Evaluating Adherence to Antiretroviral Therapy and Managing the Sub-Optimally Adherent HIV Patient

OVERVIEW: What every practitioner needs to know

All practitioners should measure adherence routinely at each visit. It is important for practitioners to accept the concept that non-adherence is typically ordinary behavior and to address it in a non-judgmental way. If non-adherence is identified, practitioners should query patients as to what their reasons for non-adherence might be and then problem solve with them to overcome the barriers. In follow-up, monitor for improved adherence and if non-adherence persists, implement more intensive strategies using social supports and technology.

Introduction: Imperfect adherence is the rule, not the exception

All HIV practitioners must recognize that even the most adherent patient misses doses of their antiretroviral medication. The perfectly adherent patient is a myth borne by our lack of measurement of adherence and our inattention to the issues in patients experiencing treatment success. The keys to managing this issue are having accurate measures of adherence, or at least, sensitive measures to open the conversation regarding adherence, avoiding being judgmental or critical when non-adherence is acknowledged or uncovered, and working in collaboration with the patient, their social supports and the clinical staff to identify potential solutions to the problem.

Adherence is a dynamic phenomenon – it must be maintained over time and may drop off in previously adherent patients. Providers must remain vigilant to non-adherence and avoid complacency with respect to asking about non-adherence at every encounter. Providers often use undetectable viral loads as a measure of adherence. But, the occult non-adherence that leads to treatment failure precedes virologic failure. Therefore, non-adherence can be present even with an undetectable viral load. If not identified and ameliorated, virologic failure with treatment resistance may then ensue.

Evidence for importance of adherence to treatment response

Dozens of studies over the last decade have demonstrated a strong dose-response relation between magnitude of adherence and virologic suppression.

These studies have been performed using various measures of adherence (microelectronic monitors, pharmacy refill data, self-reports, etc.) and with various regimens (unboosted and boosted protease inhibitors-based, NNRTI-based) and various clinical settings (developed and developing worlds).

Several themes have emerged from these studies. There is no clear threshold effect of adherence on outcome. That is, there is no amount of non-adherence that should be considered benign and no amount of adherence that should be considered useless. Rather, adherence should be viewed as a continuum with the more adherence, the more likely the treatment success, irrespective of method of measurement or regimen. The target of “95% adherence” is based on an arbitrary maximum amount in one of the original studies and should not be the focus of clinician’s efforts.

Rather, encourage maximal adherence, but acknowledge that missing doses can happen to anyone.

How much adherence is needed to avoid resistance?

The relation between adherence and resistance is complex and likely differs by pharmacokinetic/pharmacodynamic parameters and differences in host and virus biology. Empiric data are limited. Most thinking about resistance is informed by modeling exercises. For protease inhibitors, the relation between adherence and resistance is “U” shaped; that is, at the highest and lowest adherence rates, the amount of resistance is lowest, but for moderate amounts of adherence, drug pressure in the face of viral replication results in selection pressure for resistant clones which become dominant. For NNRTIs, modeling data suggest that the risk of resistance increases with increasing amounts of adherence. Data on other drugs is lacking. Since the threshold amount of adherence needed to prevent resistance for any regimen in any individual is unknown, the clinical message is unchanged. Encourage maximal adherence.

Which adherence problems have been minimized with more modern regimens?

The first highly active antiretroviral regimens were more difficult to take than current regimens; barriers to optimal adherence were common, including higher frequency of doses per day (e.g., thrice daily protease inhibitors), special instructions related to meals (e.g., indinavir on an empty stomach), special instructions related to storage (e.g., ritonavir refrigeration), suboptimal formulations (e.g., didanosine powder), and side effects (e.g., nelfinavir associated diarrhea).

Most patients can now be prescribed once daily regimens, typically with two or more of the drugs co-formulated, with less regard for food-drug interactions. As such, regimen dependent barriers have eased and more focus can be placed on personal and system-based barriers than drug-based barriers.

Answers to common questions raised by patients regarding adherence

What should I do if I forget to take my medication on time?

Data are scant. Recommend patients to take the medication as soon as they realize they are late, unless they are within 1/4 of the dosing interval before the next dose. For example, for a bid regimen, if they remember their missed dose anything longer than 4 hours before the next dose, recommend taking that dose and the next one. If it is 3 hours or less to the next dose, skip the forgotten dose and take the next dose on time. For a q24 hour regimen, this amounts to 6 hours before the next dose.

Should I take my medication if I am drinking?

Many patients fear “mixing drugs and alcohol” extrapolating from the risk of benzodiazepines or barbiturates potentiating the effects of alcohol. While counseling about keeping alcohol use to safe ranges, recommend that patients take their medications even if they are going to be drinking alcohol.

How many doses can I safely miss? Is it ok if I miss a dose every once in a while?

Avoid answering this question directly and redirect conversation to encourage plans for not missing any doses whilst acknowledging that everyone misses a dose of medication in the course of their treatment. While any missed doses increase the risk for treatment failure, we do not know the threshold for any one patient or regimen. Avoid predictions. Reemphasize that the goal is to avoid any missed doses, any amount of adherence should be praised and any amount of non-adherence should be addressed.

Measurement of adherence in clinical practice

The “adherent” patient can become “non-adherent” when new issues arise. Only by asking about adherence at EACH visit can these events be detected before they result in virologic failure.

How to measure adherence?

Two main modalities are available to the busy practitioner for measuring adherence-patient self-report and refill data.

Self-report questions:

“How many doses of your HIV medications did you miss this past week?”

“About what amount of your HIV medication did you take this past month – a few doses, about half, about 3/4ths (most), almost all (a couple of missed doses), all (no missed doses)?”

Caveat: self-reports questions must be administered in a permissive way, that is, non-adherence must be an acceptable answer.

Pros:

Simple, fast, opens dialog about non-adherence when admitted

Cons:

Recall inaccuracy even when patient is trying to be honest (i.e. most patients)

Harder to be non-judgmental than it might seem-but data inaccurate otherwise

Specific for non-adherence, but not sensitive. If non-adherence reported, very likely present, but if good adherence reported, may be inaccurate.

If patient wants to be deceptive, data useless (relatively uncommon).

Pharmacy refill data:

Obtain 2 fill dates from pharmacy or from label of patient’s pill bottles, or other records

Determine number of days supply provided by refill. Days supply equals number of tablets dispensed, divided by number of tablets prescribed to be taken per day.

Divide days supply by number of days between fills.

Multiply by 100 to get percent of doses taken over interval.

(Addendum: Above formula assumes empty bottle at time of refill. If pill count available and doses left over, days supply needs to be adjusted down. Subtract number of tablets remaining from number of tablets dispensed to adjust numerator in days supply formula above.)

Caveat: pharmacies typically provide the data to practitioners who call, but permission of the patient may be required by the pharmacist.

Pros:

This metric is more objective than self-report. It can be obtained without the patient present.

Cons:

This metric can miss prescriptions if filled at different pharmacies. It can overestimate adherence if refills are automated, or patient is not taking medications and is just getting them refilled. It is logistically more difficult than self-report.

Other potential measures of adherence

Several other modalities are available, but are more common in research than practice. However, as technology evolves, these may become more feasible for clinical implementation.

Microelectronic monitors: The fundamental approach is to use computer chips in pill caps, pill boxes or pill bottles to date and time stamp dose timing. The systems typically require maintaining the medication in the monitored packaging until a dose is taken.

Pros: Richly detailed data are captured electronically and automatically. There is the potential for electronic linkage for data feedback to patient and provider.

Cons: The costs are typically not covered by health plans and too expensive for most patients. It mandates the type of container for medication storage (e.g., it precludes patients from using their own preferred pill boxes).
Directly observed therapy:This approach is typically used as part of an intervention to improve adherence. In essence, it is self-report by proxy (i.e., the DOT administrator).

Pros: It provides more objective reports than simple self-report.

Cons: It is most feasible in the setting of an intervention. If performed by the patient’s social support, it may be susceptible to same cons as simple self-report (see above).
Provider estimates: It is strongly discouraged for providers to estimate patients’ adherence based on their own perceptions. Several studies have shown providers to be no better than “flip of a coin” in assessing patient adherence. Of course, a patient with a detectable viral load is more likely to have been non-adherent than a patient with a detectable viral load. However, non-adherence resulting in virologic failure can occur at least 3 months BEFORE the viral load breakthrough occurs. Therefore, undetectable viral load is an inadequate marker of adherence. Similarly, a detectable viral load can be due to causes other than non-adherence. Therefore, the above methods for measuring adherence must be implemented irrespective of provider estimate or viral load data.

When to intervene for non-adherence

Overarching Principle: there is no threshold for “optimal adherence”. Any missed doses should be addressed, although current regimens are relatively forgiving of a few missed doses per month.

If only a few missed doses – briefly explore reasons for missed doses and problem solve to prevent in future. Reassure but continue to encourage avoidance of doses have been missed.
If only taking 3/4ths of doses or fewer – spend more time explaining that they are at increased risk of virologic failure and resistance and then explore reasons for missed doses and problem solve to prevent in future.

Barriers to adherence

General Issues

Barriers differ by patient, and each patient may have multiple barriers. Each dose (AM versus PM) may have multiple and different barriers. Barriers that weren’t present before may arise. You must query about barriers.

Forgetting is ubiquitous and should simply be assumed to happen.
For other potential barriers, it is important to tailor recommendations to specific barriers.
Always ask specifically about: depression, drugs and alcohol use, difficulty getting to pharmacy.
Questions must be phrased in a neutral, non-judgmental way to obtain accurate information.

Specific barriers

Depression:

During the past month:

Have you been bothered by feeling down, depressed or hopeless?

Have you often been bothered by little interest or pleasure in doing things?

Drug and alcohol use:

What (alcohol) do you prefer to drink?

How many drinks per day?

When was the last time you have more than (4 for women, 5 for men) drinks in one day? (Never, in past 3 months, >3 months ago)

Ask open-ended question to prompt them for barriers:

“What do you believe gets in the way of your adhering to the HIV medications?”

Adherence intervention strategies

General strategies

Enlist social supports

Query patients regarding family and friends who may be helpful as reminders and supporters who encourage persistent adherence. NB: social supports may be negative (scolding, belittling, nagging); avoid these supports. Think only of positive (or mostly positive supports).
Allow patients to bring social supports to visits.
Teach social supporters about medications and what is expected of the patient.
Encourage social supporters to stay positive (“Did you take your medications today? (If yes-“great job”, if no-“ok, let me get them for you”). Avoid negative (“You’re going to get sick if you don’t take them! Now come on!!”)

Link dosing to key daily events

Draw a timeline of the typical day – wake up, brush teeth, eat breakfast, lunch, dinner, watch TV program, go to sleep, etc.
Set pill taking for memorable times (e.g., when brushing teeth before breakfast or before bed).

Use tools to remember and keep track of medications

Pill boxes filled weekly by patient or pharmacist (where possible)
Use reminder device-set cell phone alarm, alarm clock or alarm watch.

Strategies targeted to specific barriers

Falling asleep

Move dose up to earlier in the evening or recommend they take it on the way to bed.

Frequency of dosing or number and/or size of tablets

Change regimen to less frequent dosing, change to fixed dose combination, split/crush tablets/mix with water (as pharmacologically permissible)

Stigma/need to hide medications

Help create cover story of non-stigmatizing condition – “These are my medications for hypertension, hypothyroidism, etc.”

Depression

Start anti-depressant
Counsel
Refer to psychiatrist/psychologist for counseling

Intoxication

Recommend cutting back on drug/alcohol use, refer for drug/alcohol counseling
Recommend taking HIV medications before drugs/alcohol to avoid forgetting

Being away from home

Recommend keeping pill case with spare supply of one or two doses in case of unexpectedly being away

Running out at end of prescription/Inability to get to pharmacy

Set a date on the calendar after each prescription to identify when a refill is due
Call the pharmacy to see if they will call to remind patient of a refill being due
Use a near-empty bottle as a prompt to obtain a refill
Use a mail order pharmacy or pharmacy delivery service to remove logistics of getting to pharmacy

Emerging Strategies

SMS (text messaging)

In most settings, SMS is currently the only feasible technological strategy.
SMS alone only serves as an alert to address forgetting. It is ineffective for most barriers.
SMS with linkage to a problem solver can be effective. Patients are required to reply to an SMS querying about “medication problems.” If the patient replies “Yes” or fails to reply, a problem solver is alerted and implements the “Strategies Targeted to Specific Barriers” (see above).

Directly observed therapy (DOT)

Many approaches to DOT exist including patient-nominator DOT supporters and system-designated DOT supporters. Like SMS-based interventions, DOT must be linked to problem solving solutions.
DOT does not improve outcomes and is not recommended for patients without clearly identified poor adherence in the past or obvious current adherence barriers.
For patients with drug abuse issues, several studies suggest a role for DOT. This is delivered in specialized settings.

Assess effect of intervention

Once intervention is implemented, provider should see patient at 1-2 weeks after the visit (depending on severity of non-adherence) and reassess if:

Intervention implemented (suggestions were enacted)
Adherence improved in response to implementation

If both yes- provide encouragement to continue. If either a = yes, b = no, OR a = no, brainstorm for other solutions to the barrier and repeat step above for assessment.

Medical monitoring of the non-adherent patient

If non-adherence resolved and viral load undetectable: check viral load at regularly scheduled interval
If non-adherence resolved and viral load detectable: determine if new regimen is warranted by obtaining resistance test
If non-adherence persists:

Ask the patient for concrete details of the potential harm the non-adherence will likely cause. This will demonstrate their level of understanding of the seriousness of non-adherence and detectable viremia. If knowledge lacking, re-educate (in a non-judgmental manner).

Determine if a new approach to intervention is warranted – i.e., repeat problem-solving activities.

What's the evidence?

Thompson, M, Mugavero, MJ, Amico, R, Cargill, V, Chang, L, Gross, R. “Guidelines for Improving Entry into and Retention in Care and Antiretroviral Adherence for Persons with HIV: Evidence-Based Recommendations from an IAPAC Panel of International Experts”. Annals of Internal Medicine. vol. 156. 2012. pp. 817-33. (These guidelines summarize the major work in the fields of adherence measurement and intervention since 1995.)

Gross, R, Bellamy, S, Chapman, J, Han, X, O’ Duor, J, Palmer, S. “Managed Problem Solving for Antiretroviral Therapy Adherence: A Randomized Trial Running Head: MAPS Adherence Trial”. JAMA Internal Medicine. 2013. pp. 1-7. (This study demonstrated the effectiveness of a problem-solving-based HIV adherence intervention.)

Lester, RT, Ritvo, P, Mills, EJ, Kariri, A, Karanja, S. “Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): a randomised trial”. Lancet. vol. 376. 2010. pp. 1838-1845. (This study demonstrated the effectiveness of querying patients about their adherence and responding with problem solving strategies when non-adherence was identified.)

de Bruin, M, Viechtbauer, W, Schaalma, HP, Kok, G, Abraham, C. “Standard care impact on effects of highly active antiretroviral therapy adherence interventions: A meta-analysis of randomized controlled trials”. Arch Intern Med. vol. 170. 2010. pp. 240-250. (This study demonstrated the importance of the quality of standard of care in clinical settings on HIV treatment adherence and outcomes.)

Ford, N, Nachega, JB, Engel, ME, Mills, EJ. “Directly observed antiretroviral therapy: a systematic review and meta-analysis of randomised clinical trials”. Lancet. vol. 374. 2009. pp. 2064-2071. (This study demonstrated the lack of effect of DOT in populations without a history of poor adherence and/or specific adherence barriers identified.)

No sponsor or advertiser has participated in, approved or paid for the content provided by Decision Support in Medicine LLC. The Licensed Content is the property of and copyrighted by DSM.

Infectious Diseases