The Association of Clinical Follow-Up Intervals in HIV-Infected Persons with Viral Suppression on Subsequent Viral Suppression

Introduction

P ersons with HIV infection are living longer with the disease as a result of effective antiretroviral therapy (ART). ¹ ART is a life-long treatment that requires repeated laboratory monitoring and HIV primary care clinic visits, and which results in substantial costs for patients and healthcare systems. How often HIV-infected persons with viral suppression need to have clinical follow-up is difficult to study. Treatment failure rates are low, and randomizing persons to defined intervals would be difficult to justify. To attempt to answer this question, Reekie et al. ² examined virologic failure rates in a cohort of 2240 participants on stable and fully suppressive ART regimens for at least 1 year. After 3, 6, and 12 months, 0.1%, 1.5%, and 4.6% of participants had virologic failure, respectively. The authors concluded that it might be reasonable to extend monitoring to 6 months, given the low failure rate at 3 months. ² A small study of persons who had been on ART for 15 months and had viral suppression found that virologic failure was not associated with time between viral load tests. In fact, in unadjusted analyses, persons with failure had more monitoring, not less. ³

Health care providers likely use objective and subjective criteria to determine when a patient ought to be scheduled to return for a follow-up visit. The current U.S. Department of Health and Human Services treatment guidelines state that in persons on a stable ART regimen, viral load assessments should be repeated every 3–4 months. However, the guidelines do acknowledge that some clinicians may extend the interval to every 6 months for adherent individuals who have maintained suppressed viral loads for more than 2–3 years and who are clinically stable. ⁴ This is a B-III recommendation (i.e., it is based on expert opinion).

One limitation of previous observational studies is that they are based on when individuals are actually seen, not when they are scheduled to be seen. Retention in HIV care is often poor, ⁵ no-show rates are high, ⁶ and unintended gaps in care are frequent. ⁷ On the other hand, individuals may present earlier for care than scheduled because of drug side effects or inter-current illnesses. Thus, estimates based on actual visits may be biased. For the clinician seeing a patient, the immediately relevant question is, when should I ask the patient to return for his or her next visit? To our knowledge, no study has attempted to answer whether clinicians can safely extend the requested follow-up interval.

We conducted a multi-center study using clinic-wide data on scheduled appointments, regardless of whether the appointment was attended or missed. Using a retrospective cohort design, we sought to estimate virologic failure rates based not on when individuals attended clinic next, but on when they were scheduled to attend the next visit by their provider. This unique dataset therefore would allow us to answer an important question: Do providers in routine care safely extend follow-up intervals for their patients with viral suppression? More directly, we wanted to determine if scheduling a follow-up interval at 4 or 6 months is associated with an increased risk of virologic failure at 1 year, compared to the more standard 3-month follow-up interval.

Methods

Results

There were 7770 participants who completed a visit in May, June, or July 2008. Thirty-one percent of these participants had viral suppression but had scheduled visit intervals that were outside the 3, 4, and 6 month windows (most had an interval that was <3 months); 9% met the visit interval criterion but did not have a viral load of ≤400 c/mL at their index visit; and 32% failed to meet both criteria. As a result, 2171 participants (28%) were included in the analysis. Mean (SD) participant age was 47 (10) years (Table 1). Seventy percent of the participants were male and 52% reported their HIV risk factor as heterosexual sex. Similar numbers of participants had private health insurance, Medicare, Medicaid, and no insurance. The median (IQR) CD4 cell count at the index visit was 497 (345, 692) cells/mm³. Sixty-six percent of participants had a follow-up appointment scheduled at 3 months, 26% at 4 months, and 8% at 6 months. The majority of the participants scheduled at a 6-month follow-up interval came from a single site.

Participants who were scheduled a follow-up appointment in 3 months were more likely to be non-Hispanic black (62%), while participants who were scheduled an appointment at 6 months were more likely to be non-Hispanic white (54%, p<0.01, Table 2). Participants scheduled for follow-up at 6 months were more likely to have private insurance (54%) than participants scheduled at 3-month (18%) or 4-month (19%) intervals (p<0.01). Participants scheduled a next follow-up visit at 6 months were more likely to be men who have sex with men (MSM; 53%) compared to participants scheduled at 3-month (31%) and 4-month (39%) intervals (p<0.01). Longer intervals between the index and next scheduled visit were also associated with higher index CD4 cell counts (p<0.01). Seventeen percent of participants cancelled and 18% missed their next appointment following the index visit. Fewer persons scheduled for a 6-month follow-up appointment missed that visit (7%), compared to persons scheduled follow-up at 3 months (19%) and 4 months (20%; p<0.01).

Counting missing viral loads as virologic failure, 25% of participants had virologic failure at 12 months. Participants with a scheduled follow-up appointment at 3 months were as likely to have virologic failure at 12 months (25%) as participants with scheduled follow-up at 4 months (25%) or 6 months (24%, p=0.95). Male participants were more likely to have virologic failure at 12 months (26%) compared to females (22%, p=0.03) (Table 3). Participants with a CD4 count<200 cells/mm³ at the index visit were more likely to have virologic failure at 12 months (33%) compared to participants with CD4 cell counts between 200 and 350 cells/mm³ (26%) and >350 cells/mm³ (23%; p=0.02). Forty-two percent of participants who missed their next follow-up appointment had virologic failure and 29% of persons who cancelled their next follow-up appointment had virologic failure compared to 19% of persons who attended their follow-up appointment (p<0.01).

The majority of the virologic failures (79%) were due to missing viral loads at 12 months. There were no statistically significant differences between participants who had available 12-month viral loads and participants whose 12-month viral loads were missing on the following characteristics: gender, age, race/ethnicity, HIV risk factor, HIV duration, or CD4 count at the index visit. Not unexpectedly, a higher percentage of participants with missing 12-month viral loads had cancelled (21%) or missed (32%) their follow-up appointment compared to participants who had available 12-month viral loads (17% and 15%, respectively, p<0.01).

In analyses in which persons with missing 12-month viral loads were excluded, 6.5% of participants experienced virologic failure at 12 months, including 7.1% of persons with a 3-month follow-up interval, 5.7% of persons with a 4-month interval, and 4.5% of persons with a 6-month interval (p=0.35). Among the other variables we studied, virologic failure was statistically significantly predicted by insurance status (3.0%, 6.5%, 7.3%, and 8.5% for persons with private health insurance, Medicaid, Medicare, and no health insurance, respectively [p=0.01]), CD4 cell count at index visit (12.0%, 9.2%, and 5.3% for persons with CD4 count<200 cells/mm³, between 200 and 350 cells/mm³, and >350 cells/mm³, respectively [p<0.01]), and disposition of the next scheduled follow-up appointment (5.1%, 7.9%, and 11.6% for persons who kept, cancelled, and missed the appointment, respectively [p<0.01]).

Using a multivariate GEE analysis with site as a cluster variable and where missing 12-month viral loads were considered failures, we determined the factors associated with virologic failure at 12 months (Table 4). Compared to participants scheduled follow-up at a 3-month interval, participants scheduled to follow-up at either 4 months (aOR 1.04, 95% CI 0.88–1.24, p=0.64), or 6 months (aOR 0.97, 95% CI 0.57–1.65, p=0.91) were as likely to have virologic failure at 12 months. Female sex was associated with a lower odds of having virologic failure (aOR 0.79, 95% CI 0.71–0.88, p<0.01) as was Hispanic ethnicity (aOR 0.65, 95% CI 0.56–0.76, p<0.01, compared to white, non-Hispanic participants). Index CD4 cell count <200 cells/mm³ was associated with a higher odds of virologic failure (aOR 1.77, 95% CI 1.28–2.44, p<0.01, compared to an index CD4 cell count >350 cells/mm³). Cancelling the next scheduled follow-up appointment (aOR 1.60, 95% CI 1.29–1.98, p<0.01), and missing the next scheduled follow-up appointment (aOR 2.92, 95% CI 2.71–3.15, p<0.01) predicted virologic failure compared to attending that follow-up appointment.

In the multivariate GEE analysis in which persons with missing 12-month viral loads were excluded, participants with a scheduled follow-up appointment at 4 months were as likely as participants scheduled a 3-month follow-up appointment to have virologic failure at 12 months (aOR 0.88, 95% CI 0.62–1.27, p=0.50), as were participants with scheduled follow-up at 6 months (aOR 0.85, 95% CI 0.36–1.97, p=0.70). The following variables were associated with higher odds of virologic failure: not having private health insurance (Medicaid [aOR 2.72, 95% CI 1.68–4.41, p<0.01], Medicare [aOR 2.59, 95% CI 1.54–4.35, p<0.01], and having no insurance [aOR 3.82, 95% CI 2.11–6.93, p<0.01] compared to private insurance); CD4 cell count at index visit <200 cells/mm³ (aOR 2.41, 95% CI 1.92–3.03, p<0.01) compared to >350 cells/mm³, and missing the next scheduled follow-up appointment (aOR 2.23, 95% CI 1.61–3.07, p<0.01) compared to attending that appointment.

Discussion

In this multisite retrospective cohort study of 2171 HIV-infected persons with viral suppression in routine care, participants who were scheduled a next HIV primary care clinic visit in 4 or 6 months were not at increased risk of virologic failure at 12 months compared to persons with follow-up scheduled at 3 months. Furthermore, missing or cancelling the next follow-up appointment was independently associated with 12-month virologic failure, and, among the factors we could examine, was the strongest predictor of failure. Even when missing 12-month viral loads were excluded from the analyses, missing the follow-up appointment still predicted virologic failure. These observations lead us to conclude that clinicians providing routine care are able to safely extend scheduled next follow-up intervals beyond 3 months in persons with viral suppression, at least in the short-term. Further, retention in care is strongly predictive of viral suppression, regardless of scheduled follow-up interval.

Few data exist to guide clinicians regarding when to schedule follow-up visits for virally suppressed individuals. Our data extend the previous literature by using the intended follow-up interval as the primary explanatory variable, rather then the observed follow-up interval. In prior work, virologic failure rates in a population with at least 1 year of viral suppression were 0.1% at 3 months, 1.5% at 6 months, and 4.6% at 12 months in one large study that did not count missing values as failures. ² The rate of failure we found at 12 months after removing missing data was 6.5%. We were unable to assess how long participants had been suppressed because the dataset did not archive that data. However, clinicians almost certainly factored that information as well as other objective and subjective data into their recommended follow-up interval. Persons who were scheduled at a 6-month interval may have been suppressed longer than persons scheduled at a 3-month interval. Importantly, our data suggest that providers are able to make these decisions safely for their patients, at least in the short-term. The long-term effects of lengthening the intervals between visits are unknown. Efforts to extend the follow-up interval should be balanced against possible negative impacts on the management of comorbidities, early detection of virologic failure, the patient–provider relationship, and retention in care, and so must be done with deliberate clinical judgment. ^{9 –11}

In addition to clinical implications, our study has other implications. Primary care costs (including laboratory but excluding pharmacy costs) in one study were an estimated $3,470 per person per year. ¹² Outpatient care costs constitute about 9% of the lifetime medical costs for HIV-infected persons. ¹³ As more HIV-infected persons live longer, associated outpatient healthcare costs will likely increase. If clinicians safely schedule some patients to be seen less frequently than every 3 months, as our study indicates is possible, cost savings could be realized. In addition, many HIV clinics are overburdened, and lengthening the interval between visits may provide capacity for patients recruited through routing HIV testing and retention in care efforts. This strategy would also benefit resource-constrained settings whose health systems may be over capacity. Finally, less frequent visits would benefit HIV-infected individuals by decreasing their direct costs of care (e.g., co-pays and transportation costs) as well as their opportunity costs (e.g., income lost from time off from work to attend appointments). The cost and capacity savings of extending the follow-up interval will only be realized if providers can extend follow-up intervals without compromising patients' health. Reekie et al. clearly demonstrated that consistency of viral suppression predicted subsequent risk of failure: persons who were suppressed at least 80% of the time before inclusion in their study were less likely to fail than persons suppressed less consistently. ² Additional studies are needed to offer specific guidance as to when to extend follow-up intervals and for what patients.

Participants in our study who had missed or cancelled their next follow-up appointment had greater odds of virologic failure at 12 months. A number of earlier studies have found similar results. ^{14 –17} Notably, our study is among the first to evaluate the impact of cancelled appointments on HIV outcomes. Most prior studies have focused on missed or “no show” visits. Study findings suggest cancelled appointments are similarly associated with virologic failure, albeit with a less strong association than observed for missed appointments. Future research is needed to delineate the impact of cancelled visits on HIV health behaviors and outcomes, but efficacious interventions to improve retention in care are needed.

Persons in our study with a CD4 cell count <200 cells/mm³ were more likely to have virologic failure at 12 months. CD4 cell count at entry into our cohort may be at least a partial surrogate for duration of viral suppression prior to entering the cohort (a variable that we could not measure). Of course, initiating ART in persons with low CD4 cell counts is associated with poorer treatment outcomes, including less durable viral suppression. ^{18 –21} Low CD4 cell counts are associated with an increased risk of death even after viral suppression, making it imperative that HIV-infected persons receive prompt diagnosis and treatment. ²²

In the analyses that excluded persons with missing 12-month viral loads, participants without private health insurance were more likely to have virologic failure. The same trends were observed when missing viral loads were considered failures (Table 4). Persons without private insurance have well-documented difficulty accessing medical care and obtaining ART consistently ^{23 –25} and as a population have higher viral loads ²⁶ and mortality rates. ²⁷

Our study has several limitations. It is an observational study that included only six clinics. We had a limited number of participants who were scheduled a 6-month follow-up appointment and most of those persons were concentrated at one of the six sites. Our adjusted analyses relied on GEE with site as a cluster variable, but residual confounding may be present. We do not have data on how long our participants had viral suppression prior to their index visit, or the specifics of their ART regimens. We do note that among the 12,690 persons seen at the clinics during 2008, with a mean (SD) of 3.37 (2.0) VL results per patient, 50.0% had viral suppression at all measurements, 26.5% never had viral suppression during the year, and 23.5% had mixed VL results.

We also do not have information on our participants' adherence to ART, and individuals who miss clinic visits are more likely to have poor adherence to ART. ¹⁶ Data on depression and other co-morbidities are also not available. These co-morbid conditions could impact adherence and some of the decisions about when individuals should return for follow-up. Our results may not be generalizable to clinics that are not academically affiliated.

This multicenter retrospective cohort study suggests that clinicians in routine care can safely extend follow-up intervals in persons with viral suppression, at least in the short-term. Increasing the length of time between follow-up visits in the right individuals has the potential to decrease healthcare costs and relieve overextended clinics operating near or above capacity. Regardless of scheduled visit interval, maximizing appointment adherence remains critical for maintaining viral suppression.