Report
Estimations and Projections of HIV Epidemic – Republic of Moldova
Chisinau 2010
EPP2009v data input process – Republic of Moldova
EPP2009v and Spectrum 3.47 are epidemiological software for the estimation of prevalence and incidence trends over time. Taking into account the definition of epidemics provided in the EPP2009v software manual, Republic of Moldova has a Concentrated Epidemic pattern, as all criteria below apply:
· HIV has spread rapidly in at least one defined sub-population, but is not well-established in the general population. Most often more than one subpopulation at higher risk is affected.
· HIV prevalence is consistently over 5% in at least one defined subpopulation. HIV prevalence is below 1% in pregnant women in urban areas.
· This epidemic state usually suggests active networks of risk exist within and between the sub-populations. (UNAIDS Reference group on Estimates, Modeling and Projections, 2009)
First there was a dataset created in EPP, using the recommendations from the manual; Republic of Moldova template , with a population, age 15+, of 2,996,180 in 2009, has been selected for the creation of the working file ,then updated to correspond to 2009 official statistics and the concentrated epidemic pattern was selected. The population size is consistent with official statistics for the right bank of Nistru, population of Transnistria is not included. Although there is a possibility to change the population in EPP2009v, the national technical team has opted for entering the data and producing estimates only for the Right Bank of Dniester river due to data quality and availability issues where the left bank is concerned (for example, there are no estimates of sizes of populations most at risk for the left bank of Dniester river in the Republic of Moldova). Based on the data available, 5 subpopulations were created for mapping the epidemic patterns (Table 1). Data related to HIV prevalence among clients of FSW and migrants, groups that, according to anecdotal evidence, may play an important role in determining the epidemic, are missing:
Table 1 General characteristics of sub-population used for EPP2009v
IDU / MSM / CSW / General Population Male / General Population women15 + Population Size / 25000 / 2200 / 6000 / 1,380,334 / 1,544,557
Proportion of male (0 to 1) / 0.8 / 1 / 0 / 1 / 0
Excess mortality in IDUs %/year / 1.07
Average time spent in group / 10.4 / 13.1 / 6.7 / 50 / 50
The subpopulations are further defined below:
IDU – the risk group of injecting drug users, with an estimated size of 25 000 on the right bank of the Nistru river, Republic of Moldova, with the proportion of male to female of 0.8, an excess mortality in IDUs of 1.07% per year, and the average time spent in group of 10.4 years. Following the recommendation from the EPP2009v manual, the option of risk turnover was selected and the prevalence in this risk group was added to the General Population – Male prevalence (UNAIDS Reference group on Estimates, Modelling and Projections, 2009). In the HIV data page, the routine statistics data on prevalence among IDU registered at the Republican Narcological Dispensary, who are tested as a recommended general practice twice per year, was entered for 1995 – 2009 years (Table 2).
MSM – the risk group of men who have sex with men, with an estimated number of 2 200 on the right bank of the Nistru river, Republic of Moldova, with the proportion of male to female of 1.0, and an average time spent in the risk group of 13.1 years. The option of risk turnover was turned off. In the HIV data page the prevalence among MSM was entered, the sources of data being the Behavioral Surveillance Surveys 2003/2004 and 2007 (Table 3).
CSW – Commercial Sex Workers risk group, with an estimated number of 6 000 on the right bank of the Nistru river, Republic of Moldova, the proportion of male to female of 0.0, and an average time spent in the risk group of 6.7 years. The turnover option was selected and replace prevalence option to the general population – women was set (this indicates that some of the HIV+ individuals in the target population, assumed to come from former at-risk group, may be captured in prevalence estimations for the general population; in the case of Republic of Moldova, some ex-CSWs may have been captured in the antenatal care prevalence data). In the HIV data page, the prevalence among CSW was entered, the sources of data being the Behavioral Surveillance Surveys 2003/2004 and 2007 (Table 4).
General Population – male – the remaining low risk population of 1,380,334 males, from the right bank of the Nistru river, Republic of Moldova, the proportion of male to female being 1, and the average time spent in group -= 50 years (the maximum value for EPP; life expectancy in Moldova being 69 years, the respective average duration in group is largely consistent with the population 15+ EPP is working with).
General Population – female – the remaining low risk population of 1,544,557 women, from the right bank of the Nistru river, Republic of Moldova, the proportion of male to female being 0, and the average time spent in group -= 50 years.
HIV data entered general populations (both male and female) was the prevalence among pregnant women (ANC data of women tested HIV+ during their pregnancy and women that became pregnant knowing their HIV+ status, from the right bank of Nistru river, Republic of Moldova for 2000 – 2009 years) (Table 5).
The ART treatment data was entered for the right bank of Nistru river, Republic of Moldova; in setting targets for 2015, the enrolment in treatment targets for 2010 – 2012, specified in the grants funded by the Global Fund fighting AIDS/Tuberculosis and Malaria, were taken into account. The data on enrollment of MARPs in treatment has limitations, as currently in Moldova records are kept of PLHIV on ART, disaggregated by mode of transmission rather then by belonging to a risk group; hence data on IDU on treatment are available, and with some limitations due to underreporting, data on MSM on treatment are available. The enrollment of CSW has been estimated based on other MARPS, using the following formula: the share of MARPS on treatment out of MARPs that are HIV+. The same proportions of coverage of subpopulations with treatment existent in 2008-2009 have been extrapolated until 2015. (Table 6)
In the “Survey data” page, the information regarding the prevalence among IDUs was entered for 2001, 2007 and 2009 from the Behavioral Surveillance Surveys held in the respective years, to enhance the precision of national projection and to complement the routine statistics which are not fully representative of the whole IDU population. This data was used afterwards for the calibration of prevalence-incidence curves of the given risk group. For the remaining sub-populations, there was no information added in this page, because the prevalence data from BSS surveys were used as baseline HIV prevalence data (Table 7).
In the projection page, the analysis were made for every sub-population; the projections were made using the ”Fit Data” option (as recommended by EAD unit, UNAIDS HQ) and the curves were fitted to the data available.
In the calibration page:
IDU – the trend was calibrated to the results of the surveys in order to be adjusted to the prevalence found in the BSS Surveys (at the moment there are no other estimates of prevalence among IDU made in Republic of Moldova)
MSM – no adjustments were made
CSW – prevalence was calibrated to a value of 10.9% in 2007. BSS 2007 had a larger sample with more data collections sites compared to other surveys and hence more representative nationally.
General Population - male – the trend was scaled down by the factor of 0.7, as recommended by EAD unit, UNAIDS HQ, as ANC data overestimate prevalence along males.
General Population - women – the trend was scaled down by the factor of 0.6 because usually ANC data overestimate the prevalence in the General Population. ANC data is representative of women of reproductive age that had unprotected sex, hence overestimating prevalence in the general population.
After that the projections were made in EPP and Spectrum output file was created in order to import the incidence estimated by EPP2009v into Spectrum.
Spectrum V3 data input process – Republic of Moldova
The default data file was created based on the template for the Republic of Moldova which contains the demographic information for the population of the Republic of Moldova. The population in Spectrum is estimated at 3 921 323; the National Bureau of Statistics of the Republic of Moldova report the total population at 3 572 703 on the right bank of Nistru. The difference in about 350.000 accounts for the popualiton on the left bank of Nistru – Transnistria region. While EPP default population can be amended, since estimate size of subpopulations for the left bank of Nistru is missing, we opted for producing estimates representative only for the righ bank of Nistru in EPP2009v. Because EPP2009v calculates the incidence in proportions, it was assumed that the incidence distribution for the right bank is similar to the incidence distribution for the left bank and was applied for both banks (the demographic pattern in Spectrum) (Stover, 2009).
The steps made for the projections were as follows:
1) In the Epidemiology section, the HIV incidence for 15 – 49 years was imported from EPP2009v (Table 8),
2) In the HIV Age Distribution Page of AIM, there was selected the pattern for Concentrated IDU epidemic pattern,
3) In the MTCT Page there were entered the data regarding:
a) Distribution of HIV+ pregnant women by treatment regimen (Table 9)
b) Distribution of HIV+ pregnant women by infant feeding method (Table 10)
c) Percent of pregnancies of HIV+ women terminated by abortion (Table 11)
4) In the ART Treatment Page there were entered the data regarding:
a) Number of adults in need for ART receiving ART: The information was imported from EPP2009v and adjusted to data for both banks (In EPP2009v the data regarding ART Treatment only for the right bank was entered). The enrollment targets in ART treatment of Global Fund to fight Tuberculosis, AIDS and Malaria were entered till 2012 and the option “Linear projection of past 3 years” was used with 2010 as initial point. (Table 12).
5) In the Child Treatment section there were entered the data regarding:
a) The number of children (0 – 14) who received ART (Table 13)
b) The number of children (0 – 14) who received cotrimoxazole (Table 13)
c) The percentage of infants diagnosed with PCR (Table 14)
6) In the Treatment Cost Section in AIM there were entered the data regarding:
a) Treatment costs per patient per year (Table 15)
b) Service delivery costs (per in-patient day / out-patient visit), (Table 16)
c) Service delivery requirements (per patient per year) (Table 17)
7) In the Impacts Section there were entered the data regarding (Table 18):
a) TB incidence without HIV (per 1000)
b) Percent pop with latent TB
c) TB incidence with HIV (%)
After the data were entered, the Uncertainty Analysis in Spectrum with 1,000 curves were made
5
Table 2 Prevalence among officially registered IDU (1995 - 2009 Years)
1995 / 1996 / 1997 / 1998 / 1999 / 2000 / 2001 / 2002 / 2003 / 2004 / 2005 / 2006 / 2007 / 2008 / 2009Median Prevalence / 0.1 / 2.3 / 15.4 / 20.4 / 22.3 / 19.9 / 20.5 / 19.3 / 19.2 / 20.9 / 17.2 / 16.3 / 12.6 / 24.9 / 25.2
Mean Prevalence / 0.1 / 2.3 / 15.4 / 20.4 / 22.3 / 19.9 / 20.5 / 19.3 / 19.2 / 20.9 / 17.2 / 16.3 / 12.6 / 24.9 / 25.2
National (%) / 0.1 / 2.3 / 15.4 / 20.4 / 22.3 / 19.9 / 20.5 / 19.3 / 19.2 / 20.9 / 17.2 / 16.3 / 12.6 / 24.9 / 25.2
(N) / 1350 / 1725 / 2098 / 2916 / 3240 / 3492 / 3954 / 4336 / 4523 / 4456 / 4555 / 4465 / 4827 / 3444 / 3354
Source: Republican Narcological Dispensary
Table 3 Prevalence among MSM (1995 – 2009 years)
2002 / 2003 / 2004 / 2005 / 2006 / 2007 / 2008 / 2009Median Prevalence / - / 1.7 / 2.5 / - / - / 4.8 / - / -
Mean Prevalence / - / 1.7 / 2.5 / - / - / 4.8 / - / -
Chisinau (%) / - / 1.7 / 2.5 / 4.8 / - / -
(N) / - / 118 / 121 / 83 / - / -
Source: National Centre of Health Management, BSS Surveys 2003/2004, 2007
Table 4 Prevalence among CSW (1995 - 2009)
2003 / 2004 / 2005 / 2006 / 2007 / 2008 / 2009 / 2010Median Prevalence / 4.6 / 8.5 / - / - / 2.9 / - / - / -
Mean Prevalence / 4.6 / 8.5 / - / - / 10.87 / - / - / -
Chisinau (%) / 4.6 / 8.5 / - / - / 2.9 / - / - / -
(N) / 150 / 151 / - / - / 243 / - / - / -
Balti (%) / - / - / - / - / 32.8 / - / - / -
(N) / - / - / - / - / 122 / - / - / -
Source: National Centre of Health Management, BSS Surveys 2003/2004, 2007
5
Table 5 HIV prevalence among general population (2000 – 2009)
2000 / 2001 / 2002 / 2003 / 2004 / 2005 / 2006 / 2007 / 2008 / 2009Median Prevalence / 0 / 0 / 0 / 0 / 0.05 / 0.08 / 0.1 / 0.13 / 0.16 / 0.15
Mean Prevalence / 0 / 0 / 0 / 0 / 0.11 / 0.11 / 0.15 / 0.13 / 0.22 / 0.2
Chișinău (%) / 0 / 0 / 0 / 0 / 0.05 / 0.08 / 0.1 / 0.13 / 0.16 / 0.15
(N) / 975 / 2218 / 4586 / 10868 / 13018 / 10578 / 11553 / 13182 / 13918 / 13410
Bălti (%) / - / - / - / - / 0.53 / 0.37 / 0.63 / 0.25 / 0.72 / 0.81
(N) / - / - / - / - / 1502 / 1615 / 1903 / 1982 / 2217 / 2104
Anenii - Noi (%) / - / - / - / - / 0 / 0.11 / 0.4 / 0.09 / 0 / 0.21