Limits...
Examining assumptions regarding valid electronic monitoring of medication therapy: development of a validation framework and its application on a European sample of kidney transplant patients.

Denhaerynck K, Schäfer-Keller P, Young J, Steiger J, Bock A, De Geest S - BMC Med Res Methodol (2008)

Bottom Line: To test internal validity, we examined if (1) EM equipment functioned correctly, (2) if all EM bottle openings corresponded to actual drug intake, and (3) if EM did not influence a patient's normal adherence behavior.Internal validity was determined by assessing the prevalence of nonfunctioning EM systems, the prevalence of patient-reported discrepancies between cap openings and actual intakes (using contemporaneous notes and interview at the end of the study), and by exploring whether adherence was initially uncharacteristically high and decreased over time (an indication of a possible EM intervention effect).Our analysis revealed that some assumptions were not fulfilled: 1) one cap malfunctioned (0.4%), 2) self-reported mismatches between bottle openings and actual drug intake occurred in 62% of the patients (n = 155), and 3) adherence decreased over the first 5 weeks of the monitoring, indicating that EM had a waning intervention effect.

View Article: PubMed Central - HTML - PubMed

Affiliation: Institute of Nursing Science, University of Basel, Bernoullistrasse 28, 4056, Basel, Switzerland. kris.denhaerynck@unibas.ch

ABSTRACT

Background: Electronic monitoring (EM) is used increasingly to measure medication non-adherence. Unbiased EM assessment requires fulfillment of assumptions. The purpose of this study was to determine assumptions needed for internal and external validity of EM measurement. To test internal validity, we examined if (1) EM equipment functioned correctly, (2) if all EM bottle openings corresponded to actual drug intake, and (3) if EM did not influence a patient's normal adherence behavior. To assess external validity, we examined if there were indications that using EM affected the sample representativeness.

Methods: We used data from the Supporting Medication Adherence in Renal Transplantation (SMART) study, which included 250 adult renal transplant patients whose adherence to immunosuppressive drugs was measured during 3 months with the Medication Event Monitoring System (MEMS). Internal validity was determined by assessing the prevalence of nonfunctioning EM systems, the prevalence of patient-reported discrepancies between cap openings and actual intakes (using contemporaneous notes and interview at the end of the study), and by exploring whether adherence was initially uncharacteristically high and decreased over time (an indication of a possible EM intervention effect). Sample representativeness was examined by screening for differences between participants and non-participants or drop outs on non-adherence.

Results: Our analysis revealed that some assumptions were not fulfilled: 1) one cap malfunctioned (0.4%), 2) self-reported mismatches between bottle openings and actual drug intake occurred in 62% of the patients (n = 155), and 3) adherence decreased over the first 5 weeks of the monitoring, indicating that EM had a waning intervention effect.

Conclusion: The validity assumptions presented in this article should be checked in future studies using EM as a measure of medication non-adherence.

Show MeSH
Observed course of non-adherence over time.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2275282&req=5

Figure 4: Observed course of non-adherence over time.

Mentions: The random-intercepts logistic regression analysis confirmed an increase in both taking and timing non-adherence over time (Table 3). The odds on non-adherence increased over one month by about 30% for taking (OR: 1.31; 95%CI: 1.17–1.46) and 25% for timing adherence (OR: 1.26; 95% CI: 1.17–1.35). In addition, the nonlinear regression lines showed that the increase in both dimensions mainly occurred during the first 5 weeks of monitoring (Figure 4). After day 35, the taking dimension of non-adherence stabilized. The average percentage of correctly dosed days was 96.7% when including the entire 3-month measurement period, but slightly decreased to 96.3% when the first 35 days were excluded. The timing dimension of non-adherence stabilized after about day 50. The average percentage of correctly timed intakes was 91.8% when including all data points, and slightly decreased to 91.4% when only considering the stable phase between day 50 and 75. A post hoc analysis identifying potential interactions between exposure to EM and perception of the EM-intervention effect, showed a stronger EM-intervention effect in patients acknowledging an intervention effect than in patients stating that they experienced no intervention effect (p = 0.003).


Examining assumptions regarding valid electronic monitoring of medication therapy: development of a validation framework and its application on a European sample of kidney transplant patients.

Denhaerynck K, Schäfer-Keller P, Young J, Steiger J, Bock A, De Geest S - BMC Med Res Methodol (2008)

Observed course of non-adherence over time.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2275282&req=5

Figure 4: Observed course of non-adherence over time.
Mentions: The random-intercepts logistic regression analysis confirmed an increase in both taking and timing non-adherence over time (Table 3). The odds on non-adherence increased over one month by about 30% for taking (OR: 1.31; 95%CI: 1.17–1.46) and 25% for timing adherence (OR: 1.26; 95% CI: 1.17–1.35). In addition, the nonlinear regression lines showed that the increase in both dimensions mainly occurred during the first 5 weeks of monitoring (Figure 4). After day 35, the taking dimension of non-adherence stabilized. The average percentage of correctly dosed days was 96.7% when including the entire 3-month measurement period, but slightly decreased to 96.3% when the first 35 days were excluded. The timing dimension of non-adherence stabilized after about day 50. The average percentage of correctly timed intakes was 91.8% when including all data points, and slightly decreased to 91.4% when only considering the stable phase between day 50 and 75. A post hoc analysis identifying potential interactions between exposure to EM and perception of the EM-intervention effect, showed a stronger EM-intervention effect in patients acknowledging an intervention effect than in patients stating that they experienced no intervention effect (p = 0.003).

Bottom Line: To test internal validity, we examined if (1) EM equipment functioned correctly, (2) if all EM bottle openings corresponded to actual drug intake, and (3) if EM did not influence a patient's normal adherence behavior.Internal validity was determined by assessing the prevalence of nonfunctioning EM systems, the prevalence of patient-reported discrepancies between cap openings and actual intakes (using contemporaneous notes and interview at the end of the study), and by exploring whether adherence was initially uncharacteristically high and decreased over time (an indication of a possible EM intervention effect).Our analysis revealed that some assumptions were not fulfilled: 1) one cap malfunctioned (0.4%), 2) self-reported mismatches between bottle openings and actual drug intake occurred in 62% of the patients (n = 155), and 3) adherence decreased over the first 5 weeks of the monitoring, indicating that EM had a waning intervention effect.

View Article: PubMed Central - HTML - PubMed

Affiliation: Institute of Nursing Science, University of Basel, Bernoullistrasse 28, 4056, Basel, Switzerland. kris.denhaerynck@unibas.ch

ABSTRACT

Background: Electronic monitoring (EM) is used increasingly to measure medication non-adherence. Unbiased EM assessment requires fulfillment of assumptions. The purpose of this study was to determine assumptions needed for internal and external validity of EM measurement. To test internal validity, we examined if (1) EM equipment functioned correctly, (2) if all EM bottle openings corresponded to actual drug intake, and (3) if EM did not influence a patient's normal adherence behavior. To assess external validity, we examined if there were indications that using EM affected the sample representativeness.

Methods: We used data from the Supporting Medication Adherence in Renal Transplantation (SMART) study, which included 250 adult renal transplant patients whose adherence to immunosuppressive drugs was measured during 3 months with the Medication Event Monitoring System (MEMS). Internal validity was determined by assessing the prevalence of nonfunctioning EM systems, the prevalence of patient-reported discrepancies between cap openings and actual intakes (using contemporaneous notes and interview at the end of the study), and by exploring whether adherence was initially uncharacteristically high and decreased over time (an indication of a possible EM intervention effect). Sample representativeness was examined by screening for differences between participants and non-participants or drop outs on non-adherence.

Results: Our analysis revealed that some assumptions were not fulfilled: 1) one cap malfunctioned (0.4%), 2) self-reported mismatches between bottle openings and actual drug intake occurred in 62% of the patients (n = 155), and 3) adherence decreased over the first 5 weeks of the monitoring, indicating that EM had a waning intervention effect.

Conclusion: The validity assumptions presented in this article should be checked in future studies using EM as a measure of medication non-adherence.

Show MeSH