Appendix X: Advantages and Disadvantages of Adherence Measures

Updated June 2005

Advantages and Disadvantages of Adherence Measures

Table 1: Advantages and Disadvantages of Adherence Measures

Method	Advantages	Disadvantages
Self-report	Easily obtained using patient interview or questionnaire (report of non-adherence is more reliable than report of adherence) Inexpensive	Overestimates adherence Correlation is dependent on patient’s relationship with staff Individuals may give providers what they perceive as socially desirable, “right” responses
Pill counts	Useful adjunct to self-report Unannounced pill counts may be more accurate Direct costs minimal	Tends to overestimate adherence because of “pill dumping” before visit Casts provider in the role of medication monitor and not ally or advocate Indirect costs may be a concern due to time constraints Does not prove that patient actually took medication
Electronic monitoring	Best correlation with virologic outcomes Allows more detailed view of non-adherence patterns Most accurate measure	Expensive and generally reserved for clinical trials Precludes use of pillbox Fails if multiple medications are kept in a single bottle or if multiple doses are taken out at one time Requires carrying the container Subject to “pocket doses” (removing more than one dose at a time) Does not guarantee that the patient took the medication
Pharmacy refill monitoring	Easy, minimal time commitment Timely refilling of prescriptions correlates well with adherence Most successful when limited to patient using one pharmacist Is a useful adjunct to self-report Effective in understanding adherence behavior in large populations	Patients may use more than one pharmacy Does not equate with medication-taking
Therapeutic drug monitoring	Low drug levels confirm non-adherence, but therapeutic drug levels do not confirm adherence	Pharmacokinetic levels for most drugs have not been well established Only confirms the pre-measurement adherence, long-term adherence still unknown
Hematologic monitoring using either complete blood counts or expanded chemistry panels	Confirms patient reporting	Only effective for certain drugs: zidovudine, stavudine (increased MCV); indinavir (increased bilirubin) Not always reliable
Directly observed therapy	100% adherence, in theory Ideal method for institutional settings (prisons, nursing homes, residential treatment programs, etc.)	Labor intensive Not practical for complex regimens with multiple doses and/or dietary restrictions May compromise confidentiality
Modified directly observed therapy (observation of most but not all medication doses)	100% adherence, in theory Ideal method for ambulatory settings	Labor intensive Concern for development of resistance if plan not followed
Viral load	Can correlate with adherence	Does not necessarily indicate non-adherence; not all individuals with virologic failure will be poor adherers May overestimate adherence Virologic failure can be indicative of drug resistance
Provider estimation	None	Most poorly correlated with actual adherence