age strata

TB n

otific

ation

s

HIV negative HIV unknown HIV positive

The numbers of tuberculosis notifications, stratified by 5-year age groups and HIV-status

67%

47%

22%

8%

Perc

en

tag

e o

f p

ati

en

ts s

tart

ing

HA

AR

T

TB Burden Prior to Initiation of HAART

A:TB Incidence by CD4 without HAARTB: TB Incidence by CD4 with HAART

A: Cape Town AIDS Cohort B: Cape Town ART Cohort

R2 = 0.9702

0

5

10

15

20

25

30

0 100 200 300 400 500

CD4 cell count

TB

in

cid

en

ce

ra

te (

ca

se

s/1

00

py

s)

R2 = 0.9643

0

5

10

15

20

25

30

0 100 200 300 400 500

CD4 cell count

TB

in

cid

en

ce

ra

te (

ca

se

s/1

00

py

s)

A: Holmes, Wood, Badri, et al JAIDS 2006B: Lawn, Myers, Edwards Bekker, Wood. AIDS 2009

Good to get TB/HIV positive people onto ART

01

020

30

40

50

60

70

80

90

100

Perc

en

tag

e o

f p

ati

en

ts w

ith

CD

4 b

elo

w c

on

tou

r

0 4 8 12 16 20 24 28 32 36 40 44 48

Duration of ART (months)

1000 cells/ul

500 cells/ul

200 cells/ul

TB rate 9.3-16.8

TB rate >4.2-5.5

TB rate = 1.5

Presentation of TB and HIV co-infection:

1. TB diagnosed before starting ART

2. A patient develops TB while on ART

3. A patient who has defaulted ART develops TB

Reason for urgency

• Patients known to be HIV positive who develop TB and are not diagnosed or not treated – morbidity and mortality

• In addition they are a TB risk to others• Patients known to have TB who are diagnosed

HIV+ need ART (recommend: <8weeks)• Delays result in morbidity and mortality

– Number of RCTs – lower CD4 groups

  2009 2010 2011 Total

TB cases (n) 25,841 26,104 25,554 77,49

9

HIV Positive (%) 49.7 50.4 50.9 50.3

HIV Negative (%) 44.9 46.8 47.1 46.3

HIV status unknown (%) 5.4 2.9 2.0 3.4

HIV prevalence in the PHC TB service-CT 2009-2011

0.0

00

.25

0.5

00

.75

1.0

0

Pro

babi

lity

of s

urv

iva

l (%

)

0 50 100 150 200 250Time (days)

No ART On ART at TB diagnosisStarted ART during TB treatment

0.8

50

.90

0.9

51

.00

0 50 100 150 200 250

Survival stratified by ART status for patients with CD4 counts < 350

Time (Days)

Median Time to DeathStarted ART 71days (IQR: 38-119)

On ART at TB diagnosis 60 days (IQR: 26-118)

No ART54 days (IQR: 25-104)

Half of the deaths in patients who do not start ART occur within 8 weeks

Summary: In the PHC TB service in Cape Town

• 50% of adult TB patients are HIV positive

• 82% have CD4 counts below 350

• 91% of deaths in HIV+ve patients occur in patients with CD4 counts below 350

• 32% of patients with CD4 counts <350 did not start ART during TB treatment (2009 – 2011)

• Mortality for patients on ART was 50% less than patients not on ART for CD4 counts <350

• Median time to death for patients not on ART is +/- 8 weeks

• A median delay of over 16 weeks between start of TB treatment and start of ART was noted for patients referred from TB facilities to the Gugulethu ART clinic.

• This was reduced to 41 days if TB diagnosis was made at the ART clinic. 1

• Median delay of 2.66 months (+/- 74 days) between start of TB Rx and start of ART in clinics in CT (Masiphumulele, Gugulethu and in Khayelitsha) between 2002 and 20082

1Lawn et al BMC Infect Dis 2011,

2Lawn et al JAIDS 2011,

Delays occur moving between the facilities

Primary Health care setting

TB clinic HIV CT Referral to ART services

HIV/ART clinic TB ACTIVE CF

Referral to TB services

Missed opportunities: HCT, ACF, IPT, CTXInherent DELAYS and LOSS TO FOLLOW UP

HIV only HIV/TB TB only

ART for life ART for life

TB treatment for 6-9 months

TB treatment for 6-9 months

The primary health care settingAdh

eren

ce s

uppo

rt

Adh

eren

ce s

uppo

rt DOTS support

DOTS support

Overburdened clinics and overburdened patients

By “integrating” services (the 4th I)

• Tackle the 3 Is in HIV+: – Intensified case finding (delay freeTB Rx)

– (INH) Prophylaxis – Infection control

• But other advantages in TB suspects and patients:– Test all for HIV– Offer ART without delays (5th I !!)- ? All CD4s?– Streamline services to provide both medications

hassle free- quality of services (?reduce LTFU).

Variety of Models:• Co-location of services in the same clinic with

referral between services on site.• Co-location of services with shared

management discussions and shared adherence support services.

• Integration of services with HIV/ART services managing TB

• Integration of services with TB services managing HIV/ART

• HIV/ART/TB clinics de novo.

Nurse run ART Nurse run HIV Nurse run TB

Weekly Interdisciplinary

meeting

FIELD SUPPORTERSART

TB

VCT

DOCTOR

Facility basedcounsellors

Home VisitSupportmentorship

Treatment readinessSessions and ad hoccounselling

A model for service integration

DOTS visits: Post integration of services

Jan-06 Feb-06 Mar-06 Apr-06 May-06 Jun-06 Jul-06 Aug-06 Sep-06 Oct-06 Nov-06 Dec-06 Jan-07 Feb-07 Mar-07 Apr-07 May-07 Jun-07 Jul-07 Aug-07 Sep-07 Oct-07 Nov-07 Dec-07 Jan-08 Feb-08 Mar-080

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

Systematic review :2010

• 136 papers describing models of integration• None RCTs and very few with TB or HIV

outcomes • Models based on referral only easiest to

implement – Referrals may fail, communication key and often

poor.• More integration needs more staff buy in and

training.

Grant, et al 2010)

Barriers and enablers to integration:• Service users unconvinced of need for more testing! • Those referred battling to find referral services • Fragmentation of services• Poor communication between services• Data systems inadequate for coordinated care• Infrastructure poor for privacy • Staff not motivated to take on “more”• Supply of drugs and test kits unreliable

• Joint staff training and support• Identifying staff “champion”

• In the Nyanga CHC which has co-located TB and ART clinics1

– 19.7% of ART eligible TB patients did not start ART– Median delay of 51 days from TB Rx start to ART start

• In an integrated ART/TB clinic in Khayelitsha2

– 34/100 TB patients ART eligible patients did not start ART– Median delay of 58 days from TB Rx start to ART start

• Town 2 Clinic after ART was introduced into a TB service3

– Median delay of 75 days between TB and ART initiation post integration (decreased from 147 days)

1Nglazi et al S Afr Med J 2012, 2Pepper et al PLoS One 2011, 3Kershberger et al PloS One 2012

Simply Integrating TB and ART services doesn’t simply solve uptake and delays.

Pilot analysis in 5 clinics in CT vs Standard TB program

Caldwell et al, 2010

Qualitative data

• Preferred by field adherence supporters– Better relationships with patients and better

outcomes

• Preferred by patients– Less transport and “hassle” factors– Better understanding of both diseases

• Preferred by nursing staff.– Less “DOTS” burden and improved outcomes.

TB outcomes in facilities in CT: A: 13 integrated and B: 4 single service facilities. N= 13 542 newly registered patients (66% HIV+).

Kaplan, et al 2014

Conclusions

• There is a burden of co-infection especially in areas where TB and HIV are hyperendemic

• Both diseases require long term adherence to medications that can interact.

• Outcomes are improved when TB/HIV is managed actively: 3 Is – ICF, IPT, IC (and now add IS)

• Delays in treatment of both HIV and TB leads to increased morbidity and mortality.

• Pragmatic to co-locate and move toward integration paying attention to infection control.

• More research required that measures outcomes including TB cure, viral control and any risks (IC).

Thanks

• Sten and Paolo• Richard Kaplan (DTHC)• Judy Caldwell and CoCT• Steve Lawn