A Theoretical Framework for STD Epidemiology: From Transmission Dynamics to Program Design

A Theoretical Framework for STD Epidemiology: From Transmission Dynamics to Program Design Kimberley Fox, MD, MPH Director, Global AIDS Program Thailand MOPH U.S. CDC Collaboration Bangkok, Thailand

Approaches to prevention of infectious diseases Biomedical Examples: vaccines, preventive therapy Behavioral Examples: handwashing, self-quarantine (ill persons avoid exposure to well persons) Structural Examples: negative pressure air flow in rooms for TB patients, beer tax to lower STD rates

Approaches to STD prevention (examples) Biomedical Screening and treatment Vaccines Behavioral Reduce risky sexual exposures (condoms, changes in norms for sexual behaviors, etc.) Structural Ensuring that stores in high-risk areas sell condoms

Gonorrhea Reported rates: United States, 1970 2000 and the Healthy People year 2010 objective Rate (per 100,000 population) 500 400 Gonorrhea 2010 Objective 300 200 100 0 1970 73 76 79 82 85 88 91 94 97 2000 Note: The Healthy People 2010 (HP2010) objective for gonorrhea is 19.0 cases per 100,000 population.

Sir, the following paradigm shifts occurred while you were out.

First we must consider Why do STDs persist in our communities? A mathematical model Ro = β c D Ro = initial rate of secondary cases arising from a case β = efficiency of transmission c = measure of sexual behaviors related to partner change D = duration of infectiousness

Determinants of Transmission Efficiency (Ro=BcD) Infectivity or organism load of pathogen Co-infections Circumcision, cervical ectopy (susceptibility of host) Condoms, microbicides, other barriers

Estimates of Key Epidemiologic Parameters Necessary to Sustain Transmission of Five Different STDs (c=1/ßd) AGENT DURATION OF INFECTIOUSNESS (D) IN YEARS TRANSMISSION EFFICIENCY (ß) PER PARTNERSHIP MEAN PARTNER CHANGE RATE (c) PER YEAR N. gonorrhoeae No control 0.5 Control 0.5 C. trachomatis 0.2 T. pallidum No control Control HIV African parameters American parameters 0.3 0.3 0.1 0.01 Haemophilus ducreyi 0.8 Produced from data collected by Brunham and Plummer, 1990

A model of sexual transmission of HIV 5 4 3 1/10,000-1/1000 HIV RNA in Semen (Log10 copies/ml) 2? Transmission Threshold? 1/500-1/3 Stage of Disease Risk of Transmission 1/1000-1/100 1/50 Seroconversion Asymptomatic HIV Progression AIDS

High viral load in primary infection: a triple threat? Early Infection Chronic Infection Late Infection Biologic: HIV viral load ++++ + ++++ Behavioral: unprotected sex ++++ +++ + Social: networks of high-risk persons ++++ ++ + According to mathematical models (Fraser 2006), the proportion of transmission that results from each stage of the disease is: Primary infection: 12% Asymptomatic chronic infection: 71% AIDS: 17% Although infectiousness is higher during early and late infection, the asymptomatic period is much longer.

Biologic evidence for STIs as co-factors for HIV transmission Presence of STI Increase in HIV infectiousness Increase in viral load in genital secretions Increase in susceptibility to HIV Disruption of epithelial barrier Increase in HIV-receptive cells Increase in receptors expressed per cell

Magnitude of the risk in epidemiologic studies 2- to 10-fold increased risk of HIV infection in persons with syphilis 2- to 8.5-fold increase in persons with HSV-2 2- to 9-fold increase in persons with gonorrhea, chlamydia, or trichomonas

Levine, et al. J Infect Dis, 1998

A model of sexual transmission of HIV 5 4 3 1/10,000-1/1000 HIV RNA in Semen (Log10 copies/ml) 2? Transmission Threshold? 1/500-1/3 Stage of Disease Risk of Transmission 1/1000-1/100 1/50 Seroconversion Asymptomatic HIV Progression AIDS

Effect of other STDs on sexual transmission of HIV (theoretical) 10 8 6 4 STD Antibiotic Therapy HIV RNA in Semen (Log10 copies) Seroconversion Asymptomatic HIV Progression AIDS 2 Stage of Disease

Biological mechanisms for HIV transmission in uncircumcised men Foreskin susceptible to tears and abrasions Foreskin provides a mucosal environment conducive to survival of HIV and contains high densities of HIV target cells Circumcised glans penis protected by keratinized epithelium

Determininants of Sexual Behaviors (Ro=BcD) Cultural and contextual norms Substance use (alcohol, illicit drugs) Economic pressures Gender-based hierarchical relationships

Lifetime STIs, by Number of Sex Partners since Age Eighteen (cases per 1,000) Partners since age 18 0 1 2-4 5-10 11-20 21+ Gonorrhea 0 11 30 83 148 206 Syphillis 0 1 4 13 9 21 Chlamydia 0 14 20 44 64 70

Distribution of Number of New Sexual Partners in Last Year

Individual-Level Factors that May Contribute to c Number of partners Rate of acquiring new partners Casual partners Sexual practices (dry sex, anal sex)

Core Groups for Selected STDs Chlamydia Gonorrhea Syphilis Chancroid

Gonorrhea as a Social Disease J. Potterat It s not how many partners, but who they are that counts

Source: MC Boily et al, 2000. The Network Matters

The Network Structure Can accelerate risk or protective behaviors for some people It can make the same behavior more or less risky depending on where an individual is located in the network. Provides a roadmap for where infection is going to travel Source: T. Valente, 2002.

Assortativeness Matters, Too Source: MC Boily et al, 2000.

And Then You Do the Calculations Source: MC Boily et al, 2000.

What do Asian HIV epidemics look like? FSW Clients Low or no risk males Multiple interlinked epidemics in higher risk populations MSM Low or no risk females IDUs Source: T. Brown, 2003

Males visiting sex workers in last year Country Percent clients Year Thailand 22% 10% 1990/1993 Cambodia 13% 2000 Japan 11% 1999 Philippines 7% 2000 Hong Kong/ 5% Early 1990s Singapore Conclusion: from 7 to 25% at risk of HIV Source: T. Brown, 2003

The number of clients largely explains differences in speed & severity Asian Epidemic Model 10 8 6 4 2 0 Percent of adults living with HIV 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 5% 10% 20% male clients Source: T. Brown, 2006

Why were Thailand & Cambodia so serious? Asian Epidemic Model: 20% male clients with 0.8% females FSW Percent HIV+ 80 60 40 20 By 2030, 23% of males, 8.6% of females HIV+ 0 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 IDU FSW Gen male Gen female MSM Source: T. Brown, 2006

Determinants of Duration of Infectiousness (Ro=BcD) Natural history of infection Health care services quality accessibility Referral to services by partner or health care provider Health care seeking behaviors

STD Clinic Patient Survey, North Carolina: Duration of Symptoms Before Visit All Male Female 0 1 2 3 4 5 6 7 Median number of days

Barriers to Access: STD Clinic in North Carolina 25 Percent of patients 20 15 10 5 0 Waiting time Transp. Clinic hours Can't get appt. Not comf. Child care

STD Clinic Patient Survey, North Carolina Waiting Time Before Talking to Nurse or Doctor Percent of patients 45 40 35 30 25 20 15 10 5 0 0-15 16-30 31-45 46-60 61-90 91-120 121-180 Median waiting time: 20 minutes 181-240

STD Clinic Patient Survey, North Carolina: Duration of Symptoms According to Mode of Transportation Car Median 5 days 27.9% waited more than one week Friend/Bus/Taxi/Walk Median 7 days 38.5% waited more than one week

Prevalences of Gonorrhea, Chlamydia, Syphilis, and HIV among NC Migrant Farmworkers 12 10 8 6 4 2 0 Gonorrhea Chlamydia Syphilis (RPR &FTA+) Syphilis, ever (FTA+) HIV

Prevalences of Gonorrhea and Chlamydial Infection among Incarcerated Male Youth in NC 10 8 6 4 2 0 Gonorrhea Chlamydia Syphilis (RPR & FTA+)

Quality of Services STD treatment with the wrong antibiotic doesn t reduce D Getting treatment right involves good history-taking, examination, and testing (when available), and using recommended treatment Antibiotic resistance may need to be considered

Gonococcal Isolate Surveillance Project (GISP) Percent of Neisseria gonorrhoeae isolates with decreased susceptibility or resistance to ciprofloxacin, 1990 2000 Percent 2.0 1.6 Decreased susc. Resistance 1.2 0.8 0.4 0.0 1990 91 92 93 94 95 96 97 98 99 2000 Note: Resistant isolates have ciprofloxacin MICs >1 μg/ml. Isolates with decreased susceptibility have ciprofloxacin MICs of 0.125-0.5 μg/ml. There were sixty-one (61) resistant isolates: one in 1991, one in 1993, two in 1994, eight in 1995, two in 1996, five in 1997, four in 1998, nineteen in 1999, and nineteen in 2000. Susceptibility to ciprofloxacin was first measured in GISP in 1990.

Estimates of Key Epidemiologic Parameters Necessary to Sustain Transmission of Five Different STDs (c=1/ßd) AGENT DURATION OF INFECTIOUSNESS (D) IN YEARS TRANSMISSION EFFICIENCY (ß) PER PARTNERSHIP MEAN PARTNER CHANGE RATE (c) PER YEAR N. gonorrhoeae No control 0.5 0.5 Control 0.15 0.5 C. trachomatis 1.25 0.2 T. pallidum No control Control HIV African parameters American parameters 0.5 0.25 2.0 8.0 0.3 0.3 0.1 0.01 Haemophilus ducreyi 0.8 0.8 Source: Brunham and Plummer, 1990

Estimates of Key Epidemiologic Parameters Necessary to Sustain Transmission of Five Different STDs (c=1/ßd) AGENT DURATION OF INFECTIOUSNESS (D) IN YEARS TRANSMISSION EFFICIENCY (ß) PER PARTNERSHIP MEAN PARTNER CHANGE RATE (c) PER YEAR N. gonorrhoeae No control 0.5 0.5 4 Control 0.15 0.5 13 C. trachomatis 1.25 0.2 4 T. pallidum No control 0.5 0.3 7 Control 0.25 0.3 13 HIV African parameters American parameters 2.0 8.0 0.1 0.01 5 13 Haemophilus ducreyi 0.8 0.8 15 Source: Brunham and Plummer, 1990

Without Antiretroviral Therapy HIV incidence HIV prevalence HIV-related deaths

With Antiretroviral Therapy HIV incidence HIV prevalence HIV-related deaths

ARV treatment scale-up (NAPHA) will Number of infection increase the prevalence of HIV 800,000 700,000 600,000 500,000 400,000 300,000 200,000 100,000 0 1985 1990 1995 2000 2005 2010 2015 2020 2025 Source: T. Brown, 2006 NO NAPHA Baseline (with NAPHA)

So now we re back to Why do STDs persist in our communities? (What makes Ro >1?) Ro = β c D Ro = initial rate of secondary cases arising from a case β = efficiency of transmission c = measure of sexual behaviors related to partner change D = duration of infectiousness

Core Groups for Selected STDs Chlamydia Gonorrhea Syphilis Chancroid

Geographic Core Areas Source: J. Potterat, 1985

Spatial Analysis The Spread of Disease Is Unavoidably Spatial EE Holmes Infection moves from individual to individual following a network of contacts within a population Local transmission Global (long-distance) transmission Bridge contacts

Geographic Information Systems Spatial data are stored in a map layer Data referenced to location on the earth's surface Source: D. Law, 2001

Geographic Distribution of STDs Case distribution No context Source: D. Law, 2001

Source: D. Law, 2001

Source: D. Law, 2001

Source: D. Law, 2001 Syphilis Distribution

Demographic and Social Factors Affecting c (behaviors) and D (services) Age structure and sex ratio of population Rapid demographic change Population growth Migration Shift from agricultural to wage-based economy Poverty Unstable power hierarchies

OUTBREAK INVESTIGATION Syphilis in Dade County

Infectious Syphilis in Miami-Dade County, Florida 1998-2001 175 150 125 100 75 50 25 0 1998 1999 2000 2001 through June MSM Heterosexual men Women

Outbreak Investigation Objective(s) The investigation Case-control design: who are controls? What other sources of information might be useful? Laboratory studies? Questionnaire: what do you want to know?

A Theoretical Framework for STD Epidemiology: From Transmission Dynamics to Program Design Part II

Role of STD Outbreak Investigations in STD Control Confirm the existence of a suspected outbreak Assess the extent of under reporting Determine risk factors for increased transmission Assess extent of emerging antimicrobial resistance Describe STD/HIV interrelationships in the setting of acute increases in morbidity Provide data for improving local STD/HIV prevention and control strategies

Methods for STD Outbreak Investigations Patient interviews, focus groups, and qualitative community assessment Detailed review of medical and laboratory records Retrospective and prospective case-control studies Intensive use of laboratory testing, including new diagnostic tests

Using a Theoretical Framework to Design STD and HIV Prevention Programs

Approaches to STD prevention: Decreasing transmission efficiency (Ro=βcD) Male condoms (HIV, gonorrhea) Female condoms (maybe) make condoms available, affordable, and appealing (normative) Microbicides (maybe) Reduce vaginal douching, which washes away normal flora Decrease efficiency of HIV transmission by treating other STDs

Condoms only work if they re used Must be: Available Affordable Appealing (normative)

Topical microbicides: 2 nd generation products and the future Viral Target Free virus Fusion/Attachment Reverse transcription Microbicide Savvy BufferGel PRO2000 Cellulose sulfate Tenofovir 30-40 candidates in preclinical development 14 products in early safety trials 5 products in large-scale efficacy trials (first results expected in 2007) Microbicide could be commercially available by 2010

Randomized trial of circumcision in South Africa (ANRS trial) Number of subjects (ages 15-24y) Number of HIV acquisitions Acquisition rate per 100 personyears Control group 1590 49 2.1 Circumcision 1538 20 0.85 Relative risk = 0.4 (60% protection) Aubert et al, 2005.

Circumcision: Unresolved issues Efficacy in other populations and settings Impact of hygiene on transmission Age at circumcision Acceptability Feasibility Operations Safety

Genital herpes (HSV-2) and HIV prevention HSV-2 may account for up to 74% of new HIV infections in men and 22% in women in some populations (del Mar Pujades Rodriguez M et al. AIDS 2002;16:451-62) Two large placebo-controlled trials of HSV suppression for HIV prevention are ongoing: NIH study at 9 global sites, high-risk women and MSM, results 2007 Partners in Prevention, 14 sub-saharan African sites, discordant couples, results 2008

Antiretroviral (ARV) therapy as prevention Mechanisms In HIV-infected persons, reduces viral load In HIV-negative persons, prevents HIV infection of cells or replication in cells Uses PMTCT (for mother and child) proven effective HIV-infected persons some evidence Pre- and post-exposure prophylaxis under study

Rationale for pre- and post-exposure prophylaxis (PrEP) trials with tenofovir Nucleotide reverse transcriptase inhibitor Licensed by U.S. FDA for treatment of HIV Long half-life allowing once daily dosing Phase II/III studies safe and effective Prevents HIV infection in animal models Minimal drug-drug interactions Few resistance mutations Studies underway USA: gay men Thailand: injecting drug users West Africa: high-risk women

What s next for tenofovir PrEP? Combination tenofovir and FTC (Truvada) is being studied in: Botswana (young men and women) Peru (gay men) Scale-up issues Expansion beyond motivated study participants (how to find target group, adherence with/without DOT) Cost (who will benefit most from it?) Potential for resistance (how to treat persons who become infected while on tenofovir PrEP?)

Biomedical interventions: A note of caution Randomized controlled trial in South Africa has shown 60% lower incidence of HIV in circumcised men (reported 2005) According to transmission models, the benefit of circumcision to the man would be overcome by a 30% increase in risk behavior Similar concerns have been raised for other incompletely protective interventions: vaccines pre-exposure prophylaxis (maybe)

Approaches to STD prevention: Changing sexual behaviors (Ro= βcd) Desirable behavior changes increasing condom use delaying initiation of sexual intercourse reducing number of sexual partners Types of interventions individual-focused (enhanced counseling) community-level (peer opinion leaders) mass media messages

Increasing condom use turns epidemics around 4 3 2 1 0 Percent of adults living with HIV 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 30% 40% 50% 60% 70% 80% condom use in sex work Source: T. Brown, 2006

In Thailand, focused prevention altered the course of the HIV epidemic C urrent H IV Infections in m illions 10 8 6 4 2 0 Red line represents what might have been if behaviors had not changed in Thailand Infections prevented 1985 1990 1995 2000 2005 2010 Source: T. Brown, 2006 With behavior change Without prevention

Sexual Network:Syphilis Among MSM in Miami-Dade County, Florida H H Syphilis Status H Uninfected Primary Secondary Early Latent H Known HIV positive

Can we use network information for health promotion? With complete information we could implement prophylaxis Can use it to change community norms Can use it to decide who should be the focus of interventions and how to segment the audience Source: T. Valente, 2002.

Network Data Types Network Data Type Instrument Result Survey Ego centric Sequential/ snowballing Sociometric Had sex the past 6 months? Provide first names of people had sex with? Contact trace recent sexual partners Ask for names of sexual partners Dichotomous Risk - Yes/NO Characterize sexual partners Identify specific individuals at risk Map sexual/ social network Locations Where had sex? Locate sites of opportunity & infection

Opinion Leaders: Individuals Who Receive the Most Nominations 33 32 4 10 28 13 14 30 1 6 15 8 2 3 5 29 9 7 11 12 31

Opinion Leaders Promoting Positive Behavior Change

Approaches to STD prevention: Reducing duration of infectiousness (Ro= βcd) Treatment of infected patients Treatment of sexual partners and suspects Screening to identify asymptomatic or latent infections (better for curable bacterial infections than for viral infections) Screening of pregnant women to prevent transmission to fetus/neonate Health services must be: high quality, accessible, affordable

Treatment of Exposed Persons (Sexual Partners) Patient or health care provider tells partner(s) that they need treatment Pre-emptive treatment Patient-provided partner therapy reduces rates of reinfection (azithromycin for chlamydial infection; Schillinger et al)

Screening Works But Has Limits

Region X Chlamydia positivity in 15 24 year old women in family planning clinics, 1988-2005 Percent Positive 15 15.1 AK 10 5 11.8 11.4 9.5 8.2 6.1 5.7 5.1 4.9 4.9 5.6 6.0 6.1 6.2 6.7 7.5 7.7 7.4 WA 0 OR ID 1988 89 90 91 92 93 94 95 96 97 98 99 2000 01 02 03 04 05 Unadjusted Adjusted

Rescreening vs. Routine Testing of STD Patients Rescreening hin a recent study, 12% of rescreened patients had an STD (gonorrhea, chlamydia, syphilis, or HIV) hanother large study found 15% rate of reinfection with chlamydia among young women within 4 months (Schillinger et al) Routine testing hgonorrhea positivity rates 0.1%-4.5% (young women) hchlamydia positivity rates 2.8%-15.8% (young women)

Source: D. Law, 2001

Applications of Kriging to STD Prevention Estimate rate of STD over a continuous surface Evaluate spatial spread of STD through time Refine geographic descriptions of core, adjacent, periphery, etc. in urban areas Identify locations for resource allocation and interventions

STD intervention in Mwanza trial STD reference center in Mwanza town Training of health workers in syndromic treatment, health education, condom promotion Regular supply of effective STD drugs Regular supervisory visits to health centers Periodic visits to villages by health educators to promote treatment for symptomatic STDs

Factors possibly affecting impact of STI treatment on HIV infection Stage of the HIV epidemic Differences in proportional incidence of curable STIs (i.e., bacterial versus viral) Differences in incidence of STIs most strongly associated with HIV transmission (gonorrhea, syphilis, herpes, other ulcerative diseases) Symptomatic versus asymptomatic infection

Epidemics Evolve

Extensive prevention alters the course & creates a moving target HIV cases in Thailand 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% Extramarital Female from Husband Sex worker Male from wife Male from sex worker Male sex with male Injection Drug User 0% 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: T. Brown, 2006

Preventing STDs and Sexually Transmitted HIV: Summary Three key factors maintain the circulation of STDs and sexually transmitted HIV Efficiency of transmission Sexual behaviors Duration of infectiousness Interventions can be designed to address these factors Biomedical interventions tend to address B Reducing risk behaviors and addressing access to quality health services remain critical in controlling STDs and preventing HIV spread