Online supplemental data1
SEARCH STRATEGY
Concept Table: Medline (Ovid SP) terms presented were adapted for each database
Medication / Medication Non-compliance / Choice behaviour (outcome) / Method – stated preferenceMeSH terms/subheadings / exp therapeutic uses/ / exp Medication Adherence/ / Patient Preference/or
decision making/ or choice behavior/
Text Words / (medication* or treatment) adj1 (adheren* or persisten* or complian*)).tw. / patient* or consumer*) adj1 preference*.tw or
(choice* adj1 behavio*r*) or (decision* adj1 making).tw. / (Stated adj1 preference*).tw or
(best adj1 worst).tw
Or
(standard adj1 gamble).tw
Or
(time adj1 trade adj1 off)
or
(conjoint adj1 analysis).tw.
(discrete adj1 choice adj1 experiment*).tw.
(contingent adj1 (rating* or valuation*)).tw.
Medline (Ovid SP)
1. (best adj1 worst).tw.
2. (Stated adj1 preference*).tw.
3. (standard adj1 gamble).tw.
4. (time adj1 trade adj1 off).tw.
5. (conjoint adj1 analysis).tw.
6. (discrete adj1 choice adj1 experiment*).tw.
7. (contingent adj1 (rating* or valuation*)).tw.
8. 1 or 2 or 3 or 4 or 5 or 6 or 7
9. exp Patient Satisfaction/
10. decision making/ or choice behavior/
11. ((patient* or consumer*) adj1 preference*).tw.
12. (choice* adj1 behavio*r*).tw.
13. (decision* adj1 making).tw.
14. 9 or 10 or 11 or 12 or 13
15. exp Medication Adherence/
16. (medication* adj1 (adheren* or persisten* or complian*)).tw.
18. 15 or 16
19. 8 and 14
20. (treatment* adj1 (complian* or adheren* or persisten*)).tw.
21. 18 or 20
23. exp Patient Compliance/
24. 16 or 20 or 23
25. 19 and 24
Online Supplemental data 2
Which drug attributes matter?
Safety [54]
The study among GIST patients evaluated preferences for the extent and risk of specified treatment side effects. There were distinct preferences between the different side effects and a reduction from a severe side effect to a moderate side effect more important than moderate to mild. Heart related disease was the most important side effect. All toxicities were stated as likely to influence adherence to treatment.
Dosing characteristics [44, 53]
In the study among acne patients [44], the mode of storage was found to be the more important attribute than dosing frequency, formulation, application method or shelf-life on treatment choice. However, this study demonstrated that strength of preference for dosing attributes may change with product experience, and may even change in direction (e.g. application method).
The study among parents of children suffering acute otitis media demonstrated a willingness to pay substantiallymore for more convenient dosing frequency[53]. In this study, prior experience with treatment and the scope of health gain compared to the hypothetical treatments presented influenced the estimated willingness to pay.
Treatment attributes were assumed to influence medication adherence without explicitly modelling adherence with stated preferences.
Efficacy versus safety[43]
In the study among bipolar patients, safety-related attributes pre-dominated treatment preferences. In particular, weight gain within three months was considered most important. The risk of life-threatening side effect was not considered as important as either efficacy or other safety attributes.
In this study, adherence to current medications was correlated with stated adherence to hypothetical medications.
Efficacy versus dosing versus cost versus other [42]
In the study among ulcerative colitis patients, the extent and risk of specified treatment inefficacy, dosing, cost and ease of administration were evaluated. In this case, the risk of inefficacy (e.g. reduced flare risk) was the most statistically significant attribute influencing treatment. Preference differences were noted between self-reported good versus poor adherers, with good adherers reporting a preference for clinical benefits of therapy rather than convenience and cost attributes.
Efficacy versus safety versus dosing versus cost [56]
Within a diabetes population, the extent of a specified efficacy attribute (i.e. reduced average glucose) was the most influential attribute on treatment choice, despite convenience, safety and cost significantly influencing choice. However, with respect to likely adherence, efficacy and safety were not important to respondents; instead only dosing and cost were influential.
Efficacy versus Safety versus Dosing versus Cost versus Other [41]
When all broad categories of drug-related attributes were compared within osteoporosis, the presence/absence of a specified side effect (nausea) was the most influential attribute on treatment choice. The expected duration of the side effect was described (i.e. 2 hours after treatment). In this study, the “other” attribute was the length of time treatment would be needed. For this study, the selection of a treatment as opposed to the no treatment option was inferred to mean adherent behaviour.
1