Background
In the past decade in developed countries, highly active antiretroviral therapies have dramatically decreased HIV infection morbidity and mortality
In most European countries, risk-factor-based HIV testing strategy alone shows limits to detect HIV-infected people because 1) people do not consider themselves at risk and 2) healthcare providers fail in risk assessment
Few studies have evaluated patient consultations during the possibly-HIV-infected period
Methods
Study design, setting and population
A cross-sectional study was conducted between June 2009 and October 2010 in 69 ANRS centers providing care to patients with HIV throughout France (metropolitan France and French overseas departments). The study was proposed to all the ANRS centers (around 160) on a voluntary basis; 64 centers accepted to participate and enrolled effectively patients. The two mains reasons for not participating were lack of manpower and too few new HIV-infected patients per year. However, centers from smaller hospitals were recruited to ensure geographic coverage of French territories.
Patients were eligible for this study if they were over 18-years-old, had an initial HIV-positive test (diagnosis) in France, and sought care in a participating center within the 6-month period following HIV diagnosis. Patients diagnosed in a foreign country or who had been living for ≤1 year in France were excluded, since opportunities for testing were not in France. Eligible patients were consecutively enrolled during the study period.
The study received approval from two French data protection authorities (CCTIRS and CNIL) and did not require formal ethical approval according to French regulations. All patients received information and gave written consent before their inclusion in the study.
Definition
We defined situations that would lead care providers to suggest HIV testing, i.e. to persons belonging to a high-risk group for HIV acquisition and who presented with HIV-related clinical conditions. Risk groups for acquiring HIV infection were (by descending risk)
Possibly HIV-related conditions included symptoms more frequent during chronic stage of HIV infection and diseases associated with a high HIV prevalence (>1%)
General symptoms: fever unexplained and/or lasting ≥1 month; diarrhea recurrent and/or lasting for ≥1 month; weight loss ≥10%;
Cutaneous/mucous symptoms: seborrheic dermatitis; oral herpes; oral hairy leukoplakia; oral candidiasis; varicella zoster; onychomycosis; unexplained prurigo. This class also included generalized lymphadenopathy;
Bacterial infections: community-acquired pneumonia, pulmonary tuberculosis, recurrent bacterial infections;
Diseases associated with high prevalence of HIV infection: viral hepatitis A, B or C, and sexually transmitted infections (STIs): syphilis; gonorrhea; chlamydia; genital herpes; genital condyloma / human papilloma virus; lymphogranuloma venereum proctitis infection; genital mycoplasma; trichomoniasis.
AIDS-defining opportunistic infections, except for pulmonary tuberculosis, were not considered in this analysis. Indeed, we hypothesized that HIV testing would be performed during the diagnostic process surrounding these events.
Main outcomes
Patients were asked about their encounters with healthcare settings and HIV testing history during the three years prior to HIV diagnosis. For patients belonging to a high-risk group, we determined whether, in the 3-year period prior to HIV diagnosis, they had seen a healthcare provider and, if so, mentioned that they belonged to a high-risk group. For patients presenting with possibly HIV-related conditions, we determined whether they had seen a healthcare provider. Among patients who had seen a healthcare provider from both groups, we determined whether an HIV test had been proposed during the first visit. Medical encounters between patients from high risk groups or with a possibly HIV-related condition and any healthcare provider, which did not lead to a test proposal, were defined as missed opportunities for HIV testing.
When evaluating patient medical history during the past three years, we considered that a patient to whom a test had not been proposed was probably HIV-infected at the time of this contact if: (i) he/she did not report an HIV-negative test after this contact; or (ii) he/she was not diagnosed at an acute HIV infection stage (defined as an incomplete western blot, detectable Agp24, or detectable plasma viral load, with a negative or weakly ELISA, or an interval <3 months between a negative and positive ELISA test
Data collection
The face-to-face interview used a standardized questionnaire that anonymously collected information on patient socio-demographic characteristics, risk factors for HIV acquisition, date of the first HIV-positive test, clinical and immunological characteristics at HIV diagnosis, history of HIV testing, contacts with healthcare settings, and clinical events. Possibly HIV-related conditions, occurring during the three years prior to HIV diagnosis, were to be detailed. Patients belonging to high-risk groups and/or presenting with possibly HIV-related conditions and who had seen a healthcare provider were questioned regarding which healthcare facility they had visited and whether or not an HIV test had been proposed.
Sample size estimate
Age has been found to be associated with missed opportunities for HIV testing in US studies
Statistical analysis
Descriptive statistical methods were used to describe the study population and opportunities for HIV testing. Continuous data were compared using the Wilcoxon rank test and categorical data using the Fisher exact test. P-values ≤0.05 (two-tailed) were considered significant.
To differentiate between missed opportunities for testing in patients “belonging to a high risk group” and those “presenting with possibly HIV-related conditions”, we considered only patients without a past history of possible HIV-related conditions when we evaluated opportunities for HIV testing in patients belonging to a risk group.
HIV-related conditions occurring within the 3-month period prior to HIV diagnosis were considered to be directly related to diagnosis. Testing opportunities were therefore evaluated only in patients who had presented HIV-related conditions between 3 years and 3 months prior to HIV diagnosis. When multiple HIV-related conditions had occurred during that period, we included in the analysis only the earliest HIV-related conditions.
Center effect impact on missed opportunities rate for HIV testing was studied by regrouping centers base HIV low of high new diagnosis rate (cut-off: 80 HIV positive tests per million inhabitants
Statistical analyses were performed using SAS 9.2 software (SAS Institute Inc., Cary, NC, USA).
Results
During the study period, 2,009 newly diagnosed HIV-infected patients sought care in the 69 participating centers throughout France. Overall, 659 (33%) were excluded because they did not fulfill inclusion criteria: 288 had been living in France for <1 year, 95 had had the initial HIV-positive test performed abroad, 148 sought care ≥6 months after HIV diagnosis, 18 were aged ≤18 years, and 110 were excluded for other reasons (mainly because they did not speak French). The study was not proposed to 227 patients mainly because of time pressure and lack of availability of care providers, and 115 more patients refused to participate in the study. Finally, 1,008 patients (50%) were enrolled. The overall 2,009 patients who sought care in study centers were not different from newly diagnosed HIV-infected patients in France during the same period
Patient’s median age was 39.5 years; 793 (79%) were males. More than 80% belonged to a high risk group, 12 (1%) were IDU, 530 (53%) were MSM, 155 (15%) were heterosexuals with risky behavior and 124 (12%) were migrants from Sub-Saharan Africa (Table
History of HIV testing
Except for the test leading to HIV diagnosis, 407 (41%) participants had been tested within the last three years, 273 (27%) had been previously tested but more than three years ago and the remaining 325 (32%) had never been tested for HIV. Among MSM, 94 (18%) had never been tested. In contrast, 45% of sub-Saharan African migrants, 44% of heterosexuals with risky behavior, and 57% of patients who stated not belonging to a risk group had never been tested (p < 0.0001 versus MSM). Among migrants from sub-Saharan Africa, 43 (35%) had a history of HIV testing in France.
Clinical and immunological characteristics at HIV diagnosis
HIV was diagnosed during acute infection in 153 (15%) patients (Table
Contact with the healthcare system during the three years prior to HIV diagnosis
Among the 1,008 patients enrolled in the study, 994 (99%) had had ≥1 medical encounter during the 3 years prior to HIV diagnosis: 922 (93%) with a general practitioner and 329 (33%) with an emergency department (Table
Opportunities for HIV testing in patients belonging to high risk groups (MSM)
As asymptomatic patients from high-risk groups other than MSM were not in sufficient numbers, description of opportunities for HIV testing in people belonging to high-risk groups were restricted to MSM. Among the 530 MSM enrolled in the study (Figure
Opportunity for HIV testing proposal
Opportunity for HIV testing proposal. Opportunity for HIV testing proposal based on (a) group at high risk of acquiring HIV (asymptomatic MSM) and (b) any possibly HIV-related condition, among the 994 patients who had contact with the healthcare system in the three years prior to HIV diagnosis. When considering opportunities for HIV testing based on a high-risk group, only patients with no HIV-related conditions in the three years prior to HIV diagnosis were included in the analysis. When considering opportunities for HIV testing based on a HIV-related condition, the first (oldest) symptom reported in the 3-year to 3-month period prior to HIV diagnosis was included in the analysis.aPatients were
Upon HIV diagnosis, median CD4 count tend to be lower in patients who did not state the healthcare provider that they were MSM compared to those who did so (400 vs. 468 cells/mm3, p = 0.08). Among those who stated, median CD4 cell count at diagnosis was not different in MSM whether the test was proposed or not.
Opportunities for HIV testing in patients with possibly HIV-related conditions
Overall, 364 out of 588 patients who reported a possibly HIV-related conditions during the 3-year to 3-month period prior to HIV diagnosis sought care (Figure
An HIV test was proposed by the healthcare provider upon first contact to 65 (18%) of these patients. Thus, 299 patients (82%) had a missed opportunity for HIV testing, and 233 (78%) of them were probably infected at the time of this contact (no subsequent HIV test and not diagnosed at an acute HIV infection stage). At HIV diagnosis, median CD4 count was 274 cells/mm3 (IQR, 111–507) in patients with a missed opportunity for HIV testing vs. 432 cells/mm3 (IQR, 308–621) in those to whom testing was proposed (p < 0.0001).
Results were similar when detailed by HIV-related condition (Table
The proportions of missed opportunities for HIV testing in any circumstances were not different between clinical centers when classified following epidemiological characteristics of their belonging regions (data not shown).
Discussion
This study, which included more than one thousand newly HIV-diagnosed patients in France, demonstrated that 99% of the patients had had at least one medical encounter during the 3-year period prior to HIV diagnosis, and 89% had seen a general practitioner at least once a year during that period. Moreover, results revealed the high proportion of missed opportunities for HIV testing among MSM and those seeking care for possibly HIV-related conditions. Seventy nine percent of overall MSM who visited a healthcare provider did not receive an HIV test proposal. Likewise, HIV testing was not proposed by the healthcare provider to 82% of patients who sought care for possible HIV-related conditions upon first contact.
This study had some strengths and limitations. We enrolled 1,008 newly-HIV-diagnosed patients, who accounted for 16% of new HIV diagnoses in France during the same period
We evaluated the patient course through the healthcare system in the years prior to HIV diagnosis. That may be subject to recall bias, leading to an underestimation of the number of missed opportunities. To limit this bias, patients diagnosed >6 months previously were excluded, and the period during which retrospective data were collected was restricted to three years. Another limitation may be the misperception of being tested. On one hand, some patients may be tested without knowing the test was performed, without the consent process, leading to overestimate missed opportunities proportion. On another hand, some patients may think having been tested for HIV because a blood analysis was performed during an inpatient stay for example, which may this time, underestimate missed opportunities proportion.
In MSM or patients presenting with possibly HIV-related conditions, we determined whether an HIV test was proposed upon first contact with the healthcare setting. This may underestimate the HIV test proposal rate, since HIV infection may be brought up at the second or a later visit. First, the effect of this bias was reduced by accounting only for visits that had occurred >3 months before HIV diagnosis, ensuring these events did not lead to HIV diagnosis. Next, a high proportion of patients with missed opportunities for HIV testing did not undergo any other HIV test prior to the test that diagnosed the HIV infection.
Only a few asymptomatic people belonging to a high-risk group other than MSM, i.e. sub-Saharan African migrants or heterosexuals with risky sexual behavior, were enrolled in the study and therefore we could not determine missed opportunities for them.
Our results are cause for concern and are in agreement with those reported in the medical literature that overall estimated high proportions of missed opportunities for HIV testing. However, a comparison of our results with other studies on the topic is difficult. Our study is unique in that it investigated the overall patient trajectory though the healthcare system; thus, it evaluated HIV testing strategies based on risk factors and symptoms such as applied currently. Most other studies
Our study illustrates the failure to identify persons at high risk of being HIV-positive. In the past three years, >80% of the 191 asymptomatic MSM had seen a healthcare provider. However, only 48% of them informed the provider, either spontaneously or upon questioning, that they belonged to a risk group. Indeed, patients may be wary of the healthcare setting, fearing moral judgment of their behavior, leading to unwillingness to disclose sexual practices to care providers
Even in patients identified as being at high risk of HIV infection, our study illustrates the failure to propose an HIV test: only one test proposed out of two MSM identified. Barriers to HIV testing on the part of the healthcare provider include a lack of training in HIV testing and a lack of self-confidence when proposing it
Thus, these findings suggest that risk-factor-based HIV testing is insufficient because its application is insufficient. In this situation, routine HIV testing may be attractive because it does not need risk assessment and follows an opt-out process. However, in US as in France where it is recommended, it is not widely applied because of implementation difficulties, and high costs
In patients with STIs, a test was only proposed to one out of two patients although a STI diagnosis is an indication for an HIV test
Conclusions
This study illustrates that the rate of missed opportunities for HIV testing remains unacceptable and highlights the low efficiency of risk-factor- and symptom-based HIV testing as it is currently applied. We must therefore reinforce and implement new HIV screening strategies. Testing initiated by the healthcare provider should be promoted. Because risk-factor based testing is not sufficient, routine screening should be considered in particular in high prevalence geographical areas and/or specific populations. In our study, most persons had contact with healthcare facilities and particularly general practitioners who should play an important role in future HIV testing strategies. But we must train physicians to better recognize HIV-related conditions and develop tools such as behavioral and clinical instruments to better identify those at high risk of HIV infection in whom HIV testing should be regularly performed. HIV testing has become a major tool and the basis of the new prevention strategies such as the Test and Treat strategy; it should be promoted urgently.
Ethical statement
The study received approval from two French data protection authorities (CCTIRS, Comité Consultatif sur le Traitement de l’Information en matière de Recherche dans le domaine de la Santé, and CNIL, Commission Nationale de l’Informatique et des Libertés) and did not requite formal ethical approval according to French regulations. All patients signed an informed consent form before taking part to the study.
Abbreviations
CCTIRS: Comité Consultatif sur le Traitement de l’Information en matière de Recherche dans le domaine de la Santé; CNIL: Commission Nationale de l’Informatique et des Libertés; ELISA: Enzyme-linked immunosorbent assay; HIV: Human immunodeficiency virus; IDU: Injected drug user; MSM: Man who have sex with man; UK: United Kingdom; US: United States.
Competing interest
No authors declare association that might pose a conflict of interest with the submitted work. However, LC received travel grants, participated in advisory boards, or received fares for public presentations from Gilead, Boehringer Ingelheim, Janssen-Tibotec, BMS, ViiV, MSD, and Pfizer. SDB received grants from Roche, Janssen-Cilag and Schering-Plough, and received consultancy honoraria from Merck and GlaxoSmithKline. YY received travel grants, honoraria for presentations at workshops and consultancy honoraria from Abbott, Bristol-Myers Squibb, Gilead, Merck, Roche, Tibotec and ViiV Healthcare.
Authors’ contribution
All authors contributed to the study. KC, SLV, SDB, CS and YY had the idea for the study. All authors substantially contributed to the conception, design and feasibility of the study. KC wrote the protocol and questionnaires used in the study. LC, KL and YY tested the questionnaires. KC coordinated the study. PB was responsible for management of data. KC and AC performed statistical analyses and presented results. All authors participated in interpretation of results. KC and YY drafted the article. All authors critically revised the manuscript for important intellectual content and approved the final version.