Declines in adult HIV mortality in Botswana, 20035: evidence for an impact of antiretroviral therapy

1
Declines in adult HIV mortality in Botswana,
2003-5 evidence for an impact of antiretroviral
therapy programs

• Rand Stoneburner, Dominic Montagu, Cyril
  Pervilhac,
  
• Boga Fidzani, Wayne Gill, Gail Kennedy,
• Hilary Spindler, George Rutherford.
• HIV/AIDS Programme, World Health Organization,
• National AIDS Co-ordinating Agency Botswana,
• Joint United Nations Programme on HIV/AIDS,
• Institute for Global Health, University of
  California San Francisco
  
• .

2
Background/purpose

• Antiretroviral therapies (ART) reverse
  immunodeficiency and mortality risk among
  patients with severe HIV disease
  
• ART program indicators provide evidence of
  therapeutic benefits in an increasing population
  of severely ill HIV/AIDS patients
  
• However, impacts of ART programs in reducing HIV
  mortality at the population level have yet to be
  documented in Africa and are of strategic
  importance in assessing the ultimate intervention
  effectiveness
• Botswana MOH and National AIDS coordinating
  agency asked the simple question. Is there an
  impact of ART on HIV- mortality at the
  population level and can it be measured using
  available data sources?

3
Aim and approach

• To assess ARV impact on adult mortality
• Hypothesis increasing ARV use among HIV ill will
  reduce mortality
  
• Approach analyze trends in adult mortality and
  ARV uptake over time and place
  
• First, do mortality trends reasonably reflect HIV
  dynamics over time?
  
• Are there declines in suspected HIV deaths?
• Are declines in suspected deaths related to
• The general temporal distribution of ARV uptake
• The population coverage rate of ARV uptake by
  district
  
• The timing or initiation date of ARV coverage by
  district.
  

4
Methods

• Analyses of mortality data from vital
  registration Ministry of Health (MOH)
  
• Reported deaths by age sex and district from
  1991-June 2005. Deaths in 2005 were annualized
  based on deaths occurring through June 2005 and
  reported by November 2005.
• Analyses of ART program monitoring data-- ( MOH)
• Numbers of patients enrolled and receiving ART
  by district site and year from 2002 through June
  2005

5
Methods (cont.)

• Mortality
• Vital registration (1)
• Institutional discharges (dead alive) from
  district hospitals reported to Central
  Statistical Office (routinely available early
  1990s -2003) but 2 year lag in collation and
  public release. This required rapid review of
  deaths reports from 2004 to June 2005
• Non-institutional deaths reported from district
  health adm.
  
• Validation 95 - 98 mortality capture compared
  to census 1991 and 2001
  
• Hospital midnight census (2)
• Nightly record of institutional occupancy,
  admissions discharges (alive/dead), births,
  neonatal deaths ( through 2004)

6
Crude mortality and deaths by age in
institutional settings Botswana 1991-2003

• Crude mortality rates increase 3 fold
• of non-institutional deaths decreased from 26
  to 15
  
• AIDS as cause of death increased from 4 to
  27,1992-2003.
  
• Age-specific institutional deaths increased 8
  fold in ages 25-54 3 fold in 0-4
  
• Trends in institutional deaths mirror trends in
  HIV mortality.

7
ARV programme indicators, 2002 to February 2006.

8
Association of increasing ART uptake and
declines in adult mortality overall and by age
2003-2005

• Coincident with increasing patients on ARV
  between 2002 June 2005
  
• Institutional deaths declined 8 20 between
  2003-4 and 2004-5, respectively.
  
• Declines were greatest in 25-54 year olds
  remained relatively stable in older and younger
  ages.
  
• Similar declines not evident in 0-4 year olds
  between 2003-4 (? 05) .
  
• Recent mortality declines are temporally
  associated with increasing ARV uptake

annualize based on deaths through June, reported
by November 2005.
9
Geographic association between ARV coverage
rates and initiation dates by district, Botswana,
2003-05

• Districts with early initiation and highest ARV
  coverage rates by 2003
  
• Francistown
• Gaborone
• Swerowe
• Maun (Ngami)

10
Association of district level declines in deaths
ages 20-49 between 2003-04 and ARV treatment
rates by July 2003

• Gaborone and Francistown earliest opening dates,
  highest ARV coverage rates and 27 and 17
  mortality declines, respectively.
  
• After weighting for population size declines in
  district level mortality were significantly
  correlated with district
  
• date of initiation of ART coverage (Plevel ART coverage (P

11
Comparing institutional mortality using vital
registration and mortality derived from the
midnight bed census in Gaborone and Francistown
districts,1994-2004

• Trends in mortality through 2004 from the
  midnight bed census data are similar to those
  from vital registration providing independent
  corroboration of observed declines.

12
Conclusions

• Our preliminary analysis provides evidence of any
  early association between ART uptake and declines
  in adult mortality in Botswana.
  
• These findings must be interpreted with caution
  until further verified by updated and validated
  2005 and 2006 mortality statistics and results
  from cohort based studies.
• Alternative hypotheses?
• Natural dynamics, effects of other interventions
  surveillance artifacts or other biases are less
  plausible.
  
• The absence of similar declines in children is
  disturbing and needs to be explained urgently.
  
• Mortality registration systems like that in
  Botswana can be an important adjunct to routine
  ART monitoring systems in informing questions
  related to population impact of ART as well as
  other health interventions.

13
Thank you

• Cyril Pervilhac, HIV/AIDS Programme, World
  Health Organization,
  
• Boga Fidzani, National AIDS Co-ordinating
  Agency Botswana,
  
• Wayne Gill, Joint United Nations Programme on
  HIV/AIDS,
  
• Dominic Montagu
• Gail Kennedy,
• Hilary Spindler,
• George Rutherford
• . Institute for Global Health, University of
  California San Francisco