Presented at the annual meeting of the Canadian Association of Emergency Physicians (CAEP), Victoria, BC, June 2–4, 2007.
The Consent and Prescription Compliance (COPRECO) Study: Does Obtaining Consent in the Emergency Department Affect Study Results in a Telephone Follow-up Study of Medication Compliance?
Article first published online: 22 SEP 2008
© 2008 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 15, Issue 10, pages 932–938, October 2008
How to Cite
Campbell, S. G., McCarvill, E. M., Magee, K. D., Cajee, I. and Crawford, M. (2008), The Consent and Prescription Compliance (COPRECO) Study: Does Obtaining Consent in the Emergency Department Affect Study Results in a Telephone Follow-up Study of Medication Compliance?. Academic Emergency Medicine, 15: 932–938. doi: 10.1111/j.1553-2712.2008.00234.x
- Issue published online: 1 OCT 2008
- Article first published online: 22 SEP 2008
- Received November 20, 2007; revisions received March 31, June 23, and July 11, 2008; accepted July 13, 2008.
- informed consent;
- medication compliance;
- telephone study;
Objectives: The objectives were to determine whether mandated research requirements for consent in the emergency department (ED) falsely distorts the results of a survey of patient-reported compliance with ED prescriptions and, in addition, to ascertain the level of patient compliance to medication instructions and find out the degree of displeasure expressed by patients called without prior consent.
Methods: Patients given new prescriptions for a medicine to be taken regularly over a period of less than 30 days were eligible. A convenience sample of eligible patients was randomized to having consent obtained during their ED visit or at the time of telephone follow-up. Patients were called 7–10 days after their ED visit to determine their compliance with the prescription. Compliance rates between the two groups were compared, as was the prevalence of displeasure expressed by patients called without prior consent.
Results: Of 430 enrolled patients, 221 were randomized to receive ED consent for telephone follow-up, and 209 received telephone follow-up without prior ED consent. Telephone follow-up was successful in 318 patients (74%). The rate of noncompliance was slightly higher in the group without ED consent, 74/149 (50%; 95% confidence interval [CI] = 41% to 58%) than the group who gave ED consent for telephone follow-up, 67/169 (40%; 95% CI = 32% to 42%; p = 0.07). Among the two groups, 141/318 (44%) did not fill the prescription (n = 42) or took it incorrectly (n = 99). Only 1 (0.7%) of the 149 patients with successful telephone follow-up without prior ED consent expressed displeasure at this telephone call.
Conclusions: Medicine noncompliance is a significant issue for patients discharged from the ED in this study. Although there was a trend toward greater compliance in patients who consented to the follow-up call, this did not reach statistical significance. ED patients do not object to receiving telephone follow-up for a research survey without giving prior consent.
The treatment plan for patients discharged from the emergency department (ED) frequently involves the prescription of medications for the patient to take at regular intervals for a specific period of time. The treating physician generally believes that compliance with the prescription will result in a better patient outcome than noncompliance, and presumes that the patient will take the medication as directed.
Medication noncompliance, where a prescription is not filled, taken correctly, or taken at all, is a complex problem that continues to burden our health care system. One study attributed almost 10% of all ED visits to medication noncompliance.1 Rates of nonfilling of prescriptions dispensed in adult EDs of 12%–22% have been reported.2,3 Studies assessing compliance, however, are limited in several ways. Mail-in surveys that provide the patient the most anonymity have notoriously low rates of return.4 Institutional review boards (IRBs) frequently insist that only the health care provider who treated the patient may contact the patient for follow-up without obtaining prior consent.5 Patients being called by a treating physician may perceive that they are being judged as to their obedience to the instructions of a respected medical professional, so may be reticent to admit to noncompliance. Obtaining prior consent for a follow-up call might affect compliance, since patients who are aware that they will be “checked up on” are more likely to be “obedient.” If treating physicians themselves obtain consent, the Hawthorne effect might come into play, as physicians who are aware that their instructions will be followed-up on may change the way they give instructions.
These issues can result in studies of compliance being performed as internal continuous quality improvement projects, without IRB approval, and therefore remaining unpublished or not being performed at all in realization that the results might be unreliable. Nonetheless, numerous other ED studies involve follow-up calls to patients, and the potential influence of the consent process on findings is frequently overlooked. We argue that it is less ethical to expose patients to a study that might have unreliable results than it is to obtain consent only at the time of follow-up call.
The primary objective was to determine whether the act of obtaining consent in the ED for telephone follow-up affects patient-reported compliance with prescriptions given to them at that visit. Considering reported prescription nonfilling rates of 12%–22%,2,3 we estimated that approximately 30% of patients would not comply fully (either not filling or not taking as directed) with their prescriptions. We hypothesized that the consent process would result in an absolute 20% apparent improvement in overall compliance rates when compared to a nonconsenting group, indicating that the informed consent requirement may be falsely distorting survey results. Our secondary objectives were to ascertain the level of compliance with ED prescriptions, to assess symptom outcomes using basic self-reporting variables (better, same, worse), and to determine if patients object to being called for research purposes without prior consent having been obtained.
This was a randomized, prospective study designed to illustrate the effects of obtaining informed consent prior to telephone follow-up on medication noncompliance. IRB approval was granted for the study.
Study Setting and Population
This study was conducted at the ED at the New Infirmary site of the QEII Health Sciences Center in Halifax, Nova Scotia, Canada, a 1,028-bed urban academic hospital that sees about 65,000 adult (age > 16 years) ED visits per year.
From a convenience sample of patients, subjects were enrolled if they were prescribed a new prescription (not an alteration of a pre-existing prescription), from an emergency physician that required a regular dose (not one with instruction to take ‘as required’) for a time period of less than one month. Patients with prescriptions were considered ineligible if they were not responsible for taking their own medications or filling their prescriptions, if they had cognitive impairment that might preclude informed consent, if they did not have access to a telephone, or if they were unable to communicate in English.
All patients deemed eligible to participate by the treating physician were randomly assigned to one of two groups. The first group had informed consent obtained by the treating physician in the ED for a follow-up call to be performed in the next 7–10 days to ask about the course and treatment process of their illness. These patients were informed that the purpose of the study was to assess compliance rates. It was also explained that other patients in the study would not be asked for consent and they were asked to not refer to the ED consent process at the time of follow-up call, to maintain the blinding of the surveyor. Patients in the second group were not approached for consent in the ED, but their name and telephone numbers were recorded along with details of the prescription given. The number of consents and refusals in the ED was recorded. Patient randomization was performed by “Consent in ED” and “No Consent in ED” adhesive labels being assigned to each of the numbers one through 450 by an electronic random number generator6 and then sealed in opaque envelopes.
Data Collection and Processing
The initial follow-up call was made 7 days after discharge from the ED (with the day of visit taken as Day 1) by a single medical student investigator (EM) who was blinded to group allocation and had no previous contact with the study participants. If the subject was not successfully contacted on this day, subsequent attempts were made over the following 2 days. The surveyor solicited consent for follow-up at the time of the call for all patients. In an effort to maximize the reliability of self-reported behavior of patients, they were assured at follow-up that any response would be kept in complete confidence and that written data would not be linked to their name. Specifically, they were informed that neither the physician who prescribed the medication nor their family physician would be informed as to whether the patient complied with the prescription or not. Data were entered onto an electronic database developed by the database designer of the Dalhousie University ED. Patients were excluded from further analysis if they refused consent either in the ED or at the time of the follow-up call. They were considered lost to follow-up after at least three unsuccessful contact attempts on three different days, with at least 3 hours time difference between each call.
Patients were surveyed according to a standard questionnaire (Data Supplement 1, available as supporting information in the online version of this paper). After surveying each patient, the surveyor informed patients that part of the objective of the study was to see if the consent process affected the study outcome and were asked if they were still prepared to have their answers included after being informed about the initial deception.
The surveyor was blinded to the group allocation of all subjects called. Physicians were blinded as to the allocation of the patients until after the prescription had been written, and telephone-consented patients were unaware that they were study subjects until they were called.
Regarding the filling of the prescription, patients were assessed as “compliant” if they obtained the medication within 2 days of the ED visit, a definition adapted from previous studies.3,7 In addition, we considered missing two or more doses, or taking a different dose to that prescribed, as being “noncompliant.”“Complete compliance” was defined as patients having both filled the prescription within 2 days and taken the full course of medication as prescribed.
All patients were enrolled over a 9-month period—the length of time necessary for us to achieve the numbers required based on our initial power calculations. Assuming a noncompliance rate of 30%, we calculated that a sample size of 213 subjects in each group would enable us to detect an absolute reduction in the rate of apparent noncompliance of 12%, or a relative reduction of 40%, given a power of 0.80 and a significance level of 0.05.
Compliance was compared between groups (consented in the ED versus consented only by telephone). Further breakdown between “medication obtained and completed as prescribed” and “medication obtained but not completed as prescribed” was performed. The proportions complying with the prescription were compared by chi-square analysis. Demographic and clinical characteristics of the participants in each group were compared. Basic self-reported outcome variables (better, worse, same) were compared between those who complied with their prescribed medication regime and those who did not comply.
Analysis of the primary outcome on all randomized consented subjects was done on an intention-to-treat basis. For all outcome events, 95% confidence intervals (CI) were calculated. Intervention comparisons were made using chi-square analysis.
Four-hundred thirty subjects were enrolled between July 2005 and April 2006. The breakdown of allocation and subject destiny is illustrated in Figure 1. Thirty-five patients (8.0% of all approached) refused to provide follow-up. The telephone refusal rate was slightly higher in the group that did not give consent in the ED (8.6% vs. 5.0%; p = 0.1). Six patients who were approached in the ED refused at that time and were excluded from subsequent analysis. However, when added to the group to which they had been allocated but were not evaluable because of their refusal for follow-up, the total refusal rate for those who were approached for consent in the ED (17/227) was 7.5%, similar to that found in the group that was asked for consent only at the time of telephone follow-up (p = 0.7). A comparison of variables in each group is presented in Table 1.
|Characteristic||Initial Consent in ED||Initial Consent at Telephone Follow-up|
|Gender: female||120 (54.3%)||111 (53.1%)|
|Gender: male||101 (45.7)||98 (46.9)|
|Age (±SD), years||44 (±19)||43 (±17)|
|Class of medication prescribed|
|Antimicrobials||104 (61.5)||95 (63.8)|
|Nonsteroidals||16 (9.5)||9 (6.0)|
|Proton pump inhibitors/H2 antagonists||9 (5.3)||12 (8.0)|
|Antispasmodic/antinauseants||5 (3.0)||2 (1.3)|
|Antiepileptics||2 (1.2)||3 (2.0)|
|Other||33 (19.5)||28 (18.8)|
|Duration of regime (days)||9.5||8.5|
|Frequency of administration|
|Once daily||21 (12.4)||26 (17.4)|
|Twice a day||66 (39.1)||64 (42.9)|
|Three times a day||35 (20.7)||16 (11.0)|
|Four times a day||47 (27.8)||43 (28.8)|
Total noncompliance (prescription either not filled or not taken as directed) was found to be 39.6% in the group that consented in the ED versus 49.7% in the “nonconsented” group (p = 0.07). The relative risk (RR) of noncompliance in patients not consented in the ED was 1.2 (95% CI = 1.0 to 1.5). Among all consenting subjects, the nonfilling rate was found to be 13.2% and total noncompliance rate was 44.3%. Reasons cited for noncompliance are illustrated in Figure 2.
With reference to the symptomatic outcomes reported by patients at the time of telephone follow-up, compliant patients were not statistically more likely to report that they had improved by the time of follow-up (see Table 2).
|Back to normal||101 (57.1)||71 (50.3)||172|
|Somewhat improved||65 (36.7)||54 (38.3)||119|
|Same||10 (5.6)||15 (10.6)||25|
|Worse||1 (0.6)||1 (0.7)||2|
Of the 318 patients who completed the survey, there was only one individual, from the nonconsented group (0.3% of the total, or 0.7% of the nonconsented patients), who indicated that he or she was not pleased about receiving a telephone call for research purposes.
To the best of our knowledge, we present the first study investigating the effect of obtaining consent in the ED on subsequent findings of a survey of compliance with ED prescriptions. These data also represent the first Canadian study of adult compliance rates with ED prescriptions.
Studies conducted in the primary care or clinic setting report rates of noncompliance with filling prescriptions of 5% to 20%,7–9 while such rates in adult EDs are reported at 12% to 22%.2,3 Medication noncompliance rates in pediatric EDs have been found to range from 11% to 93%.10
Levels of compliance vary with the definition of compliance used by the specific authors.2,3,7,11 For the purposes of this study, we adapted the definition of compliance with filling the prescription from that used by Saunders3 and Matsui et al.7 Fewer authors define compliance with following the instructions of a filled prescription.11,12 The definition of noncompliance with regard to taking medication was decided by consensus of the investigators as “having missed two or more doses” or having exceeded the frequency of daily dosages.
We did not find a significant difference in self-reported outcomes between those who were fully compliant with their prescriptions and those who were not, a finding likely the result of our strict definition of noncompliance. Despite the absence of any differences in this study, it has been shown in the past that medication noncompliance is a significant issue and contributes to a large number of ED visits each year.1 The definition used in this study was chosen in an attempt to pick up differences in the medication-taking practices between those patients who knew they were going to receive follow-up and those who were oblivious to this fact. While we can appreciate that missing two or three doses of an antibiotic may not have had an effect on treatment success or failure, we felt that this strict definition was necessary to adequately measure our primary outcome.
Reasons cited for noncompliance include limited access or transportation to pharmacies, inability or unwillingness to pay for expensive medication,12 the poor palatability of the medication, and confusion from multiple medications and complex regimens. Several studies have demonstrated an inverse relationship between adverse effects and compliance.10 Failing eyesight and/or hearing can prevent some from understanding instructions; childproof containers offer an additional obstacle for those with arthritic or aging hands.11 Communication between physicians and patients is a critical factor in ensuring compliance,11,13 and patient perceptions of the issues involved, such as fear of dependence on medication or fear of long-term effects,14 need to be explored by prescribing physicians. Patients may simply stop taking medications because they “feel better” and “no longer need them.”
There are several limitations to research in the area of medication noncompliance. Researchers are generally forced to rely on telephone, e-mail, or mail-in surveys to gather data. Several studies have examined how best to contact patients for follow-up after discharge from the ED. While contact rates of 54.9 and 28.6% by telephone have been reported,15,16 Isaacman et al.17 showed that verifying the best contact number for a follow-up call before discharging the patient could improve contact rates to 93%; Ferrigno et al.18 reported a contact rate of 77% using a similar strategy. Electronic mail has been shown to result in a disappointing 41% response rate from ED patients.19
Telephone assessment of compliance is limited by the fact that telephone-solicited self-reported compliance rates are likely an overestimate of the actual rate of compliance. In a study by Ginde et al.,20 patients were given a prescription for azithromycin to be filled at no charge to the patients at the hospital pharmacy. A subsequent telephone survey found that 96.8% of subjects claimed to have filled their prescriptions, although this was confirmed by the designated pharmacy in only 74.2%. While they concede that it is possible that some patients filled their prescriptions elsewhere, they argue that this seems unlikely, considering the proximity of the designated pharmacy and the fact that they would have to pay for it themselves elsewhere. In an attempt to compare self-reported vs. measured medication compliance, Bachmann et al.21 used data from Medication Event Monitoring System (MEMS) caps and found that only 16% of subjects took their medication correctly, despite a self-reported compliance of 90%.
Institutional review board stipulations that researchers may not contact a patient for follow-up without obtaining prior written consent5 further limit studies of compliance. One could argue that by obtaining consent for a follow-up call, the Hawthorne effect will come into play. This phenomenon has as its basic premise, that study subjects may be influenced by the fact that their behavior is being observed, therefore altering the validity and reliability of study results.22 The Hawthorne effect has been successfully used to improve oral hygiene compliance,23 and other studies have also shown the Hawthorne effect’s ability to alter subject behavior, resulting in an improvement in the outcome variable.24 The Hawthorne effect may therefore be considered a significant confounder when assessing compliance, so ED patients who have been informed that their compliance to a prescription will be evaluated would, according to this premise, be expected to be more compliant than if no such warning had been given.
The effects of obtaining informed consent have been studied in other contexts as well. A group at the University of Pennsylvania found that the written informed consent process was associated with a higher refusal rate in screening for violence in the home. Despite this high refusal rate, however, they did not find that the use of written consent effected the detection of such violence.25
While the U.S. Code of Federal Regulations states that some surveys may not require informed consent, this does not cover research where “information obtained is recorded in such a manner that human subjects can be identified, directly or through identifiers linked to the subjects.”26 In studies such as ours, the telephone numbers and names of the patients being surveyed are likely to be considered “identifiers” in this context. One could argue that the requirements mandating health professionals to obtain written, informed consent from study subjects prior to them being telephoned for follow-up may actually be impeding proper compliance research. The informed consent process, along with the determination that the potential benefits outweigh the potential risks, constitutes the criterion standards of ensuring patient safety. The deception in our study posed “minimal risk” to the patient. Considering that obtaining written informed consent from those not consented in the ED would be impossible without jeopardizing its purpose, our IRB approved the study. It could still be argued that this form of deception, and also the practice of “cold calling,” may not be ethical. Because all subjects were made aware of any deception and the reasons for this at the conclusion of the study, and were given the opportunity to remove their information from the study after learning this, we feel that there was no ethical lapse. All but 1 of the 318 subjects (0.3%) who were contacted by telephone said that they did not mind being called at home for medical research purposes. Of patients not consented in the ED, this percentage is 0.5%. We submit that if 99.5% of patients do not mind being “cold-called” for research, then one could reasonably doubt the necessity for the requirement for ED consent in the first place.
While there was a trend toward statistical significance, we did not show a statistically significant difference between the two groups in terms of overall noncompliance. Larger studies may still prove the hypothesis that the Hawthorne effect represents a confounding variable in research of this type.
The fact that the refusal rate was slightly higher in the group that did not consent initially in the ED suggests that it may be easier to refuse consent over the telephone than it is in person, especially when one thinks of the vulnerable state of some patients while in the ED. With this in mind, one could consider that approaching patients in the security of their homes could be considered a more ethical option than approaching them in the ED.
Discharge and prescription instructions given to patients were not standardized, and we did not compare variables such as specific physician or patient characteristics. We asked physicians to enroll suitable patients at their convenience, which may have contributed to a selection bias. We did not consider the effect of different socioeconomic factors on compliance, nor did we compare the effects of the indication for the prescription, the different types of medication, the payer for the medications, or length of course or frequency at which the medication was to be taken.
Our definition for noncompliance with taking the (filled) regimen as prescribed was somewhat arbitrary, at “missing two or more doses.” We chose this definition because we believed that most patients could be excused for missing at least one dose of a prescription. Allowing fewer, or more “misses” may have changed our findings somewhat. As mentioned in the Discussion, a clinical impact of minor “noncompliance” was not demonstrated.
Our results are based on patient-reported compliance rates. As discussed previously, it has been shown that this is clearly not an ideal way to accurately ascertain a true picture of noncompliance, as it frequently leads to an underestimation of true rates. More precise methods of measuring compliance can be both expensive and cumbersome, however, and because of this, despite its limitations, obtaining data through self-reporting is a methodology commonly employed in research of this type. In this study we were primarily interested in whether there was a describable difference in compliance rates between the two groups—the actual rate of noncompliance was a secondary outcome.
There was no significant difference in compliance between the groups we studied; however, we argue that there is a definite trend toward better compliance in patients who had given consent for follow-up in the ED. Our reported nonfilling rate is within the range reported previously in U.S. studies. There were no significant differences noted in patient-reported outcomes between the two groups. In spite of the opinion that self-reported compliance may be an overestimation of true levels, we still report a high overall rate of noncompliance among all subjects. The vast majority of ED patients called did not object to being called at home without prior consent.
- 5Capital District Health Authority. Initial Submissions Form: Ethics Approval Submission Form. Available at: http://www.cdha.nshealth.ca/default.aspx?page=87&category.Categories.1=310&docCategory.Categories.0=325¢erContent.Id.0=14960. Accessed Oct 1, 2007.
- 6Research Randomizer. Available at: http://www.randomizer.org/form.htm. Accessed Jan 31, 2005., .
- 11Medication compliance problems in general practice: detection and intervention by pharmacists and doctors. Aust J Rural Health. 2002; 10:33–8., .Direct Link:
- 23Intentional use of the Hawthorne effect to improve oral hygiene compliance in orthodontic patients. J Dent Educ. 2002; 66:1129–35., , , , .
- 26U.S. Department of Health & Human Services. Code of Federal Regulations. Available at: http://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.htm#46.101. Accessed May 12, 2008.
Data Supplement S1. COPRECO study.
Please note: Wiley Periodicals Inc. are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
|ACEM_234_sm_DataSupplementS1.pdf||73K||Supporting info item|
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.