| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the The Toronto Hospital (J.I., B.B., J.S., S.J.), Departments of Psychiatry (J.I., B.B.) and Medicine (P.D.), University of Toronto, and St. Michael’s Hospital (M.P., P.D.), Toronto, Ontario; Dean, Faculty of Medicine, University of British Columbia, Vancouver, British Columbia (J.C.); and Departments of Medicine (S.C.) and Clinical Epidemiology and Biostatistics (R.R., M.G.), McMaster University, Hamilton, Ontario, Canada.
Address reprint requests to: Jane Irvine, DPhil, CPsych, Psychology Department, cw-2-330, The Toronto Hospital, 200 Elizabeth St., Toronto, Ontario, Canada, M5G 2C4. Email: jane.irvine{at}utoronto.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Subjects were 1141 patients participating in the Canadian Amiodarone Myocardial Infarction Arrhythmia Trial. Poor adherence to study medication (amiodarone or placebo), measured by pill count over 2 years, was defined as the lower 20th percentile of the pill count distribution. Predictors of adherence were also studied and included demographic and cardiac variables and, in a subset of participants (N = 671), measures of depression, distress, hostility, and social support.
RESULTS: In survival analysis controlling for cardiac and demographic variables, poor adherence in the placebo and amiodarone groups was associated with an increased risk of sudden cardiac death (relative risk (RR) = 2.11, 95% confidence interval (CI) = 1.03–4.56, p < .05; and RR = 3.15, 95% CI = 1.34–7.44, p < .01, respectively), total cardiac mortality (RR = 2.04, 95% CI = 1.12–3.72, p < .02; and RR = 2.49, 95% CI = 1.32–4.72, p < .01, respectively), and all-cause mortality (RR = 2.25, 95% CI = 1.27–3.97, p < .001; and RR = 2.34, 95% CI = 1.32–4.17, p < .004, respectively). Logistic regression analysis identified two predictors of poor adherence to placebo: age > 70 years (odds ratio = 2.18, 95% CI = 1.11–4.29, p < .03) and social activities in the month before the index heart attack (odds ratio = 1.02, 95% CI = 1.00–1.04, p < .05).
CONCLUSIONS: Poor adherence is associated with a greater risk of mortality. The relationship between adherence and social activities suggests a higher motivation to adhere to treatment in individuals more engaged in enjoyable activities.
Key Words: adherence • placebo therapy • mortality
Abbreviations: CAMIAT = Canadian Amiodarone Myocardial Infarction ArrhythmiaTrial; CHF = congestive heart failure; CI = confidenceinterval; MI = myocardial infarction; OR = odds ratio; RR = relative risk; SCD = sudden cardiac death; VPD =ventricular premature depoloarization.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Two studies have found that adherence to pharmacological therapy, even if it is adherence to a placebo, is associated with better survival. The first was the Coronary Drug Project Research Group study of lipid-lowering therapy (4). Poor adherence was defined as taking less than 80% of pills prescribed. Thirty-three percent of the placebo group and 34% of the active drug group were classified as poor adherers by this measure. Mortality rates during follow-up were significantly higher in poorly adherent than adherent patients in both the active treatment group (15.0% vs. 24.6%, p < .0001) and the placebo treatment group (15.1% vs. 28.3%, p < .0001). Differences between the adherent and poorly adherent groups remained significant even after controlling for 41 baseline characteristics. A similar finding was observed in the Beta Blocker Heart Attack Trial (5). Five percent of the sample was classified as poorly adherent, defined as taking less than 76% of the pills prescribed. Poor adherence was associated with a threefold greater risk of mortality in the active treatment group (p = .08) and a 2.5 times greater risk of mortality in the placebo group (p = .10). A follow-up publication reported similar findings in women (6). The results of these studies emphasize the health benefit of adherence. However, the underlying mechanisms are not understood.
It has been suggested that the mechanisms underlying the nonspecific adherence effect may be similar to the mechanisms underlying the placebo effect (7). For instance, the placebo effect is thought to be a function of the nonspecific components of treatment, such as expectancy and belief in the benefits of therapy. Alternatively, it is also possible that certain characteristics of individuals who adhere to treatment predispose them to benefit from the nonspecific components of treatment (eg, optimists may be more likely to adhere to treatment because they believe that treatment will be beneficial). It is also possible that adherence is a marker of a more general orientation to healthy behaviors such that patients who adhere to their medications may also be those who follow healthy lifestyle practices that reduce their risk of mortality.
Neither of the two previous trials identified characteristics of patients who were less likely to adhere to medication. The Coronary Drug Project Research Group did not report any efforts to identify characteristics of adherence (4). Characteristics of poor adherence were investigated in the Beta Blocker Heart Attack Trial and included disease severity factors, demographics, and psychosocial variables (eg, life stress, social isolation, self-reported Type A characteristics) (5). Smoking status was the only variable found to relate to adherence, with smokers being less likely to adhere to therapy.
The aims of our study were 1) to retest the association between adherence to placebo therapy and mortality rates in a contemporary acute MI population and 2) to identify medical, demographic, and psychosocial characteristics of poor and good adherence. This study builds on our initial study of psychosocial predictors of SCD in patients participating in CAMIAT (8). CAMIAT was a multicenter double-blind, placebo-controlled trial of amiodarone treatment for the prevention of SCD in patients with frequent or repetitive VPDs after an acute MI (9).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
1 mm and S 1 mm in V1). A subsample of the CAMIAT sample was recruited into our study of psychosocial predictors of SCD. The results of the psychosocial substudy have been published in abstract form (8). The only exclusion criteria used in the latter study were inability to read English or French well enough to complete the psychological questionnaires and death before the 2-week postrandomization clinic visit for CAMIAT. Written consent for the psychosocial substudy was obtained at the 2-week postrandomization clinic visit. At that time, patients were administered a brief rating scale to assess symptoms of dyspnea and fatigue (Yale Scale) (10), and they were given the psychological questionnaire battery to complete at home and return by mail within 2 weeks. Patients who had not returned questionnaires within this period were given two reminder calls. The full CAMIAT sample of 1202 patients was examined to test the association between adherence and mortality, whereas the investigation of markers of poor adherence was necessarily restricted to the psychosocial substudy sample of 671 patients. Not all CAMIAT centers agreed to implement the psychosocial study (31 of 36 CAMIAT hospital centers participated in the psychosocial assessment of patients), and not all patients approached to fill out the psychosocial questionnaire agreed or returned completed questionnaires (671 (79.2%) of 847 eligible patients).
Measures
Adherence.
As in other studies (4, 5), adherence was assessed by a pill count at each clinic visit, calculated as the proportion of pills not returned, multiplied by 100, and averaged over the 2 years of follow-up. During the first 2 weeks, patients in both the amiodarone and placebo groups generally took 3 to 5 tablets per day. At 2 weeks, this was reduced to 1.5 to 2 tablets a day, and by 8 months, most patients were taking only 1 tablet a day, although some continued with 2 tablets per day. There was no specific study protocol for informing patients about whether their pills were being counted. Patients were simply asked to return their pill bottle at each CAMIAT clinic visit and were issued a new pill bottle at each visit. Centers varied with respect to collecting and dispensing pill bottles. Some centers had the study nurse take care of collecting and dispensing pills, whereas others used the hospital pharmacy to perform this function. Because there was no study protocol for informing or not informing patients that their pills were being counted, there is no way to know what the study nurses told patients about the pill count. Regardless of how many pills were returned, patients were given a new bottle of pills for the next follow-up interval. Because the study nurses did not need to know how many pills the patient had returned to dispense the new bottle and because pill count was not specifically used as an adherence-enhancing strategy in CAMIAT, it is likely that the nurses did not count the returned pills in the presence of the patient.
Patients who had an average pill count below the 20th percentile of the pill count distribution were classified as poorly adherent. The 20th percentile cutoff point was used to define poorly adherent vs. adherent patients for two reasons. First, this cutoff point was approximately midway between the 33rd percentile used in the Coronary Drug Project Research Group study (4) and the fifth percentile used in the Beta Blocker Heart Attack Trial (5). Secondly, we wanted to select a cutoff point low enough to enable identification of characteristics associated with poor adherence. The 20th percentile cutoff point was <66% (ie, taking <66% of pills dispensed).
Three percent of patients had a pill count of >100%. If patients did not return their bottle of pills at a follow-up visit (eg, they reported having lost their bottle of pills), adherence was recorded as if they had taken all of the lost pills. Thus, the previous number, when added to the number of newly dispensed pills, could total more than 100%. Because there was no way of knowing whether these patients were truly adherent, these subjects were dropped from the analysis.
Primary Outcome Variable.
The primary outcome event was SCD during the 2 years of follow-up. Outcome events were reviewed by a blinded external validation committee. SCD was defined as arrhythmic death or resuscitated ventricular fibrillation. The criteria for arrhythmic death or resuscitated ventricular fibrillation are reported in the main CAMIAT article (9). Secondary outcomes were total cardiac deaths and all-cause mortality.
Predictors of Adherence.
Although the psychological questionnaire battery had been originally chosen to assess markers of increased risk for SCD, the measures include most of the variables that have been associated with poor adherence in previous studies. We used the Beck Depression Inventory, which measures depressive symptoms (11); the Symptom Checklist-90 (SCL-90), which measures general psychological distress (12); the Cook-Medley Hostility Scale (13); the Multidimensional Scale of Perceived Social Support, which measures perceptions of support from a patient’s significant other, family, and friends (14); and the Social Network Contacts and Social Participation subscales of the Health and Daily Living Form (15).
Patients’ reports of dyspnea and fatigue were assessed using the Yale Scale (10), which was administered at baseline. This scale has been shown to be sensitive to drug treatment for CHF (10) and was shown to correlate satisfactorily with the 12-minute walking test (r = 0.60) in patients with respiratory disease (16).
Data on the following cardiac prognostic variables were obtained from the CAMIAT files (9): study treatment status (ie, amiodarone or placebo); age, dichotomized at 70 years or <70 years; previous MI; history of CHF (defined by history and assignment of New York Heart Association functional class or a documented previous history of interstitial or airspace pulmonary edema); CHF at the time of hospitalization for the index acute MI (defined as pulmonary edema evidenced by interstitial or airspace disease on at least one in-hospital chest x-ray); VPDs on the baseline 24-hour Holter monitoring record; high VPD rate, defined as 20 or more VPDs per hour; history of diabetes mellitus; and randomization to CAMIAT within 28 days of the index MI. Cardiac history variables were extracted from patients’ medical records by CAMIAT study nurses.
Statistical Analysis
The primary analysis was conducted using Cox proportional hazards models for survival to SCD. The first set of analyses identified biological variables associated with SCD, with inclusion defined by a level of significance of p 
.20. Separate analyses were run for the amiodarone and placebo groups because they had different mortality outcomes and because the distribution of adherence differed significantly between these two groups. Patients were retained in the amiodarone group regardless of whether they stopped taking the medication (ie, intention-to-treat analysis). Variables were assessed by stepwise backward elimination of variables after including all biological variables in an initial model. After identifying the subset of associated biological variables, models including sex and age at Step 1 and adherence at Step 2 were tested. Sex and age were included in the hazards models because they were found to be associated with adherence.
Stepwise logistic regression analysis, with forward inclusion of variables, was used to investigate variables associated with poor vs. good adherence. The amiodarone and placebo groups were analyzed separately. 
2 analysis and independent t tests were used to examine univariate associations between predictor variables and adherence.
Lastly, descriptive analyses (2 test for categorical variables and Student’s independent t test for continuous variables, ie, psychosocial questionnaire variables) were conducted to explore associations between adherence to pills and reported adherence to other health behaviors (eg, smoking, diet, and exercise) and associations with psychosocial questionnaire variables. Both SPSS-PC (17) and SAS (18) were used for statistical analyses.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Of the 36 hospitals participating in CAMIAT, 31 agreed to implement the psychosocial substudy. Of the 969 CAMIAT patients available for the psychosocial study, 847 met the psychosocial study eligibility criteria, 144 (17%) refused to participate, and 703 gave written consent. Of the 703 who consented, 671 (95.4%) returned the psychological questionnaire battery. Missing data on some of the questionnaires reduced the sample size for multivariate analysis predicting adherence to 634 (90% of those who consented).
Cardiac and demographic variables were compared among three subgroups: patients who declined participation in the psychosocial substudy, patients who consented but failed to return completed questionnaires, and patients who participated in the psychosocial substudy. The analyses revealed that patients who declined participation were more likely to be 70 years old (2(1) = 12.06, p < .008) and to have been randomized to CAMIAT within 28 days of their qualifying acute MI (2(1) = 13.40, p < .003). Importantly, however, mortality rates were not found to differ among the three groups (p values > .20). Furthermore, adherence to medication did not differ among these groups (see Table 1).
|
Adherence
The placebo group had a significantly higher pill count than the amiodarone group (mean ± SD = 78.3 ± 17.4%, median = 80.3% vs. mean = 75.3 ± 18.0%, median = 77.7%; Mann-Whitney U test = 143,779.5, N = 1141, p < .001). The distribution of adherence was negatively skewed in both treatment groups. As described in the methodology, poor adherence was defined as taking <66% of pills prescribed (ie, lower than the 20th percentile). Fifteen percent of the placebo group and 22.0% of the amiodarone group took <66% of their pills. Had we defined a separate cutoff point for poor adherence within the treatment group, the cutoff points for the 20th percentile would not have been very different from <66%. The 20th percentile was <69% in the placebo group and <65% in the amiodarone group. Given that the adherence variables differed significantly between the two treatment groups, subsequent analyses relating adherence to mortality or examining correlates of adherence were conducted separately on the two treatment groups. Poor adherence was defined in two ways. First, in accord with our methodological plan, poor adherence was defined as taking <66% of pills prescribed; this cutoff point was the combined-group 20th percentile cutoff point. This definition ensured that a patient in either treatment group would be classified as poorly adherent by the same absolute pill count threshold. Second, to verify that the results would be the same if we used within-group 20th percentile cutoff points to define poor adherence, the analyses were repeated using a cutoff point of <69% for the placebo group and <65% for the amiodarone group. As shown below, the pattern of results does not differ regardless of whether the combined- or a within-group cutoff point is used to define poor adherence. Therefore, unless otherwise specified, the results are presented for analyses using the combined-group cutoff point of <66% of pills taken to define poor adherence.
Adherence and Mortality
In both the placebo (N = 568) and amiodarone groups, mortality rates were significantly higher in poorly adherent patients than in adherent patients. In the placebo group, poorly adherent patients had a higher 2-year SCD mortality rate than adherent patients (11.0% vs. 5.5%, 2(1) = 3.95, p < .04). Poorly adherent patients also had higher rates of total cardiac deaths (16.5% vs. 8.4%, 2(1) = 5.73, p < .01) and all-cause mortality (18.7% vs. 8.8%, 2(1) = 8.0, p < .01) than adherent patients. Likewise, poorly adherent patients in the amiodarone group (N = 573) had higher rates of mortality due to SCD (7.0% vs. 2.9%, 2(1) = 4.55, p < .04), higher total cardiac deaths (11.7% vs. 5.8%, 2(1) = 5.17, p < .02), and higher all-cause mortality (14.8% vs. 7.4%, X2(1) = 6.65, p < .01) than adherent patients.
Although the magnitude difference in SCD rate between the poorly adherent and adherent groups was similar between the full placebo sample and the subsample of 309 placebo patients who participated in the psychosocial assessment study (11.1% SCD in the poorly adherent group and 5.7% in the adherent group), the result of the test of the poor adherence effect in the psychosocial subsample was not significant (2(1) = 2.33, p = .13). Similarly, the difference in the SCD rate between the poorly adherent (5.8% SCDs) and adherent groups (2.2% SCDs) in the subsample of amiodarone-treated patients who participated in the psychosocial assessment study was not statistically significant (2(1) = 2.47, p = .12). Because the effect of poor adherence was not statistically significant in the subsample of patients who participated in the psychosocial assessment study, the multivariate analysis predicting SCD was confined to variables available on the full CAMIAT sample (ie, the cardiac prognostic variables).
The multivariate Cox proportional hazards model identified two cardiac variables as significant predictors of SCD in the placebo group: previous MI (RR = 3.11, 95% CI = 1.54–6.30, p < .002) and history of CHF (RR = 3.15, 95% CI = 1.57–6.34, p < .002). Although gender and age were not significant predictors of SCD, they were retained in the hazards model because of their association with adherence. When adherence was added to the model, poor adherence was found to be independently associated with SCD (RR = 2.11; 95% CI = 1.03–4.56, p < .05) (see Table 2). As shown in Figure 1, the survival curves were clearly different between the adherent and poorly adherent groups.
|
|
|
|
Characteristics of Poor Adherence
Table 4 displays the relationship between patient characteristics and adherence in the placebo and amiodarone treatment groups. Table 5 displays the relationship between the psychosocial variables and adherence in both treatment groups. Poor adherence in the placebo group was significantly associated with being more than 70 years old (2(1) = 4.93, p < .05), being less likely to participate in a heart-healthy diet program (2(1) = 4.73, p < .03), and participating in fewer social activities in the month before the index MI (t(307) = 2.69, p < .01). There was also a trend for women to be more poorly adherent than men (2(1) = 3.31, p < .07). Adherence to pills was not significantly related to participation in an exercise program or smoking status at the time of the index MI. In the analyses of the amiodarone group, none of the cardiac, demographic, or psychosocial variables were found to be associated with poor adherence. As the numbers reflect in Table 4, these analyses were done on the subsample of subjects for whom psychosocial data were available. When the analyses were repeated on the full CAMIAT sample, focusing only on the cardiac prognostic variables and demographic variables (ie, age and sex) that were available for the full sample, the pattern of the results was identical to that produced by the analyses on the psychosocial subsample.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
For example, none of the cardiac variables were associated with poor adherence. Thus, within the context of an acute illness, adherence to placebo seems to be unrelated to severity of disease or illness. Our results do provide some limited support for the possibility that adherence to pills may be a marker of a more general health-oriented behavior pattern. Specifically, adherence to placebo was associated with a greater likelihood of reporting consumption of a heart-healthy diet. However, the lack of a similar association with exercise behavior or smoking raises doubts about a general health-oriented behavior pattern as a viable explanation for the health benefits of adherence. Inconsistencies in the degree to which adherence among different health behaviors intercorrelate is congruent with what has been reported by others. For example, in the Medical Outcomes Study of 1828 patients recruited from five medical specialties in three large US cities, medication adherence was only moderately correlated with exercise adherence (r = 0.26, p < .01) and diet adherence (r = 0.39, p < .01) (19). In the Beta Blocker Heart Attack Trial (5), smoking was related to medication adherence; however, many studies assessing adherence across a range of health behaviors have not found significant correlations (20), even when studied within a single disease state (21). Thus, the results from our study, taken together with those reported by others, provide only limited support for the suggestion that adherence to placebo is a marker of a general health-oriented behavior pattern.
Only one psychosocial variable was found to be associated with adherence. Compared with patients who adhered poorly to placebo, adherent patients participated in approximately twice the number of pleasurable social activities in the month before their index MI. Greater social support has been previously associated with better adherence (22–25). However, because neither perceived social support nor social network contacts were related to adherence in this study, greater social support per se may not be the sole explanation for the positive association between social participation and adherence. We speculate that people who are more actively engaged in pleasurable life activities may have a stronger desire for good health and thus may be more motivated to adhere to health-promoting treatment. Adhering better, in turn, may engender a greater treatment expectancy effect, thereby strengthening the placebo effect (7).
It is also possible that related factors that were not specifically measured in this study, such as optimism (26), may predispose a patient to both greater social participation and better adherence. For example, a generalized optimistic outcome expectancy may motivate one to adhere to treatments because one expects these treatments to work. As suggested by others (7), a higher treatment outcome expectancy may be related to the placebo effect of treatment, such that a higher outcome expectancy strengthens the placebo effect and ultimately leads to better health outcomes. Additional research is clearly needed to better understand the psychosocial and cognitive determinants of adherence and how these determinants lead to better health outcomes.
In contrast to the lack of association between depressive symptoms and adherence in our study, Carney et al. (27) observed that depression predicted poor adherence to cardiac medications. At least two differences between their study and our study may explain the inconsistency in findings. First, in the Carney et al. study (27), patients were diagnosed as having a major depressive disorder, whereas in our study depression was measured by self-reported symptoms. Patients can have elevated depressive scores on a questionnaire without meeting the criteria of diagnosis for a major depressive disorder (28). Thus, it may be that depressive symptomatology per se is not associated with poor medication adherence. Second, in the Carney et al. study (27), medication adherence was measured over a brief time interval of 3 weeks using an electronic medication monitor. Thus, their measure of adherence may have been more precise than our measure.
The reason for the lack of concordance in the predictors of adherence to placebo therapy vs. amiodarone therapy in unclear. However, amiodarone-treated patients took significantly fewer pills than placebo-treated patients, suggesting that taking amiodarone was itself associated with poorer adherence. As reported in the main CAMIAT article (9), side effects were more common in patients taking amiodarone than in patients taking the placebo. Thus, amiodarone-related side effects may have contributed to poorer adherence in the amiodarone group, thereby obscuring associations between adherence and psychosocial variables in this group. We did not have the side effects data to test this issue directly.
Finally, in the analyses of predictors of adherence in the placebo group, older age was associated with poor adherence, and there was a trend for women to be less likely to adhere than men. Previous studies show an inconsistent association between older age or female gender and adherence (29). The relationships observed between age, sex, and adherence in our study may be specific to this sample.
Limitations
First, adherence was measured by pill count. This method tends to overestimate adherence because patients may not ingest all the pills they remove from the bottle (29, 30). Biases inherent in the pill count measure may have obscured associations between adherence and the cardiac, demographic, and psychosocial characteristics. Second, the cutoff point for defining poor adherence was arbitrary. It did not seem reasonable to use the cutoff points used in the other studies relating pill count adherence to mortality because half of our sample would have been classified as poorly adherent by these cutoff points. Third, the psychosocial variables were available only for a subgroup of the CAMIAT sample. Therefore, we could not directly test for a mediating role of psychosocial factors (ie, the social participation variable) between adherence and mortality. We previously reported that the social participation variable was associated with a lower risk of SCD in the combined placebo- and amiodarone-treated samples of CAMIAT (8). This observation suggests the possibility of a mediating role; however, without a direct test, it is impossible to know whether the association between poor adherence and risk of SCD is mediated by social participation.
Conclusion
Poor adherence to placebo therapy is associated with a twofold greater risk of mortality, although the reason remains unknown. Our data provide only partial support for the hypothesis that adherence to other health-promoting behaviors (eg, healthy dietary practices) accounts for the association between adherence to placebo therapy and lower risk of mortality in post-MI patients. Moreover, like other researchers, we did not find evidence to suggest that disease severity factors explained the association between poor adherence to placebo and increased mortality risk. The relationship between adherence and social activities might indicate a higher motivation to adhere to treatment in individuals who are more engaged in enjoyable activities.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This study was funded by a research grant from the Heart and Stroke Foundation of Ontario and by a scholarship award from the Heart and Stroke Foundation of Canada (to J.I.).
Received for publication July 24, 1998.
Revision received May 3, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  P. M. Ho, D. J. Magid, S. M. Shetterly, K. L. Olson, P. N. Peterson, F. A. Masoudi, and J. S. Rumsfeld Importance of Therapy Intensification and Medication Nonadherence for Blood Pressure Control in Patients With Coronary Disease Arch Intern Med, February 11, 2008; 168(3): 271 - 276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. K. Gehi, S. Ali, B. Na, and M. A. Whooley Self-reported Medication Adherence and Cardiovascular Events in Patients With Stable Coronary Heart Disease: The Heart and Soul Study Arch Intern Med, September 10, 2007; 167(16): 1798 - 1803. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Mitchell and T. Selmes Why don't patients take their medicine? Reasons and solutions in psychiatry Advan. Psychiatr. Treat., September 1, 2007; 13(5): 336 - 346. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Olshansky Placebo and Nocebo in Cardiovascular Health: Implications for Healthcare, Research, and the Doctor-Patient Relationship J. Am. Coll. Cardiol., January 30, 2007; 49(4): 415 - 421. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. C. Ziegelstein and S. D. Miller Just a Spoonful of Sugar J. Am. Coll. Cardiol., December 5, 2006; 48(11): 2223 - 2224. [Full Text] [PDF]
|
|
|
|
|
|
P. M. Ho, J. S. Rumsfeld, F. A. Masoudi, D. L. McClure, M. E. Plomondon, J. F. Steiner, and D. J. Magid Effect of medication nonadherence on hospitalization and mortality among patients with diabetes mellitus. Arch Intern Med, September 25, 2006; 166(17): 1836 - 1841. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. M. Ho, J. A. Spertus, F. A. Masoudi, K. J. Reid, E. D. Peterson, D. J. Magid, H. M. Krumholz, and J. S. Rumsfeld Impact of medication therapy discontinuation on mortality after myocardial infarction. Arch Intern Med, September 25, 2006; 166(17): 1842 - 1847. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Spertus, R. Kettelkamp, C. Vance, C. Decker, P. G. Jones, J. S. Rumsfeld, J. C. Messenger, S. Khanal, E. D. Peterson, R. G. Bach, et al. Prevalence, Predictors, and Outcomes of Premature Discontinuation of Thienopyridine Therapy After Drug-Eluting Stent Placement: Results From the PREMIER Registry Circulation, June 20, 2006; 113(24): 2803 - 2809. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Gehi, D. Haas, S. Pipkin, and M. A. Whooley Depression and Medication Adherence in Outpatients With Coronary Heart Disease: Findings From the Heart and Soul Study Arch Intern Med, November 28, 2005; 165(21): 2508 - 2513. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C M Dickens, L McGowan, C Percival, J Douglas, B Tomenson, L Cotter, A Heagerty, and F H Creed Lack of a close confidant, but not depression, predicts further cardiac events after myocardial infarction Heart, May 1, 2004; 90(5): 518 - 522. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A Vik, C. J Maxwell, and D. B Hogan Measurement, Correlates, and Health Outcomes of Medication Adherence Among Seniors Ann. Pharmacother., February 1, 2004; 38(2): 303 - 312. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Inouye, S. T. Bogardus Jr, C. S. Williams, L. Leo-Summers, and J. V. Agostini The Role of Adherence on the Effectiveness of Nonpharmacologic Interventions: Evidence From the Delirium Prevention Trial Arch Intern Med, April 28, 2003; 163(8): 958 - 964. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. P. McDonald, A. X. Garg, and R. B. Haynes Interventions to Enhance Patient Adherence to Medication Prescriptions: Scientific Review JAMA, December 11, 2002; 288(22): 2868 - 2879. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Scott and M. Pope Self-Reported Adherence to Treatment With Mood Stabilizers, Plasma Levels, and Psychiatric Hospitalization Am J Psychiatry, November 1, 2002; 159(11): 1927 - 1929. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
