At a time when the treatment of acute HIV infection is expanding , it is important to recognize that most patients infected with HIV in the first decade of the epidemic presented for care with advanced immunosuppression [2–4]. Reasons for patient delays in seeking medical care for HIV infection have not been extensively examined. A conceptual framework of an HIV-infected individual's path to primary medical care is depicted in Fig. 1. Of those patients who become infected with HIV, some are aware of their risk of infection whereas others are not. For patients aware they are at risk, the decision to be tested is difficult. Even after testing and learning of a positive result, patients must then link to medical care in order to receive treatment. The pathway described above for HIV-infected patients to connect with primary care is a complex one. It is not surprising that physicians are often faced with individuals who present to medical care with advanced immunosuppression and disease progression, patients we have called ‘long-term non-presenters'. After seeking primary care, patients must remain connected to care throughout the course of HIV disease to reap the benefits of new therapies. These four steps: awareness of HIV risk; HIV testing; linkage to medical care; and maintenance of primary care depend upon factors that are not well understood. In this paper, we focus on the first two steps.
We had several objectives in examining the delay to medical care of HIV-infected individuals. Our primary objective was to estimate the duration of an individual's HIV infection and determine the factors associated with CD4 cell counts at presentation to primary care. Our secondary objectives were to determine the duration of individuals’ HIV risk awareness before testing and patient characteristics associated with awareness of HIV risk before testing.
Materials and methods
Study site and population source
This is an analysis of data from a cohort that has previously been described with regard to the delay in presentation to primary medical care after receiving a positive HIV test report . Patients were interviewed after prospective enrolment from two sites: the Boston City Hospital (BCH) HIV Diagnostic Evaluation Unit from February 1994 to April 1996, and the Rhode Island Hospital (RIH) primary care HIV Clinic from December 1994 to March 1996. Using a prospective study design, a questionnaire exploring barriers and facilitators to medical care for HIV infection, was administered to each subject. The BCH HIV Diagnostic Evaluation Unit and the RIH HIV clinic are weekly clinics designed for the initial assessment and triage of all new patients presenting with HIV infection and seeking primary care . Referrals to each site come from a wide variety of sources, including inpatient hospital services, hospital outpatient clinics, the emergency department and urgent care clinic, community health centers, drug treatment programs, HIV testing sites, local correctional institutions, as well as self-referrals. The study was approved by the institutional review boards of both hospitals.
Subjects were from an inception cohort, a study design that identifies patients with a specific disease at a uniform point of its course. In this cohort, the uniform point for these patients was their first presentation to primary care for HIV. We used explicit criteria to identify patients who had not received previous primary medical care for HIV infection. New to primary care was defined as: (i) an initial positive HIV test within 4 calendar months of the evaluation; or (ii) an initial positive HIV test more than 4 months before presentation and the absence of the following history, determined by medical record review and patient report: specific previous HIV primary care, past zidovudine or other antiretroviral use, or two or more previous CD4 cell counts. Patients who reported these types of care only while incarcerated were included in the study, as care while incarcerated was not construed as an active effort to seek care. The criteria of two, rather than one, previous CD4 cell count was used as some patients have a CD4 cell count obtained at the time of HIV testing. Only patients fluent in English, Spanish, or Haitian Creole were eligible. Each patient provided written informed consent before entering the study.
Patients were asked to participate in this study after their initial clinical care was performed, including history, physical examination, and laboratory tests. At RIH this was at the initial encounter, and at BCH this was at a clinical appointment generally one week after the first visit. Those who met the entry criteria and agreed to participate underwent a 60–90 min standardized interview. One of three research associates carried out all interviews concerning behavioral, medical, and social history. Interviews were administered in Spanish or Haitian Creole when appropriate, by interpreters working with the research associates. Interview instruments were translated into Spanish and Haitian Creole, back-translated into English to check for accuracy, and then corrected.
The initial CD4 cell count (cells/μl) at presentation was the primary outcome examined in this study, whereas awareness of risk at testing and the delay between awareness and testing were the secondary outcomes. By reviewing the clinical records we examined CD4 cell counts obtained at outpatient encounters within 3 calendar months of the initial medical evaluation. When two initial CD4 cell counts before antiretroviral therapy were available we used the mean count. Subjects in our sample with CD4 cell counts greater than 1000 μl at the time of the study (n = 3) were truncated at 1000 cells/μl.
We also measured the awareness of HIV risk before testing and the delay before seeking testing after becoming aware of this risk. Our definition of awareness status (aware or unaware) was derived from the question, ‘Before you tested positive for HIV, did you consider yourself to be at risk for HIV?’ An affirmative answer confirmed their awareness status. Those aware of their risk, were asked, ‘For how many months/years did you consider yourself at risk before you were tested?’ The ‘awareness’ period before seeking testing is reported in years.
Independent variables were selected to address the patient's status at the time of the interview or HIV testing. The specific variables examined are listed in Table 1, and included demographics, drug and alcohol abuse, social support, sexual beliefs and practices, and medical and psychiatric issues. The following standardized questionnaires were used: CAGE*, Center for Epidemiology Scale–Depression (CES–D) , Cleary AIDS Health-Related Quality of Life , and the Miller Hope scale .
The Cleary Health-Related Quality of Life Scales were chosen because they specifically assess the health-related quality of life in people infected with HIV. These scales are validated scales based on a battery of items that measure HIV-infected symptoms, as well as life satisfaction, general health perception, physical functioning, emotional well-being, fatigue, disability, pain, memory problems and illness severity. For our purposes, we examined the following scales: quality of life, pain, overall symptoms, fatigue, perception of best health, and perception of worst health.
The Miller Hope scale is a validated scale based on 40 items that measure hope in adults. Each item is measured on a 6-point Likert scale from 1, very strongly disagree, to 6, very strongly agree, and the possible range of the composite scale is 40 to 240, with higher scores indicating more hope . The symptom scale is a part of the Cleary scales as described above.
We adopted the Centers for Disease Control and Prevention hierarchical classification of HIV risk factors, with the exception of the combined category of men who have sex with men and injection drug users (IDU) . If injection drug use was positive, then this was the designated risk factor even in the presence of other risk factors.
Descriptive statistics were generated and analysed for all study variables. Bivariate analyses were conducted to assess the relationship between each independent variable and the CD4 cell count using analysis of variance and Pearson correlations for discrete and continuous independent variables, respectively. Bivariate analyses were conducted to assess the relationship between each independent variable and the awareness of HIV risk using chi-square analysis and analysis of variance for discrete and continuous independent variables, respectively. Variables that were either significant at the P < 0.10 level in the bivariate analyses or were clinically significant (i.e. were important to consider as covariates) were considered candidates for the multivariable analysis. One of each pair of variables with Pearson correlation coefficients greater than 0.4 was excluded to avoid problems with co-linearity in the multivariable analysis . The multivariable analyses, multiple linear regression and multiple logistic regression examined the dependent variables, CD4 cell count and awareness of HIV risk status before HIV testing, respectively.
The approximation of time delay was based on a CD4 cell decline of 60/μl per year, and an initial CD4 cell count of 800/μl. The 60 cells/μl per year is a marginal mean decline based on a published viral load distribution. Mellors et al.  reported CD4 cell count decline to be a function of viral load concentration (copies/ml) as listed in Table 2. The authors published the mean (and 95% confidence intervals; CI) CD4 cell count decline at each of five different levels of plasma HIV-RNA concentration. Decline was calculated by averaging over viral load concentrations using the published distribution of viral load, resulting in a mean decline of 60 cells/μl per year.
Data were collected at two sites, but sample size limitations did not allow for the development of separate models within each site. Relationships between significant independent variables and the dependent variable were evaluated separately for each site. As the direction and magnitude of the associations were consistent in all cases, the data were pooled.
A two-tailed P-value less than 0.05 was considered statistically significant in multivariable analyses. Data were analysed by SAS statistical software .
Enrolled patients included 74% (203/276) of all eligible patients at initial presentation for primary care for HIV infection from both sites during the period of study [68% (150/222) of patients at BCH and 98% (53/54) at RIH]. At the BCH site, 72 patients not enrolled in the study included 37 who refused to participate, 25 who agreed to participate but never returned for the initial interview, and 10 who were never contacted. There were no significant differences between patients who enrolled in the study at BCH (n = 150) and those who were not enrolled (n = 72) with respect to age, sex, and HIV risk group category. There was a significant difference with respect to race/ethnicity (P < 0.05). Disproportionately fewer Haitian (5/27, 19%) and more white individuals (44/54, 81%) enrolled in the study compared with Hispanic/Puerto Rican (37/54, 69%) and African American individuals (64/87, 74%).
We interviewed a total of 203 patients, of whom 74% were from BCH and 26% from RIH. The mean age was 37 years. Other characteristics of the study population are described in Table 3. One-quarter of the study patients were women and 71% were non-white. The two most common primary HIV risk factors were injection drug use and heterosexual intercourse. Nearly half the subjects had two or more positive responses to the CAGE questionnaire, a screening tool that indicates past or present alcohol problems . Although 38% had spent time in jail in the past 10 years, only 5% were included in the study despite having possibly received medical care for HIV infection in prison.
The median CD4 cell count of the subjects was 280/μl (range 1–1710/μl) with three subjects with counts greater than 1000/μl. Thirty-seven per cent presented with CD4 cell counts of 200 μl or less, 43% presented with counts between 201 and 500/μl, and 20% presented with counts greater than 500/μl.
The estimate of the mean duration of HIV infection before medical presentation for HIV-related medical care, extrapolated from the CD4 cell count as described in the Methods section, was 8.1 years (95% CI 7.5, 8.6) with a range of 0–13.3 years. The interquartile range (the difference between the 25th and 75th percentiles) was 6.0–11.6 years.
In the bivariate analysis, the following characteristics were associated (P < 0.10) with lower CD4 cell counts at the initial presentation for medical care: having 0–1 close friends (P = 0.05), not being in jail in the past 10 years (P = 0.05), voluntary testing (P = 0.03), lower hope (P = 0.08), poor quality of life (P = 0.08), higher symptom score (P = 0.03), and older age (P = 0.02) (Table 4). We entered these variables along with sex and CAGE score, potentially important covariates considered clinically relevant, into a multiple linear regression model. No co-linearity was found between these variables. The only characteristics that were significantly associated with lower CD4 cell counts were being male (P = 0.05), not having been in jail over the past 10 years (P = 0.02), and older age (P = 0.02). Only 15% of the variation in CD4 cell counts at the initial presentation to primary care could be accounted for in this model (R2 = 0.15).
Sixty-six per cent of subjects (128/194) were aware of their HIV risk before testing. In the bivariate analysis, the following characteristics were associated (P < 0.10) with awareness of HIV risk at the time of HIV testing: hospital site (BCH) (P = 0.09), in a residence for less than 6 months (P = 0.02), belief that self-care avoids illness (P = 0.06), fatalistic viewpoint (P = 0.07), heard of zidovudine (P = 0.08), HIV risk group of IDU and gay/bisexual man (P = 0.001), fatigue (P = 0.05), symptoms (P = 0.07) and pain (P = 0.09). No co-linearity was found between these variables and they were all entered into a logistic regression model. As shown in Table 5, hospital site, duration in residence, the belief that self-care avoids illness, and HIV risk behaviors of injection drug use and gay/bisexual men were significantly and independently associated with the awareness of HIV risk at the time of HIV testing.
The median time that aware subjects felt at risk for HIV prior to receiving a test was 1 year (mean = 2.5 years) with a range from 0–10 years. No factor was significantly associated with being aware of HIV risk without testing for greater than one year (Data not shown).
Reasons for HIV testing among individuals aware and unaware of their HIV risk status are displayed in Fig. 2. Interactions with medical care (either during hospitalization or on a physician's recommendation) accounted for 42% of the unaware patients being HIV tested.
Engaging HIV-infected persons in medical care yields personal and public health benefits. However, in the second decade of the HIV epidemic, the initial presentation of patients with HIV infection to medical care continues to occur at a stage of advanced immunosuppression, years after infection with the virus [2–4]. This phenomenon has been demonstrated using various outcome measures including CD4 cell counts, the duration of time between HIV testing and AIDS diagnosis, and AIDS diagnosis at initial presentation for medical care [2–4,16,17]. The median CD4 cell counts we found at presentation between 1994 and 1996 (280 cells/μl), were similar to those we reported from 1990–1992 (300 cells/μl) . As a result of the asymptomatic nature of early HIV infection and the slow rate of disease progression, a certain amount of delay to care is inevitable. However, we found striking delays in this urban population.
Our data show that 80% of patients initially present to medical care with counts less than 500/μl, and 37% present with counts of 200/μl or less. The benefit of antiretroviral therapy, prophylaxis for opportunistic infections, immunizations, and behavioral interventions will not be maximized in patients who delay presentation [18–23]. Neither will the potential public health benefits be achieved by decreasing infectivity via sexual or needle use behavior or decreased vertical transmission.
To illustrate more clearly the phenomenon of delayed presentation to medical care for HIV infection, we transformed the CD4 cell counts of patients into an approximate period of delay between viral acquisition and initial medical care. The initial CD4 cell count represents for a population the period between acquiring HIV and the initial presentation and linkage to medical care, as noted by the time period, T5, in Fig. 1. Overall, this time period represents the delay to medical care from the time of infection, thus the CD4 cell count was chosen as the primary outcome measure. Because the rate of decline of the CD4 cell count is not constant and is influenced in each individual by multiple factors (e.g. viral load), we estimated the period of time elapsed from initial infection for each subject, and expressed this period as the group mean. Our approximation of a 6.0–11.6 year mid-range period (interquartile approximation) of delay is alarming. This finding demands that attention be focused on the earlier engagement of HIV-infected individuals into medical care.
Many components contribute to the delay to medical care for HIV infection. In theory, delays can occur at the stages depicted in Fig. 1 : the time between acquiring HIV infection and testing (T1, T2, T3); the time between receiving a positive test and initiating medical care (T4); and the period after initially establishing care for HIV infection (T6).
The major portion of the delay between acquiring HIV infection and the linkage to medical care (T5) is the period before HIV testing (T3 or T1 plus T2). Even though 39% of subjects in this population were previously found to delay 1 year or more after testing positive for HIV, the estimated average 8 year delay emphasizes the importance of delay before HIV testing .
The fact that over one-third of patients were unaware of their HIV risk before HIV testing suggests that some public health HIV messages are not being communicated. The high odds ratio for awareness among IDU and gay/bisexual men reflects the fact that patients with heterosexual HIV risk behaviors are often unaware of their HIV risk. Even patients who are aware of their risk of HIV delay on average 2.5 years before getting tested. Unfortunately, we were unable to provide insight on this striking delay (T2).
The delay between acquiring HIV and getting HIV tested has received limited attention. In 1994, Wenger et al.  reported results from a sample of 227 men, predominantly (81%) men who had sex with men, who were consecutive outpatients at an urban HIV clinic. They evaluated how individuals became aware of their infection, when they first suspected infection, and the factors associated with suspecting infection. They found that nearly all (99%) infected patients acknowledged personal risk factors for HIV, but only 40% indicated that they suspected they were positive before testing. Forty-eight per cent of those who suspected they were positive before testing waited a year or more before obtaining an HIV test . Although our results substantiate the findings of Wenger et al. , it is important to note that their population differed from the urban population we studied. Delay to HIV testing is both a result of unawareness (or lack of acceptance) of one's own risk of HIV infection and inaction after realizing that one is personally at risk.
Physicians can play a major role in patients getting HIV tested, a finding revealed in the responses to why subjects tested (`physician recommended it’ and ‘during hospitalization') (Fig. 2). This physician role should be expanded , but the fact is that the majority of HIV testing in the USA is performed for blood donation or life insurance [26,27]. Our data suggest the need for aggressive public health campaigns that enhance HIV risk awareness and personalize the problem.
This study has several limitations. We do not have data on subjects’ HIV-RNA levels at presentation. Our assumptions about the distribution of viral load are based on published data in a cohort of AIDS patients, which excludes the small minority of patients who are long-term non-progressors . This would make our estimate of the decline in CD4 cell count overly rapid. A small number of patients may die unaware that their illness is HIV related. Both these limitations may bias our results by understating the duration. However, it does not affect the overall message of our findings. We feel that our estimate of the range of duration of delay to medical presentation extrapolated from CD4 cell counts is realistic.
In addition, the population evaluated came from medical centers located in the urban northeastern United States and may not reflect those patients receiving HIV test results at sites in other geographical regions either within or beyond the USA. However, initial presentation for medical care with HIV frequently occurs at a time of advanced immunosuppression, as reported by other sites in the USA and abroad [2,16,28].
Among those eligible, disproportionately fewer Haitians enrolled in the study, a particularly vulnerable population for delayed presentation . Fear of disclosure and discussion of one's HIV infection may account for the lower participation in this group.
HIV-infected patients present for medical care years after acquiring HIV infection. Despite examining a broad array of sociodemographic and psychosocial factors, we could not determine specific distinguishing factors associated with populations at risk of presentation for HIV medical care with advanced immunosuppression. Over one-third of HIV-infected patients in this study were not cognisant of their HIV risk before testing. Even those aware of their HIV risk frequently delayed more than a year before seeking HIV testing. The ‘long-term non-presenter', one who enters HIV care years after initial infection, is the rule, not the exception, among HIV-infected individuals. Important steps necessary for patients to take maximal advantage of new advances in HIV therapy are raising HIV risk awareness and promulgating early HIV testing among those aware of their risk. Physician action and public health initiatives can play important roles in each of these areas.
The authors appreciate the contributions of our clinical staffs, Colleen LaBelle, RN, Susan Hart, MSW, and Kristin Jhamb, MD, as well as other contributors to the project: Margaret Marisi, A. Kate Karter, Annette Diehl, and Patricia Takash.
1. Carpenter CC, Cooper A, Fischl MA. et al. Antiretroviral therapy in adults.
:Updated recommendations of the International AIDS Society – USA panel.
JAMA 2000, 283: 381 –390.
2. Katz MH, Bindman AB, Keane D, Chan AK. CD4 lymphocyte count as an indicator of delay in seeking HIV-related treatment.
Arch Intern Med 1992, 152: 1501 –1506.
3. Samet JH, Retondo MJ, Freedberg KA, Stein MD, Heeren T, Libman H. Factors associated with initiation of primary medical care for HIV-infected persons.
Am J Med 1994, 97: 347 –353.
4. Porter K, Wall PG, Evan BG. Factors associated with lack of awareness of HIV infection before diagnosis of AIDS.
BMJ 1993, 307: 20 –23.
5. Samet JH, Freedberg KA, Stein MD. et al. Trillion virion delay: time from testing positive for human immunodeficiency virus to presentation for primary care.
Arch Intern Med 1998, 158: 734 –740.
6. Samet JH, Libman H, Labelle C. et al. A model clinic for the evaluation and establishment of primary care for persons infected with human immunodeficiency virus.
Arch Intern Med 1995, 155: 1629 –1633.
7. Mayfield D, McLeod G, Hall P. The CAGE questionnaire: validation of a new alcohol screening instrument.
Am J Psychiatry 1974, 131: 1121 –1123.
8. Radloff LS. The CES–D scale: a self report depression scale for research in the general population.
Appl Psychol Meas 1977, 1: 385 –401.
9. Cleary PD, Fowler FJ, Weissman J. et al. Health-related quality of life in persons with acquired immune deficiency syndrome.
Med Care 1993, 31: 569 –580.
10. Miller JF, Powers MJ. Development of an instrument to measure hope.
Nurs Res 1988, 37: 6 –1.
11. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report.
Nurs Res 1996, 8: 32 –33.
12. Concato J, Feinstein AR, Holford TR. The risk of determining risk with multivariable models.
Ann Intern Med 1993, 118: 201 –210.
13. Mellors JW, Munoz A, Giorgi JV. et al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection.
Ann Intern Med 1997, 126: 946 –954.
14. SAS Institute Inc. SAS/STAT Guide for personal computers
, 6th ed. Cary, North Carolina: SAS Institute; 1987.
15. Kitchens J. Does this patient have an alcohol problem?
JAMA 1994, 272: 1782 –1787.
16. Poznansky MC, Coker R, Skinner C. et al. HIV positive patients first presenting with an AIDS defining illness: characteristics and survival.
BMJ 1995, 311: 156 –158.
17. Vlahov D, Munoz A, Solomon L. et al. Comparison of clinical manifestations of HIV infection between male and female injecting drug users.
AIDS 1994, 8: 819 –823.
18. Centers for Disease Control and Prevention. 1997 USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with human immunodeficiency virus.
MMWR 1997, 46: 1 –46.
19. Hammer SM, Katzenstein DA, Hughes MD. et al. A trial comparing nucleoside monotherapy with combination therapy in HIV-infected adults with CD4 cell counts from 200 to 500 per cubic millimeter.
N Engl J Med 1996, 335: 1081 –1090.
20. Eron JJ, Benoit SL, Jemsek J. et al. Treatment with lamivudine, zidovudine, or both in HIV-positive patients with 200 to 500 CD4+ cells per cubic millimeter.
N Engl J Med 1996, 333: 1662 –1669.
21. Kaplan JE, Masur H, Jaffe HW, Holmes KK. Reducing the impact of opportunistic infections in patients with HIV infection.
JAMA 1995, 274: 347 –348.
22. Bozzette SA, Finkelstein DM, Spector SA. et al. A randomized trial of three antipneumocystis agents in patients with advanced human immunodeficiency virus infection.
:NIAID AIDS Clinical Trials Group.
N Engl J Med 1995, 332: 693 –699.
23. Jacquez JA, Koopman JS, Simon CP, Longini IM. Role of the primary infection in epidemics of HIV infection in gay cohorts.
J Acquire Immune Defic Syndr 1994, 7: 1169 –1184.
24. Wenger NS, Kusseling FS, Beck K, Shapiro MF. When patients first suspect and find out they are infected with the human immunodeficiency virus: implications for prevention.
AIDS Care 1994, 6: 399 –405.
25. Freedberg KA, Samet JH. Think HIV: why physicians should lower their threshold for HIV testing.
Arch Intern Med 1999, 159: 1994 –2000.
26. Anderson JE. CDC data systems collecting behavioral data on HIV counseling and testing.
Pub Health Rep 1996, 3 (Suppl. 1) : 129 –132.
27. Stein MD, Leibman B, Wachtel TJ. et al. HIV-positive women: reasons they are tested for HIV and their clinical characteristics on entry into the health care system.
J Gen Intern Med 1991, 6: 286 –289.
28. Gilleat S, Mallal S, French M, Dawkins R. Epidemiology of late presentation of HIV infection in Western Australia.
Med J Aust 1992, 157: 117 –118.