HIV in Alameda County Annual lepidemiology i Dt Data Presentation ti to the CCPC July 22, 2015 Prepared By Richard Lechtenberg & Neena Murgai HIV Epidemiology and Surveillance Unit HIV in Alameda County: An Overview 220 new diagnoses, on average, in 2011 2013 14.3 per 100,000 population per year 70.4% 82.5% linked to care within 90 days 5,649 people living with HIV disease (PLHIV) at yearend 2013 363.7 per 100,000 population 70.1% received any HIV care in 2013 55.7% were virally suppressed at last measurement in 2013 7/17/2015 1
On the agenda 1. Diagnosis and Prevalence A. By Demographics B. By Social Determinants of Health 2. The Continuum of HIV Care 1. Identify Epi 101 Recap A. the population/denominator B. the sub population/numerator 7/17/2015 2
Most Measures = What defines the subpopulation? # # What defines the population?? Numerator Denominator May be called different names Fraction Proportion Percentage Rate May be expressed in different ways ½ 0.5 50% 5per 100,000 (instead of 0.005%) To understand the measure you need to understand both! Surveillance Database Right for PLHIV who are patients at Clinic A Right for all PLHIV (bigger denominator) Right for PLHIV who are patients at Clinic B 7/17/2015 3
1. Identify Epi 101 Recap A. the population/denominator B. the sub population/numerator 2. Beware A. the prosecutor s fallacy! The Prosecutor s Fallacy Defined The assumption that: The chances of A among B = The chances of B among A 7/17/2015 4
The Prosecutor s Fallacy Illustrated The fallacy: Most striped squares are grey. Most grey squares are striped. Chances that a square is grey among striped squares: 4/9 = 44.4% Chances that a square is striped among grey squares: 4/25 = 16% 1. Identify Epi 101 Recap A. the population/denominator B. the sub population/numerator 2. Beware A. the prosecutor s fallacy! B. association vs. causation 7/17/2015 5
Confounding: An example Question: Is Alaska s mortality rate different than Florida s? Mortality rate: 399 per 100,000000 50% 40% 30% 20% 10% 0% State <5 5 19 20 44 45 64 >64 Age The observed difference in mortality rates is confounded by age!? Not fair to compare them directly! 50% 40% 30% 20% 10% 0% Death <5 5 19 20 44 45 64 >64 Mortality rate: 1,069 per 100,000 SOURCE: http://sphweb.bumc.bu.edu/otlt/mph Modules/EP/EP713_StandardizedRates/EP713_StandardizedRates3.html 1. Identify Epi 101 Recap A. the population/denominator B. the sub population/numerator 2. Beware A. the prosecutor s fallacy! B. association vs. causation C. small numbers 7/17/2015 6
Less Data Less Confidence in What the Data Says A caveat: Gender identity is not reliably captured in surveillance data because only sex assigned at birth is routinely captured in the medical record. To avoid underestimating the burden of HIV in the transgender community, breakdowns will be provided by sex assigned at birth. 7/17/2015 7
On the agenda 1. Diagnosis and Prevalence A. By Demographics B. By Social Determinants of Health 2. The Continuum of HIV Care 7/17/2015 8
HIV in Alameda County by the Numbers # new diagnoses, regardless of stage #new AIDS diagnoses # of PLHIV (at year end) 2010 238 74 5,465 2011 206 67 5,560 2012 238 88 5,585 2013 215 64 5,649 On the agenda 1. Diagnosis and Prevalence A. By Demographics B. By Social Determinants of Health 2. The Continuum of HIV Care 7/17/2015 9
Annual Diagnosis Rate per 100,000 Trends in New HIV Diagnosis Rates by Sex, Alameda County, 2006 2013 30 20 10 Sex All Male Female 2006-2008 2007-2009 2008-2010 2009-2011 2010-2012 2011-2013 NOTE: (1) Rates are 3 year average annual rates. (2) Sex refers to sex assigned at birth. (3) Grey areas are 95% confidence bands. Trends in New HIV Diagnosis Rates by Race/Ethnicity, Alameda County, 2006 2013 All races African American White Race/Ethnicity Hispanic/Latino API Annual Diagnosis Rate per 100,000 60 40 20 2006-2008 2007-2009 2008-2010 2009-2011 2010-2012 2011-2013 NOTE: (1) Rates are 3 year average annual rates. (2) Grey areas are 95% confidence bands. NOT SHOWN: Other/unknown race (rates not calculable). 7/17/2015 10
Trends in New HIV Diagnosis Rates by Sex & Race/Ethnicity, Alameda County, 2006 2013 Race/Ethnicity All races African American White Hispanic/Latino API Male Female Ann nual Diagnosis Rate per 100,000 90 60 30 2006-2008 2007-2009 2008-2010 2009-2011 2010-2012 2011-2013 2006-2008 2007-2009 2008-2010 2009-2011 2010-2012 2011-2013 NOTE: (1) Rates are 3 year average annual rates. (2) Sex refers to sex assigned at birth. (3) Grey areas are 95% confidence bands. NOT SHOWN: Other/unknown race (rates not calculable). Trends in New HIV Diagnosis Rates by Age & Race/Ethnicity, Alameda County, 2006 2013 NOTES: Analysis done by Poisson regression assuming a linear effect of time (on the log scale) and allowing for all 2 way interactions. 7/17/2015 11
7/17/2015 12
Key takeaways: Overall, diagnosis rates have decreased since 2006 The most notable declines have occurred among African American women African Americans and whites in their 30s and 50s Although increases have been seen in API in their 20s and 40s, rates among them remain low compared to other groups On the agenda 1. Diagnosis and Prevalence A. By Demographics B. By Social Determinants of Health 2. The Continuum of HIV Care 7/17/2015 13
Social Determinants of Health Factors such as Poverty Unemployment Education level Can have individual as well as community effects E.g., an individual s id health may be impacted dby their own wealth as well as that of their community Diagnosis Rates by Neighborhood Poverty Level, Alameda County 2010 2011 9.8 17.2 16.0 25.5 27.0 % of Census Tract Residents Living Below Poverty 36.1-100% 27.8-36.0% 19.7-27.7% 11.4-19.6% 5.1-11.3% 0.0-5.0% 5.1 0 10 20 30 Annual Diagnosis Rate per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. 7/17/2015 14
Prevalence by Neighborhood Poverty Level, Alameda County, Year End 2011 224.4 290.8 398.2 488.7 612.0 597.3 723.7 % of Census s Tract Residents Living Below Poverty 45.7-51.5% 38.8-45.6% 31.5-38.7% 24.6-31.4% 16.5-24.5% 10.1-16.4% 4.5-10.0% 0.0-4.4% 137.5 0 200 400 600 800 Prevalence per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. Diagnosis Rates by Neighborhood Unemployment, Alameda County 2010 2011 21.2 22.9 74 7.4 11.4 18.0 % of Census Tract Residents who are Unemployed 22.3-39.7% 17.7-22.2% 12.7-17.6% 8-12.6% 3.9-7.9% 0.0-3.8% 92 0 10 20 30 Annual Diagnosis Rate per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. 7/17/2015 15
Prevalence by Neighborhood Unemployment, Alameda County, Year End 2011 233.22 306.3 576.0 773.9 638.0 498.4 456.3 % of Census s Tract Residents who are Unemployed 29.4-39.7% 26.6-29.3% 22.3-26.5% 18.1-22.2% 13.3-18.0% 8.5-13.2% 4.2-8.4% 0.0-4.1% 0 227 6 300 600 900 Prevalence per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. Diagnosis Rates by Neighborhood Education Level, Alameda County 2010 2011 9.7 12.7 13.0 17.8 20.7 23.7 % of fcensus Tract Residents with Less than a High School Education 38-52.3% 31.1-37.9% 24.2-31.0% 17.3-24.1% 10.4-17.2% 4.5-10.3% 0.0-4.4% 7.0 0 10 20 30 Annual Diagnosis Rate per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. 7/17/2015 16
Prevalence by Neighborhood Education Level, Alameda County, Year End 2011 248.1 227.6 347.9 412.4 542.1 519.8 441.2 % of Census Tract 461.9 Residents with Less than a High School Education 45-52.3% 38.6-44.9% 33.2-38.5% 27.1-33.1% 19.6-27.0% 11.3-19.5% 4.9-11.2% 0048% 0.0-4.8% 0 200 400 600 Prevalence per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. Diagnosis Rates by Neighborhood Insurance Status, Alameda County 2010 2011 21.7 20.4 6.7 13.3 17.9 % of Census Tract Residents who are Uninsured 29.3-50.0% 22.9-29.2% 16.6-22.8% 10-16.5% 4.6-9.9% 0.0-4.5% 7.8 0 10 20 30 Annual Diagnosis Rate per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. 7/17/2015 17
Prevalence by Neighborhood Insurance Status, Alameda County, Year End 2011 409.1 528.5 652.1 185.6 289.0 390.3 547.77 % of Census s Tract Residents who are Uninsured 33.7-36.6% 28.6-33.6% 23.7-28.5% 18.7-23.6% 13.8-18.6% 8.5-13.7% 4-8.4% 0.0-3.9% 140.8 0 200 400 600 Prevalence per 100,000 NOTE: (1) Bar widths proportional to the fraction of the underlying population in the category (with the exception categories comprising <2% of the population, for which bars are enlarged for visibility). (2) A clustering algorithm was used to determine optimal category cut points. (3) The dashed line indicates the overall rate for the county as a whole. Prevalence by Neighborhood Poverty Level and Race/Ethnicity, Alameda County, 2011 revalence per 100,000 Pr 2000 1500 1000 500 0 Race/Ethnicity African American API Hispanic/Latino White 0% 10% 20% 30% 40% 50% % of Census Tract Residents Living Below Poverty NOTE: A clustering algorithm was used to determine optimal category cut points. EXCLUSIONS: N=24 PLHIV <18 years of age. NOT SHOWN: N=183 PLHIV with other or unknown race/ethnicity. 7/17/2015 18
Prevalence by Neighborhood Unemployment and Race/Ethnicity, Alameda County, 2011 Pr 1500 1000 500 revalence per 100,0001500 Race/Ethnicity African American API Hispanic/Latino White 0 0% 10% 20% 30% % of Census Tract Residents who are Unemployed NOTE: A clustering algorithm was used to determine optimal category cut points. EXCLUSIONS: N=24 PLHIV <18 years of age. NOT SHOWN: N=183 PLHIV with other or unknown race/ethnicity. Prevalence by Neighborhood Education Level and Race/Ethnicity, Alameda County, 2011 2000 Race/Ethnicity African American API Hispanic/Latino White revalence per 100,000 Pr 1500 1000 500 0 0% 10% 20% 30% 40% 50% % of Census Tract Residents with Less than a High School Education NOTE: A clustering algorithm was used to determine optimal category cut points. EXCLUSIONS: N=24 PLHIV <18 years of age. NOT SHOWN: N=183 PLHIV with other or unknown race/ethnicity. 7/17/2015 19
Prevalence by Neighborhood Insurance Status and Race/Ethnicity, Alameda County, 2011 2000 Race/Ethnicity African American API Hispanic/Latino White revalence per 100,000 Pr 1500 1000 500 0 0% 10% 20% 30% % of Census Tract Residents who are Uninsured NOTE: A clustering algorithm was used to determine optimal category cut points. EXCLUSIONS: N=24 PLHIV <18 years of age. NOT SHOWN: N=183 PLHIV with other or unknown race/ethnicity. Key takeaways: Diagnosis rates and prevalence generally increase with increasing neighborhood poverty and unemployment and with decreasing rates of insurance and education These associations appear to vary by race/ethnicity Appear to be less prominent among Latinos 7/17/2015 20
On the agenda 1. Diagnosis and Prevalence A. By Demographics B. By Social Determinants of Health 2. The Continuum of HIV Care The Continuum of HIV Care in Alameda County Among N=669 new diagnoses in 2010 2012* Among N=5,370 PLHIV in Alameda Co. for the entirety of 2013** 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 82.5% 70.4% 70.1% 55.7% 44.5% Linked Retained Virally Suppressed *Of 682 total diagnoses, 13 died within 90 days and are excluded from analysis **Of 5,585 PLHIV at year end 2012, 42 are known to have died and an additional 173 to have moved out of Alameda County in 2013 1) Linkage defined as having a reported CD4 or VL ordered within 90 days or less of diagnosis; 2) Retention calculated using labs ordered in 2013; 3) Viral suppression defined as most recent VL in 2013 < 200 copies/ml 7/17/2015 21
Linkage to HIV Care in 90 days of Diagnosis by Sex, Alameda County, 2010 2012 Including labs on the date of diagnosis? No Yes All (N=669) 70.4% 82.5% Male (N=573) 71.0% 82.9% Female (N=96) 66.7% 80.2% 0% 25% 50% 75% 100% Percent linked in 90 days or less NOTES: (1) Linkage defined as having CD4 and viral load tests. (2) Sex refers to sex assigned at birth. EXCLUSIONS: N=13 patients who died within 90 days of diagnosis. Linkage to HIV Care in 90 days of Diagnosis by Race/Ethnicity, Alameda County, 2010 2012 Including labs on the date of diagnosis? No Yes All races (N=669) African American (N=277) White (N=161) Hispanic/Latino (N=149) 70.4% 70.4% 67.1% 73.2% 82.5% 81.2% 80.7% 86.6% API (N=65) 70.8% 83.1% 0% 25% 50% 75% 100% Percent linked in 90 days or less NOTE: Linkage defined as having CD4 and viral load tests. EXCLUSIONS: N=13 patients who died within 90 days of diagnosis. NOT SHOWN: N=17 patients with other/unknown race. 7/17/2015 22
Linkage to HIV Care in 90 days of Diagnosis by Age at Diagnosis, Alameda County, 2010 2012 All ages (N=669) 13-19 (N=34) 20-29 (N=188) 30-39 (N=158) 40-49 (N=182) 50-59 (N=84) 60 & over (N=22) Including labs on the date of diagnosis? No Yes 58.8% 70.2% 67.1% 70.4% 70.6% 70.3% 72.7% 82.5% 82.4% 81% 82.4% 81% 81.8% 90.5% 0% 25% 50% 75% 100% Percent linked in 90 days or less NOTE: Linkage defined as having CD4 and viral load tests. EXCLUSIONS: N=13 patients who died within 90 days of diagnosis. NOT SHOWN: N < 5 patients aged 0 12. Engagement in HIV Care in 2013 by Sex Among PLHIV at Year End 2012, Alameda County Measure 1+ visit 2+ visits 90+ days apart All (N=5,370) 44.5% 70.1% Male (N=4,402) 45.6% 70.3% Female (N=968) 39.5% 69.1% 0% 20% 40% 60% 80% NOTE: (1) Care visits defined as having CD4 and viral load tests. (2) Sex refers to sex assigned at birth. EXCLUSIONS: PLHIV at year end 2012 who died (N=42) or moved (N=173) during 2013. 7/17/2015 23
Engagement in HIV Care in 2013 by Race/Ethnicity Among PLHIV at Year End 2012, Alameda County Measure 1+ visit 2+ visits 90+ days apart All races (N=5,370) African American (N=2,280) White (N=1,787) Hispanic/Latino (N=906) 44.5% 40.6% 48.8% 44.2% 70.1% 68.2% 72.9% 67.0% 51.1% API (N=231) 78.8% 0% 20% 40% 60% 80% NOTE: Care visits defined as having CD4 and viral load tests. EXCLUSIONS: PLHIV at year end 2012 who died (N=42) or moved (N=173) during 2013. NOT SHOWN: N=166 PLHIV with other/unknown race/ethnicity. Engagement in HIV Care in 2013 by Age Among PLHIV at Year End 2012, Alameda County Measure 1+ visit 2+ visits 90+ days apart All ages (N=5 5,370) 44.5% 70.1% 13-19 (N=26) 42.3% 73.1% 20-29 (N=392) 41.3% 71.4% 30-39 (N=782) 38.1% 65.3% 40-49 (N=1,646) 43.2% 68.7% 50-59 (N=1,664) 47.4% 72.7% 7% 60 & over (N=848) 48.7% 71.2% 0% 20% 40% 60% 80% NOTE: (1) Care visits defined as having CD4 and viral load tests. (2) Age is at year end 2011. EXCLUSIONS: PLHIV at year end 2012 who died (N=42) or moved (N=173) during 2013. NOT SHOWN: N=12 PLHIV aged 0 12. 7/17/2015 24
Most Recent Viral Load in 2013 by Sex Among PLHIV at Year End 2012, Alameda County Virologic Status Undetectable Suppressed Unsuppressed Only CD4 reported No CD4s or VLs reported All (N=5,370) Male (N=4,402) Female (N=968) 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% NOTE: VL categories are defined as follows: Undetectable = 0 75 copies/ml, Suppressed = 76 199, Unsuppressed = 200+; Sex refers to sex assigned at birth EXCLUSIONS: PLHIV at year end 2012 who died (N=42 ) or moved (N=173) during 2012 Most Recent Viral Load in 2013 by Race/Ethnicity Among PLHIV at Year End 2012, Alameda County Virologic Status Undetectable Suppressed Unsuppressed Only CD4 reported No CD4s or VLs reported All races (N=5,370) African American (N=2,280) White (N=1,787) Hispanic/Latino (N=906) API (N=231) 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% NOTE VL categories are defined as follows: Undetectable = 0 75 copies/ml, Suppressed = 76 199, Unsuppressed = 200+ EXCLUSIONS: PLHIV at year end 2012 who died (N=42 ) or moved (N=173) during 2012 NOT SHOWN: N=166 PLHIV with other/unknown race/ethnicity 7/17/2015 25
Most Recent Viral Load in 2013 by Age Among PLHIV at Year End 2012, Alameda County Virologic Status Undetectable Suppressed Unsuppressed Only CD4 reported No CD4s or VLs reported All ages (N=5,370) 13-19 (N=26) 20-29 (N=392) 30-39 (N=782) 40-49 (N=1,646) 50-59 (N=1,664) 60 & over (N=848) 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% NOTES: VL categories are defined as follows: Undetectable = 0 75 copies/ml, Suppressed = 76 199, Unsuppressed = 200+ EXCLUSIONS: PLHIV at year end 2012 who died (N=42 ) or moved (N=173) during 2012 NOT SHOWN: N=12 PLHIV aged 0 12 Key takeaways: Linkage Lowest among women and whites Highest among those in their 50s Retention in (any) care Lower among African Americans, Latinos, and those in their 30s in continuous care Lower among women, as well as the above groups Viral Suppression Lower among women, African Americans, and Latinos Increasingly common in older age groups 7/17/2015 26
Thank you! Contact Richard.Lechtenberg@ACgov.org withany questions or comments 7/17/2015 27