The clinical management of HIV disease relies on CD4+ T-lymphocyte (CD4) and plasma HIV-1 RNA (i.e., viral load) testing to guide the initiation of treatment and monitor care. The Department of Health and Human Services (DHHS) Panel on Antiretroviral Guidelines for Adults and Adolescents recommend CD4 count and viral load (VL) testing for a new patient during the initial visit and every 3 to 4 months after HIV diagnosis. Among patients who are clinically stable, CD4 may be monitored less frequently (every 6-12 months) [1]. The reporting of CD4 and VL results to health departments enhances local and national HIV case surveillance data and is used to identify cases, stage disease at diagnosis, and monitor disease progression. CD4 and VL data can also be used to determine entry and retention in care, measure viral load suppression, and assess unmet healthcare needs; however these measures are only as accurate as the quality and completeness of data reported.

AIDS was a reportable condition by the mid-1980s in all 50 states and the District of Columbia; however AIDS surveillance was limited to the collection of clinical data. With the advent of antiretroviral medications, which have helped HIV-infected persons live longer, and the increased availability of HIV tests, the national focus has shifted to integrate AIDS surveillance and surveillance of HIV infection. In the early 1990s, surveillance programs began collecting CD4 test results as part of routine surveillance activities. This was in part a result of the expansion of the AIDS case definition in 1993 to include an immunologic definition of AIDS; CD4 counts of less than 200 cells/mm³ or CD4 percentages less than 14% of total lymphocytes [2]. For the first time, the AIDS case definition could be met based exclusively on a positive HIV test and a low CD4 count or percentage. The most recent revision of the HIV surveillance case definition, in 2008, highlights the central role of CD4 results by using CD4 counts and percentages to define three stages of HIV infection, increasing in severity from stage 1 through stage 3 (AIDS) [3]. An unknown stage was built into the case definition to account for cases that do not have CD4 results or information on AIDS-defining conditions. The 2008 case definition also incorporated a new role for viral load test results in surveillance, as a detectable viral load became sufficient criteria for establishing a case for surveillance purposes [3].

The majority of U.S. states have policies or regulations that require laboratories to report CD4 and VL results to health departments. However, the level at which these laboratory results are reported varies within and across jurisdictions. In addition, barriers to maintaining complete and timely laboratory data in surveillance systems exist and may include the management of a large volume of reports and the receipt of paper vs electronic reports.

We conducted enhanced data collection in five state HIV surveillance programs for CD4 and VL laboratory test results and OIs among newly diagnosed HIV-infected persons. Using these data, we evaluated the completeness of CD4 and VL laboratory reports collected through surveillance, determined stage of disease at diagnosis, the most common OIs reported at diagnosis, and linkage to medical care within 3 months of diagnosis using laboratory tests as a marker for receipt of care.

Five surveillance jurisdictions (Colorado, Indiana, Louisiana, Michigan and New York [excluding New York City]) were selected to participate in the project as part of a competitive announcement. Analyses were based on data collected in the five states and reported to the Centers for Disease Control and Prevention (CDC) on HIV-infected persons aged 13 years or older diagnosed during April 2005 through March 2006, with the exception of New York, where persons were diagnosed from June 2005 through May 2006. Enhanced data collection of CD4 results, VL results, and OIs was conducted for all cases reported to the surveillance programs by 6 months after the end of the 12-month diagnosis period, except for Michigan where two out of three cases were sampled.

The five surveillance programs collected information according to routine surveillance procedures. Information was obtained on age, sex, race/ethnicity (white, black or African American, Hispanic or Latino, American Indian or Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, and multiple races), transmission category (men who have sex with men [MSM], injection drug use [IDU], MSM and IDU [MSM/IDU], high-risk heterosexual contact, and other), CD4 result, VL result, and OI information. All analyses adjusted for reporting delays and missing risk factor information [4, 5].

Methods for the enhanced collection of data included: a) updating data in the surveillance software from existing auxiliary laboratory databases, b) strengthening active surveillance efforts through increased medical record abstraction of laboratory data, c) identifying and collecting data from laboratories not previously reporting to HIV surveillance, and d) increasing the number of laboratory reports being sent to the surveillance programs electronically. Enhanced medical record abstraction was conducted up to 6 months following the 12-month diagnosis period using the Adult HIV Case Report Form. Electronic and paper-based laboratory results received up to 6 months after the 12-month diagnosis period and all medical record data were entered into the surveillance software and transferred to CDC as part of routine reporting of national HIV surveillance data. Although there may have been differences in how reports were transmitted and collected across the states (e.g., passive transmission of paper or electronic reports from laboratories vs medical record abstraction), the goal was for each health department to receive all HIV-related laboratory data.

We determined the distribution of stage of disease at diagnosis based on CD4 results or OIs within 3 months of diagnosis and the number and percentage of persons diagnosed with HIV who were linked to care within 3, 6, and 12 months based on CD4 and VL tests within these time frames. We also determined the number and percentage of persons diagnosed with HIV who were linked to care within 3 months by demographic and transmission categories. Finally, we determined the OIs diagnosed within 3 months of HIV diagnosis overall and by level of immune-suppression (CD4 count ≥200 cells/µL or ≥14%, and <200 cells/µL or <14%).

To assess the impact of enhanced data collection, we compared the project data on two measures (stage of disease and linkage to care) with (1) data from the same five states on cases diagnosed within 12 months prior to the initiation of the project (population A), and (2) data from 25 states on cases diagnosed during April 2005 through March 2006 (population B). The 25 states were selected for the analysis because they met the criteria outlined in the Technical Guidance for HIV/AIDS Surveillance Programs of at least 50% of newly diagnosed persons having an initial CD4 or VL result within 3 months of diagnosis reported to the national HIV surveillance system. Population B did not include any of the states with enhanced data collection.

SAS software version 9.1 (SAS Institute., Cary, NC) was used to perform univariate and bivariate analyses with the X² test. Chi-square p-values less than or equal to 0.05 were considered significant. To describe the completeness of data, we included all cases whether or not they were alive at the end of the observation period.

Of the 3,065 persons diagnosed during the 12-month diagnosis period in the five states, 9.6% were diagnosed with stage 1, 29.1% with stage 2, and 35.5% with stage 3 disease; 25.8% were stage unknown (Table 1A). The proportion of persons who had a result from at least one CD4 or VL increased as more time was allowed for tests to be performed and reported (Table 2A); however, the majority (78.5%) had at least one CD4 or VL test within 3 months of diagnosis. Additionally, we assessed the impact of CD4 and VL tests initiated at the time of testing. When we removed persons who had the same diagnosis date and CD4 or VL collection date, the percentage of persons with a CD4 or VL test within 3 months dropped to 69.3%, data not shown.

Blacks/African Americans were less likely to have a CD4 or VL test performed (73.6%) within 3 months of HIV diagnosis than whites (82.1%, p<0.001) (Table 3). Persons aged 13-29 were also less likely to have a CD4 or VL test performed (72.3%), as compared to persons aged 30 and above (30-39 years: 76.7%, 40-49 years: 83.2%, ≥50 years: 86.6%). Across the five states, there were significant differences in the percentage of persons with a result from at least one CD4 or VL test. The percentage of persons with a CD4 or VL test result also varied by the facility of diagnosis; 93.0% at adult HIV clinics to 52.9% at HIV counseling and testing sites. No significant differences were found between the percentage of persons with at least one CD4 or VL test result among males and females or across transmission categories.

The most frequently reported OI was Pneumocystis jiroveci pneumonia (PCP) (38.6%, 197/511), followed by Esophageal Candidiasis (16.4%, 84/511), and Wasting Syndrome (7.2%, 37/511) (Table 4). Of the 378 persons with an OI diagnosed within 3 months of HIV diagnosis, a larger percentage of OIs was found among persons with a CD4 result of <200 cells/µL or <14% (98.3%) as compared to persons with a CD4 result of ≥200 cells/µL or ≥14% (1.7%).

Enhanced data collection appears to have resulted in more complete data and therefore a more accurate measurement of stage of disease and linkage to care. The proportion of persons classified as stage unknown within 12 months of diagnosis was larger in the comparison populations, with 25.4% and 30.3% in Population A and B, respectively, compared to 17.0% among cases with enhanced data collection (Table 1). As the time between HIV diagnosis and the first CD4 test performed increased from 3 months to 6 and 12 months, the percentage of persons classified as stage unknown decreased in all populations.

Persons in the five states with enhanced data collection (Table 2A) were more likely to have at least one CD4 or VL test result at 12 months after diagnosis (86.4%) compared with persons in Population A (79.8%, p<0.001, Table 2B) or Population B (80.4%, p<0.001, Table 2C). Within 3 months of HIV diagnosis, 78.5% of persons with enhanced data collection had a result from at least one CD4 or VL test, as compared to 68.1% and 74.7% in Population A and B respectively. Persons in the five states were also more likely to have an OI reported to surveillance within 3 months of diagnosis during the time of enhanced data collection than during the previous year (Population A) (12.3% vs 10.6%, p=0.02), data not shown.

Enhanced collection of laboratory data through the utilization of data in existing databases, strengthened active surveillance efforts, the identification of labs not reporting to surveillance, and an increased number of electronic reports, improved the completeness of CD4, VL, and OI data in national HIV surveillance. Although no true standard was available for comparison, the percentages of persons diagnosed with HIV who had stage of disease assigned or who were linked to care were up to 10% higher with the enhanced data collection.

The collection of more complete data provides a more accurate picture of disease severity and linkage to care. Based on CD4 and VL laboratory tests performed within 3 months following diagnosis, we estimated that 78.5% of persons diagnosed in the five states were linked to care. This finding is somewhat higher than what was observed through an analysis of persons newly diagnosed with HIV infection in New York City, where the time between the first positive Western blot test and the first CD4 and/or VL result reported to surveillance was used to indicate the time period from the initial HIV diagnosis (non-AIDS) to the first HIV-related medical care visit [6]. Of 1,928 newly diagnosed persons in New York City in 2003, an estimated 63.7% initiated care within 3 months of diagnosis.

Overall, linkage to care needs strengthening to reach the goal outlined in the National HIV/AIDS Strategy of 85% of newly diagnosed patients linked to clinical care within 3 months of diagnosis [7]. We found a significantly higher proportion of persons with no CD4 or VL test results among blacks/African Americans as compared to whites which suggests that a racial disparity may exist among newly diagnosed persons who receive care. This finding is consistent with other studies that have documented an association between racial factors and disparities in HIV-related healthcare [8-11] and may be related to a complex interaction between health care, public health, and social factors [12]. These data also show that younger persons aged 15-29 were less likely to have a CD4 or VL test, as compared to persons aged 30 and above. The disparity seen in age may be a function of the large proportion of young adults that are uninsured [13], and serves to highlight the potential gap in accessing care between age groups. Additional interventions are needed to meet the goals of the National HIV/AIDS Strategy on linkage to care and the reduction of HIV-related disparities.

The initiation and frequency of laboratory testing and interpretation of laboratory results can be used to make inferences about the quality of health care that HIV-infected persons receive. Laboratory testing that occurs shortly after HIV diagnosis implies the successful linkage to health care. Serial CD4 and VL testing suggests utilization of ongoing care compared with a one-time visit. If the frequency of serial laboratory tests meet clinical management testing guidelines (e.g., every 3-4 months after initial diagnosis baseline measurement), it suggests receipt of higher quality care than less frequent laboratory testing would.

Persons without CD4 or VL testing following HIV diagnosis may represent persons with unmet healthcare needs. To address unmet healthcare needs and develop effective interventions, it is important to understand the barriers to accessing care such as lack of health insurance coverage [14, 15], unsuccessful patient notification of positive HIV test results [16-18], denial, poverty, mental illness, lack of transportation, and homelessness [19-21]. States have a responsibility to provide unmet health care need estimates to the Health Resources and Services Administration (HRSA), which oversees the Ryan White CARE Act. Surveillance data could be used to provide estimates of persons not in care and, therefore, assist states in meeting this federal reporting obligation.

OI data can be a useful indicator of clinical outcomes for HIV-infected persons in care. Based on OI data reported within 3 months following diagnosis, we estimated that 12.3% of persons diagnosed in the five states had at least one OI; PCP was the most frequently reported. A similar finding was observed through an analysis of a population-based survey of persons in HIV-related medical care in King County Washington, selected health districts in Louisiana, and the state of Michigan [22]. Of all HIV-infected persons in care in these areas in 1998, 11.3% (CI, 8.8–13.9) had at least one OI diagnosis and PCP was the most commonly diagnosed.

There are some limitations to consider when interpreting our findings. First, if laboratory data reported to surveillance are incomplete, the methods outlined may underestimate the prevalence of persons with access to medical care. In addition, the contribution of individuals who may be diagnosed in a jurisdiction reported to a state HIV program but then leave the state is unknown. From the local surveillance perspective, these individuals may not have begun care (i.e., evidence of lab testing) but received care after they moved out of jurisdiction. In Louisiana, the number of cases reported to surveillance and the completeness of laboratory data collected after August 2005 was impacted by Hurricane Katrina, as several clinical sites were closed and complete medical record abstraction was not possible.

In jurisdictions that have laws or regulations for laboratory reporting of all HIV-related tests, private and public laboratories must report all test results to the respective health departments. During the data collection period, changes in CD4 and VL reporting laws may have contributed to improvements seen in the completeness of laboratory reporting. For example, the laboratory reporting requirements were significantly broadened in New York State in June 2005, the start of the 12-month diagnosis period, to require the reporting of any VL result and all CD4 counts and percentages. Prior to this change, only detectable VLs and CD4 counts less than 500 cells/µL or percentages less than 29% were reportable. In July 2005, Michigan also revised their regulations to require the reporting of all CD4 and VL results.

In general, surveillance programs believe they are receiving the vast majority of HIV test results; but, unless a state is actively monitoring the number of laboratories reporting to them and the volume of reports, they cannot know for certain. Instances where laboratory testing may change (e.g., the provider decides to use another laboratory or the primary laboratory contracts with a new laboratory for specialized testing) may not necessarily be identified by a surveillance program. Active surveillance and medical record abstraction can help to obtain more complete laboratory data and identify laboratories that may not be reporting to HIV surveillance. The electronic transmission of HIV-related laboratory test results enhances the completeness, timeliness, and accuracy of reporting to surveillance programs [23]. Although many surveillance programs have received data electronically for years, many still need improvements or enhancements to implement and maintain the system. HIV surveillance programs are currently using software called eHARS that has the capacity for storage of all laboratory results and the ability for electronic laboratory data to be imported. To ensure that all laboratory results are reported to surveillance at the national level, state and local surveillance programs must ensure that all laboratory results, including those stored in auxiliary databases, are entered into eHARS software and transmitted to CDC.

The National HIV/AIDS Strategy proposes to reduce new HIV infections, increase access to care and improve health outcomes for people living with HIV, and reduce HIV-related disparities [8]. Data collected through the national HIV surveillance system can be used to monitor the outcomes of the national strategy; however the validity of these measures is dependent upon the completeness and quality of surveillance data. Monitoring outcomes such as linkage to care and racial/ethnic disparities among persons who are virally suppressed is particularly dependent upon complete reporting of all HIV-related laboratory results to surveillance.