Innovative Uses of Surveillance Data to Respond to

Innovative Uses of Surveillance Data to Respond to the HIV Epidemic Amanda D. Castel, MD, MPH The George Washington University School of Public Health and Health Services

National Minority AIDS Council • Established in 1987, NMAC is the premier national organization dedicated to develop leadership in communities of color to END the HIV/AIDS epidemic. Program services: v. Capacity Building Assistance (CBA) v. Conferences and Meeting Services v. Legislative and Public Affairs v. Treatment, Education, Adherence, and Mobilization

Webinar Series Getting to Data Driven Outcomes Jan. 16 Feb. 20 Illinois May 15 Innovative Uses of Surveillance Data for Nontraditional Purposes to End the Epidemic Reducing Perinatal Transmission: An Perspective Pay for Performance in HIV Testing Services

Objectives • Understand how to assemble STD and HIV surveillance data to measure the effectiveness of HIV providers and grantees • Name at least three ways data triangulation can be used to evaluate HIV programmatic outcomes • Understand how quantitative data can be used to make programmatic decisions in this new HIV environment

Background and Historical Context

HIV/AIDS Surveillance. . . is the ongoing and systematic collection, analysis, interpretation, dissemination and evaluation of population-based information about persons infected with HIV or diagnosed with AIDS (CDC)

History of CDC AIDS Case Definition for Surveillance in the U. S. * • 1981 –Reports of Pneumocystis carinii pneumonia (PCP) and Kaposi’s sarcoma (KS) in young gay men in SF, NY and LA • 1982 – CDC clinical AIDS case definition developed including 20 opportunistic illnesses, including infections (bacteria, fungi, protozoa) and cancers (KS and lymphoma) • 1987 – Case definition expanded to include TB, wasting syndrome, and dementia • 1993 – Expanded again to include cervical cancer, bacterial pneumonia, pulmonary TB, and HIV+, CD 4<200 *www. thebody. com/encyclo/aids. html

Pathway of HIV Infection High risk behaviors HIV incidence HIV prevalence HIV Case Surveillance National HIV Behavioral Surveillance HIV Incidence Surveillance Molecular HIV Surveillance/ VAHRS HIV Prevalence Estimates Advanced HIV/AIDS Death AIDS Case Surveillance Morbidity and Monitoring Project Death Surveillance using vital registration data

Importance of HIV/AIDS Surveillance • Describes the epidemiology and magnitude of the HIV/AIDS epidemic • Monitors trends in affected populations • Targets HIV prevention and treatment services • Guides the development of new interventions and approaches to HIV prevention • Provides data upon which funding decisions for HIV/AIDS programs (e. g. Ryan White) are based

Key Changes to HIV/AIDS Surveillance Systems • Integration of HIV and AIDS surveillance • Transition from code to name-based HIV (and AIDS) surveillance • Data capture is now longitudinal due to move from HARS to e. HARS • Measurement of HIV continuum of care • Use of surveillance data to drive and measure the impact of programmatic and clinical outcomes

Measuring the HIV Continuum of Care: The Treatment Cascade 79% 62% 41% 36% 28% Using data from 3 different surveillance systems, able to measure engagement in care Source: MMWR Vital Signs, Dec. 2, 2011, Vol 60 (47).

Prevention and Testing

Missed Opportunities for HIV Testing • Objective: To provide evidence for missed opportunities for HIV testing in South Carolina • Methods: Linked surveillance registry with state-wide all payer health database • Looked for visits and diagnoses occurring before the 1 st HIV positive test • Results: 73% of newly diagnosed persons had visited a health care facility >= 1 time prior to being HIV tested. Source: Duffus WA et al. AIDS Patient Care and STD, 2009.

Impact of South Carolina Study • Routine, voluntary HIV screening for all persons 13 -64 in health care settings, not based on risk • Repeat HIV screening of persons with known risk at least annually • Opt-out HIV screening with the opportunity to ask questions and the option to decline

National HIV Behavioral Surveillance Project (NHBS) Heterosexual HIV Behavior Survey, 2010 Characteristics of Relationships, by Sex, DC Behaviors Study, 2010, N=482 Male Female 70. 0% 60. 4% 60. 0% 52. 8% 55. 3% 50. 0% 40. 8% 40. 0% 30. 0% 20. 0% 10. 0% Felt partner definitely or probably had concurrent sex Self reported a concurrent sex partnership in last 12 partnerships in the last 12 mo mo.

National HIV Behavioral Surveillance Project (NHBS) Heterosexual and MSM Survey, 2006 -2009

Social Marketing Being in a Relationship Isn’t Always Easy: Know Where You Stand

HIV Testing, Social Mobilization, Condoms and MSM in DC Dissemination of media messages via billboards, METRO ads, and bus ads

Expansion of Municipal Condom Distribution, DC, 2007 -2010

Data Triangulation • Triangulation is a method used to check and establish validity in studies by analyzing a research question from multiple perspectives. • Data triangulation involves using different sources of information in order to increase the validity of a study. • Triangulation has proved to be an effective tool for reviewing and corroborating findings in the surveys, assessments, appraisals, etc. , that are an essential part of effective monitoring and evaluation Sources: www. unaids. org; Lisa Guion at https: //edis. ifas. ufl. edu/fy 394

Routine HIV Testing Scale-up in DC 1) June 2006, Testing Campaign >50 Partners Rapid Test Expansion DC Jail 2) Focus on Medical Settings: Ask for the Test Offer the Test 3) Preliminary Positive? Go directly to HIV care

Impact of Routine Testing in Washington, DC 2005 -2009 Source: Castel AD et al. CROI 2010 Presentation

The Impact of Community-Wide HIV Testing in NYC Source: Myers JE, Braunstein SL, Shepard CW, et al. JAIDS, 2012 Sep.

Syndemic Analyses and Co. Morbidities

Poll Question #1 Does your organization/health dept. have an integrated surveillance system?

Syndemics • Defined as two or more diseases, or conditions, that interact and increase transmission probability or worsen the health outcomes of people and communities • Can assess syndemics by linking and matching cases of from each disease registry

Syndemics in DC, 2010 TB Syphili s Chlamydia 3. 3% 1. 1% HBV 3. 7% HIV n=835 Gonorrhea 2. 1% HIV 52. 2% HCV 0. 7% Chlamydia 7. 5% Gonorrh ea TB HBV Syphilis 6. 0% 3. 0% n=134 HCV 7. 3% Chlamydia 0. 3% Gonorrhea 0. 2% HIV 8. 9% HCV n=2, 494 HBV 3. 8% TB 0. 1% Syphilis

Assessing the Role of Syndemics in San Francisco • Assessed the prevalence of co-occurring infections and their impact on persons living with HIV/AIDS (PLWHA) by matching 7 disease registries • • • Results: Syndemics highest among certain populations Co-infected PLWHA affected diverse geographic areas, regardless of socioeconomic status Conclusions: PH Impact: Underscores need to address multiple conditions in tandem in an integrated health system Source: Impact of syndemics on people living with HIV in San Francisco. P. L. Chu, G. -M. Santos, A. Vu et al. IAS Conference, 2012

HIV and non-Infectious Conditions: National HIV/AIDS Cancer Match Study Number of Cases Known vs. Unknown by Year of Cancer Diagnosis, 1996 -2006 Cancer Type by HAART Era of Diagnosis ADCs

Linkage, Retention, Care, and Treatment

Measuring the HIV Continuum of Care Source: Linkage, retention, ART use and viral suppression in four large cities in the United States. N. Benbow, S. Scheer, A. Wohl et al. IAS Conference 2012.

Comprehensive Community Approach to HIV : The 4 R’s Recruitment n. Health System Navigator n. Red Carpet Entry Recapture/Reengagement n. Blitz! Retention n. Acuity Scale & MCM Guidelines Results n. Linkages 32 n. Treatment Promotion

DC Recapture Blitz Purpose: To re-engage people Recapture Blitz Summary living with HIV in care who are 2009 ‘loss to care’ Number of clients lost to care 1, 365 Define: Loss to care: Not in care (reported) for more than 6 months 1, 008 Number of clients in care (actual) Methods: Primary Medical Care Providers send list of clients Number of clients contacted 404 Average number of contacts per not seen in their clinics for 2 to 4 person greater than 6 months. HAHSTA “matched” these lists Number of appointments made 207 to e-HARS, labs surveillance and ADAP databases. 186 Number of appointments kept 2010* 585 488 203 2 to 10 132 109

Louisiana Public Health Information Exchange (La. PHIE) • LA surveillance data demonstrate that ~1/3 of HIV infected persons are not receiving HIV care • In 2007, the OPH partnered with LSU, the largest HIV care provider in the state • La. PHIE is a ‘secure bi-directional public health informatics application which links public health surveillance data with patient-level EMR data’. • LSU ER, primary care, specialty ambulatory care, and inpatient units participate in real-time Herwehe, J. et al. Implementation of an innovative, integrated electronic medical record (EMR) and public health information exchange for HIV/AIDS. J Am Med Inform Assoc 2012; 19: 448 -452

La. PHIE • Between 2/1/09 -1/31/11, La. PHIE processed registration messages for 488 patient encounters. • Identified 345 unduplicated, HIV-positive patients in need of treatment. • Of those identified, 82% had at least one CD 4 or VL test during the study follow-up period. Herwehe, J. et al. Implementation of an innovative, integrated electronic medical record (EMR) and public health information exchange for HIV/AIDS. J Am Med Inform Assoc 2012; 19: 448 -452

Development of a City-Wide Cohort of HIVInfected Patients: The DC Cohort • Objective: to contribute to improving the quality of care and treatment of HIVinfected patients in DC • 12 participating sites with 10 -15, 000 patients • Prospective, multicenter longitudinal cohort • Collect socio-demographics, risk factors, treatments, diagnoses, lab and procedures from EMR and data abstraction • Data will be linked to DC DOH surveillance data to improve completeness and accuracy

Poll Question #2 Do laws or regulations exist in your jurisdiction prohibiting the sharing of data across surveillance programs?

Spatial and Social Determinants of Health

AIDSVu • Created by Emory University, CDC, and local/state surveillance and prevention bureaus • Interactive online maps illustrating the prevalence of HIV in the United States. • Maps can be filtered by race/ethnicity, sex and age, and social determinants of health, such as educational attainment and poverty • Visually explore the HIV epidemic alongside critical resources: – HIV testing center locations – HIV treatment center locations – NIH-Funded HIV Prevention & Vaccine Trials Sites

Comparing HIV Data with Social Determinants of Health HIV Prevalence Rates, 2009 % Living in Poverty, 2009

Gentrification and Impact on HIV Rates in DC • Definition of gentrification: The process of renewal and rebuilding accompanying the influx of middle class or affluent people into deteriorating areas that often displaces earlier usually poorer residents • Purpose: investigate the association between gentrification and HIV case density • Neighborhood level comparison of gentrification • Gentrification measure: normalized median household value adjusted for neighborhood population Source: Ahmed T. et al. Gentrification and its effects on HIV/AIDS rates in D. C. IAS Conference 2012.

Reported Case Density 2000 Source: Ahmed T. et al. Gentrification and its effects on HIV/AIDS rates in D. C. IAS Conference 2012. Results 2010 Gentrification Index by Neighborhood • The correlation coefficient of r = 0. 51 shows a significant positive association between gentrification and HIV rates (p< 0. 05). Approximately 31% of the variation in neighborhood rates of HIV/AIDS in DC can be explained by gentrification and demographic shift alone. • Gentrification is a significant factor in explaining how HIV/AIDS rates vary by neighborhood. • Findings underscore the need to consider gentrification when using M&E to advise programmatic priorities to accurately how public health interventions affect new diagnoses.

National Responses: Programmatic and Research-Related Initiatives

What is Community Viral Load? • A population-based measure of the concentration of plasma HIV-1 RNA (viral load) in HIV-infected individuals • Represents the level of viremia in a community in a geographic area • Community viral load (CVL) is a potential biomarker for HIV transmission and quality of HIV care and treatment • Hypothesis: – By measuring CVL, can assess progress in treatment and therefore reductions in a community's level of viremia – Declining CVL should correlate with a reduction in incident cases • Recent studies have supported CVL as a means of measuring HIV incidence: (Vancouver, British Columbia, San Francisco) • CVL is being used as an outcome measure for: – National HIV/AIDS Strategy – Testing and Linkage to Care Plus Source: CDC CVL Guidance Document

Mean CVL and HIV Incidence, 2006 -08 Mean CVL copies/ml 20, 000 935 (CI 658, 1212) 1100 792 (CI 552, 1033) 621 (CI 462, 781) 15, 000 10, 000 Mean CVL & HIV-incidence p=0. 3 900 700 5, 000 0 1300 2006 Mean CVL 2007 2008 HIV Incidence Source: Das M. UCHAPS Presentation, June 2010. 100 Number of HIV Cases 25, 000 Overall Mean CVL: 23, 348 copies/ml

Figure 2. Geographic Distribution of Mean and Total CVL by Census Tract, Ward, and SES Indicators, 2008 Areas with higher mean and total CVLs appear to correspond to those areas with the worst SES indicators. Source: Castel AD, Befus M, Willis S, et al, AIDS, 2012 Jan.

Impact of CVL • Useful marker for – Assessing HIV/AIDS epidemic trends – Measuring access and impact of care and treatment – Serving as indicators of the viral burden in the population • Geospatial and subgroup analyses may be useful for informing targeted interventions • Methodologies vary but CDC has attempted to standardize • Being used to evaluate prevention activities on a nationallevel

HIV Prevention Trials Network (HPTN) 065: TLC-Plus Study • Purpose: To evaluate the feasibility of an enhanced community-level HIV test, link-to-care plus treat strategy in the U. S. • Five components: I. Testing II. Linkage to care III. Viral suppression IV. Positive prevention V. Patient and provider survey • Two intervention communities (NYC, DC) compared to 4 comparison communities (Houston, Phila, Chicago, Miami)

TLC+ Outcomes Measured Using Surveillance Data • HIV surveillance data utilized to determine: – Newly diagnosed HIV infections – Linkage to care within 3, 4 -7, 7 -12 months – Continuous HIV care – Viral suppression (VL <400 copies/m. L) • Compare FI vs. non-FI, intervention cities, and intervention vs. control cities

The National HIV/AIDS Strategy • On July 13, 2010 the White House released the National HIV/AIDS Strategy • This is the nation’s first ever comprehensive HIV/AIDS strategic plan Goals: • Reducing HIV infections • Increasing access to care and improving health outcomes among people living with HIV • Reducing HIV related health disparities • Achieving a more coordinated national response

Progress in Achieving the National HIV/AIDS Strategy Goals

Population-Based Metrics to Measure NHAS Progress-Seattle King County Source: Dombrowksi J et al. AIDS 2012.

Conclusions • Use of surveillance data has evolved over time • Will the Walls Came Tumbling Down (Fairchild and Bayer) – CDC: “ Once data are in hand, it is a failure to use those data for public health purposes that must be justified” – NYC; “conceive [surveillance] registry as a kind of ‘universal’ electronic medical record…” • Now being used to drive, direct, and determine the impact of prevention, care and treatment efforts • Challenges: regulatory, ethical, confidentiality, data access and sharing • Provides opportunities to highlight surveillance as a foundation of public health efforts, leads to development of data-driven policies and programs, provides a more complete understanding of the HIV epidemic Source: Fairchild AL, Bayer RB. NEJM 2011.

Acknowledgements • • • Tashrik Ahmed Moupali Das Wayne Duffus Bruce Furness Alan Greenberg Angelique Griffin Michael Kharfen Irene Kuo Manya Magnus • • • Jenevieve Opoku Gregory Pappas Melissa Taylor Arpi Terzian Tiffany West Sarah Willis A Public Health/Academic Partnership between the District of Columbia Department of Health and The George Washington University School of Public Health and Health Services Department of Epidemiology and Biostatistics: Contract Number POHC-2011 -C-0073

Poll Question #3 Do you think that your jurisdiction is currently or could potentially conduct some of these activities?

Poll #5

Q&A

Upcoming Webinars Getting to Data Driven Outcomes Feb. 20, 2 pm EST Reducing Perinatal Transmission: An Illinois Perspective Mildred Williamson, MD, MPH, Illinois Dept. of Health May 15, 2 pm EST Pay for Performance in HIV Testing Services Mary Orticke, RN, MPH, Los Angeles County Dept. of Health

National Minority AIDS Council 1931 13 th Street NW Washington, DC 20009 (202) 483 -6622 linc@nmac. org www. nmac. org

Additional Information The slides following this one provide more information about the previously mentioned topics.

TLC+ Site Randomization • ~20 HIV test sites and ~20 HIV care sites were selected to participate in TLC+ in each intervention city. • Selected sites were randomized to FI/non-FI • TLC+ intervention (FI) began 4/2012

Treatment as Prevention (Tas. P) • Scientific studies have shown that achieving an undetectable VL can prevent transmission (HPTN 052) • Test and Treat Concept: Universal HIV testing and initiating ART immediately may significantly reduce HIV incidence and prevalence to <1 case per 100, 000 population per year and within 50 years the overall prevalence • Domestically: HPTN designed and implemented study to assess feasibility of modified test and treat approachwoul <1 • %.

Social Determinants of Health and HIV • Denning et al used NHBS and HIV prevalence data to study association between poverty and HIV • Found an inverse relationship between SES and HIV prevalence rates (Denning, 2010. Communities in Crisis. Available at http: //www. cdc. gov/hiv/topics/surveillance/resources/other/pdf/poverty_poster. pdf

Matching Registries for Integrated Partner Services: Gonorrhea and HIV • • Conducted a probabilistic match to the HIV 4 site GC/HIV Co-infection, 2000 -2008 and STD surveillance databases for GC Site Total GC GC cases % Cocases diagnosed 2000 -2008 cases with HIV infect (range per ed Public Health Implications: year) • STDs occurring among HIV-infected New York 107, 786 5, 930 5. 5 City (367 -778) patients represent a public health priority for intervention Miami/Dade 40, 214 1, 504 4. 0 (91 -211) • Health departments should consider real-time STD/HIV registry Washington, 17, 910 1, 312 7. 3 DC crossmatches or integrated HIV/STD surveillance databases to identify high Arizona 39, 779 725 2. 0 (31 -136) risk HIV/GC co-infected patients for: Total 205, 689 9, 471 4. 6 • Partner services • Risk Reduction Counseling • HIV care referral Source: M. Taylor et al. Gonorrhea infections diagnosed among persons living with HIV: cross matching surveillance registries to identify potential opportunities for integrated partner services - New York City, Washington D. C. , Miami/Dade County and Arizona, IAS Conference 2012.

Spatial distribution of AIDS in San Francisco Source: Das M. UCHAPS Presentation, June 2010.

Treatment Promotion: ADAP Advertising Media Outlets (Print, TV, Radio, Metro) 25 # of Ads >9, 000 Ave Audience 6. 5 Million Amount Spent $830, 000 Increase ADAP Enrollment 50%

DC ADAP Enrollment: 2008 -2009