Sensitivity and Specificity of the National Death Index for Multiple Causes of Death in People With HIV.

OBJECTIVES: Inaccuracies in cause-of-death information in death certificates can reduce the validity of national death statistics and result in poor targeting of resources to reduce morbidity and mortality in people with HIV. Our objective was to measure the sensitivity, specificity, and agreement between multiple causes of deaths from death certificates obtained from the National Death Index (NDI) and causes determined by expert physician review. METHODS: Physician specialists determined the cause of death using information collected from the medical records of 50 randomly selected HIV-infected people who died in San Francisco from July 1, 2016, through May 31, 2017. Using expert review as the gold standard, we measured sensitivity, specificity, and agreement. RESULTS: The NDI had a sensitivity of 53.9% and a specificity of 66.7% for HIV deaths. The NDI had a moderate sensitivity for non-AIDS-related infectious diseases and non-AIDS-related cancers (70.6% and 75.0%, respectively) and high specificity for these causes (100.0% and 94.7%, respectively). The NDI had low sensitivity and high specificity for substance abuse (27.3% and 100.0%, respectively), heart disease (58.3% and 86.8%, respectively), hepatitis B/C (33.3% and 97.7%, respectively), and mental illness (50.0% and 97.8%, respectively). The measure of agreement between expert review and the NDI was lowest for HIV (κ = 0.20); moderate for heart disease (κ = 0.45) and hepatitis B/C (κ = 0.40); high for non-AIDS-related infectious diseases (κ = 0.76) and non-AIDS-related cancers (κ = 0.72); and low for all other causes of death (κ < 0.35). CONCLUSIONS: Our findings support education and training of health care providers to improve the accuracy of cause-of-death information on death certificates.

Gender differences in causes of death among persons with HIV/AIDS in San Francisco, California, 1996-2013.

The objective was to examine gender differences in causes of death using the San Francisco HIV/AIDS and death registries. Data from San Francisco residents diagnosed with HIV/AIDS who died from 1996 to 2013 were analyzed. Age, race/ethnicity, year, and gender-adjusted standardized mortality ratios and Poisson 95% confidence intervals were calculated for underlying causes of death. Among the 6268 deaths, deaths attributed to drug use, mental disorders due to substance use, cerebrovascular disease, chronic obstructive pulmonary disease, renal disease, and septicemia were more likely among women than among men. Compared to the California population, women had elevated standardized mortality ratios for drug overdose (25.37), mental disorders due to substance abuse (27.21), cerebrovascular disease (2.83), chronic obstructive pulmonary disease (7.37), heart disease (2.37), and liver disease (5.54), and these were higher than the standardized mortality ratios for the men in our study. Men, but not women, had elevated standardized mortality ratios for suicide (2.70), undetermined intent (3.88), renal disease (2.29), and non-AIDS cancer (1.68) compared to population rates. Continued efforts to reduce HIV-related illnesses and an increased emphasis on diagnosing and treating preventable causes of death, including substance use, heart disease, and mental health disorders, are needed as part of comprehensive HIV care.

Disparities and Trends in Viral Suppression During a Transition to a "Test and Treat" Approach to the HIV Epidemic, San Francisco, 2008-2012.

BACKGROUND: In 2010, 2 years before national recommendations to provide antiretroviral therapy for HIV-infected persons regardless of CD4 count, the San Francisco Department of Public Health (SFDPH) implemented a "test and treat" strategy that expanded HIV testing and offered antiretroviral therapy to persons at all publicly funded HIV clinics. We used the SFDPH registry of HIV-infected persons to measure trends in the time to viral suppression of persons suppressed. METHODS: The Kaplan-Meier product limit method was used to assess trends in time from HIV diagnosis to viral suppression (HIV RNA <200 copies/mL) among persons diagnosed from 2008 to 2012. The annual proportion of persons living with HIV who were virally suppressed was measured for the years 2008 to 2012. Independent predictors of viral suppression were determined using proportional hazards regression for newly diagnosed cases and Poisson regression for living cases. RESULTS: Of the 2349 persons newly diagnosed, the median time from diagnosis to suppression decreased from 13 months in 2008 to 5 months in 2012 (P < 0.0001). Among the 11,787 persons living with HIV, the annual proportion suppressed increased from 69% in 2008 to 78% in 2012. African Americans, persons who inject drugs, persons without private insurance, and persons with nadir CD4 counts above 350 cells per cubic millimeter, were less likely to be virally suppressed. DISCUSSION: We found a decrease in time and overall population-level increases in viral suppression under a test and treat strategy and highlight sociodemographic disparities that may hamper the full benefit of this approach.

Difference in Survival between Housed and Homeless individuals with HIV, San Francisco, 2002-2011.

San Francisco (SF), a city with large HIV-infected and homeless populations, expanded supportive housing for HIV-infected people in 2007. We used the SF HIV/AIDS registry to compare survival between people who were homeless and who were housed at time of HIV diagnosis from 2002 through 2011. Housing status was obtained from medical records and deaths from local, state, and national vital registration. Survival was estimated using the Kaplan-Meier product-limit method. Ten percent of the 5,474 cases were homeless. Among people diagnosed between 2002 and 2006, the five-year survival was worse for people who were homeless at HIV diagnosis than for housed individuals (79% vs. 92%, p<.0001), but not for those diagnosed between 2007 and 2011 (92% vs. 93%, p=.3938). The improved survival among HIV-infected homeless people occurred during the time of increased supportive housing for this population. Our findings support including housing as an essential component of HIV care.

Changes in causes of death among persons with AIDS: San Francisco, California, 1996-2011.

The increased life expectancy among HIV-infected persons treated with combination antiretroviral therapy (ART), risk behaviors, and co-morbidities associated with ART place HIV-infected persons at risk for non-HIV-related causes of death. We used the San Francisco HIV/AIDS registry to identify deaths that occurred from January 1996 through December 2011. Temporal trends in AIDS- and non-AIDS-related mortality rates, the proportion of underlying and contributory causes of death, and the ratio of observed deaths in the study population to expected number of deaths among California men aged 20-79 (standardized mortality ratio [SMR]) of underlying causes of death were examined. A total of 5338 deaths were identified. The annual AIDS-related death rate (per 100 deaths) declined from 10.8 in 1996 to 0.9 in 2011 (p<0.0001), while the annual death rate from non-AIDS-related causes declined from 2.1 in 1996 to 0.9 in 2011 (p<0.0001). The proportion of deaths due to all types of heart disease combined, all non-AIDS cancers combined, mental disorders resulting from substance abuse, drug overdose, suicide and chronic obstructive pulmonary disease increased significantly over time. The SMRs for liver diseased decreased significantly over time but remained elevated. Our data highlight the importance of age-related causes of death as well as deaths from causes that are, at least in part, preventable.

Trends in early initiation of antiretroviral therapy and characteristics of persons with HIV initiating therapy in San Francisco, 2007-2011.

In 2010, the San Francisco Department of Public Health offered antiretroviral therapy (ART) to all its patients with human immunodeficiency virus (HIV) regardless of CD4 count. We assessed trends in time from diagnosis to ART initiation and factors associated with ART initiation among San Francisco residents living with HIV between 2007 and 2011. Time to ART initiation decreased among those diagnosed with higher CD4 count. ART initiation rate was significantly higher in recent years and lower among African Americans, men who have sex with men who also inject drugs, and persons aged ≥50 years. We found a trend toward early treatment. However, racial and social disparities persist.

Identifying barriers to HIV testing: personal and contextual factors associated with late HIV testing.

Late diagnosis of HIV is associated with increased morbidity, mortality, and health care costs. Despite the availability of HIV testing, persons continue to test late in the course of HIV infection. We used the HIV/AIDS case registry of San Francisco Department of Public Health to identify and recruit 41 persons who developed AIDS within 12 months of their HIV diagnosis to participate in a qualitative and quantitative interview regarding late diagnosis of HIV. Thirty-one of the participants were diagnosed with HIV because of symptomatic disease and 50% of the participants were diagnosed with HIV and AIDS concurrently. Half of the subjects had not been tested for HIV prior to diagnosis. Fear was the most frequently cited barrier to testing. Other barriers included being unaware of improved HIV treatment, free/low cost care, and risk for HIV. Recommendations for health care providers to increase early diagnosis of HIV include routine ascertainment of HIV risk behaviors and testing histories, stronger recommendations for patients to be tested, and incorporating testing into routine medical care. Public health messages to increase testing include publicizing that (1) effective, tolerable, and low cost/free care for HIV is readily available, (2) early diagnosis of HIV improves health outcomes, (3) HIV can be transmitted to persons who engage in unprotected oral and insertive anal sex and unprotected receptive anal intercourse without ejaculation and from HIV-infected persons whose infection is well-controlled with antiretroviral therapy, (4) persons who may be infected based upon these behaviors should be tested following exposure, (5) HIV testing information will be kept private, and (6) encouraging friends and family to get HIV tested is beneficial.

Estimation of unmet need for HIV primary medical care: a framework and three case studies.

Reauthorization of the Ryan White Comprehensive AIDS Resources Emergency Act requires that each jurisdiction estimate the number of people living with HIV/AIDS who have unmet need for care. Past assessments of unmet need have used various definitions of care, relied on qualitative evaluations, or examined nonrepresentative subpopulations. This article outlines a single, flexible framework designed to quantitatively estimate unmet need in varied settings. The framework adopts a definition of unmet need which focuses on HIV primary medical care (CD4 test, viral load test, or antiretroviral therapy in a 12-month period), employs a standard analytic structure to integrate population and care data, and allows use of locally available data. We report on three field tests (Louisiana, Atlanta, and San Francisco). The field tests suggest that the unmet need framework provides an approach which can be used by states and metropolitan areas to estimate the number of individuals with unmet need for HIV primary medical care.

Determinants of survival among HIV-infected chronic dialysis patients.

Over 100 HIV-infected patients have initiated chronic dialysis at San Francisco General Hospital (SFGH) since 1985. This study employed retrospective analysis to identify determinants of and trends in survival among HIV-infected patients who have initiated chronic dialysis at SFGH from January 1, 1985 to November 1, 2002 (n = 115). Cohort patient survival was compared with survival after an AIDS-opportunistic illness in all HIV-infected patients in San Francisco during the study period. Higher CD4 count (hazard ratio [HR], 0.86 per 50 cells/mm(3) increase; 95% confidence interval [CI], 0.80 to 0.93) and serum albumin (HR, 0.53 per 1 g/dl increase; CI, 0.36 to 0.78) at initiation of dialysis were strongly associated with lower mortality. Survival for those initiating dialysis during the era of highly active antiretroviral therapy (HAART) was 16.1 mo versus 9.4 mo for those initiating dialysis before this time, but this difference was not statistically significant. In adjusted analysis, only a non-statistically significant trend toward improved survival during the HAART era was noted (HR, 0.59; CI, 0.34 to 1.04). By comparison, survival for all HIV-infected patients after an AIDS-opportunistic illness in San Francisco increased from 16 mo in 1994 to 81 mo in 1996. The dramatic improvement in survival that has occurred since the mid-1990s for patients with HIV appears to be greatly attenuated in the sub-group undergoing dialysis. Although this may partly reflect confounding by race, injection drug use and HCV co-infection, future attempts to improve survival among HIV-infected dialysis patients should focus on barriers to the effective use of HAART in this group.

Trends in causes of death among persons with acquired immunodeficiency syndrome in the era of highly active antiretroviral therapy, San Francisco, 1994-1998.

To understand recent temporal trends in acquired immunodeficiency syndrome (AIDS) mortality in the era of highly active antiretroviral therapy (HAART), trends in causes of death among persons with AIDS in San Francisco who died between 1994 and 1998 were analyzed. Among 5234 deaths, the mortality rate for human immunodeficiency virus (HIV)-related or AIDS-related deaths declined after 1995 (P<.01), whereas the mortality rate for non-HIV- or non-AIDS-related deaths remained stable. The proportion of deaths of persons with AIDS associated with septicemia, non-AIDS-defining malignancy, chronic liver disease, viral hepatitis, overdose, obstructive lung disease, coronary artery disease, and pancreatitis increased (P<.05). The standardized mortality ratio was high for these causes in both pre- and post-HAART periods, except for pancreatitis, a possible complication of HAART, which demonstrated an increasing standardized mortality ratio trend after 1996. With increasing AIDS survival, prevention of chronic diseases, assessment of long-term toxicity from HAART, and surveillance for additional causes of mortality will become increasingly important.