Transcript HIV and Non-communicable Diseases Pre-Conference, July 15

HIV and NCDs: models of chronic care
delivery in Africa
Shabbar Jaffar
London School of Hygiene and
Tropical Medicine
1
This talk
• The problem and why AIDS and NCDs together
• What’s the evidence on delivery of chronic care in
Africa?
• How can we integrate NCD and HIV care?
• Where are the research gaps?
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
2
The Problem
• NCDs are rising rapidly and affecting young adults.
Miranda, Trop Med Int Health. 2008;13:1225-34; Lopez, Lancet 2006; 367: 1747
Addo,. Hypertension. 2007;50:1012; Mbanya, Lancet 2004;364:900.
• Limited services for the detection or treatment of hypertension,
diabetes and other chronic conditions.
• HIV is the first large chronic care programme in Africa. From a
health service perspective, HIV and NCD control have similarities.
• ART increases CVD risk. Links between TB and smoking, TB and
diabetes. Lifestyle changes greater among patients on ART?
Bates, Arch Intern Med 2007;167: 335; Wen, BMC Infect Dis 2010; 10:156
Jeon, Plos Med 2008; 5: e152
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
3
WE MUST THINK ABOUT AIDS AND NCD
CONTROL TOGETHER
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
4
Why is chronic care such a challenge?
• Very severe shortage of doctors (<1 in 10,000
population in many countries).
• HIV care models are hospital, doctor and nurse time
intensive. Little research to inform delivery of care.
• Delivering NCD care will be particularly challenging:
– Knowledge of hypertension, diabetes etc is limited among patients
– Conditions are generally silent. Patients should start treatment when
healthy
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
5
10km
Karonga District, Malawi
120km long. 240,000 population
and 1 government doctor
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
8
Jinja District, Uganda. Population ~0.5m and 8 doctors
in the district + an additional 6-8 at the regional hospital
9
Experiences from ART programmes.
•
In rural /peri-urban areas Uganda, 25% of people eligible for ART either die or
drop out before treatment can be started (primary reason – can’t afford transport)
Amuron, BMC Public Health. 2009;9:290
•
Median CD4 count at ART initiation ~ 130 /µl. Getting people into care earlier has
proved difficult even with strong community relations.
•
Patients are often helped financially by relatives during the early months.
•
Survival and retention are poor in the first 6-12 months after starting ART.
Fox, TMIH 2010; 15 suppl 1:1
Lawn, AIDS 2008;22:1897
•
Sustained care is a challenge:
– In Uganda, each clinic visit in costs >10% of a man’s and 20% of woman’s monthly salary.
– Patients want to “normalise”, work, have relationships and lose the HIV badge
Jaffar, Lancet 2009; 374: 280
Allen, Cult Health Sex 2011; 13: 529
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
10
13
Models of care for NCDs
• Little evidence from Africa. The few studies show poor
retention even in trial settings.
Labhardt TMIH 2011 (epub)
Mendis, Bull WHO 2010; 88: 412
• Feasibility study done in Cambodia with HIV/AIDS, diabetes
and hypertension services offered from doctor-run clinics.
Retention of diabetics< 75% at 12 months
Janssens, Bull WHO 2007; 85:880 -885
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
14
Models of HIV care
• Vary considerably between settings:
– Hospital based, doctor and nurse time intensive: common
– Nurse-led management from primary care centres:
becoming more common:
– Home care using lay-workers: still rare
• Policies are not driven by evidence
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
15
HIV delivery models – the evidence
•
Nurses versus doctor management in SA:
–
Similar clinical outcomes
–
Only 800 patients followed for 2y.
Sanne, Lancet 2010; 376: 33-40
•
Home care vs clinic care in Uganda:
–
–
Similar survival. Home care slightly cheaper for health service and hugely cost saving for patients.
Only 1453 patients studied for 2.5y.
Jaffar, Lancet 2009; 374: 280-89
•
Mobile phones and adherence in Kenya:
– Text messaging led to better adherence and virological suppression.
– Only 538 patients followed for 12m.
Lester, Lancet 2010; 376: 1838-45
+
15 other small studies with weak endpoints
Lazarus, GFATM; personal communication
+
evidence from observational studies on task shifting, mostly doctors to nurses.
Callaghan Human Resour Health 2010, 8:8
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
16
17
MUCH MORE RESEARCH IS NEEDED ON HOW TO
DELIVER AND ORGANISE CARE
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
18
Minimal model: clinic-based NCD and ART services
•
Integrate detection and treatment for hypertension, diabetes
and other NCD related conditions into hospital-run ART
programmes.
•
Provide services from the same clinic.
–
–
–
•
Adherence counselling, drug procurement etc. common to both.
Easier for patients with both conditions
May lower stigma
But will add pressure to severely limited human resources.
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
19
Nurse managed HIV and NCD care
• Task shifting from doctors to nurses is now common but nature of tasks
shifted varies considerably. Need a more standardised model.
Callaghan Human Resour Health 2010, 8:8
• Nurse-managed HIV and NCD care:
– Integrated services for ART and NCDs delivered from primary care.
– Drug initiation, management, monitoring done by nurses with referral to doctors as
necessary.
– Use of trained lay workers supporting nursing staff.
– Targeted intensive support for some patients (e.g. Home care for patients presenting
with very advanced disease or those with multiple conditions).
– More support at the beginning of treatment initiation and minimal support for patients
stable on therapy.
• Policy to integrate NCD care into such models will be slow unless rigorous
evaluation is done.
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
20
Evaluating nurse-led integrated care
• Roll-out into clinics sequentially over a number
of time periods. By the end, all participants
receive the intervention.
• Randomise the sequence of introduction
• Called a “step-wedge” design
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
21
Example
Brown, BMC methodology 2006; 6:54
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
22
Step-wedge designs
• Ideal for short-term endpoints: e.g. retention,
mortality.
• Has many uses – for example evaluate:
– E.g. ART at 350 compared to previous guidelines
– geneXpert for TB diagnosis
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
23
Home care for HIV/NCDs using lay workers
• Particularly suited to rural settings where access is difficult.
• Lay-workers lead to better outcomes in some service delivery and in some
settings; but evidence from Africa is limited.
Lewin, Cochrane Database Syst Rev. 2010:CD004015.
• In Uganda, HIV home care was delivered by lay workers who :
– were paid, trained and supported
– were responsible for drug delivery, adherence support and clinical monitoring
using checklists and mobile phones for discussions with clinic staff
– gave patients more time and built a stronger bond than that possible in clinic
settings.
• Rates of death, adherence etc were similar in the home and clinic care
models; but home care was cheaper for the health service and
considerably cheaper for patients.
Jaffar, Lancet 2009; 374: 280-89
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
24
25
26
Lay-workers and chronic care?
• Can lay-workers provide integrated HIV and NCD care in the home?
• Services which could be provided:
• Drug delivery
• Adherence support
• Clinical monitoring using a checklist. Limited laboratory sample
collection
• Information on diet and lifestyle
• Which type of lay-worker is best (e.g. can one lay worker do all?).
How long do we need to provide home care for? What are the costs
and benefits of lay-worker care?
• Needs evaluation of costs and benefits on a large-scale to shift
policy, probably using cluster-randomised trials.
Amuron Open AIDS J. 2007; 1: 21-7
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
27
Expanding the lay-worker model
• In Uganda, home-based HIV VCT offered to households of people starting
ART had much higher uptake than clinic models. Other community-based
HIV VCT and TB testing approaches have had good uptake.
Lugada et al JAIDS 2010; 55: 245
Corbett, Lancet 2010; 376: 1244
Sweat, Lancet Infect Dis 2011; 11: 525
• Could we visit homes of patients starting ART – index patients - and offer
testing to family members for blood pressure, diabetes testing, HIV, TB
etc?
• “Family care” might lead to better adherence and retention; but needs
evaluation, probably using cluster-randomised trials.
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
28
Chronic care for people without HIVinfection?
• Likely to be the big challenge
• Is community-based testing for multiple conditions feasible?
(e.g. door to door methods of detection versus mobile van at
different locations)
• Integrate blood pressure testing in HIV VCT activities?
• Likely to need community-based models of care with greater
roles for lay-workers than in HIV care.
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
29
Other important questions
• Research on health care workers: costs and benefits of
packages aimed at increasing retention of health care workers
and quality of service provision.
– E.g. audit and constructive feedback.
Haines, Bull WHO 2004; 82: 724-31
• Effectiveness of health information provided to communities
on diet and lifestyle? Information could be provided by layworkers, local radio, and other means.
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
30
Evaluation methods
• Research needs to be kept simple with minimal impact on
service delivery.
• Needs large studies, ideally in 2-3 sites, to aid generalisability.
• Research needs to be integrated into health services and
done in close to real-life conditions.
Jaffar, TMIH 2008; 13: 795
• Needs a genuine partnership between policy makers,
programme mangers, researchers and with patient groups.
• Needs strong process evaluation and trial sites which act as a
place of learning for other parts of the country.
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
31
Big barriers
• Researchers, review bodies, funders think
about single diseases and single problems.
• Many HIV and NCD programme managers
don’t communicate.
• High impact research will drive the change
HIV and Non-communicable Diseases Pre-Conference,
July 15-16, 2011, Rome
32