Use of a national database as a tool to identify primary medication non-adherence: The Estonian ePrescription system

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Research in Social and Administrative Pharmacy journal homepage: www.elsevier.com/locate/rsap

Use of a national database as a tool to identify primary medication nonadherence: The Estonian ePrescription system Ott Laiusa,b,∗, Heti Pisarevc, Daisy Volmerd, Sulev Kõkse, Aare Märtsona,f, Katre Maasalua,f a

University of Tartu, Department of Traumatology and Orthopedics, L.Puusepa 8, Tartu, Estonia Estonian State Agency of Medicines, Nooruse 1, Tartu, Estonia c University of Tartu, Institute of Family Medicine and Public Health, Department of Epidemiology and Biostatistics, Ravila 19, Tartu, Estonia d University of Tartu, Institute of Pharmacy, Nooruse 1, Tartu, Estonia e University of Tartu, Institute of Biomedicine and Translational Medicine, Department of Pathophysiology, Ravila 19, Tartu, Estonia f Tartu University Hospital, Clinic of Traumatology and Orthopaedics, L. Puusepa 8, Tartu, Estonia b

A R T I C L E I N F O

A B S T R A C T

Keywords: Primary medication non-adherence Bisphosphonates Osteoporosis Electronic prescribing

Background: Medication adherence can be divided into primary and secondary adherence. Primary medication non-adherence (PMN) occurs when a patient does not obtain medicine with their initial prescription. Secondary non-adherence measures prescription refills among patients who previously filled their first prescription. While secondary non-adherence has been studied thoroughly, PMN has been assessed less extensively, due to lack of available data. Estonian ePrescription system might prove a valuable tool for this. Objectives: The aim of this study was to evaluate PMN and the interval between prescribing and dispensing of medicines using the Estonian ePrescriptions database to establish its potential use for this purpose and for other qualitative drug utilization research measures. Osteoporosis medicines were used as an example. Methods: The Estonian Prescription Centre was used to evaluate if patients purchase medicines after initial prescription of osteoporosis medicine. Prescriptions from 2012 to 2015 of all patients over 18 were included. PMN was defined as the first prescription not being dispensed before it expired (60 days). The rate of PMN was calculated. Results: Estonian ePrescription System enabled fast evaluation of PMN of osteoporosis patients based on data about prescribing, dispensing and time intervals in-between. Of patients who started osteoporosis treatment 13.1% were primary non-adherent. Of primary non-adherent patients 42% still started treatment at some point during the study. Of patients who did purchase their first prescription 80.4% did so within a week and 95% within 25 days. Conclusion: The Estonian ePrescription system is a useful tool for monitoring PMN. The PMN of osteoporosis medicines was identified as lower than previously reported. More similar type of studies about other groups of medicines would be needed to understand the pattern of PMN and give valuable information to healthcare specialists about how to increase initiation of treatment.

1. Introduction Adherence to medication is defined as the process by which patients take their medicines as prescribed.1 Medication adherence is crucial for achieving the expected clinical effect hoped for when a drug is prescribed.2 In the real world adherence of majority of patients is suboptimal,3 thus countries might not be spending their limited healthcare resources on as cost-effective interventions as they think they do.4 Improving adherence helps to gain maximum effectiveness from medicines and refrain from the outcomes of medical conditions we are trying to prevent.5

∗

Adherence can be divided into primary and secondary adherence.6 Primary medication non-adherence (PMN) occurs when a new medication is prescribed for a patient, but the patient does not obtain the medication within an acceptable period of time after it was prescribed. Secondary non-adherence measures prescription refills among patients who previously filled their first prescriptions.7 Aspects of non-adherence that occur after first dispensing of a medicine have been studied thoroughly but in contrast pre-initiation non-adherence has not. The reasons for this are primarily resource related, as most research is done based on claims databases that capture post initiation medication adherence rather well but lack information about whether the

Corresponding author. Nooruse 1, 50411 Tartu, Estonia. E-mail address: [email protected] (O. Laius).

http://dx.doi.org/10.1016/j.sapharm.2017.10.003 Received 2 August 2017; Received in revised form 4 October 2017; Accepted 4 October 2017 1551-7411/ © 2017 Elsevier Inc. All rights reserved.

Please cite this article as: Laius, O., Research in Social and Administrative Pharmacy (2017), http://dx.doi.org/10.1016/j.sapharm.2017.10.003

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

prescribing of medicines took place or not.8 ePrescribing in general may have different benefits: economic benefits, health benefits and social benefits.9 In addition, the increasing use of electronic prescribing systems10 is enhancing the possibilities to study primary medication non-adherence as both - the act of prescribing and dispensing are saved at the same database or are easily collated from different databases.11 Electronic prescribing was implemented in Estonia in the beginning of 2010 and it is a part of the Estonian eHealth system12 that combines Electronic Health Record, Digital Registration, Digital Image, and Digital Prescription projects and is the central storage for all various medical records in Estonia.13 Estonian ePrescribing system is one of the most comprehensive in Europe.14 The system is centrally maintained by the Estonian Health Insurance Fund (EHIF) that runs the Prescription Centre. Both - prescribing and dispensing of ambulatory medicines are carried out using the Prescription Centre. Physicians and pharmacists use third party software for their interfaces, but these systems are all connected to the central Prescription Centre. Patients can choose any pharmacy in Estonia from where they want the medicine to be dispensed. At the moment over 99% of ambulatory medicines in Estonia are prescribed digitally and 100% are dispensed using the Prescription Centre as paper prescriptions are inserted in the Prescription Centre in the pharmacy while dispensing.12 For every prescription issued using the ePrescription system the date of prescribing, prescription number and type, patient's identifier, age, sex and diagnosis code, medicines active substance(s), strength(s), dosage form, instructions for administration and rate of reimbursement and the doctors name and speciality are saved in the database. If a digital prescription or a paper prescription is dispensed the date of dispensing, package details and number of packages delivered, the name of the pharmacist and pharmacy, amount paid by EHIF, the patient and in total and comments by the pharmacist if there are any are added to the data of the prescription (Table 1). The rate of primary medication non-adherence varies substantially between drug classes15 with osteoporosis medicines showing one of the highest rates.16 Though the treatment of osteoporosis has to last for several years it does not cause pain or other symptoms until a fracture occurs. Thus patients tend to underestimate its severity which leads to lower adherence to treatment.17 Poor adherence to these medicines could result in a disabling fracture of the hip or other major site.18 Because of these reasons osteoporosis medicines were selected to serve as an example of the possible use of Estonian ePrescription system to assess PMN. The aim of this study was to introduce the possibilities of the Estonian ePrescriptions database. The more specific objectives were firstly to evaluate PMN with osteoporosis medicines and the time interval between prescribing and dispensing of these medicines and secondly building from the undertaken analysis assess database usability

for adherence research. 2. Methods 2.1. Setting and study cohort Data was extracted from the EHIF Prescription Centre about all prescriptions issued for osteoporosis medicines in Estonia from 2012 to 2015 with information linked to each patient if one had had a prescription of osteoporosis medication in the prior year or not. Dispensing data of medicines was extracted for the same time period with an additional 60 days after the end of the year so all prescriptions were followed up until dispensing or expiration. Data on other than the initial prescription were obtained in order to assess whether patients purchase subsequent prescriptions after failing to purchase the first. For every prescription date of prescribing and dispensing; patients' (age and gender) and doctors' (speciality) data; the active substance prescribed, the diagnosis code, the medicine dispensed and the amount paid by EHIF and the patient were extracted from the Prescription Centre. Start of treatment was defined as not having a prescription of osteoporosis medicines at least 1 year before. Patients older than 18 years of age were analysed. EHIF uses encrypted patient identifiers for data extraction, meaning a patient can be linked to all of ones prescriptions but the patients actual identity is not revealed therefore ethics committee approval was not needed to conduct this study. 2.2. Osteoporosis medicines Osteoporosis medicines were defined as belonging to ATC group M05B - drugs affecting bone structure and mineralization. Parathyroid hormones and selective estrogen receptor modulators (SERMs) that could also be used to treat osteoporosis do not belong to the same ATC group but they are not used in Estonia.19 2.3. Primary medication non-adherence (PMN) and time to dispensing PMN was defined as the first prescription not being dispensed within 60 days after prescribing as in Estonia prescriptions are valid for 60 days. This means that if a patients' first osteoporosis medicine prescription expired without being dispensed one was considered primary non-adherent. Patients who died less than 60 days after their first prescription were excluded from the study. In addition to PMN the time from prescribing to dispensing was calculated. If a patient failed to purchase the first prescription and was considered primary non-adherent they were still followed up until the end of the study to analyse whether they started treatment later.

Table 1 Data in the Estonian Prescription Centre about every prescription. Data inserted when prescribed

Data added when dispensed

Date of prescribing Prescription number and type

Date of dispensing Details of package dispensed (which preparation, number of tablets etc.) Number of packages dispensed Name of the pharmacist and pharmacy Amount paid by EHIF, the patient and in total Comments by the pharmacist if there are any

Patient's identifier Age and sex of the patient Diagnosis code Active substance(s)

2.4. Data analysis Statistical analyses were done using MS Excel and Stata v12 (StataCorp LP). Comparison of primary adherent/non-adherent patients by background variables was done by multiple logistic regression. Results are presented by adjusted odd ratio (OR) and 95% confidence interval. P-values less than 0.05 were considered statistically significant. 3. Results 3.1. Patients

Dosage form Instructions for administration Rate of reimbursement Prescribing doctors name and speciality

During this study 172,011 prescriptions of osteoporosis medicines to 15,629 patients were issued in Estonia. Out of all patients 7124 had a prior prescription of osteoporosis medicines and 101 patients were 2

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

Fig. 1. Identification of study cohort.

primary adherent than younger patients with the odds increasing 7% (95% CI 3–10, p-value < 0.001) with every 5 years of age.

excluded from the study because of age or death within 60 days after first prescription. This left 8404 patients who were prescribed for the first time and were included in the study (Fig. 1). Majority of the patients starting treatment were women (93.2% were women and 6.8% were men).

3.4. Time to dispensing Out of the 7304 primary-adherent patients who were dispensed their first prescription 3340 (45.7%) purchased it on the same day when being prescribed. 5873 (80.4%) purchased it within a week, 89.0% within two weeks and 95.0% of patients who were going to purchase their initial prescription did it within 25 days from being prescribed (Fig. 2). The proportion of patients who were prescribed with diagnosis code M80 and purchased the medicine on the day of prescribing was 47.5%, with diagnosis code M81 the proportion was 42.7% and with other diagnosis codes 43.1%. Within two weeks 90.9% of prescriptions with diagnosis codes M80 were dispensed, while 88.1% of prescriptions with diagnosis code M81 and 89.8% with other diagnosis codes were (Fig. 3A). 39.3% of primary adherent patients under 51 purchased medicine on the same day when they were prescribed, 43.7% of patients aged 51–60, 46.9% of patients aged 61–70, 45.7% of patients aged 71–80 and 46.8% of patients over 80 years old did so. Within two weeks 87.9% of primary adherent patients under 51, 85.7% of patients aged 51–60, 90.2% of patients aged 61–70, 91.2% of patients aged 71–80 and 90.8% of patients over 80 were dispensed their medicine (Fig. 3B).

3.2. Primary medication non-adherence Out of the 8404 patients 1100 (13.1%) were not dispensed their first prescription and were considered primary non-adherent. Eighty two patients had more than one prescription issued which they did not purchase. Three patients had 5 prescriptions issued to them but none were dispensed. Another 4 patients had 4 prescriptions annulled without being dispensed and 18 patients had 3 such prescriptions. The number of patients who were prescribed osteoporosis medicine but never started treatment was 638 patients (7.6% of the total number of patients and 68% of PMN patients). The number of patients who did not purchase the first prescription but still started treatment at some point during the study was 462 (5.5% of the total number of patients and 42.0% of primary non-adherent patients). The median time between first prescribing and actual dispensing for the patients who failed to purchase their first prescription but still initiated treatment during the study was 92 days. 3.3. Patient characteristics associated with primary medication nonadherence

4. Discussion

The characteristics of primary adherent and primary non-adherent patients are presented in Table 2. There was no difference in PMN according to gender. Patients with a prior fracture (ICD-10 code M80) were more likely to be primary adherent than patients with osteoporosis but without a fracture (M81) and with other diagnosis'. Prescriptions issued by orthopaedists or rheumatologists were more frequently dispensed than prescriptions issued by general practitioners. Differences with other specialities compared to the GPs were statistically non-significant. Prescriptions of strontium ranelate and denosumab were less likely to be dispensed compared to prescriptions of plain alendronic acid. There were no differences comparing oral bisphosphonates or combinations of oral bisphosphonates with vitamin D and/or calcium to one another. Older patients were more likely to be

There are several methods in use to assess medication adherence. These methods can be divided into direct methods (e.g. measuring drug concentration in blood) or indirect methods (e.g. pill counts or database research).1 Direct methods require a doctor's visit and blood or other body fluid to be collected and are therefore costly and unpractical for adherence assessment in every day practice.20 The indirect methods of adherence assessment can be self-reports, pill counts, prescription and dispensing databases and electronic monitoring systems.1 Self-report measures have been shown not to be sufficiently precise and to be in general unreliable in comparison with other more objective tools.21 Pill counts tend to overestimate the number of doses actually taken as patients can easily censor the number of tablets left in stock.3 A medication event monitoring system (MEMS) is a medicines smart package 3

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

Table 2 Characteristics of primary adherent and primary non-adherent patients.

Gender Male Female Diagnosis M80 M81 Other Doctor speciality Orthopedist GP Rheumatologist Other Active substance alendronic acid ibandronic acid risedronic acid zoledronic acid alendronic acid and colecalciferol risedronic acid, calcium and cole-calciferol, sequential eptotermin alfa strontium ranelate denosumab Age mean (sd)

Primary adherent (n = 7304)

%

Primary non-adherent (n = 1100)

%

Adjusted ORa

95% CI

p-value

488 6816

85.8 87.0

81 1019

14.2 13.0

1 (ref) 0.97

0.75

1.25

0.813

4527 2395 382

88.9 85.2 76.6

568 415 117

11.2 14.8 23.5

1 (ref) 0.78 0.46

0.68 0.36

0.90 0.59

0.001 < 0.001

1000 1768 3500 1036

88.4 83.9 88.5 85.6

132 340 453 175

11.7 16.1 11.5 14.5

1.69 1 (ref) 1.47 1.19

1.35

2.12

< 0.001

1.25 0.97

1.71 1.46

< 0.001 0.093

990 835 324 32 4676 0

85.0 85.8 86.6 72.7 88.2 0

175 138 50 12 625 1

15.0 14.2 13.4 27.3 11.8 100.0

1 (ref) 1.04 1.08 0.58 1.13 –

0.81 0.77 0.29 0.93

1.33 1.52 1.18 1.36

0.751 0.663 0.133 0.215

0 177 270

0 80.5 83.1

1 43 55

100.0 19.6 16.92

– 0.67 0.71

0.46 0.50

0.98 1.00

0.037 0.05

1.07b

1.03

1.10

< 0.001

70.3 (10.3)

68.4 (12.2)

OR - odds ratio. GP - General Practitioner. sd - standard deviation. M80 - osteoporosis with pathological fracture. M81 - osteoporosis without pathological fracture. a (ref) is the subgroup whose rate of adherence is used as the reference and other subgroups are compared against this subgroup. Adjusted OR shows the odds of this patient group's adherence being the same as that of the reference group. Adjusted OR below 1 means worse adherence and over 1 means better primary adherence. b Odds ratio presented to every 5-year change in age.

day practice.22 Prescription or dispensing databases can be considered the gold standard to measure adherence in the community setting. They have proved to be a valid proxy to establish patients' medication

that records the time and date of every opening of the package. Adherence data resulting from such packages are reliable and detailed but the MEMS are still too expensive and labour intensive to use in every

Fig. 2. The number of days between prescribing and dispensing of first prescriptions to patients who started osteoporosis treatment in Estonia in 2012–2015.

4

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

Fig. 3. The differences in the number of days between prescribing and dispensing of first prescriptions of patients who started osteoporosis treatment in Estonia in 2012–2015 during the first 14 days according to diagnosis codes (A) or patients' age (B).

adherence.23 While contained in separate databases both prescribing and dispensing data have their problems though. Initiation of treatment cannot be assessed using only prescribing data as it is unknown whether the patient is also dispensed the medicine. Also initiation is hard to interpret using dispensing database as it is unknown whether medicine was purchased with the initial prescription or not.1 Estonian ePrescribing system neutralizes these shortcomings as the act of prescribing and the act of dispensing are saved in the same database and can be used to assess whether a prescribed prescription was also dispensed and the time interval in-between the two acts. Experience with osteoporosis medicines from the current study suggest that PMN can easily be assessed using the Estonian ePrescribing

system as there is a note on every prescription whether and when it was dispensed. The authors are confident the system can also be used for other drug classes as the system contains data about all prescriptions issued in Estonia regardless of healthcare facility, doctor's speciality, diagnosis etc. The definition of a new user should be reassessed though depending on the diagnosis as for some diseases medicines from different ATC groups can be used. It would also be possible to distinguish new users using the diagnosis code if it proves to be more robust than ATC groups. Also the definition of adherence has to be revisited as with some other classes of medicines combination therapy and dose titration has to be taken into account while deciding on patients' adherence to medicines. 5

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

colleagues.29 One possible explanation could be that in Estonia patients with fragility fractures get medicines reimbursed at the rate of 75% if they are under 63 years old and at 90% if they are over 63. The price of medicines has been shown to influence treatment initiation34 but unfortunately the data did not allow to analyse reimbursement rate as a variable for primary non-adherence in this study. Denosumab has been suggested to have better secondary adherence than oral bisphosphonates as it is administered once over six months35 but current study showed that PMN is lower for denosumab than for alendronic acid, meaning special attention should be given to the initiation of therapy with denosumab. One of the possible explanations is that denosumab is more expensive for the patient than oral bisphosphonates suggesting the patient's willingness to pay should be addressed before prescribing denosumab. Other factors that influenced PMN were the diagnosis code and the prescribing doctor‘s speciality. Patients with a fracture were unsurprisingly more primary adherent and also patients with treatment started by orthopaedists or rheumatologists were more primary adherent but as patients with fractures tend to visit specialists rather than their GP these results are influenced by one another. The time from prescribing to dispensing was a little shorter in this study than reported before for osteoporosis medicines29 and more similar to results shown with other drug classes.36 80% of patients with their first prescription are dispensed within a week. The prescribing doctor and the dispensing pharmacist see which prescriptions are still active for the patient in the Estonian Prescription Centre during doctor's appointment or while dispensing other medicines at the pharmacy. The study results suggest that if they see prescriptions older than a week a discussion with the patient on whether he or she is adhering to therapy might be in order. For example it has been shown that proactive pharmacy based interventions to improve patients adherence can be effective.37 The differences in time to dispensing were not very substantial according to patients' diagnosis code or age. Almost half (47.5%) of patients with diagnosis code M80 purchased medicine on the very first day of being prescribed which is 4.8% more than with M81 and 4.4.% more than with all other diagnosis codes. The differences levelled though within two weeks with 90.9% of prescriptions purchased with diagnosis code M80, 88.1% with M81 and 89.8% with other diagnosis codes. This indicates that different approach to patients to improve their PMN according to their diagnosis code is not warranted. Patients' age seems to have a bit larger impact on the time to dispensing as 46.5% of patients over 61 were dispensed on the same day and 41.5% of patients under 61 were done so. Within two weeks 90.7% of older patients were dispensed and 86.8% of patients under 61 were. But still the differences are not critical and do not give reason to develop different strategies in dealing with patients of different age what comes to time to dispensing of medicines. As the majority of patients are dispensed within two weeks there seem to be no need for interventions to quicken dispensing of osteoporosis medicines as these medicines are not as time critical as some others, say antibiotics, and two weeks could be considered acceptable for the start of treatment. The effectiveness of other interventions to improve adherence is debatable with for instance telephone based interventions showing some efficacy in some studies27,28,38 and failing in others.39,40 Recently it has been recommended that screening for bone turnover markers might help identify poor adherence and serve as a starting point for intervention.41 Monitoring ePrescription databases to identify poor adherence is probably less resource demanding if such databases exist and could be used instead. There seem to be an agreement that the interventions implemented should be multidisciplinary and tackle different aspects and reasons for non-adherence.42 Establishing the rate of primary medication non-adherence gives another piece of information to base those interventions on. The foremost limitation of the current study is that patients could not be differentiated according to their socio-demographic characteristics. Only their age and gender are linked to the prescription data, thus some possible explaining factors on patients' primary non-

The Estonian ePrescribing system could also be used for other types of qualitative drug utilization research. The ideal database for qualitative research has been described as including information on all patients who received a specific drug in a specific time frame.24 As the Estonian system contains data on all prescriptions issued in ambulatory care it comes pretty close to this definition for drugs that are used for chronic diseases. The database does not include in-hospital use but the medicines used in secondary care are in large for acute conditions. These other qualitative studies could include prescribing indicator studies, compliance with additional risk minimization measures of drugs, interactions research etc. Because all prescriptions of one patient can be linked to one another then the prescription and dispensing pattern of every patient can be assessed in full. The proportion of patients that never purchase their first prescription has been shown to be between 2.4 and 30.7% depending on the definition of primary non-adherence and the drug class studied.25 The time period of 60 days was used in this study in which the patient should be dispensed after the initial prescription to be considered primary adherent. The 60 days suited well in the Estonian setting because the initial prescription is valid for 60 days before being annulled in the EHIF Prescription Centre. Of the patients who were prescribed osteoporosis medication for the first time in 2012–2015 13.1% did not purchase their medication from the pharmacy before the prescription expired. Comparing with prior research across all drug classes it is an average result.26 It is substantially better than suggested before for osteoporosis medication though as antiosteoporotic medicines have been shown to have one of the highest PMN rates amongst different drug classes with results ranging from 22.4% to 37.0%.16,27–29 The most recent study by Reynolds and colleagues found the PMN of bisphosphonates to be almost 30%.29 The number of patients in the current study was similar and Reynolds and colleagues also used a 60-day window to establish primary non-adherence. Though they included women over 55 in the study and this study included all patients over the age of 18 the number of men in this study was small (6.8%) because of the nature of the disease and the number of patients under 50 was also negligible with 3.1%, suggesting the different patient selection criteria are not likely to be the reason for differences seen. One of the possible explanations could be that in Estonia after a digital prescription is issued the patient can purchase the medicine from any pharmacy of choice. With some other digital prescribing systems the patient can use some or name certain pharmacies from where the dispensing can be done.14 Also in Reynolds' study patients had to use certain pharmacies to purchase medicines and if they used other ones they were misclassified as being primary non-adherent.29 There is a pharmacy in most healthcare centres in Estonia meaning it is very convenient for the patient to purchase medicines. Pharmacies being more accessible could improve primary adherence. The fact that a patient does not purchase ones first prescription does not necessarily mean that the patient will not initiate treatment later on. In this study it is shown that 462 (42%) patients out of the 1100 who were not dispensed their first prescription initiated treatment later on and half of those patients did so within 3 months after the very first prescription. This result indicates that concentrating only on the fact of whether the first prescription is dispensed or not overestimates the proportion of patients who do not start treatment. Yet again the fact that a patient purchased the drug with the first prescription does not mean one will carry out the whole treatment course necessary for it to be effective. The optimal duration of treatment with osteoporosis medicines is at least 3 years30,31 but the majority of patients do not persist with treatment for that long.32,33 In order to have a full understanding of patients' behaviour towards medication adherence the full treatment cycle should be studied from the initial prescribing until when the patient stops treatment. Patients' age was identified as a characteristic that influences primary adherence with older patients having higher odds to be primary adherent. This result is the opposite to what was found by Reynolds and 6

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

1097/MLR.0b013e31829b1d2a. 7. Adams AJ, Stolpe SF. Defining and measuring primary medication nonadherence: development of a quality measure. J Manag Care Spec Pharm. 2016;22(5):516–523. http://dx.doi.org/10.18553/jmcp.2016.22.5.516. 8. Hutchins DS, Zeber JE, Roberts CS, Williams AF, Manias E, Peterson AM. Initial medication adherence—review and recommendations for good practices in outcomes research: an ISPOR medication adherence and persistence special interest group report. Value Health. 2015;18(5):690–699. http://dx.doi.org/10.1016/j.jval.2015.02. 015. 9. Deetjen U. European E-Prescriptions: Benefits and Success Factors. 2016; 2016https:// ora.ox.ac.uk/objects/uuid:440a8fe6-6421-4b62-9e5e-cb0f559667d6, Accessed date: 25 July 2017. 10. Stroetmann KA, Artmann J, Stroetmann V. Developing national eHealth infrastructures–results and lessons from Europe. AMIA Annual Symposium Proceedings. vol. 2011. American Medical Informatics Association; 2011:1347http:// pubmedcentralcanada.ca/pmcc/articles/PMC3243126/, Accessed date: 18 April 2017. 11. Mabotuwana T, Warren J, Harrison J, Kenealy T. What can primary care prescribing data tell us about individual adherence to long-term medication?-comparison to pharmacy dispensing data. Pharmacoepidemiol Drug Saf. 2009;18(10):956–964. http://dx.doi.org/10.1002/pds.1803. 12. Estonian Health Insurance Fund. http://www.haigekassa.ee. Accessed 10 April 2017. 13. Estonian eHealth Foundation Website. http://www.e-tervis.ee/index.php/en/. Accessed 25 May 2017. 14. Kierkegaard P. E-Prescription across Europe. Health Technol. 2013;3(3):205–219. http://dx.doi.org/10.1007/s12553-012-0037-0. 15. Fischer MA, Stedman MR, Lii J, et al. Primary medication non-adherence: analysis of 195,930 electronic prescriptions. J Gen Intern Med. 2010;25(4):284–290. http://dx. doi.org/10.1007/s11606-010-1253-9. 16. Shin J, McCombs JS, Sanchez RJ, Udall M, Deminski MC, Cheetham TC. Primary nonadherence to medications in an integrated healthcare setting. Am J Manag Care. 2012;18(8):426–434. 17. Bianchi ML, Duca P, Vai S, et al. Improving adherence to and persistence with oral therapy of osteoporosis. Osteoporos Int. 2015;26(5):1629–1638. http://dx.doi.org/ 10.1007/s00198-015-3038-9. 18. Rietbrock S, Olson M, van Staa TP. The potential effects on fracture outcomes of improvements in persistence and compliance with bisphosphonates. QJM. 2009;102(1):35–42. http://dx.doi.org/10.1093/qjmed/hcn130. 19. Laius O, Maasalu K, Kõks S, Märtson A. Use of drugs against osteoporosis in the Baltic countries during 2010–2014. Medicina (Mex). 2016;52(5):315–320. http://dx.doi. org/10.1016/j.medici.2016.10.001. 20. Strauch B, Petrák O, Zelinka T, et al. Precise assessment of noncompliance with the antihypertensive therapy in patients with resistant hypertension using toxicological serum analysis. J Hypertens. 2013;31(12):2455–2461. http://dx.doi.org/10.1097/ HJH.0b013e3283652c61. 21. Ramsey RR, Ryan JL, Hershey AD, Powers SW, Aylward BS, Hommel KA. Treatment adherence in patients with headache: a systematic review. Headache. 2014;54(5):795–816. http://dx.doi.org/10.1111/head.12353. 22. Vrijens B, Urquhart J, White D. Electronically monitored dosing histories can be used to develop a medication-taking habit and manage patient adherence. Expert Rev Clin Pharmacol. 2014;7(5):633–644. http://dx.doi.org/10.1586/17512433.2014.940896. 23. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. J Clin Epidemiol. 1997;50(1):105–116. 24. Strom BL, Kimmel SE, Hennessy S, eds. Pharmacoepidemiology. 5. ed. Chichester: Wiley-Blackwell; 2012. 25. Thengilsdóttir G, Pottegård A, Linnet K, Halldórsson M, Almarsdóttir AB, Gardarsdóttir H. Do patients initiate therapy? Primary non-adherence to statins and antidepressants in Iceland. Int J Clin Pract. 2015;69(5):597–603. http://dx.doi.org/ 10.1111/ijcp.12558. 26. Linnet K, Halldorsson M, Thengilsdottir G, Einarsson OB, Jonsson K, Almarsdottir AB. Primary non-adherence to prescribed medication in general practice: lack of influence of moderate increases in patient copayment. Fam Pract. 2013;30(1):69–75. http://dx.doi.org/10.1093/fampra/cms049. 27. Cook PF, Emiliozzi S, McCabe MM. Telephone counseling to improve osteoporosis treatment adherence: an effectiveness study in community practice settings. Am J Med Qual. 2007;22(6):445–456. http://dx.doi.org/10.1177/1062860607307990. 28. Waalen J, Bruning AL, Peters MJ, Blau EM. A telephone-based intervention for increasing the use of osteoporosis medication: a randomized controlled trial. Am J Manag Care. 2009;15(8):e60–e70. 29. Reynolds K, Muntner P, Cheetham TC, et al. Primary non-adherence to bisphosphonates in an integrated healthcare setting. Osteoporos Int. 2013;24(9):2509–2517. http://dx.doi.org/10.1007/s00198-013-2326-5. 30. Karpf DB, Shapiro DR, Seeman E, et al. Prevention of nonvertebral fractures by alendronate: a meta-analysis. JAMA. 1997;277(14):1159–1164. http://dx.doi.org/ 10.1001/jama.1997.03540380073035. 31. Stakkestad JA, Lakatos P, Lorenc R, Sedarati F, Neate C, Reginster J-Y. Monthly oral ibandronate is effective and well tolerated after 3 years: the MOBILE long-term extension. Clin Rheumatol. 2008;27(8):955–960. http://dx.doi.org/10.1007/s10067007-0824-6. 32. Netelenbos JC, Geusens PP, Ypma G, Buijs SJE. Adherence and profile of non-persistence in patients treated for osteoporosis—a large-scale, long-term retrospective study in The Netherlands. Osteoporos Int. 2011;22:1537–1546. http://dx.doi.org/10. 1007/s00198-010-1372-5. 33. Laius O, Pisarev H, Maasalu K, Kõks S, Märtson A. Adherence to osteoporosis medicines in Estonia-a comprehensive 15-year retrospective prescriptions database study. Arch Osteoporos. 2017;12(1):59. http://dx.doi.org/10.1007/s11657-017-

adherence could not be identified. Also osteoporosis medicines use was not correlated for technical reasons with the use of other prescription medicines by the patient that could have an impact on patients overall attitude towards medication. The main strength of the study is that the Estonian ePrescription system has complete coverage of prescriptions of medicines. The prescriptions and dispensings are saved in the same database so there are no correlation issues. The number of patients and prescriptions included in the study was rather substantial compared to typical PMN studies and all adult patients were included to the study giving a good range of patients, although because of the nature of the disease, unsurprisingly elderly women prevailed as patients. 5. Conclusion The ePrescription System of Estonia is a useful tool for monitoring patient behaviour with regard to PMN. Although osteoporosis medicines were used as an example in this study the system could be used for other drug classes and qualitative research purposes as well. The PMN rate discovered in the current study was considerably smaller than reported before and almost half of the patients still started treatment after failing to purchase their initial prescription. Majority of the patients were dispensed within the week, which means that when active prescriptions are identified in the Prescription Centre by healthcare professionals that are older than a week it should be discussed with the patient to identify potential adherence issues. Offering further balanced information about the importance of taking ones medicines and their possible benefits and risks could improve adherence. Similar type of studies about other groups of medicines in the future would help to understand the pattern of PMN and provide additional information to healthcare specialists. Contributions Ott Laius, Heti Pisarev, Daisy Volmer, Katre Maasalu, Sulev Kõks and Aare Märtson declare that all authors read and approved the manuscript. Funding The research leading to these results has received institutional research funding (IUT 20–46) of the Estonian Ministry of Education and Research and the European Union Seventh Framework Programme FP7/2007–2013 under grant agreement No. 602398 (collaborative project HypOrth). Conflict of interests Ott Laius, Heti Pisarev, Daisy Volmer, Sulev Kõks, Aare Märtson and Katre Maasalu declare that they have no conflict of interest. References 1. Elseviers M, ed. Drug Utilization Research: Methods and Applications. Chichester, West Sussex: Hoboken, NJ: John Wiley & Sons Inc; 2016. 2. Sabaté E. Adherence to Long-Term Therapies: Evidence for Action. Switzerland: World Health Organization; 2003. 3. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353(5):487–497. http://dx.doi.org/10.1056/NEJMra050100. 4. Hiligsmann M, Rabenda V, Gathon H-J, Ethgen O, Reginster J-Y. Potential clinical and economic impact of nonadherence with osteoporosis medications. Calcif Tissue Int. 2010;86:202–210. http://dx.doi.org/10.1007/s00223-009-9329-4. 5. Imaz I, Zegarra P, González-Enríquez J, Rubio B, Alcazar R, Amate JM. Poor bisphosphonate adherence for treatment of osteoporosis increases fracture risk: systematic review and meta-analysis. Osteoporos Int. 2010;21(11):1943–1951. http://dx. doi.org/10.1007/s00198-009-1134-4. 6. Raebel MA, Schmittdiel J, Karter AJ, Konieczny JL, Steiner JF. Standardizing terminology and definitions of medication adherence and persistence in research employing electronic databases. Med Care. 2013;51(8 0 3):S11. http://dx.doi.org/10.

7

Research in Social and Administrative Pharmacy xxx (xxxx) xxx–xxx

O. Laius et al.

10.1007/s00198-015-3116-z. 39. Fischer MA, Jones JB, Wright E, et al. A randomized telephone intervention trial to reduce primary medication nonadherence. J Manag Care Spec Pharm. 2015;21(2):124–131. http://dx.doi.org/10.18553/jmcp.2015.21.2.124. 40. Solomon DH. Osteoporosis telephonic intervention to improve medication regimen adherence: a large, pragmatic, randomized controlled trial. Arch Intern Med. 2012;172(6):477. http://dx.doi.org/10.1001/archinternmed.2011.1977. 41. Adherence Working Group of the International Osteoporosis Foundation and the European Calcified Tissue Society. Diez-Perez A, Naylor KE, et al. International osteoporosis foundation and European calcified tissue society working group. Recommendations for the screening of adherence to oral bisphosphonates. Osteoporos Int. 2017;28(3):767–774. http://dx.doi.org/10.1007/s00198-017-3906-6. 42. Balkrishnan R. The Importance of Medication Adherence in Improving Chronic-Disease Related Outcomes: What We Know and What We Need to Further Know. LWW; 2005http://journals.lww.com/lww-medicalcare/Citation/2005/06000/The_ Importance_of_Medication_Adherence_in.1.aspx, Accessed date: 18 April 2017.

0354-z. 34. Modi A, Yu J, Brenneman S, Sazonov V. Reasons for not initiating osteoporosis therapy among a managed care population. Patient Prefer Adherence. June 2015:821. http://dx.doi.org/10.2147/PPA.S81963. 35. Inderjeeth CA, Inderjeeth A-J, Raymond WD, others. Medication selection and patient compliance in the clinical management of osteoporosis. Aust Fam Physician. 2016;45(11):814. 36. Pottegård A, Christensen R dePont, Houji A, et al. Primary non-adherence in general practice: a Danish register study. Eur J Clin Pharmacol. 2014;70(6):757–763. http:// dx.doi.org/10.1007/s00228-014-1677-y. 37. Stuurman-Bieze AGG, Hiddink EG, van Boven JFM, Vegter S. Proactive pharmaceutical care interventions decrease patients' nonadherence to osteoporosis medication. Osteoporos Int. 2014;25:1807–1812. http://dx.doi.org/10.1007/s00198-0142659-8. 38. Cizmic AD, Heilmann RMF, Milchak JL, Riggs CS, Billups SJ. Impact of interactive voice response technology on primary adherence to bisphosphonate therapy: a randomized controlled trial. Osteoporos Int. 2015;26(8):2131–2136. http://dx.doi.org/

8