Abstract
Background: Insight into patient adherence is needed to enable an effect evaluation of medication for dyspepsia.
Objectives: Adherence was explored by investigating two adherence outcome measures (completeness and intake fidelity) using data from the DIAMOND-study.
Methods: The DIAMOND-study is a pragmatic RCT comparing a ‘step-up’ with a ‘step-down’ treatment strategy. In step 1 participants (n =653) were instructed to use five pills/day for maximally 30 days: an antacid 4dd plus a placebo 1dd (‘step-up’) or a proton pump inhibitor 1dd plus a placebo 4dd (‘step-down’). If the complaints persisted, step 2 was started (H2-receptor antagonist 2dd), and subsequently step 3 (five pills/day, placebo and verum vice versa from step 1). Completeness was assessed by pill counts, intake fidelity by patient questionnaires measuring the degree to which patients adhered to specific instructions concerning timing, frequency, dose and way of intake.
Results: In step 1, patients used on average 3.9 pills/day (78% of the prescribed doses), in step 2, 1.7 pills/day (85%) and in step 3, 3.6 pills/day (72%). For the four times daily pills, half of the patients used less than 80% of the prescribed pills per day. This was one third of the patients for the twice daily pills and one quarter for the once daily pills. There were no completeness differences between active or placebo medication and no differences between the study arms. As regards intake fidelity, 70% of the patients made one or more errors in the medication intake.
Conclusion: There is room for improvement in adherence rates for dyspepsia medication.
Key words::
KEY MESSAGE(S):
Valid treatment evaluation depends on patient adherence.
One in two patients was non-adherent for their four times daily pills for dyspepsia, one third for their twice daily pills and one quarter for their once daily pills.
70% of the patients made one or more errors in the medication intake.
Introduction
To evaluate the effectiveness of a drug, it is essential to know the degree to which the drug is actually used according to the instructions. In determining adherence two aspects are substantial: intake completeness, i.e. the percentage of prescribed doses taken on average per day, and intake fidelity, i.e. concordance with instructions concerning dosing, timing in relation to meals, and way of medication intake (Citation1).
The present study focuses on adherence to acid-suppressive medication for newly diagnosed dyspepsia. Adherence is crucial, since acid-suppressants is often used as a diagnostic aid (Citation2): the relief of symptoms makes it more likely that the complaints are acid-related, but if symptoms are not relieved, the diagnosis needs to be reconsidered; this often results in prolonged treatment or referral for further diagnostic evaluation (Citation2).
Treatment failure may, however, be caused by non-adherence. When patient non-adherence is identified as a possible cause of treatment failure, physicians may be prevented from wrongly assuming that the prescribed medication was not effective. However, when it is determined that a patient was adherent and the medication was not effective, one can more confidently switch treatment instead of prolonging it.
Evidence concerning the effectiveness of how, and how often, medication was taken revealed the following. As to completeness, continuous as against on-demand use of acid-suppressants is associated with fewer recurrences and better outcomes for newly diagnosed dyspepsia (Citation3). Furthermore, the more days the medication is used, the higher the chance of achieving symptom resolution and the longer the symptom free period (Citation4). In regard to intake fidelity, proton pump inhibitors (PPIs) used with a meal are more effective than PPIs used without a meal, even when taken at a fixed time during the day (Citation5), twice daily dosing with H2-receptor antagonists (H2RAs) is more effective than once daily (Citation6), and chewing on antacid tablets is more effective than just swallowing the tablets (Citation7).
We elaborately investigated completeness and intake fidelity in a pragmatic randomized trial: the Dutch study of initial management of newly diagnosed dyspepsia (DIAMOND) (Citation8). In this RCT in primary care, a ‘step-up’ treatment strategy was compared with a ‘step-down’ treatment strategy. The design of this study is described in detail in Fransen et al., (Citation8) The results of the cost-effectiveness study comparing the two treatment strategies is described in elsewhere (Citation9). This paper focuses on providing insight into patient adherence to acid-suppressive medication that can be used as an aid to judge adherence in everyday practice.
Methods
The intervention
Treatment comprised maximally three antacid treatment steps of each maximally four weeks: if complaints persisted or recurred after step 1, step 2 was initiated, and if necessary, subsequently step 3.
In step 1, patients were instructed to use five pills a day, spread out over the day: four placebo pills and one PPI pill per day for patients randomized to the ‘step-down’ treatment strategy, four antacid pills and one placebo pill per day for patients in the ‘step-up’ treatment strategy. Treatment allocation was concealed from patients, GPs and researchers.
The antacids and placebo antacids are referred to as four times daily pills (q.i.d. pills), the PPIs and placebo PPIs as once daily pills (o.d. pills), and the H2RAs (step 2) as twice daily pills (b.i.d. pills).
The medication instructions () were described in the trial information leaflet and on all medication jars. Additionally, the GPs were asked to instruct the patients verbally as they would do in everyday practice.
Recruitment
GPs recruited patients during consultations. Adult patients presenting with a new episode of dyspepsia (i.e. they had not used any prescribed acid-suppressants during the previous three months and had not had an endoscopy during the last year) were considered eligible if they were able to complete (Dutch) questionnaires and if there were no contra-indications for using acid-suppressants. Patients giving a written informed consent received the trial information leaflet and the study medication for 30 days (120 q.i.d. pills and 30 o.d. pills).
Measurements
Completeness was measured by pill counts. Patients were instructed to return the medication jars to their GPs after each treatment step. To avoid counting errors, two researchers independently counted the returned pills. Based on the duration of medication use (in days) and the pill counts, the percentage of the prescribed doses taken (= total number of doses taken/total number of number of prescribed doses * 100%) was calculated. The duration of medication use was based on the consultation dates, except when other starting and stopping dates were indicated (e.g. when the medication was returned prior to the follow-up consultation or when medication was stopped prematurely because of side effects).
Intake fidelity was measured by patient questionnaires () sent out two weeks after the start of treatment, which could be returned using a pre-stamped envelope. Reminders were sent out after two weeks. Intake fidelity sum scores were calculated, rewarding correct answers with 1 point and partly correct answers with 0.5 points. The maximum score of 9 points indicating perfect intake fidelity. The GP ascertained demographic data and monitored treatment.
Table I. Intake fidelity
Data analysis
SPSS version 14.0 was used. A two-sided significance level of 0.05 was applied. In respect of completeness, means with standard deviations were calculated for each medication jar (q.i.d. pills step 1; o.d. pills step 1; b.i.d. pills step 2; q.i.d. pills step 3; o.d. pills step 3). A Friedman test was carried out to test whether there were statistically significant differences between these five types of medication. Mann–Whitney U tests were undertaken to test whether there were statistically significant differences between placebo and active pills. To check whether completeness was consistent over time, first, the Spearman's correlations between step 1 and step 2; and between step 2 and step 3 were calculated. Second, the proportions of partially adherent patients (using <80% or >120% of the prescribed pills/day) and adherent patients (using 80–120%) were calculated for each treatment step. Frequency Tables were produced on intake fidelity and the mean intake fidelity sum score with SD was calculated.
Results
Patient characteristics
664 patients agreed to participate in the DIAMOND-study, but 11 patients did not use any study medication. Thus, a total of 653 patients (mean age 47 (SD 15) years, 46% males, 6% non-Caucasians) were included; 334 randomized to the ‘step-up’ and 319 to the ‘step-down’ strategy.
For 176 patients (27%) completeness was unknown because the medication (n =105) was not returned or the treatment duration (n =81) was unknown; for 248 patients (38%) no intake fidelity sum scores could be calculated due to non-response to the questionnaire (n =147) or missing items (n =101). Pill counts were more often available for women, Caucasians and patients who started with step 2 (). Except for gender, the same accounted for the availability of the intake fidelity sum scores.
Table II. Differences between patients with and patients without missing values.
Completeness
A substantial number of patients used less than 80% of the prescribed medication. shows that partial adherence is very prevalent: for the q.i.d. pills one in two patients and for the o.d. pills one in four patients used less than 80% of the prescribed pills per day. Completeness for the q.i.d. pills was lower than for the other pills (mean % of prescribed dose taken per day: the q.i.d. pills 78% (step 1) and 68% (step 3); the o.d. pills 92% (step 1) and 84% (step 3); the b.i.d. pills 86% (step 2) (P <0.001, n =131)).
Table III. Completeness: Number of pills used per day.
Differences in completeness between the two trial arms were investigated, and we only found a small though statistically significant difference in step 1 (): completeness was lower for placebo o.d. pills than for the active PPI (means resp. 88% versus 92%, P <0.05). This may indicate that patients recognized the placebo pills. If this were so, however, then it would seem logical that they would only take the antacid pills but no differences were found for the antacids and placebo-antacids. Furthermore, no differences in completeness between the two trial arms in step 2 and 3 could be identified either. Therefore, it seems that the treatment allocation was adequately concealed.
shows that a small proportion of the patients who started all treatment steps were consistently partially adherent or completely adherent over time. This is reflected in medium correlation coefficients: the correlation between step 1 and step 2 concerning completeness was 0.38 (P <0.001, n =245) and the correlation between step 2 and step 3 was 0.23 (P <0.05, n =163).
Table IV. Adherence over time.
Box 1. Medication instructions.
Intake fidelity
shows that the most frequent deviations were not taking the q.i.d. pills four times a day (12% of the patients), and not taking the medication in relation to meals or bedtime (17% for q.i.d pills and 10% for the o.d. pills). The mean intake fidelity sum score was 8.13 (SD 0.94, n =405); 30% of the patients had a score of 9 indicating perfect adherence; another 30% scored 8.5 indicating almost perfect adherence (one partly incorrect answer). There were no differences in intake fidelity between the two trials arms.
Discussion
Summary of the main findings
Partial adherence with short-term acid-suppressants is very prevalent among newly diagnosed dyspeptic patients: one in two antacid users, one in three H2RA users, and one in four PPI users used less than 80% of the prescribed acid-suppressants. Few patients were consistently adherent or partially adherent over time. Most patients (70%) deviate on intake fidelity in one or more respects, mostly in timing of the medication intake in relation to meals.
Interpretation
There may be several reasons for the low completeness scores. First, patients may discontinue the medication because they think it is not effective; they may be unaware that it may take some time for the medication to become effective. Second, patients tend to discontinue medication as soon as distressing symptoms disappear (10,11), or use their medication on-demand, i.e. only when symptomatic (Citation11). Not completing the treatment course may lead to earlier recurrence of symptoms (Citation4), and, although on-demand use may be as effective as continuous use in chronic users (Citation3), it is generally not recommended for initial treatment (Citation2).
One might argue that partial adherence or on-demand use in the case of dyspepsia is not such a big problem. This may be true for patients for whom the acid-suppressants successfully relieves their symptoms, even though this may mean that expensive medication will be left over and may be thrown away, or symptoms may recur sooner (Citation4). However, what if the symptoms are not adequately relieved? Then it is vital to identify on-demand use and other types of non-completeness, to adequately judge treatment effectiveness and decide upon future treatment.
Few patients showed consistency in completeness in the diagnostic stage over time. This has meaningful implications for the predictability of adherence in practice: the continuation of (non-)adherence behaviour should not be assumed in prolonged treatment, and during follow-up visits adherence needs to be re-addressed.
The low intake fidelity scores were mostly caused by deviation from the instructions about timing. Perhaps patients found the suggested timing inconvenient or were not fully aware of the importance of following these instructions. The latter may be explained by a lack of attention paid to the verbal communication of these medication instructions, as revealed in a recent survey study conducted by our group (Citation12). This study showed that GPs paid little attention to explaining timing and way of taking dyspepsia medication. Explaining the rationale and stressing the importance of following all instructions may improve adherence and consequently may increase the effectiveness of the medication.
Strengths and limitations of the study
Our study thoroughly investigated patient adherence to a medication regimen in a relatively large primary care patient population. Nevertheless, some limitations concerning the internal validity of our completeness measure need to be addressed. Pill counts do not indicate whether the medication was actually taken and may result in overestimating completeness (Citation10). Furthermore, for the calculation of completeness it is important to measure treatment duration accurately. It is possible that treatment duration was overestimated for a number of patients. If patients stopped using medication days before returning their medication, this may have resulted in underestimation of completeness. However, even of patients who indicated that they had used the medication only for a short period of time, a large proportion used less than 80% of the prescribed dosage. Therefore, our results are probably a good approximation of the actual completeness of our study population.
As far as generalizability is concerned, the fact that patients had to return their medication, indicating that their completeness could be checked, may have improved completeness. Alternatively, the addition of a placebo does not reflect everyday practice. Consequently, patients may have taken fewer verum pills than they would have taken in everyday practice, and this could lower the effectiveness of our trial medication. Our results provide insight into adherence rates when five pills per day (step 1 and 3) or two pills per day (step 2) were prescribed; perhaps adherence would be better if verum pills only were prescribed. Nevertheless, our results have shown that the completeness for the q.i.d. pills was worse than for the other regimens; this implies that, when optimal adherence is required, it is perhaps best to prescribe o.d. pills. It has been confirmed that completeness decreases with increasing complexity (Citation10) and with multiple dosing in studies concerning dyspepsia (Citation13), hypertension (Citation14) and psychiatric disorders (Citation15).
Some limitations need to be discussed in respect of intake fidelity too. Intake fidelity, measured by questionnaire, reflected the patient's behaviour perception. Furthermore, the answers had to be given in general, so patients may have answered how they thought they used their trial medication most days. Patients may have given socially desirable answers. Although this requires some knowledge about what the correct answers are, this could easily be derived from the medication jars and the patient information leaflet. This means that intake fidelity may be even worse in everyday practice.
Comparison with existing literature
The limited literature available on adherence to acid-suppressants confirms our results. Van Soest et al., (Citation13) investigated adherence using a large Dutch primary care database containing prescription data and found that approximately half of the patients with at least two PPI prescriptions used less than 80% of the PPIs prescribed. In other patient groups, taking medication at the wrong time and/or omitting one or more doses were also the most common forms of deviations, e.g. among patients with hypertension (Citation16). Furthermore, one fifth to one half of elderly patients has difficulties understanding, or a lack of knowledge about, their medication instructions (Citation17,Citation18); this may explain why errors in intake were common in our study.
Implications for future research or clinical practice
These findings imply that there is room for improvement in patient adherence to medication regimens treating dyspepsia: in turn, this may increase the effectiveness of acid-suppressants and improve the diagnostic value of a course of treatment as a means of ‘testing’ the nature of the symptoms. The results imply that completeness may be improved by using o.d. or b.i.d. pills instead of q.i.d. pills for dyspepsia. Intake fidelity may be improved by clearly communicating the importance of following treatment instructions.
Conclusion
There is room for improvement in adherence rates for dyspepsia medication.
Funding
This study was part of the DIAMOND trial and financially supported by the Netherlands Organisation for Health Research and Development (ZonMw) (grant number 095–03–052). The protocol of the DIAMOND trial was approved by the ethics committees of the University Hospitals of Nijmegen, Utrecht and Maastricht.
Acknowledgements
The authors wish to thank all general practitioners and patients for participating in the DIAMOND trial.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Bartholomew LK, Parcel GS, Kok G. Intervention mapping: A process for developing theory- and evidence-based health education programs. Health Educ Behav. 1998;25:545–63.
- The Dutch Institute for Healthcare Improvement (CBO), Dutch College of General Practitioners (NHG). Multidisciplairy guideline stomach complaints (Multidisciplinaire Richtlijn Maagklachten). Alphen aan den Rijn: Van Zuiden Communications; 2004.
- Vakil N. Increasing compliance with long-term therapy: Avoiding complications and adverse events. Rev Gastroenterol Disord. 2005;5(Suppl. 2):S12–7.
- Hunt RH, Cederberg C, Dent J, Halter F, Howden C, Marks IN, . Optimizing acid-suppression for treatment of acid-related diseases. Dig Dis Sci. 1995;40:24S–49S.
- Howden CW, Chey WD. Gastroesophageal reflux disease. J Fam Pract. 2003;52:240–7.
- Thomson AB, Mahachai V, Bailey RJ, Kirdeikis P, Zuk L, Marriage B, . Twice daily nizatidine or ranitidine is superior to once daily dosing in elevating 24 h intragastric pH in patients with duodenal ulcer disease. J Gastroenterol Hepatol. 1996;11:1171–6.
- Robinson M, Rodriguez-Stanley S, Miner PB, McGuire AJ, Fung K, Ciociola AA. Effects of antacid formulation on postprandial oesophageal acidity in patients with a history of episodic heartburn. Aliment Pharmacol Ther. 2002;16:435–43.
- Fransen GA, van Marrewijk CJ, Mujakovic S, Muris JW, Laheij RJ, Numans ME, . Pragmatic trials in primary care. Methodological challenges and solutions demonstrated by the DIAMOND-study. BMC Med Res Methodol. 2007;7:16.
- van Marrewijk CJ, Mujakovic S, Fransen GA, Numans ME, De Wit NJ, Muris JWM, . Effect and cost-effectiveness of step-up versus step-down treatment with antacids, H2-receptor antagonists, and proton pump inhibitors in patients with new onset dyspepsia (DIAMOND study): A primary-care-based randomised controlled trial. Lancet 2009;373:215–25.
- Bosworth HB. Medication treatment adherence. In: Bosworth HB, Oddone EZ, Weinberger M, editors. Patient treatment adherence. Concepts, interventions and measurement. New Jersey: Lauwrence Erlbaum Associates Inc.; 2006:147–94.
- Krol N, Muris JW, Schattenberg G, Grol R, Wensing M. Use of prescribed and non-prescribed medication for dyspepsia. Scand J Prim Health Care 2004;22:163–7.
- Fransen GAJ, Mesters I, Van Marrewijk CJ, Mujakovic S, Knottnerus JA, Muris JWM. Leefstijladviezen, medicatie-instructies en het belang van therapietrouw bij maagklachten. Lifestyle changes, medication instructions and the importance of adherence in case of dyspepsia. Huisarts Wet 2007;50:446–52.
- Van Soest EM, Siersema PD, Dieleman JP, Sturkenboom MC, Kuipers EJ. Persistence and adherence to proton pump inhibitors in daily clinical practice. Aliment Pharmacol Ther. 2006;24:377–85.
- Iskedjian M, Einarson TR, MacKeigan LD, Shear N, Addis A, Mittmann N, . Relationship between daily dose frequency and adherence to antihypertensive pharmacotherapy: Evidence from a meta-analysis. Clin Ther. 2002;24:302–16.
- Granger AL, Fehnel SE, Hogue SL, Bennett L, Edin HM. An assessment of patient preference and adherence to treatment with Wellbutrin SR: A web-based survey. J Affect Disord. 2006;90:217–21.
- Rudd P. Clinicians and patients with hypertension: Unsettled issues about compliance. Am Heart J. 1995;130:572–9.
- Blenkiron P. The elderly and their medication: Understanding and compliance in a family practice. Postgrad Med J. 1996;72:671–6.
- Lowe CJ, Raynor DK, Courtney EA, Purvis J, Teale C. Effects of self medication programme on knowledge of drugs and compliance with treatment in elderly patients. Br Med J. 1995;310:1229–31.