Abstract
From a clinical point of view, the subjective preference of the patient regarding headache medication is imperative. Consequently, a new trial design for headache medication, the so-called patient preference trial, has been vividly discussed. However, some critical questions have been raised concerning preference trials, such as whether placebos should be used, the necessity of blinding, and how to assess the data if a patient does not prefer medication A over B. As patient preference is the topic, we passed these questions on to headache patients using a questionnaire, which was handed out to 1112 headache patients. Out of 612, 486 returned questionnaires were correctly completed and analysed. Complete pain relief was the most important factor for 61% of patients to qualify a therapy effective. This is in contrast to the literature, where rapid speed of onset of the drug is discussed as the most important factor for migraineurs. Regarding the study design, 80% of the patients want to decide the time-point of taking acute medication themselves. About 60% of the patients would participate in placebo-controlled clinical trials and agree that studies should be blinded. If patients had to decide between two equally effective drugs, most would vote for the drug which is available in different application forms. Furthermore, we used the same questionnaire to ask headache specialists with expertise in performing headache studies the same questions. This was mailed to and completed by 22 experts in Germany. In this article, we discuss the patient preference compared with the expert preference regarding clinic trials and drug therapy.
Introduction
In clinical trials, many outcome measures are investigated and researchers debate the optimal endpoints for the studies. In 1991, the first edition of the guidelines of the International Headache Society Clinical Trials Subcommittee was published and subsequently revised in 2000 because of the requirement ‘to improve the quality of controlled clinical trials in migraine’ (1). However, the patient's satisfaction with treatment is still not well understood. Patients develop individual strategies to cope with the headache, have opinions of requirements for a drug therapy, efficacy and side-effects and try to find out the best treatment for themselves. These individual practices may well differ from the investigator's view of a ‘good drug therapy’ and, furthermore, may differ from the outcome of clinical trials, if these trials are well designed but clinically irrelevant. For example, in modern migraine studies, patients are asked to take the drug when headache is moderate or severe (2, 3). However, in day-to-day practice, migraine patients often complain of therapy failures if the medication is not taken at the beginning of headache. Clinical research is dependent on the willingness and the agreement of our patients to take part in clinical trials. The protocols of most study designs, however, are rigid and often patients have to undergo the risk of poor treatment. This is the case in placebo-controlled trials or in trials with late onset of the treatment (4).
In a recent study investigating the optimal time-point to take the medication in the treatment of an acute migraine attack, taking rizatriptan at the onset of headache was associated with more rapid relief of headache and reversal of functional disability (5). Thus, the treatment of a migraine attack at an early stage is a worthwhile therapeutic pursuit, as success may be better before kindling effects have occurred. However, nearly all competitor trials investigating acute headache medication in the last 10 years, instructed the patients to take the study medication at an advanced-pain stage.
In this context, researchers begin to ask whether the outcome parameters of the trials are clinically relevant and whether the conclusions of the trials, e.g. our recommendations on drug therapy, reflect what the patients want (6–10). Patient satisfaction with migraine therapy was evaluated to determine the relationship between measures of treatment response in a clinical trial and patient satisfaction with treatment. Dahlöf found that rapid headache relief was ranked as ‘very important’ by 69% of patients (11, 12). Eighty-four per cent of patients perceived rapid relief to be 30 min and 16%, 1 h. No patient considered pain relief later than 1 h to be rapid. Similarly, in a study by Lipton and Stewart, the most common reasons for dissatisfaction with a migraine medication were pain relief taking too long (87%), lack of complete pain relief (84%), and inconsistent headache response (84%) (13). Interestingly, Dahlöf commented on his pilot study that the efficacy of the treatment with regard to headache relief does not necessarily relate to patient's preference (11). A recent study suggests that, next to freedom from pain within 2 hours, the absence of associated symptoms (photophobia, phonophobia, nausea and vomiting) is also important for the patient's satisfaction with treatment (3).
Researchers debate the optimal design for patient preference trials. Headache specialists discuss whether clinical trials should be placebo-controlled or not, whether an open or closed design is more suitable, at what time-point the attack should be treated, how often a patient should treat his attacks before he is thought to be able to decide which drug he prefers, and how to interpret a preference study when a patient has no preference (14–18).
With respect to this debate, we developed and handed out a questionnaire for headache outpatients, asking about their opinions on drug therapy and clinical trials. Additionally, we sent the same questionnaire to German headache experts. This article represents the results and discusses possible consequences on further study designs.
Methods
Questionnaires (1112) designed to investigate different aspects of patient's satisfaction with drug therapy, as well as opinions and preferences on clinical trials, were either handed out during the patients’ visits or sent by post. The aim of the questionnaire (for a translation see Appendix 1) was to evaluate the opinion of headache patients on different items regarding study designs. As the main topic is preference, we also included questions regarding expectancies of an ideal drug for acute therapy. To avoid undue bias in the responses, the order of the questions was randomized. Inclusion criteria was a headache diagnosis of an idiopathic relapsing or chronic headache disorder according to IHS criteria in adult patients (19). The diagnosis was made by a physician who is experienced in headache diagnosis and therapy. As the questionnaire was anonymous, we did not specify the headache diagnosis. All questions were designed in a simple ‘yes’ or ‘no’ fashion and only one answer was allowed in those questions where the participants had to declare one out of several items as the most important. The questionnaire was responded to by 328 patients from the headache outpatient clinic of the University of Regensburg and 286 patients from the headache outpatient clinic of the University of Münster. The same questionnaire was posted to 22 headache specialists in centres throughout Germany who have extensive experience in study performance and design.
Results
The questionnaire was handed out to 1112 headache patients and was returned by 612 (55%). Out of questionnaires from Regensburg, 70/328 (21.3%) were not eligible, and from Münster 58/286 (20.3%). In these cases, either the forms were not filled in completely, or more then one item was marked when only one was allowed. Therefore, 486 out 612 (79.4%) were complete and correctly filled in and subequently analysed. Twenty-two physicians took part and all their questionnaires were filled in correctly and have been analysed in the same way as the patients’ questionnaires.
The answers of the patients and of the headache experts, as a percentage, is included in Appendix 1. Cross-validation between the correctly filled-in questionnaires of headache patients in Regensburg (n = 258) and Münster (n = 228) demonstrated that the answers for each question were comparable. Figure 1 summarizes the opinions of the patients and experts regarding the study design, including placebo, blinding, point of drug intake, and whether they prefer an open or closed trial design. Of our patients, 318 out of 486 (65.4%) would take part in a placebo-controlled study and, of those, 74.5% would not consider it unpleasant to learn that their symptoms responded to placebo. Out of 486 patients, 317 (65%) support clinical trials which are blinded and 302 (62.1%) prefer to take part in an open-study design. Three hundred-and-ninety (80.3%) of the patients want to decide the time-point of drug intake in clinical trials themselves, whereas 13 out of the 22 experts (59%) think that the time-point of drug intake should be defined. Figure 2 shows the patients’ and experts’ preferences regarding drug efficacy and tolerability. For 297 out of 486 patients (61.2%) and 31.8% of the experts, the most important item regarding headache medication is the ability to render them pain free, whilst for 15% of the patients and 45.4% of the experts, respectively, pain relief is the most important item.
Patients’ and experts’ views on study design: placebo, blinding and open or closed trial design. □, Patients’ responses; ▪, experts’ responses.
Which is the most important item for efficacy and tolerability of an acute headache medication in the opinion of patients and experts? □, Patients’ responses; ▪, experts’ responses.
A slower drug with less side-effects was preferred by 78% of patients and a slower drug with longer duration by 79.8%, compared with a drug which works faster but with more side-effects or shorter duration. Of the physicians, 20 out of 22 (90.9%) would prescribe the drug with slower onset but fewer side-effects, and 72.7% the slower drug with longer duration of action. When two drugs are judged equal in efficacy and side-effect profile in a given clinical trial, for 231 (48%) of the patients, to vote for a drug, the most important item is numerous options of administration. Other possibilities were: package design; form, colour and size of tablets; subject to prescription; freely available; price of the drug. Of the experts, 68.2% decided in favour of the price of the drug (Fig. 3).
When two drugs are judged as equal in efficacy and side-effect profile, which is the most important? The most important item to vote for a drug: □, Patients’ responses; ▪, experts’ responses.
Regarding willingness to take prophylactic treatment, even if patients already have an effective and well-tolerated acute therapy, 387 (79.6%) of them favour a preventative medication, and consider four attacks per month necessary to accept taking drugs every day.
Discussion
Clinical trials for acute migraine therapy have evolved over the years from open-label, small, observational studies to highly structured, controlled trials. Recently, so-called ‘Preference trials’ have been introduced to reflect clinically relevant trial conditions.
It is common knowledge that patients do not join clinical trials for scientific reasons. Expectations of personal benefit, rather than altruism, is more often their motive for participating (20, 21). Recently it has been debated whether placebo-controlled studies are ethically acceptable if a sufficient therapy is available (22). However, headache is a condition in which over 30% of the afflicted respond to placebo (23). Pill colour, patient confidence and the means of administration seem to be important (24, 25). In a preference trial, it seems questionable whether a placebo control has to be included, as preference as such is highly subjective.
We have, therefore, asked headache patients whether they would accept placebo treatment in a trial: 65.4% of patients would take part in placebo-controlled studies and think they are necessary to unequivocally find out the efficacy and the side-effects of a drug; 34.6% voted not to take part. If we take these numbers as representative for all headache patients, this would suggest that, at least in German studies, about 30% of the headache population is not represented in headache medication trials. Interestingly, as indicated above, 74.5% do not consider it unpleasant to learn that what did help them was actually placebo. Experts clearly vote for placebo-controlled study designs and 14 out of 22 (63.6%) of them think that placebo-controlled study designs are necessary in headache research, and none has a problem telling the patient that they responded to placebo (Fig. 1). We conclude that patients and experts accept the placebo effect as a part of an effective drug therapy.
Obviously, headache patients accept the necessity to improve objectivity of clinical trials, because 65% indicate that clinical trials should be blinded to improve objectivity, even although 62.1% prefer to take part in an open study. In addition, 86.4% of the headache experts think that clinical trials should be blinded and opted for a closed-study design (Fig. 1).
In common study protocols, the time-point of drug intake is defined: the headache intensity has to be at least moderate or severe. In contrast, patients report that drug intake has to be as soon as possible, otherwise the desired drug effect is at risk. Consequently, an overwhelming number (80.3%) of the patients want to decide themselves when to take the drug. The experts, however, are unsettled: 59% think that the time-point of drug intake should be predetermined and 41% that the patients should decide themselves (Fig. 1). A recent study reported that taking rizatriptan at the onset of headache was associated with more rapid relief of headache and reversal of functional disability (5). With respect to the optimal time-point of drug intake, one can only speculate about the outcome of the published head-to-head comparisons of migraine therapy (11, 18, 26–28). For these, and other reasons, the strengths and weaknesses of conventional approaches have been discussed (3, 11, 12, 29).
Regarding the preference for drug efficacy and tolerability, 61.2% of patients decided that pain free is the most important factor and only 15% chose pain relief. The other items were, by far, less important. Interestingly, only 2.9% estimated speed of onset to be the most important item, next to a low side-effect profile (9.1%) and efficacy against concomitant symptoms (4.8%; Fig. 2). In prior studies with migraineurs, rapid onset of pain relief, lack of adverse effects and relief of associated symptoms received high ratings from the majority of patients (3, 11, 13, 30). Nearly 80% of our patients decided against rapid speed of onset if this would lead to more side-effects. Additionally, 79.8% of patients prefer a longer duration of action, even if the speed of onset is lower. Similar results were obtained from the headache experts, who rated ‘pain free’ as the most important item (31.8%), whereas 45.54% voted for pain relief (Fig. 2). Additionally, 20 out of the 22 physicians would prescribe the drug with slower onset but fewer side-effects, and 72.7% the slower drug with longer duration of action. This is in contrast to the literature, where rapid speed of onset of the drug is discussed as the most important factor for migraineurs (3, 13, 31). Consistent with earlier results suggesting that patients who experience a rapid recurrence of headache do not consider the medication to be effective (11), we conclude that rapid speed of onset is not the most important point in drug preference, if other aspects like headache relief, side-effects and duration of action are taken into account. However, we did not distinguish between headache syndromes. Therefore, based on demographic data, we must assume a substantial amount of patients who took part in the study were suffering from some form of chronic head pain. These patients may have different expectations of acute medication and this may therefore bias the result regarding speed of onset. However, consistent with the above-mentioned statement, only 2.9% of all patients voted speed of onset as important. Based on demographic data, however, most of the patients in headache outpatient clinics suffer from migraine, and we have to conclude that a substantial part of migraine patients prefer factors other than speed of onset as imperative for an ‘effective’ substance.
In a preference study, patients are asked for their preference for drug A or B. In this case, patients have to decide between items such as efficacy and side-effects. But when patients judge both drugs as equal in efficacy and side-effect profile, which item would nevertheless turn the balance? Forty-eight per cent voted for a variability of administration method. This could be a consequence of the individual strategies patients develop to cope with their headaches. None of the patients considered package design of the medication as important and only 2.9% voted for form, colour or size of the tablets (which can influence the placebo effect as discussed above and may affect a patient's preference subconsciously). Experts were more resolute: most of them (68.2%) voted for the price as being the most important item and 7 out of 22 (31.2%) for variety of administration.
In the discussion on preference trials, it has been discussed whether the efficacy of a drug therapy can be sufficiently judged after treating only one headache attack, or if it is necessary to treat several attacks (12). All of the experts and 85% of the patients think that treating a single attack is not sufficient and voted for several treatments. In this case, consistency becomes a factor in influencing the outcome of preference trials. An objection has been that, after several treatments, the patient might ‘forget’ how the first treatment performed. However, from a clinical point of view this is academic and, in any case, may be counteracted by randomizing the order of treatment.
Asked about expectations of being fully restored to working ability after drug intake, 55.8% of the patients accept a reduction in their activities of daily living and, therefore, do not count being headache-free as a conditio-sine-qua-non in headache treatment. Interestingly, 63.3% of the experts think that patients should accept a rest after drug intake. It is intriguing that more than half of the patients, and more than 60% of the experts, accept a rest as part of the therapy and obviously do not imply being pain-free (at 2 h) as the most important item of the ‘ideal’ drug.
In conclusion, we asked patients with an idiopathic headache disorder at two large supraregional headache outpatient clinics about their opinion on preference trials and their expectations regarding an effective drug therapy. Understanding factors influencing patients’ preferences will improve guidance in making rational choices in expanded symptomatic migraine treatment. However, we did not differentiate between headache diagnosis and feel that the views of migraineurs may differ from patients with tension-type headache or other types of headache. This could be a topic for further studies.
Footnotes
Acknowledgements
The authors wish to thank Tina Schneider for helpful assistance in the evaluation of the questionnaires. This study was supported by an unrestricted grant from Berlin Chemie, Germany.