1. Validating a Model for Risk-based Differentiation of HIV Prevention and Testing for Female Sex Workers in Maharashtra, India
IAS HIV SCIENCE 2019, MEXICO CITY
Jimreeves Kirubakaran, Senior Program Management Specialist, fhi360 India, 23 July 2019

2. Background and Methods
Key indicators on risk behaviors and vulnerabilities were collected and monitored every quarter by the intervention.
Program data from three FSW TIs in Maharashtra from April 2016 to March 2018 were analyzed.
Four HIV test outcomes for each FSW and linked each individual’s behavioral data from the previous quarter to the HIV test results, generating 16,228 data points
Excluded the following data points:
Known HIV-positive individuals
No HIV test conducted during the six-month period
1. Penalized and regular logistic regression analyses were used
2. Receiver operating characteristic (ROC) curve analysis
&
Youden’s index method
3. Dominance analysis

3. Weighted scores for each indicator/variableSN
Indicator/variable
Score 1
Age (in years) 5
RTI/STI 1a
18-25 5a
No RTI/STI 1b
26-35 4 5b
RTI/STI and treated 1c
> 35 15 5c
Not visited clinic 19 2
Sex acts/week 2a
0-4 6
Condom use 2b
>=5 6a
Always 3
Years in sex work 6b
Missed 3a 3b
5-15 7
Condom compromise 3c
>=16 7a No
Years associated with the TI 7b
Yes 4a
0-2 35 4b
3-6 10 4c
>=7
The cutoff with the best sensitivity for HIV positivity in the high-risk group and best specificity for HIV negativity in the low-risk group were identified.
Category
Score
High
>=39
Moderate
19-38
Low
<=18

4. Findings
Applied prospectively in two of the above-mentioned FSW TIs to study its efficiency for segmenting FSWs into high, medium, and low priority for HIV testing and differentiated service delivery
HIV testing and detection by priority group (April to September 2018)
Female Sex Workers
High Priority
Medium Priority
Low Priority
Total
Number (percentage) in priority group
684 (30.5%)
1,068 (47.7%)
487 (21.8%)
2,239 (100%)
Number tested
573
840
311
1724
Number of HIV-positive cases 11
After the model was completed, it was applied prospectively in two of the above-mentioned FSW TIs to study its efficiency for segmenting FSWs into high, medium, and low priority for HIV testing and risk-reduction services.
Among 2,239 FSWs categorized into priority groups, 684 (30.5 percent) were considered high priority. Of the 11 HIV cases detected, all were from this high-priority group (Table 3).
Among 2,239 FSWs categorized into priority groups, 684 (30.5 percent) were considered high priority. Of the 11 HIV cases detected, all were from this high-priority group

5. Conclusions and Recommendations
The model demonstrated effective and precise categorization of FSWs at increased HIV risk, corroborating the need for intensified interventions among the high-priority group.
This would include augmenting the frequency and intensity of high-threshold prevention services through differentiated strategies within the concentrated HIV epidemic in India.
Expanding the model to other typologies—men who have sex with men (MSM), transgender people, and people who inject drugs (PWID)—and building the model at the national scale would help the national program prioritize high-risk key population members for differentiated prevention services.
Employing this model would increase HIV case detection by focusing on the FSWs who are at increased HIV risk.
It is recommended to standardize and include the certain variables during the next phase model development.
The model demonstrated effective and precise categorization of FSWs at increased HIV risk, corroborating the need for intensified interventions among the high-priority group. This would include augmenting the frequency and intensity of high-threshold prevention services through differentiated strategies within the concentrated HIV epidemic in India.
It is recommended to standardize and include the following variables during the next phase model development:
Have you consumed alcohol during/before sex in the past three months?
Have you experience sexual violence in the past three months?
Have you travelled for sex out of your district/town/village in the past three months?
Have you had anal sex in the past three months?
Have you had group sex (more than one partner at the same time) in the past three months?
Expanding the model to other typologies (MSM, TG and PWID) and building the model at the national scale would help the national program prioritize high-risk KP members for differentiated prevention services.

6. Differentiated Service Delivery Model for Prevention for FSW, MSM & TG
NEW KP
Contact: 4 / month
-Intensive BCC
-Risk assessment & counselling
-Condom need analysis
-HIV testing: immediately
-Presumptive treatment
DIC / Outreach based BCC
HIGH PRIORITY
Contact: 3 / month
-Condom provision
-Risk reduction counselling
-HIV testing: Once in 6 months
-Partner testing & services
-RMC: Quarterly & as when required
Intensive Intervention at DIC, complemented with outreach
MEDIUM PRIORITY
Contact: 2 / month
Outreach / DIC based BCC
LOW PRIORITY/ STABLE
Contact: 1 / month
-Motivational counselling (for sustaining BC)
-HIV testing: At least annually
-RMC: Quarterly
BCC at the hotspot / using technology
KPLHIV
-Linkages to ART: Priority
-ART initiation: Priority
-Adherence support: monthly
-Tracking LFU: priority
-Mobilizing for CD4/VL testing: 6 monthly / annually
-Index testing
-PHDP: monthly
-Referral to other health services : need based
-Linking with social protection : need based
-Other prevention services

8. Acknowledgements
Mr. Parimal Singh, Former Project Director, MSACS, India
Dr. Srikala Acharya. Additional Project Director, MDACS, India
Dr. Bitra George, Country Director, FHI360 India
Mr. Prabuddhagopal Goswami, Associate Director, FHI360 India
Mr. G. S. Shreenivas, Country Representative, LINKAGES India
Dr. Maninder Setia, Consultant Epidemiologist, India
Mr. Srikar Panyam, Consultant Monitoring and Evaluation Expert, India
Dr. Lokesh Gabhane, Joint Director-TI, MSACS, India
Mr. Sachin Ketkar, Joint Director-TI, MDACS, India
Mr. Jagdish Patel, Team Leader, TSU, Maharashtra, India

9. THANK YOU