Secondary Logo

Share this article on:

Review of Medical Encounters in the 5 Years Before a Diagnosis of HIV-1 Infection: Implications for Early Detection

Klein, Daniel*; Hurley, Leo B.; Merrill, Deanna; Quesenberry , Charles P. Jr.CHAIR (Consortium for HIV/AIDS Interregional Research)

JAIDS Journal of Acquired Immune Deficiency Syndromes: February 1st, 2003 - Volume 32 - Issue 2 - p 143-152
Clinical Science

Early detection of HIV infection improves prognosis and reduces transmission, but 30%–40% of cases are diagnosed late. A comprehensive and systematic review of medical encounters before diagnosis has not been done. This study reviews 5 years of medical encounters before the diagnosis of HIV infection in members of a large managed care organization where access to care is reasonably good. Patient characteristics, HIV risk factors, and clinical events preceding diagnosis were examined and tested for association with late diagnosis (CD4 cell count of <200/μL at diagnosis). Of 440 HIV-infected patients, 62% had CD4 cell counts of <350/μL, 43% had CD4 cell counts of <200/μL, and 18% had CD4 cell counts of <50/μL at diagnosis. Twenty-six percent of all patients had risks documented >1 year before diagnosis. Only 22% of patients had one of eight clinical indicators suggested in the literature as reasons to test for HIV >1 year before diagnosis. In multiple logistic regression, older age, male sex, race, risk group, no prior HIV testing, physician-initiated testing, and having any of eight clinical indicators before diagnosis were each associated with late diagnosis (p ≤ .05). Late diagnosis remains a challenge despite good access to care. In our setting, effective risk assessment before symptoms arise offers greater potential for raising the mean CD4 cell count at diagnosis than does increased awareness of selected HIV-associated clinical prompts.

*Kaiser Permanente Medical Center, Hayward, and †Kaiser Foundation Research Institute, Division of Research, Oakland, California; and ‡The Kaiser Permanente, Consortium for HIV/AIDS Interregional Research, Denver, Colorado, U.S.A.

This work was supported by a grant from the Garfield Memorial Fund to establish CHAIR. Agouron Pharmaceuticals and the DuPont Pharmaceutical Company provided unrestricted funding to support the HEDS UP Project.

Address correspondence to Dr. Daniel Klein, 27400 Hesperian Boulevard, Hayward, CA 94545; e-mail: Address reprint requests to Leo Hurley, Kaiser Foundation Research Institute, Division of Research, 2000 Broadway, Oakland, CA 94612; e-mail:

Highly active antiretroviral therapy has dramatically reduced the morbidity and mortality attributed to HIV infection and has changed the view of HIV disease in the Western world from that of an inevitably progressive and fatal infection to more of a chronic manageable condition (1–3). The potential benefits of early diagnosis include reduced transmission of infection, preservation of immune function, and prolongation of disease-free survival (4). Recent reports have found an inverse relationship between CD4 cell count at first diagnosis of HIV infection and virologic suppression once a highly active antiretroviral therapy regimen has begun (5). Other studies report that patients initiating therapy at a higher CD4 cell count have a better prognosis than patients who initiate therapy at a lower CD4 cell count (6–8). Current national guidelines for the treatment of HIV infection recommend that providers consider initiating highly active antiretroviral therapy for HIV-infected patients with CD4 cell counts of <350/μL, and treatment is strongly recommended for patients with CD4 cell counts of <200μ/L (9).

Despite the potential benefits of early detection, a large proportion of newly diagnosed HIV-infected individuals are being identified late in the course of their infection, after profound immune depletion has occurred (10–13). It has been reported from various settings that 30%–40% of persons with HIV infection have already reached the point of immunologic AIDS—i.e., CD4 lymphocyte counts of <200/μL—by the time HIV infection is diagnosed (11,13–17). We also know from previously reported studies that after primary infection CD4 cell counts decrease at an estimated rate of 60–100/μL each year (18–20). This estimated rate of decline in the presence of HIV infection suggests that most HIV-infected individuals with CD4 cell counts <200/μL at diagnosis have been infected for a decade or more.

Recommendations by the Centers for Disease Control and Prevention and others encourage screening more routinely for HIV infection, particularly in areas of high prevalence (21–23). Yet despite these recommendations, there is little indication that either the mean CD4 cell count or the proportion of patients with CD4 cell counts of <200/μL at diagnosis have improved over time (10,11,14,15,17,24). In addition, several years after the Centers for Disease Control and Prevention issued recommendations for universal prenatal screening, it is estimated that only one half of pregnant women in the United States are tested for HIV (25–27).

There are several barriers to testing for HIV. Risk factors for acquiring HIV infection have been well established, and ways to improve HIV risk assessment skills have been described (28). However, limited time for provider/patient interaction and age-old taboos can prevent frank discussions of sexual practices between provider and patient (29,30). The mechanics of testing for HIV are relatively simple, and rapid testing may make it easier for some populations to be tested and made aware of their results (31). However, testing for HIV typically involves informed consent, appropriate specimen collection, and some degree of posttest counseling (23). Studies have evaluated these mechanics as well as other possible barriers to testing (15,32). Improved counseling strategies may make testing for HIV more convenient and acceptable (31,33). Other work has been done to examine the social stigma and shame associated with HIV testing and infection, the denial of risk or an altered sense of risk, and the perception of life after testing positive for HIV (34–37). These factors no doubt contribute to delayed testing for and identification of HIV infection. Access to health care is likely an additional barrier to early diagnosis for some populations (32), but the prevalence of late testing among populations with good access to care has not been described.

There are also clinical factors that make early detection of HIV infection difficult. The signs and symptoms of acute HIV infection are not easily distinguished from other more common viral illnesses and can be dismissed by patients and providers alike (22,38). Aside from the acute phase of infection, several specific clinical conditions have been suggested as potential indicators of longer-term underlying infection, and it has been suggested that these indicators may be important prompts for testing (22). However, to our knowledge, no study has been performed to describe the frequency with which such prompts occur in the period preceding the diagnosis of HIV infection or to describe the extent to which CD4 cell depletion has already occurred when patients present with these indicators. Separately, aside from general associations with male gender and increased age, attempts to identify predictors of late testing have revealed few certainties (11–13,15).

The aims of this study, the HIV Early Detection Study of Unrecognized Positives, were to gain further insight into the problem of late testing for HIV in a setting where the problem of access to care is minimized and to potentially shed light on possible strategies for bringing about earlier detection. Our approach was to conduct chart reviews to examine all of the medical encounters between patients newly diagnosed with HIV infection and their medical care system in the years preceding diagnosis. Our setting, a closed, prepaid medical system with large numbers of newly identified cases of HIV infection, is one of few settings in which such a study is possible. This may explain why our search of the literature found no other work in this area.

Our primary research questions were as follows: Does ready access to medical care result in earlier detection of undiagnosed HIV-1 infection than what has been reported for general populations? How often are HIV risk factors and clinical indicators of possible post–acute HIV infection documented before diagnosis, and are they present before severe CD4 cell depletion occurs (in essence, what is the potential benefit of more vigilant attention to signs of possible infection)? What patient characteristics or other factors that predate a diagnosis of HIV infection are associated with lower CD4 cell counts at diagnosis?

Back to Top | Article Outline


Study Setting and Study Population

The Kaiser Permanente Medical Care Program (KP) is one of the largest integrated health care delivery systems in the United States. KP seeks to provide competitively priced high-quality health care services to its members. For a fixed monthly fee, members have easy access to preventative, routine, emergency, and hospital care that is coordinated by a personal physician. KP encourages its members to seek care before a medical problem becomes serious. KP divisions in seven geographic areas (Mid-Atlantic, Ohio, Colorado, Northern California, Southern California, Northwest, and Hawaii) and Group Health Cooperative of Washington State (GHC) currently provide care for a combined total of >8 million members. Among those members are >12,000 identified and currently active HIV-positive individuals. In addition, ≈500 new cases of HIV infection are diagnosed each year. KP and GHC are “closed” care delivery systems, where members typically receive all of their medical care from health plan providers at health plan–owned and –operated facilities. Using internal case registries and clinic records, all adult KP and GHC members newly diagnosed with HIV infection in 1998 and having at least 12 months of membership before diagnosis were identified as eligible for this study. The Institutional Review Board of the Kaiser Foundation Research Institute, Oakland, California, and local Institutional Review Boards at sites not served by the Oakland Institutional Review Board approved the study.

Back to Top | Article Outline

Data Collection

A data collection form and set of instructions were designed for medical record reviewers to use in the abstraction of relevant data from patient charts. The reviewers for this study were HIV-experienced medical record analysts or HIV nurse clinicians with experience in research protocols. They were recruited from participating study sites and attended a training program involving practice reviews of HIV-positive cases first diagnosed in 1999 and thus not eligible for the study. New patients with HIV infection coming into care in 1998 were identified in each administrative region, and the medical records for these patients were assembled for review. After verifying that a patient was eligible for inclusion in the study (newly diagnosed in 1998 and having 12 months of membership preceding diagnosis), the reviewer was instructed to review the medical record for up to 5 years before the diagnosis of HIV infection. Any event or encounter noted in the chart and listed in the Medical Encounter Catalog (MEC) was captured. The MEC was developed specifically for this study and is described below. The reviewers were also instructed to review and collect data from the charts for up to 6 months after diagnosis to capture additional information about risk factors that may not have been documented before HIV infection was diagnosed and to collect results of testing for the first CD4 cell count and viral load.

A comprehensive and detailed MEC was compiled and listed >80 medical conditions possibly indicative of underlying HIV infection. The MEC was developed based on reports in the literature (21,22) and on input from KP staff clinicians experienced in the diagnosis of HIV infection. The medical conditions in the MEC ranged from headache and mononucleosis-like syndromes to sexually transmitted infections and AIDS-defining conditions. MEC entries also included behavioral and social risk factors for HIV infection (such as sexual orientation and practices, drug and alcohol use, incarceration, etc.). Also included in the MEC were relevant laboratory tests and procedures (e.g., tests for sexually transmitted infections and mononucleosis and chest radiography) and any reference to HIV testing, including mention of prior negative tests for HIV, recommendations and refusals to test for HIV, and events and circumstances at the time of the positive HIV test. Unique numeric codes were assigned to each MEC entry to facilitate data collection and entry.

The data elements captured in the course of the medical record review included patient characteristics and circumstances of each potentially HIV-related encounter (e.g., date, inpatient or outpatient, phone or face-to-face contact, department, and provider type). The reviewer also recorded the numeric code corresponding to the MEC entry found in the chart. The reviewer was also instructed to describe the nature of each encounter using an open-ended free-text field on the data collection form. Chart review began in September 1999 and continued through October 2000. Completed data entry forms from all sites were sent to the Kaiser Permanente Division of Research in Oakland, California, for editing, key data entry, and conversion into analysis data sets.

Back to Top | Article Outline

Data Analysis

Descriptive univariate summaries were prepared for patient age, sex, race/ethnicity, HIV risk group, geographic region, duration of health plan membership before HIV infection was diagnosed, whether the patient underwent HIV testing in the 5 years before testing positive, prior refusal to test, denial of risk factors, and who initiated the test at diagnosis (patient or provider). CD4 cell counts, when measured within 60 days of diagnosis, were also summarized. The frequency of individual and grouped MEC entries and the duration between those entries and the eventual diagnosis of HIV infection were examined on an exploratory basis. Of particular interest were the presence/absence of three major HIV risk indicators—male-to-male sex, injection drug use, and sexually transmitted infections (suspected or confirmed gonorrhea, syphilis, chlamydial infection, herpes simplex, or genital warts)—and the presence/absence of eight clinical indicators of possible HIV infection—oral infection, pneumonia, unexplained fever, herpes zoster, seborrheic dermatitis, night sweats, unexplained weight loss, and lymphadenopathy.

For descriptive purposes, the crude proportions of total patients having risk indicators or clinical indicators in two periods “1 to 5 years before diagnosis” and “anytime (0–5 years) before diagnosis” were calculated. In addition, the rates at which these indicators occurred in these two periods were calculated, taking person-time of observation into account. Person-time of observation was defined as the period starting with date of first membership or the date 5 years in advance of being diagnosed with HIV infection, whichever came last, and ending with the date HIV infection was first detected (date of first positive results of western blotting).

In bivariate analyses, patient characteristics (including the three major risk indicators listed above) and a variable representing the time between the first date any of the eight clinical indicators were diagnosed and the date the patient first tested positive (three strata, including a stratum for patients who had no such symptoms prior to diagnosis) were tested for association with CD4 cell count at diagnosis (<200 vs. ≥200/μL) using Pearson χ2 and Fisher exact tests where appropriate. Variables found to be associated with CD4 cell count at diagnosis at the .10 level in bivariate analyses and variables that were a priori thought to be possibly related to CD4 cell count at diagnosis (e.g., size of site: <50 new cases of HIV infection diagnosed in 1998 vs. ≥50 new cases) were entered into a multiple logistic regression model to identify which factors might be independent predictors of having a CD4 cell count of <200/μL at diagnosis. Number of years of membership before being diagnosed with HIV infection was included in the model (five strata) to adjust for variability across patients with respect to the amount of medical history available for review. All analyses were conducted using SAS software (39).

When it was determined that a substantial proportion of patients in the study had tested negative for HIV during the 5-year review period, a subanalysis was conducted to compare the incidence of individual clinical events in the “not infected period” (the period preceding the most recent negative HIV test, where one or more negative tests were noted) with the incidence of those events in the “possibly infected period” (the period between the most recent negative HIV test and the first positive test or the entire review period if no negative test was noted). The subanalysis, presented elsewhere in poster form (40), indicated that the following conditions were at least twice as likely to be present in patients later diagnosed with HIV infection than in the same patients before testing negative for HIV: oral infection, pneumonia, unexplained fever, herpes zoster, seborrheic dermatitis, night sweats, and unexplained weight loss. Interestingly, the subanalysis did not show lymphadenopathy to be an important indicator of possible underlying HIV infection. The results of this subanalysis, as well as expert advice (22), provide a basis for believing these conditions might serve as important indicators of potential underlying HIV infection.

Back to Top | Article Outline


Medical record review was completed on 440 cases, or 88% of the ≈500 total cases of newly diagnosed HIV infection that were eligible for inclusion in the study at the eight participating sites. The primary reason for not completing medical record review on the remaining 12% of eligible cases was because the chart was not available to the study reviewer after repeated attempts to order it. A total of 1,685 patient-years of medical history preceding the diagnosis of HIV infection were reviewed (per patient mean, 3.8 patient-years; median, 5.0 patient-years; range, 1.0–5.0 patient-years). Fifty-nine percent of all patients had the full 5 years of medical history available for review.

The study chart reviewers recorded >10,000 MEC conditions. These entries included encounters across the full range of MEC conditions, from headache and sore throat to AIDS-defining diagnoses, and were culled from the >3800 contacts the 440 patients had with the KP and GHC care systems in the 5 years leading up to diagnosis. The mean total number of contacts per patient was 8.6 (median, 7 contacts; range, 0–59 contacts). Entries were excluded from these counts where only patient weight was recorded.

Patient characteristics are reported in Table 1. The mean age of the patients at diagnosis was 39.5 years (median, 37.5 years; range, 17–80 years), and 84% of patients were male. Of the 87% of patients with race documented in the chart, 43% were white. Of the 66% of patients with HIV risk factors documented in the chart, 78% had a history of male-to-male sex, 14% were heterosexual, and 8% had other risks. The mean duration of health plan membership before diagnosis was 7.5 years (median, 5.7 years). As documented in the medical record during the 5-year observation period before diagnosis, 23% of patients were tested (and found negative) for HIV at least once before testing positive, 21% denied having HIV risk factors, and 12% declined testing one or more times when an HIV test was recommended by a provider. At the time HIV infection was detected, the physician recommended testing for HIV in 61% of cases, the patient requested the HIV test in 24% of cases, and it was uncertain who initiated testing in 15% of cases.



For the 388 patients (88%) who had a CD4 cell count measured within 60 days of when HIV infection was diagnosed, the mean CD4 cell count was 310/μL (median, 254/μL): 62%, <350/μL; 43%, <200/μL (immunologic AIDS); and 18%, <50/μL. Separately, 26 patients (6%) had other AIDS-defining conditions at diagnosis (11 had Pneumocystis carinii pneumonia, 5 had Kaposi sarcoma, 5 had lymphoma, and 5 had other conditions). Only three patients who had AIDS-defining conditions (Kaposi sarcoma and lymphoma) at diagnosis did not also have immunologic AIDS.

Although 86% of patients were eventually found to belong to major risk groups for HIV infection, only 26% had HIV risk indicators documented in the chart >1 year before diagnosis (Table 2). Sixty-one percent (267) of all 440 patients had one or more of the eight clinical indicators of HIV infection documented anytime in the 5 years before or at diagnosis. Of these 267 patients, 139 (52%) were tested and diagnosed with HIV infection within 6 months of their first clinical indicator. Only 22% (96) of all 440 patients had one or more of the eight clinical indicators documented >1 year before diagnosis.



In bivariate analyses (Table 2), age at diagnosis (four strata), sex, race, risk group, having tested negative for HIV in the 5 years before testing positive, having denied risks for HIV infection, who it was that requested the positive HIV test, and time between the first of eight clinical indicators and testing positive for HIV were all significantly associated with CD4 cell count at diagnosis at the .10 level. When it was the patient who requested testing, symptoms were less likely to be present in the month before testing positive, in the year before testing positive, and ever in the 5-year period of observation (all at p < .0001, data not shown). Patients with risk documented >1 year before diagnosis were less likely to have a CD4 cell count of <200/μL at diagnosis than patients who did not have risk factors documented at least 1 year before diagnosis, but this difference was not significant (37% vs. 44%, respectively;p = .20). Presence of any of the eight clinical indicators anytime in the 5 years preceding diagnosis was significantly associated with CD4 cell count at diagnosis in bivariate analysis (p < .0001), as was presence of these indicators in the 1 to 5 years preceding diagnosis (p = .04).

The results of multiple logistic regression are presented in Table 3. Patients with increased odds of having a CD4 cell count of <200/μL at diagnosis were those with older age at diagnosis, those with undocumented risks, and those having one or more of eight clinical indicators before diagnosis. Conversely, patients with decreased odds of having a CD4 cell count of <200/μL at diagnosis were those who were female, those with a prior negative HIV test, and those who requested testing as opposed to the test being recommended by the clinician. Also in this model, the overall effect of race was associated with CD4 cell count at diagnosis at a level suggestive of significance (p = .05). This effect was driven in part by the subgroup of Asians who, despite having an unstable estimate due to the small number of patients in this subgroup, were the only individual race group to have a significantly increased risk of having a CD4 cell count of <200/μL at diagnosis. Neither site size nor years of medical history available for review were associated with CD4 cell count at diagnosis.



Back to Top | Article Outline


Our data indicate that even among persons who have reasonable access to medical care and reside in geographic areas known for a high prevalence of HIV infection, where attention to the HIV epidemic is relatively high, nearly one half with newly diagnosed HIV infection had AIDS-defining CD4 cell depletion or another AIDS-defining condition at first diagnosis of HIV infection, and 62% had progressed to the point that national guidelines indicate that highly active antiretroviral therapy may be appropriate (9). This finding is in line with data reported from various other settings over the last 10 years (10–17) and suggests that reasonable access to care (i.e., care available to members enrolled in a large, integrated prepaid health plan) is not sufficient for early detection.

It is also apparent from both the data collected in this study and findings of other recently reported studies (13,17) that despite recommendations to screen more routinely (23) and to lower the threshold for testing (22), new cases of HIV infection were not being detected any earlier in 1998 than nearly a decade before (14,16,21–23).

It is important to note that although 86% of patients evaluated in this study were ultimately identified as belonging to a classic risk group for HIV (men who have sex with men or intravenous drug users), only 26% had such risk documented in the chart ≥1 year before the diagnosis of HIV infection. Clearly, improved attention to the health risks of sexual behavior and injection drug use is needed if clinicians and patients are going to make informed and appropriate decisions regarding the need to test for HIV.

Our data, in agreement with findings of previously reported studies, also indicate that increased attention to clinical indicators on the part of clinicians may lead to the earlier diagnosis of HIV infection in some patients (22). However, in our setting, the proportion of cases that might have been detected >1 year earlier as a result of more vigilant attention to clinical prompts to test for HIV was relatively small. Although 61% of patients had one or more of eight clinical indicators in the 5 years leading up to diagnosis, only 22% had these symptoms >1 year before diagnosis. Furthermore, our data indicate that aside from the clinical signs of acute infection, by the time patients present with clinical signs of possible HIV infection, CD4 cell counts have already fallen to levels warranting medical attention. Thus, increased attention to the clinical indicators studied here will not by itself lead to a substantially higher mean CD4 cell count at diagnosis. Our data suggest that this is true even when the population being served is at increased risk of infection.

The significant association we observed between age and CD4 cell count at diagnosis (younger patients tend to present with higher CD4 cell counts) and between sex and CD4 cell count at diagnosis (females tend to present with higher CD4 cell counts) has been reported in the United States as well as in Europe and Australia (11–13,19). These associations could reflect the generally more recent acquisition of HIV infection in younger patients and female patients. For example, on average, younger persons and women are more likely to have been infected more recently than older patients and males. This is simply because younger persons as a group are more likely to have had fewer years of life after infection and, separately, the epidemic has been more prevalent among males for a longer period than among females. Similarly, patients with a negative HIV test in the 5-year review period, a factor we found strongly associated with higher CD4 cell count at diagnosis, are likely to have less immune depletion than their counterparts who by contrast could have been infected for the entire 5-year period or longer. Our finding that patients with undocumented HIV risks before diagnosis are more likely to have a CD4 cell count of <200/μL at diagnosis is not surprising, since lack of documentation may reflect a lack of awareness of existing risks on the part of the provider or denial of those risks on the part of the patient. Regardless of the reason why risks were not documented, our finding that they were not underscores the need to perform an effective risk assessment and to document any risks so they can be further addressed by that provider or by other providers at subsequent visits. Finally, not only does it seems intuitive that patients in this study who had symptoms indicative of immune dysfunction before being diagnosed with HIV infection were more likely to have a CD4 cell count of <200/μL at diagnosis, but also it is consistent with findings of other work (11). It is interesting that in our data the logical and powerful associations between CD4 cell count at diagnosis and age, sex, a recent negative test, undocumented risks, and presence of symptoms did not obscure the independent associations of race and who initiated testing with late diagnosis. It would be interesting to survey patients who initiated testing at diagnosis to see if a sense of stigma was associated with a lower CD4 cell count at diagnosis.

The association between race/ethnicity and CD4 cell count at diagnosis was of borderline significance and was largely driven by the small subgroup of Asian patients. However, there are other data to suggest that there may be an association between race and a clinician's assessment of risk (26). Race deserves further study as a predictor of late testing, particularly with respect to patients of Asian descent. With regard to the association between who initiated testing and CD4 cell count at diagnosis, our data suggest that patients who recognize that they are at risk and actively seek testing are more likely to be diagnosed earlier than patients who are not tested until the clinician appreciates the situation and recommends testing.

The findings presented here have important implications for early detection strategies. Primary care providers and the public health community in general need to have continued dialogue with health care consumers regarding HIV infection, routinely inquire about and raise awareness of one's risk for HIV infection among patients and communities, and engage the patient and the public regarding readiness to test. If persons at risk for HIV infection are aware of their risk and can be encouraged to undergo testing before symptoms develop, HIV infection will likely be diagnosed before extreme CD4 cell depletion occurs and opportunities to prevent further transmissions are missed. In addition, as nearly one in four patients in our population tested negative for HIV at some point in the 5 years before testing positive, methods used in posttest counseling sessions to encourage safer sex need to be evaluated for effectiveness. Renewed efforts also are needed to encourage risk reduction strategies and serial testing for those persons with continued risky behaviors.

There are several limitations to this study that deserve mention. Although the cases reviewed in this study were drawn from multiple sites nationwide, cases of HIV infection among the KP and GHC membership may not be representative of all cases in the United States. In other settings, testing for HIV in response to risk and clinical indicators may be more or less routine. Consequently, the potential benefit from increased awareness may be less or greater, respectively.

To ensure that all patients studied had a minimum of 1 year of medical history to review, patients with <1 year of medical history were excluded. Separately, not all patients had 5 years of membership with KP or GHC. We would not expect those patients who did not have 5 years of review to, as a group, have a different natural history of disease than the patients for whom a full 5 years of review was possible. Similarly, despite KP and GHC being closed systems, it is possible that a small proportion of medical encounters among enrolled members took place at non–health plan facilities. Such encounters would not have been available for review and inclusion in the analysis presented here. In addition, data examined in this study did not include patient characteristics that were known to health plan providers or clinical events that took place at health plan facilities but were not documented in the chart. The medical record is an indicator of how care is administered (41) but may not capture important details relevant to our findings (42,43).

We believe that this is the first large study to look at the interaction of HIV-infected individuals with their health care system in the years preceding their diagnosis. Being a closed medical system, where members obtain most of their care within the purview of the health plan, this setting is among a limited number of settings uniquely qualified to provide a comprehensive review of medical care contacts that precede a diagnosis. In addition to providing needed insights into the circumstances surrounding HIV testing, the clinicians and researchers involved in this work gained an appreciation for more effective risk assessment among sexually active persons and for more vigilant attention to potential signs of HIV infection and other sexually transmitted infections.

We conclude that strategies to improve early detection of HIV infection that focus on increased recognition of the clinical indicators we studied as prompts for HIV testing will not achieve early detection and the associated benefits in most cases. Rather, more effective risk assessment in the absence of symptoms, along with fewer barriers to testing and reduced social stigma surrounding HIV infection, is more likely to bring about the detection of HIV infection before severe immune depletion, make possible the benefits of treatment, and reduce further transmission of the virus. HIV testing must become more routine in the health evaluations of asymptomatic at-risk individuals. It has recently been shown in the hospital setting that more routine testing is a viable approach to detecting unrecognized infection (44).

The Centers for Disease Control and Prevention continue to recommend more routine testing for HIV (23). Yet, even in a society that is as health conscious and well served as we are in the United States, the challenge of early detection of HIV infection is clearly a daunting one. Despite the massive resources allocated to treatment and prevention, several questions regarding the early diagnosis of HIV infection remain unanswered. Further studies are needed to better understand the complexities of readiness to test among the various at-risk populations and to develop strategies to bring these populations to the attention of a care provider. In the meantime, one would hope to see an increasing proportion of patients with either routine testing for HIV or notes in their charts reflecting accurate HIV risk assessments and recommendations to test.

Back to Top | Article Outline


The authors thank the following members of the HIV Early Detection Study of Unrecognized Positives (HEDS UP) Study Group. Kaiser Permanente (KP) Hawaii: J. Cairl and D. Kovach; KP Southern California: E. Eck, C. Gravell, M. Katz, A. Perez, B. Porter, S. Shapiro, and C. Suh; KP Northern California: M. Allerton, G. DeLorenze, K. Detels, S. Hager, M. Horberg, A. Meuller, J. Selby, S. Sidney, S. Slome, D. Snow, D. Thelen, and E. Thomas; KP Pacific Northwest (Portland, OR): D. Antoniskis, B. Berg, and J. Thonstad; Group Health Cooperative (WA): S. Carzasty and W. Dodge; KP Colorado: M. Mogyoros and J. Buchanan; KP Ohio: D. Malatesta; KP Mid-Atlantic: D. Melnick; and the Kaiser Permanente Consortium for HIV/AIDS Interregional Research (CHAIR) Administration: S. Hager, S. Spande, and T. Young.

Back to Top | Article Outline


1. del Rio C. HIV: it's beginning to look like a chronic disease. AIDS Clin Care 2001; 13:82–3.
2. Murphy EL, Collier AC, Kalish LA. Highly active antiretroviral therapy decreases mortality and morbidity in patients with advanced HIV disease. Ann Intern Med 2001; 135:17–26.
3. Schwarcz SK, Hsu LC, Vittinghoff E, et al. Impact of protease inhibitors and other antiretroviral treatments on acquired immunodeficiency syndrome survival in San Francisco, California, 1987–1996. Am J Epidemiol 2000; 152:178–85.
4. Valdiserri RO, Holtgrave DR, West GR. Promoting early HIV diagnosis and entry into care. AIDS 1999; 13:2317–30.
5. Skowron G, Street JC, Obee EM. Baseline CD4+ cell count, not viral load, correlates with virologic suppression induced by potent antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 28:313–9.
6. Tortajada C, Garcia F, Plana M, et al. Comparison of T-cell subsets' reconstitution after 12 months of highly active antiretroviral therapy initiated during early versus advanced states of HIV disease. J Acquir Immune Defic Syndr 2000; 25:296–305.
7. Miller V, Mocroft A, Reiss P. Relations among CD4 lymphocyte count nadir, antiretroviral therapy, and HIV-1 disease progression: results from the EuroSIDA Study. Ann Intern Med 1999; 130:570–7.
8. Kaplan J, Hanson D, Karon J, et al. Late initiation of antiretroviral therapy (at CD4+ lymphocyte count <200 cells/mL) is associated with increased risk of death [abstract 520]. 8th Conference on Retroviruses and Opportunistic Infections, February 4–8, 2001, Chicago, IL.
9. DHHS Panel on Clinical Practices for Treatment of HIV Infection Guidelines for Use of Antiretroviral Agents in HIV-infected Adults and Adolescents. February 2002.
10. Girardi E, Sampaolesi A, Gentile M, et al. Increasing proportion of late diagnosis of HIV infection among patients with AIDS in Italy following introduction of combination antiretroviral therapy. J Acquir Immune Defic Syndr 2000; 25:71–6.
11. Samet JH, Retondo MJ, Freedberg KA, et al. Factors associated with initiation of primary medical care for HIV-infected persons. Am J Med 1994; 97:347–53.
12. Hocking JS, Rodger AJ, Rhodes DG, et al. Late presentation of HIV infection associated with prolonged survival following AIDS diagnosis—characteristics of individuals. Int J STD AIDS 2000; 11:503–8.
13. Castilla J, Sobrino P, de la Fuente L, et al. Late diagnosis of HIV infection in the era of highly active antiretroviral therapy: consequences for AIDS incidence. AIDS 2002; 16:1945–51.
14. Katz MH, Bindman AB, Keane D, et al. CD4 lymphocyte count as an indicator of delay in seeking human immunodeficiency virus-related treatment. Arch Intern Med 1992; 152:1501–4.
15. Centers for Disease Control and Prevention. Routinely recommended HIV testing at an urban urgent-care clinic—Atlanta, GA. MMWR Morb Mortal Wkly Rep 2001; 50:538–41.
16. Wortley PM, Chu SY, Diaz T, et al. HIV testing patterns: where, why, and when were persons with AIDS tested for HIV? AIDS 1995; 9:487–92.
17. Dybul M, Bolan R, Condoluci D, et al. Evaluation of initial CD4+ T cell counts in individuals with newly diagnosed human immunodeficiency virus, by sex and race, in urban settings. J Infect Dis 2002; 185:1818–21.
18. Lang W, Perkins H, Anderson RE, et al. Patterns of T lymphocyte changes with human immunodeficiency virus infection from seroconversion to the development of AIDS. J Acquir Immune Defic Syndr Hum Retrovirol 1989; 2:63–9.
19. Samet JH, Freedberg KA, Savetsky JB, et al. Understanding delay to medical care for HIV infection: the long-term non-presenter. AIDS 2001; 15:17–85.
20. Kirschner D, Webb G, Cloyd M. Model of HIV-1 disease progression based on virus induced lymph node homing and homing-induced apoptosis of CD4 lymphocytes. J Acquir Immune Defic Syndr 2000; 24:352–62.
21. Centers for Disease Control and Prevention. Recommendations for HIV testing services for inpatients and outpatients in acute care hospital settings. MMWR Morb Mortal Wkly Rep 1993; 42( RR-2):157–8.
22. Freedberg KA, Samet JH. Think HIV: why physicians should lower their threshold for HIV testing. Arch Intern Med 1999; 159:1994–2000.
23. Centers for Disease Control and Prevention. Revised guidelines for HIV counseling, testing and referral. MMWR Morb Mortal Wkly Rep 2001; 50( RR-19):7–42.
24. Samet JH, Freedberg KA, Stein MD, et al. Trillion virion delay. Time from testing positive for HIV to presentation for primary care. Arch Intern Med 1998; 158:734–40.
25. Centers for Disease Control and Prevention. US Public Health Service recommendations for human immunodeficiency virus counseling and voluntary testing for pregnant women. MMWR Morb Mortal Wkly Rep 1995; 44( RR-7):1–15.
26. Centers for Disease Control and Prevention. Prenatal discussion of HIV testing and maternal HIV testing—14 states, 1996–1997. MMWR Morb Mortal Wkly Rep 1999; 48:401–4.
27. Lansky A, Jones JL, Frey RL, et al. Trends in HIV testing among pregnant women: United States, 1994–1999. Am J Public Health 2000; 9:1291–3.
28. Epstein RM, Levenkron JC, Frarey L, et al. Improving physicians' HIV risk-assessment skill using announced and unannounced standardized patients. J Gen Intern Med 2001; 16:176–80.
29. Epstein RM, Morse DS, Frankel RM, et al. Awkward moments in patient-physician communication about HIV risk. Ann Intern Med 1998; 128:435–42.
30. Gerbert B, Bleecker T, Bernzweig J. Is anybody talking to physicians about acquired immunodeficiency syndrome and sex? Arch Fam Med 1993; 2:45–51.
31. Peralta L, Constantine N, Griffin-Deeds B, et al. Evaluation of youth preferences for rapid and innovative human immunodeficiency virus antibody tests. Arch Pediatr Adolesc Med 2001; 155; 7:838–43.
32. Galvan FH, Bing EG, Bluthenthal RN. Accessing HIV testing and care. J Acquir Immune Defic Syndr 2000; 25:S151–6.
33. Mylonakis E, Paliou M, Lally M, et al. Laboratory testing for infection with the human immunodeficiency virus: established and novel approaches. Am J Med 2000; 109:568–76.
34. Taylor B. HIV, stigma and health: integration of theoretical concepts and the lived experiences of individuals. J Adv Nurs 2001; 35:792–8.
35. Chesney MA, Smith AW. Critical delays in HIV testing and care: the potential role of stigma. Am Behav Sci 1999; 42:1162–74.
36. Herek GM. AIDS and stigma. Am Behav Sci 1999; 42:1106–16.
37. Fortenberry J, McFarlane M, Bleakley A, et al. Relationships of stigma and shame to gonorrhea and HIV screening. Am J Pub Health 2002; 92:378–81.
38. Porter K, Wall PG, Evans BG. Factors associated with lack of awareness of HIV infection before diagnosis of AIDS. BMJ 1993; 307:20–3.
39. SAS Institute. SAS/STAT user's guide, version 8. Cary, NC: SAS Institute, 1999.
40. Klein D, Hurley L, Merrill D, et al. Early detection of HIV: the HEDS UP Study [abstract 219]. 8th Conference on Retroviruses and Opportunistic Infections, February 4–8, 2001, Chicago IL.
41. Tauber A. Putting ethics into the medical record. Ann Intern Med 2002; 136:559–63.
42. Stange K, Zyzanski S, Smith T, et al. How valid are medical records and patient questionnaires for physician profiling and health service research? A comparison with direct observation of patient visits. Med Care 1998; 36:851–67.
43. Dresselhaus T, Luck J, Peabody J. The ethical problem of false positives: a prospective evaluation of physician reporting in the medical record. J Med Ethics 2002; 28:291–4.
44. Walensky R, Losina E, Steger-Craven K, et al. Identifying undiagnosed human immunodeficiency virus: the yield of routine, voluntary inpatient testing. Arch Intern Med 2002; 162:887–92.

HIV; Early detection; Late testing; Risk assessment

© 2003 Lippincott Williams & Wilkins, Inc.