Background
Assumptions underlying valid electronic medication adherence monitoring: Internal validity
Assumption 1
Assumption 2
Assumption 3
Author & publication year | Study design | Description of the sample | n | EM | Description of the intervention | Outcome variable: medication adherence or clinical outcome | Result |
---|---|---|---|---|---|---|---|
Wagner et al. 2002 | Randomized controlled trial | A community convenience sample of adult HIV-positive patients on HAART | 117 | MEMS | Experimental group received EM; control group did not | Adherence measured with self-report 4 weeks after study start (% of prescribed pills taken) | Less adherence in the EM group (91%) than in the control group (94%; p=.73) |
Pre-post intervention study | A community convenience sample of adult HIV-positive patients on HAART monitored with EM | 60 | MEMS | EM started after baseline blood pressure measurement | Adherence measured with self-report at baseline and after 4 weeks | Less adherence after introducing EM (91%) compared to baseline (93%; p=.16) | |
Bertholet et al. 2000 | Pre-post intervention study | A convenience sample of primary care/hypertensive clinic patients with therapy-resistant hypertension | 69 | MEMS | EM started after baseline blood pressure measurement | Clinical outcome: blood pressure evaluation after 1 – 2 months | Blood pressure was lower after EM (14/9 cm Hg) compared to baseline (16/10 cm Hg; p < .001) |
Matsui et al. 1994 | Pre-post intervention study | A convenience sample of young β-thalassemia outpatients on a new iron chelator | 10 | MEMS | The purpose of EM was disclosed to patients after ± 11 months | Adherence measured by EM using the taking adherence parameter ± 18 months after disclosure | Greater adherence after disclosure (84%) compared to before (77%; p=.49) |
Yeung et al. 1994 | Quasi-experimental study | Non-equivalent study: two convenience samples of asthma patients on inhaling therapy | 21 | MDI | Intervention group given disclosure; control group not given disclosure | Adherence measured by EM using the taking adherence parameter after 2 – 3 weeks from the study start | Greater adherence in the disclosed group (81%) than in the undisclosed group (71%; p=.53) |
Elixhauser et al. 1990 | Randomized controlled trial | A convenience sample of psychiatric outpatients treated with lithium | 90 | Blister package | Experimental group received EM; control group did not | Adherence measured by self-reported, assay, % of expected prescription refills (after 2 – 4 months of study start) | Fewer expected prescription refills in the EM group (18%) than in the control group (31%; p < .01) |
Cramer et al. 1990 | Observational study | An unspecified sample of patients | 24 | MEMS | All patients received EM | Adherence measured by EM using the taking adherence parameter during the first and after a mean of 7 months from the start of the study | No difference before and after (79% vs. 79%). |
Assumption underlying valid electronic medication adherence monitoring: External validity
Purpose of the study
Methods
Design, sample, and setting
Variables and measurement
Adherence to the EM instructions
Dropped EM caps
Operational definition of EM-measured adherence
Other variables included in this study
Data collection
Data analysis
Results
Variable | Categories | Value |
---|---|---|
Age | Mean= 54 (sd = 13) | |
Gender | Male | 142 (56.8%) |
Living alone | No | 194 (77.6%) |
Employed | Yes | 130 (52.0%) |
Education | until age 11/12 years | 33 (13.2%) |
until age 12/13–14/15 years | 118 (47.2%) | |
until age 15/16–18/19 years | 26 (10.4%) | |
advanced (college) | 73 (29.2%) | |
Nationality | Swiss | 209 (83.6%) |
Immunosuppression | Cyclosporine & mycophenolate mofetil | 71 (28.4%) |
Cyclosporine | 38 (15.2%) | |
Cyclosporine & azathioprine | 37 (14.8%) | |
Azathioprine & prednisone | 18 (7.2%) | |
Azathioprine & tacrolimus | 14 (5.6%) | |
Other combinations | 72 (28.8%) | |
Monitored immunosuppressives | Mycophenolate mofetil | 103 (41.2%) |
Cyclosporine | 89 (35.6%) | |
Azathioprine/prednisone | 19 (7.6%) | |
Tacrolimus | 37 (14.8%) | |
Sirolimus | 2 (0.8%) | |
Self-reported EM influence on typical adherence | No influence | 188 (76.1%) |
Positive influence | 53 (21.5%) | |
Negative influence | 6 (2.4%) |
Assumption 1
Assumption 2
Assumption 3
Outcome variable | Parameter | Estimate | Standard error | Odds ratio (95% confidence interval) | DF | t value | p value |
---|---|---|---|---|---|---|---|
Omitted intakes | Random-intercepts variance | 2.845 | 0.445 | 241 | 6.39 | < .0001 | |
Intercept | -5.900 | 0.470 | 241 | -12.54 | < .0001 | ||
Exposure to EM (per month) | 0.270 | 0.055 | 1.31 (1.17–1.46) | 241 | 4.91 | < .0001 | |
Bottle size (per 100 ml) | 0.014 | 0.040 | 1.01 (0.94–1.10) | 241 | 0.93 | 0.35 | |
Influence perception | 0.277 | 0.328 | 1.32 (0.69–2.52) | 241 | 0.84 | 0.39 | |
Interviewer 1 vs. interviewer 4 | 0.461 | 0.356 | 1.59 (0.79–3.19) | 241 | 1.29 | 0.19 | |
Interviewer 2 vs. interviewer 4 | 0.011 | 0.365 | 1.01 (0.50–2.07) | 241 | 0.03 | 0.97 | |
Interviewer 3 vs. interviewer 4 | 0.022 | 0.438 | 1.02 (0.43–2.42) | 241 | 0.05 | 0.95 | |
Intake variability | Random-intercepts variance | 3.486 | 0.422 | 241 | 8.26 | < .0001 | |
Intercept | -3.033 | 0.414 | 241 | -7.31 | < .0001 | ||
Exposure to EM (per month) | 0.227 | 0.036 | 1.26 (1.17–1.35) | 241 | 6.67 | < .0001 | |
Bottle size (per 100 ml) | -0.106 | 0.037 | 0.90 (0.84–0.97) | 241 | -2.08 | 0.04 | |
Influence perception | 0.011 | 0.314 | 1.01 (0.54–1.88) | 241 | 0.04 | 0.97 | |
Interviewer 1 vs. interviewer 4 | 0.704 | 0.340 | 2.02 (1.04–3.95) | 241 | 2.07 | 0.04 | |
Interviewer 2 vs. interviewer 4 | 0.008 | 0.344 | 1.01 (0.51–1.99) | 241 | 0.02 | 0.98 | |
Interviewer 3 vs. interviewer 4 | 0.148 | 0.422 | 1.16 (0.50–2.67) | 241 | 0.35 | 0.73 |
Assumption 4
Variable | Subgroups | ||||||
---|---|---|---|---|---|---|---|
Non-participants | Non-participants | Participants | |||||
median | iqr b
| n | median | iqr | n | ||
Self-report a
| 0.0 | 0.0 | 65 | 0.0 | 0.0 | 284 | |
Collateral report a
| 1.0 | 0.1 | 35 | 1.0 | 0.3 | 164 | |
Assay: cyclosporine (mmol/l) | 112.5 | 50.0 | 50 | 105.0 | 56.0 | 191 | |
Assay: tacrolimus (mmol/l) | 7.6 | 3.2 | 21 | 7.2 | 3.9 | 44 | |
Assay: sirolimus (mmol/l) | 14.3 | 4.2 | 5 | 8.8 | 9.5 | 15 | |
Assay: mycophenolate mofetil (mmol/l) | 3.3 | 1.8 | 41 | 2.6 | 1.9 | 122 | |
Dropouts | Non-adherers to the EM guidelines | Adherers to the EM-guidelines | |||||
median | iqr | n | median | iqr | n | ||
Self-report | 0.0 | 0.0 | 36 | 0.0 | 0.0 | 244 | |
Collateral report | 1.0 | 0.7 | 35 | 1.0 | 0.3 | 224 | |
Assay: cyclosporine (mmol/l) | 108.0 | 43.0 | 25 | 104.5 | 58.5 | 164 | |
Assay: tacrolimus (mmol/l) | 6.2 | 6.1 | 5 | 7.2 | 3.7 | 39 | |
Assay: mycophenolate mofetil (mmol/l) | 2.7 | 1.8 | 17 | 2.6 | 1.9 | 101 |