Real-world Treatment and Associated Healthcare Resource Use Among Migraine Patients in Germany

Abstract


Background
Migraine is one of the most frequently disabling neurological diseases and is the leading cause of disability in adults under the age of 50 [1]. It often manifests in the form of unilateral pain and episodes of pulsating or throbbing headaches with symptoms including nausea and sensitivity to noise/light, as well as vomiting [2], [3]. Pharmacological treatment of migraine often includes use of acute therapies, which aim to relieve symptoms during attacks, as well as preventive or prophylactic therapies in order to decrease the severity and frequency of attacks [3], [4]. In addition, several non-pharmacological approaches such as acupuncture and psychotherapy may be considered as treatment options [5], [6].
Due to its diverse symptomology and varying levels of clinical severity, migraine often has a major impact on the quality of patients' lives. In recent years, migraine has been increasingly recognized as one of the most frequent causes of disability [1]. In 2016, migraine was found to be the leading driver of years lived with disability (YLD) worldwide in the age group of 15 to 49 years [7]. Since migraine primarily affects people in their productive years (from a labor market perspective), it leads to substantial productivity losses, and therefore constitutes a serious public-health concern [5], [8], [9].
In addition to these humanistic and social/economic burdens, migraine is also recognized as an area of signi cant unmet therapeutic need, as it relates to the underdiagnosis and widespread undertreatment of patients [2], [10]. There are a number of reasons for the undertreatment of patients with prophylactic therapies, including the misdiagnosis of many migraine patients with general headache [10]. Ultimately, correct diagnosis and assessment of disease type, as well as the intensity and frequency of attacks, is essential to facilitate appropriate management of migraine [4], [11].
Ultimately, there is very limited evidence on the treatment of migraine patients and associated healthcare resource utilization (HCRU) and costs in Germany. Nevertheless, these data are urgently needed to understand potential avenues for improving the real-world treatment of patients. This retrospective claims data analysis aimed to identify patients with headache and migraine in Germany, and to describe their acute and prophylactic drug treatment as well as HCRU, direct and indirect costs.

Study design and data source
This was a retrospective, non-interventional comparative cohort analysis of patients with migraine or headache, utilizing an anonymized routine dataset provided by AOK PLUS, a German public sickness fund insuring about 3.2 million persons in Germany. Complete statutory insurance data on all documented diagnoses (inpatient and outpatient), treatments and procedures, as well as pharmaceutical and non-pharmaceutical prescriptions for the period 2013-2017 were available. Furthermore, the dataset included information on the cost of prescriptions, hospitalizations, and rehabilitations, as well as outpatient physician visits. As the dataset was anonymized, no ethical approval was required. Approval from the data owner was provided prior to the start of the analysis, based on the submission of a study protocol. The study was led by a Scienti c Steering Board, consisting of the authors of this publication.

Patient eligibility criteria
In order to describe treatment patterns and HCRU in 2017, patients were included if they received at least one inpatient and/or two con rmed outpatient diagnoses of headache (ICD-10 R51 and G44) or migraine (ICD-10 G43) from 2013-2016, and if they were at least 18 years of age and alive on 01/01/2017. Furthermore, patients must have been continuously insured from 2013-2017. Baseline characteristics such as sociodemographic pro le, comorbidities, and type of diagnosing physicians (i.e. general practitioners, specialists) were observed from 2013-2016 in four main patient cohorts.

Analysis of treatment patterns
Both acute and prophylactic prescriptions (Rx) were analyzed in the observational year 2017 to group patients based on their medication therapy. Data on over the counter (OTC) drugs were not available. The study considered a range of acute medications including triptans, analgesics, antiemetics and ergot alkaloids, as well as prophylactic agents which included propranolol, metoprolol, bisoprolol, unarizine, topiramate, amitriptyline, onabotulinumtoxinA, and valproic acid. A list of all observed agents and their respective anatomical therapeutic chemical (ATC) codes is presented in Appendix Table 1. Patients were assigned to one of four main cohorts, based on the type of Rx treatment they received in 2017, including: neither acute nor prophylactic medication (cohort N), only acute medication (cohort A), only prophylactic medication (cohort P), or both acute and prophylactic medication (cohort AP) (Fig. 1).
The number of prescriptions and average dosage received was calculated per agent as well as per drug class and reported for the four treatment cohorts. A proxy for prescribed daily dosage was calculated based on available information about package size and de ned daily doses (DDD) for medications according to the World Health Organization (WHO) [12].
In a separate analysis, migraine/headache patients who began receiving prophylactic treatment in 2014-2015 were observed in order to identify treatment lines, switches, and treatment discontinuation in the period following initiation of treatment. Patients who began receiving prophylactic treatment (including at least one of the above prophylactic agents) between 01/01/2014-31/12/2015 and had no prophylactic treatment history in 2013 were included in the analysis. The rst prescription for prophylactic treatment was considered as the index date and all patients were observed for 24 months after initiation of therapy. Discontinuation of therapy was considered, when a patient did not re ll a prescription for his/her prophylactic medication within 180 days of the last expected drug coverage date of the last observed prescription.
In addition, an exploratory analysis was performed to estimate the number of patients with migraine/headache who did not receive prophylactic medication in 2016 but may have bene ted from such a therapy. A potential need for a prophylactic medication was de ned as having received > 8 DDDs triptans per month in 2016, or alternatively > 5 DDDs triptans per month in combination with either (i) an emergency room (ER) visit with a diagnosis of headache/migraine, (ii) at least one migraine/headacheassociated hospitalization, or (iii) an inpatient rehabilitation stay associated with headache/migraine in 2016. For all patients with an identi ed need for prophylactic therapy in 2016, prescriptions of prophylactic medications were observed in 2017.

Analysis of HCRU and cost
The following HCRU variables were observed and reported per patient-year (ppy): inpatient hospitalizations, outpatient physician visits, inpatient rehabilitation stays, physiotherapy prescriptions, prescribed aids and remedies, acupuncture prescriptions, and days absent from work. Overall costs (€) were reported ppy for the following categories: (i) outpatient visits [based on documented 'treatment points' for outpatient physician visits], (ii) outpatient prescriptions [based on o cial retail list prices for outpatient medication prescriptions], (iii) hospitalizations, including procedures and medications [based on diagnosis-related group reimbursements for inpatient hospitalizations], (iv) inpatient rehabilitations [directly covered in the database], and (v) prescriptions of aids and remedies [based on prices, which are directly covered in the database]. Since patient-speci c salaries were not available in the database, indirect costs related to loss of productivity were approximated based on the number of days absent from work, multiplied by age-and gender-speci c average gross salaries in Germany [13]. HCRU and cost analyses were repeated for the exploratory cohort of patients who did not receive prophylactic treatment in 2016, but may have bene ted from it (i.e. those receiving > 8 DDDs triptans per month or > 5 DDDs in combination with further criteria as previously described).

Patient characteristics
We identi ed 199,283 patients with headache and/or migraine in our database (mean age 49.49 years; 73.04% women; Table 1). Among patients included in the study, 6.46% did not receive an outpatient diagnosis of migraine or headache, but were diagnosed with headache/migraine at least once during an inpatient stay, while 93.54% of the patients were diagnosed in an outpatient setting (55.63% with migraine diagnoses and 37.91% with headaches but not migraine).
In addition to pharmacological treatment of migraine, other forms of non-pharmacological treatment were also analyzed. The number of patients receiving opioids varied among each group, ranging from 2.94% in cohort N to 24.57% in cohort AP. Only a small proportion of patients in the study received a prescription for acupuncture (1.29%) or for physiotherapy (6.59%). In cohorts receiving acute treatment for migraine (cohorts A and AP), a higher percentage of patients was found to receive nonpharmacological therapies (as compared to patients in cohorts N and P) ( Table 2).

HCRU and costs in 2017
HCRU For nearly all observed HCRU/cost variables, a higher utilization was found in patients receiving prophylactic treatment (cohorts AP and P), compared to patients without medication (cohort N) or with only acute Rx (cohort A). HCRU and cost were also reported for the exploratory cohort of 980 patients expected to bene t from prophylactic treatment (those who received substantial monthly doses of triptans, but no prophylactic Rx). In these patients, migraine/headache related HCRU was comparatively high.
On average patients in cohort AP were hospitalized (for any cause) around 0.90 times per year, a rate which is over three-fold higher than that which was observed in patients from cohort N who received no treatment (0.28 ppy). Furthermore, migraine/headache-related hospitalizations were 0.06 ppy in cohort AP as compared to 0.01 ppy in cohort N. In the cohort of 980 patients with no prophylactic treatment but high triptan use, the rate of migraine/headache-related hospitalizations was 0.05 ppy.
The number of patients with sick leave and the number of days of sick leave were both higher in patients from cohort A, in comparison to patients who received a prophylactic medication (cohorts AP and P).
Patients in cohort A took on average 17.04 days of sick leave (1.06 days related to migraine and/or headache), compared to an average of 7.52 days (0.54 days) in patients from cohort P, and 11.77 days (1.25 days) in patients from cohort AP. For the exploratory cohort of 980 patients, the number of patients who had at least one absence (51.22%) was higher than in all other cohorts (21.59%-47.76%). Furthermore, for this cohort, the number of days of missed work due to migraine/headache was 3.94 days ppy, and therefore signi cantly higher than in all other cohorts (0.43-0.54 days ppy).

Discussion
Summary and comparison of ndings with existing literature In this study, we analyzed patients with headache/migraine in a real-world setting in order to increase knowledge related to current treatment patterns of these patients in Germany. The goal of this research is to outline ongoing challenges related to the diagnosis and treatment of migraine patients in order to pave the way for improvements in medical care for patients with migraine.
Management of migraine and headache is multidimensional. Emphasis should be placed upon educating patients about appropriate use of OTC medication, improving the quality of advice provided to patients by pharmacists, as well as integrating, coordinating and synchronizing services offered by primary healthcare providers as well as specialists. Around 50% of migraine patients with relatively mild and infrequent attacks attempt to independently manage their disease through use of OTC remedies, and thus may not consult a physician. Nevertheless, at least 50% of patients require some degree of medical follow-up or consultation [14]. To a large extent, demand for clinical consultations among migraine patients is driven by the ine cacy of OTC drugs and the resulting need for prescription treatment. It has been previously reported that up to one-third of all migraine patients who suffer from frequent migraine attacks, are probably in need of prophylactic medication [4]. It is particularly important for these patients to receive adequate treatment to improve clinical outcomes, well-being, and productivity in the workplace [15], [16]. Our ndings indicate that, even in a relatively well-resourced country like Germany, only a small portion of patients expected to bene t from migraine medication actually receive treatment.
Our study reveals that nearly half of all patients with migraine/headache diagnoses did not receive any prescription of acute or prophylactic medication in 2017 (43.47%). Approximately one-third of the patients received only acute medications (33.81%) and just one fth (22.72%) were prescribed prophylactic agents. Our exploratory analysis of patients who heavily depend on prescribed triptans (> 8DDDs per months or > 5DDDs per months with additional indicators of high disease severity) revealed a stark unmet need for prophylactic agents, since less than 10% of these patients received subsequent prescriptions for prophylactic medication. These ndings align with those of previous studies which have demonstrated generally poor treatment patterns and disease performance in headache/migraine patients throughout Germany and elsewhere in Europe. A recent medical chart review reported that 27.5%-58.4% of migraine patients do not receive any therapy [17], while a study of 8,000 migraine patients from 10 EU countries con rmed that very few patients received adequate migraine-speci c treatment including both, acute and prophylactic medications [8]. Further to these ndings, one study in Germany demonstrated an overall low frequency of migraine-speci c prescriptions, especially among general practitioners (GPs) [17].
However, reasons for the low quality of care afforded to migraine and headache in patients throughout Germany and across the world are manifold. Despite the fact that migraine is highly prevalent and is associated with huge adverse consequences for affected people and the society, it is generally not considered a signi cant medical problem by patients, healthcare providers and health policymakers [18].
Oftentimes people affected by migraine are not aware of their disease and thus, do not consult a doctor to receive a diagnosis or proper treatment for their symptoms. One study provided evidence that awareness of migraine was low among both affected patients and healthcare providers [19]. These ndings were corroborated by a study in France which indicated that 60% of 10,000 patients with migraine were unaware of their diagnosis [20]. With this in mind, we acknowledge potential underestimation of unmet need in our patient population, since not all patients in need of therapy will have visited physicians during the study period and, therefore, were not included in this analysis. A previous study in Germany provided evidence that only 42% of patients with migraine consulted a physician, while a clear majority of patients relied exclusively on OTC medications, despite reporting symptoms [21]. This presents a limitation of the present study, as data on OTC medication use for our patient population were not available in the dataset.
In addition to the potential lack of awareness regarding migraine diagnoses, some GPs and specialists may not perceive headaches to be medically important enough to spend time diagnosing/treating symptoms, providing patients with detailed advice, or following-up with regard to the presentation of new symptoms. A study conducted in 10 headache centers in Italy revealed that only 26.8% of patients with migraine had received a previous diagnosis that was correct [22]. Another multinational study found that only 28 percent of migraine patients (who were subsequently referred to specialized headache centers) had been diagnosed with migraine by their treating GPs [23]. Furthermore, as shown by a study conducted in the UK, some GPs do not follow established treatment guidelines when patients with headache symptoms attend their practice [24]. As a consequence, patients receive either no treatment or improper treatment, as demonstrated in the Europe-wide EUROLIGHT study [8].
However, among patients who received a new prescription for prophylactic therapy in our study (N = 9,005), 60.78% discontinued prophylactic treatment within two years, without re-initiation. Ultimately, low medication compliance constitutes another key driver of poor medical care and high unmet need among patients. Poor adherence to prophylactic medication for migraine has been observed in previous research and was shown to result from insu cient effectiveness and/or treatment-related side effects [25]. One retrospective claims analysis in the US revealed that over two-thirds of chronic migraine patients discontinued their medication within six months of initiation, and that more than 80% of patients discontinued treatment within twelve months [26]. Such challenges related to treatment adherence and potential side effects signal an ongoing need for the improvement of migraine therapies, with regard to effectiveness and safety.
Our HCRU and cost analysis con rmed high levels of unmet need in a substantial proportion of observed patients, particularly in those who receive medications. According to our analysis, patients taking prophylactic medications were hospitalized more frequently and logged more sick days than patients who did not receive prophylactic treatment. Although this may be due in part to the likelihood of more severe cases among those who received prophylactic treatment, and these ndings are likely biased by indication (i.e. higher proportion of migraine patients in this cohort), patients with prophylactic migraine medications had higher utilization of acute medications too, indicating a higher headache frequency.
Unfortunately, no data on headache frequency among study participants was available in the database to con rm this. Generally, economic consequences of migraine, comprised of direct and indirect costs, are enormous.
Cost estimates are an important measure used to inform health policymaking via the estimation of expenses incurred by different healthcare services. In Europe, nancial costs attributed to migraine are estimated to range between 50 € and 111 € billion, with direct and indirect costs accounting for 7% and 93% respectively [27]. Direct costs incurred by healthcare services attributed to migraine, were found to be 1,222 € per person [27]. Another European study reported comparable average costs over three months, ranging from 373.8 € in Germany to 929.6 € in the UK [28]. Indirect costs, as shown within these studies, constitute most of the economic burden, with productivity loss accounting for two-thirds of indirect costs [27]. Our analysis con rmed that indirect costs (related to days absent from work) should not be neglected in analyses of migraine related HCRU and cost. On the contrary, it must be emphasized that average monthly costs associated with diminished productivity at work (presenteeism) are higher than those arising from sick leave (absenteeism) [29]. However, an estimation of costs related to onsite productivity loss was not possible in this study and research in this area is generally limited.

Limitations
Strengths of this study include an absence of study site/patient selection bias, coverage of all inpatient and outpatient healthcare sectors, and utilization of a large cohort of 199,283 patients. Nevertheless, we acknowledge that some limitations may apply. First, generalizability of results could be affected by the fact that the health insurance fund which provided the data only covers patients in two regions of Germany (Saxony and Thuringia). However, since health reimbursement rules are identical across Germany, considerable differences in the management of migraine patients are not expected. Second, information on the number of migraine attacks in observed patient cohorts was not available. Third, analysis of prophylactic treatment with monoclonal antibodies such as erenumab, galcanezumab and fremanezumab was not included as these medications were not available during the period of observation. Fourth, dosage as well as persistence analyses were conducted using data on lled prescriptions only. Information on un lled prescriptions as well as OTC medications was not analyzed, as those data were not available in the relevant sickness fund dataset. Ultimately, the latter may have caused an underestimation of medication use in our patient population. Moreover, HCRU among headache/migraine patients may have been underestimated if alternate therapies not covered by the participating sickness fund were accessed by patients via alternate means (i.e. supplementary insurance, out-of-pocket payments, etc.). Fifth, this analysis was exclusively descriptive, and no adjusted analyses were performed. Finally, our HCRU and cost analysis was based on reimbursements only, which exclude patient co-payments. However, these costs are generally low in Germany.

Conclusions
Can the introduction of effective medical services for migraine improve clinical outcomes and costsaving measures? The bene ts are potentially enormous, both in terms of enhancing economic outcomes as well as addressing unmet needs in one billion people who are affected globally. As presented in this paper there are multiple key elements for improving the medical care of headache and migraine patients in Germany. Firstly, patients must receive better education about the disease and available treatment options, including the effective use of OTC drugs. This education could potentially be provided by public health-education programs. In addition, there is an ongoing need for healthcare providers (and especially GPs), to increase their respective knowledge of how to recognize, diagnose, and properly treat migraine.
As such, physicians should inform themselves of and consistently adhere to recommended clinical guidelines, while educating patients on the importance of treatment adherence. Furthermore, healthcare services for patients with headache and migraine should be provided nationally in a structured, effective, and equitable manner. Finally, the development of new therapeutic options with increased effectiveness and tolerability, if adequately applied by physicians, offers targeted patients with a critical opportunity to relieve their symptoms and address ongoing challenges associated with persistent migraine. Acute medications before and after start of prophylactic 1L treatment Note: Ergot alkaloids are not reported because of the low number of patients with prescriptions. Figure 3