Diabetes medication persistence, different medications have different persistence rates

ARTICLE IN PRESS

PCD-596; No. of Pages 5

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 7 ) xxx–xxx

Contents lists available at ScienceDirect

Primary Care Diabetes journal homepage: http://www.elsevier.com/locate/pcd

Original research

Diabetes medication persistence, different medications have different persistence rates Michal Shani a,b,∗ , Alex Lustman a,b , Shlomo Vinker b a b

Department of Family Medicine Central District, Clalit Health Service, Mazkeret Batya, Israel Department of Family Medicine Sackler Faculty of Medicine, Tel Aviv University, Israel

a r t i c l e

i n f o

a b s t r a c t

Article history:

Aim: To assess the persistence of diabetic patients to oral medications.

Received 13 May 2016

Methods: The study included all type 2 diabetic patients over 40 years, members of one

Received in revised form

District of Clalit Health Services Israel, who were diagnosed with diabetes mellitus before

1 March 2017

2008 and who filled at least one prescription per year during 2008–2010, for the following

Accepted 22 March 2017

medications: metformin, glibenclamide, acarbose, statins, angiotensin converting enzyme

Available online xxx

inhibitors (ACEI) and angiotensin II receptor antagonists (ARBs). Purchase of at least 9 monthly prescriptions during 2009 was considered “good medication persistence”. We com-

Keywords:

pared HbA1c and LDL levels, according to medication persistence, for each medication; and

Medication persistence

cross persistence rates between medications.

Diabetes mellitus

Results: 21,357 patients were included. Average age was 67.0 ± 11.0 years, 48.9% were men,

Primary care

and 35.8% were from low SES. Good medication persistence rates for ARBs were 78.8%,

Family medicine

ACEI 69.0%, statins 66.6%, acarbose 67.8%, metformin 58.6%, and glibenclamide 55.3%. Good persistence to any of the medications tested was associated with a higher rate of good persistence to other medications. Patients who took more medications had better persistence rates. Conclusions: Different oral medications used by diabetic patients have different persistence rates. Good persistence for any one medication is an indicator of good persistence to other medications. Investment in enhancing medication persistence in persons with diabetes may improve persistence to other medications, as well as improve glycemic control. © 2017 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

Poor adherence with medication and the associated adverse outcomes and high costs of care are of growing concern to clinicians and healthcare systems [1]. A study of the general

population in Germany found that at least 33% of patients repeatedly fail to follow their doctors’ recommendations and only 25% described themselves as fully adherent [2]. In the management of diabetes mellitus, medication persistence is of the utmost importance. Yet, a systematic review

∗

Corresponding author at: 101 Arlozorov st. Tel Aviv, Israel. Fax: +972 3 7604838. E-mail address: [email protected] (M. Shani). http://dx.doi.org/10.1016/j.pcd.2017.03.006 1751-9918/© 2017 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: M. Shani, et al., Diabetes medication persistence, different medications have different persistence rates, Prim. Care Diab. (2017), http://dx.doi.org/10.1016/j.pcd.2017.03.006

PCD-596; No. of Pages 5

ARTICLE IN PRESS

2

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 7 ) xxx–xxx

confirmed that many patients with diabetes utilize less than the prescribed amount of oral anti-glycemic drugs, with adherence rates ranging from 36–93% [3]. For persons with diabetes, better adherence has been associated with better patient outcomes [4,5] reduced hospitalizations, and reduced health care costs [6,7]. Every 25% increase in medication adherence was found to be associated with a 0.34% reduction in glycosylated hemoglobin (HbA1c) [8]. Medication adherence refers to the degree or extent of conformity to the recommendations about day-to-day treatment by the provider with respect to the timing, dosage, and frequency. Medication persistence on the other hand refers to the act of continuing the treatment for the prescribed duration [9]. Israel has a universal health system with good and accessible primary care for all patients and generous coverage. We investigated persistence with oral medications commonly used by diabetic patients, in a cohort of diabetic patients in our district.

2.

Methods

The study was conducted in the Central District of “Clalit Health Service” (CHS) in Israel, and was approved by the local ethics committee (approval number k118/2010). In Israel, since 1995, every citizen and permanent resident receives health care provided one of four health maintenance organizations (HMO). CHS is the largest HMO in Israel, serving 54% of the population. Patient medical records in CHS have been completely computerized for over a decade and an extensive healthcare database has been created. The demographic data is updated directly from the population registry of the Ministry of Interior. All laboratory tests are free of charge and sent to a central laboratory. The results are recorded automatically in patients’ electronic medical files and reported directly to primary care physicians. All community pharmacies used by CHS are computerized and report to one central repository. CHS issues medications and requires nominal co-payments. Patients can buy chronic medications for up to 3 months at a time. This system ensures that all prescriptions are documented.

2.1.

Study population

All persons over age 40 years who were diagnosed with type 2 diabetes mellitus before 2008 and were treated by the same family physician during 2008–2010 in the Central District of CHS. From this cohort we choose patients that filled at least one prescription per year in the three consecutive years 2008–2010 for the specific study medication. This approach was used to ensure medication use and to exclude patients with changes in treatment for any reason during the study period.

2.2.

Study medications

The following medications were included: metformin, glibenclamide (the most commonly used sulfonylurea in Israel during the study period), acarbose, statins, angiotensin converting enzyme inhibitors (ACEI), and angiotensin II receptor antagonists (ARBs).

Table 1 – Characteristics of the 21,357 study patients. Age (years) mean (SD) (range)

67.0 ± 11.0 (40–100)

Gender (% men) Low socioeconomic status BMI (kg/m2 ) mean (SD) (range) Smoking (current or past) Hypertension Hyperlipidemia Ischemic heart disease s/p CVA HbA1c (%) mean (SD) (range) LDL cholesterol (mg%) mean (SD) (range) Systolic BP (mm Hg) mean (SD) (range) Seen by endocrinologist in the year 2009

48.9% 35.8% 30.2 ± 5.5 (13.4–60.6) 33.4% 76.8% 88.5% 32.5% 12.8% 7.4 ± 1.4 (4.3–17.8) 90.0 ± 28.8 (20–220) 131.7 ± 16.1 (80–240) 14.0%

We analyzed all prescriptions that were filled for the six medications from January 1st 2009 to December 31st 2009 [10]. We considered purchasing of at least 9 monthly prescriptions during 2009 as “good medication persistence”, as compared to lower persistence (purchasing of less than 9 prescriptions during 2009) [11,12]. Demographic information was accessed: age, gender and socio-economic status (SES). Patients with low SES were defined as those exempt from healthcare payments on their income by the national insurance. These patients pay reduced co-payments on chronic medications and these co-payments are a capped at $60 a month. Patients who immigrated to Israel after 1990 were considered as new immigrants. We also extracted data about body mass index (BMI), smoking and other cardiovascular diagnoses at 1st January 2009 and visits to endocrinologist. We included the last LDL-cholesterol (LDL) HbA1c levels, and blood pressure measure that were taken in 2009.

2.3.

Statistical analysis

We calculated persistence rates for each medication separately. For each medication we used logistic regression models to calculate odds ratio and to examine associations between medication persistence and age, gender, SES, immigration status, BMI, chronic diseases, insulin use, endocrinology clinic visits and the number of the investigated medications used by each patient as a proxy to overall medications use. We compared HbA1c and LDL cholesterol levels (separately) between persons with good and lower persistence for each medication, and cross persistence rates between medications, by comparing medication persistence for other medications. STATA 8.0 statistical software (Stata Corp. College Station, TX, USA) was used for statistical analysis.

3.

Results

The study included 21,357 individuals. Table 1 describes their baseline characteristics. Rates of good medication persistence varied widely among the medications investigated, ranging from 55.3% for glibenclamide to 78.8% for ARBs (Fig. 1). Good medication persistence to any of the six medications investigated was associated with a higher rate of good persistence to each other medication (Table 2).

Please cite this article in press as: M. Shani, et al., Diabetes medication persistence, different medications have different persistence rates, Prim. Care Diab. (2017), http://dx.doi.org/10.1016/j.pcd.2017.03.006

ARTICLE IN PRESS

PCD-596; No. of Pages 5

3

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 7 ) xxx–xxx

Table 2 – Cross relationship of persistence between medications. Statin high Statin low Statin ACEI ARBs Metformin Glibenclamid Acarbose

Statin ACEI ARBs Metformin Glibenclamid Acarbose

p-Value <0.001 <0.001 <0.001 <0.001 <0.001

ACEI high

ACEI low

p-Value

ARBs high

ARBs low p-Value

82.1%

36.3%

<0.001

80.0%

40.8%

<0.001

75.7% 71.0% 74.0%

28.7% 30.7% 33.8%

<0.001 <0.001 <0.001

73.5% 68.8% 64.8%

32.9% 36.2% 40.0%

<0.001 <0.001 0.076

85.3% 88.9% 77.1% 72.1% 76.6%

41.9% 58.2% 31.1% 33.6% 38.0%

Metformin high

Metformin p-Value low

Glibenclamide Glibenclamide p-Value high low

Acarbose high

Acarbose p-Value low

83.1% 86.0% 90.5%

39.8% 44.3% 62.8%

<0.001 <0.001 <0.001

80.4% 83.7% 88.6% 79.3%

44.6% 48.2% 66.5% 31.9%

<0.001 <0.001 <0.001 <0.001

75.8% 76.4%

27.6% 38.3%

<0.001 <0.001

85.7% 88.0% 88.5% 82.5% 83.9%

52.9% 56.9% 73.5% 47.3% 50.7%

73.4%

35.3%

<0.001

<0.001 <0.001 0.076 <0.001 <0.001

Table 3 – Comparison of mean LDL cholesterol and HbA1c levels between individuals with good and lower medication persistence. No. of patients

Statin ACEI ARBs Metformin Sulfonylurea Acarbose

16,236 14,647 3152 13,495 5621 382

LDL cholesterol level mean (SD) Good persistence (mg%)

Lower persistence (mg%)

p-Value

Good persistence (%)

Lower persistence (%)

p-Value

81.9 ± 22.5 85.0 ± 25.8 83.1 ± 25.1 83.7 ± 25.5 86.2 ± 25.9 83.7 ± 28.3

97.5 ± 32.1 95.1 ± 31.4 88.6 ± 29.7 94.6 ± 30.8 94.2 ± 31.5 84.8 ± 22.9

<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.72

7.3 ± 1.3 7.3 ± 1.3 7.4 ± 1.3 7.3 ± 1.2 7.6 ± 1.3 7.5 ± 1.4

7.6 ± 1.5 7.7 ± 1.6 7.6 ± 1.4 7.8 ± 1.6 7.9 ± 1.5 7.9 ± 1.4

<0.001 <0.001 0.002 <0.001 <0.001 0.009

78.8% 69.0%

66.6%

67.8% 58.6%

ARBs 3,152

ACEI 14,647

stan 16,236

HbA1c level mean (SD)

acarbose 382

55.3%

meormin glibenclamide 13,495 5,621

Fig. 1 – Proportions of diabetic individuals with good persistence for each of six medications. (The number of individuals using each medication is listed below the name of each medication.)

Mean HbA1c levels were significantly lower among persons with good compared to lower persistence to each of the medications investigated (Table 3). Similarly, good persistence was found to be associated with lower LDL levels, except for persistence to acarbose, which was found not to be associated with LDL level (Table 3). Mean LDL cholesterol was 81.9 ± 22.5 mg% for patients with good persistence to statins compared to 97.5 ± 32.1 mg% for those with lower persistence (p < 0.0001). HbA1c was 7.3 ± 1.2% for patients with good persistence to

metformin compared to 7.8 ± 1.6% for patients with lower persistence (p < 0.0001). Table 4 describes multi-variant analysis for medication persistence. Increased age was associated with increased medication persistence. Gender and low socioeconomic status were not associated with medication persistence. New immigrants had lower medication persistence rates than did non-immigrants. Increased BMI was associated with small increased medication persistence. Higher medications burden was associated with higher medication persistence. In a subgroup analysis of the 3 most commonly used medications (metformin, statins and ACEI) medication persistence increased when all three medications were taken compared to only one or two of them (Table 5).

4.

Discussion

In this study of individuals with diabetes, we found that good medication persistence differs among medications. The highest rate of good persistence was for ARBs: 78.8% compared to 58.6% for metformin, and 55.3% for glibenclamide. The rate of good persistence to metformin was lower than expected, considering the key role of this medication in diabetes management. The higher rate of adherence to metformin that was reported in a randomized controlled Canadian study, 78% [14], highlights the importance of investigating medication persistence in real life situations. We found older age to be associated with better medication persistence, similar to findings of previous studies [15,16,17].

Please cite this article in press as: M. Shani, et al., Diabetes medication persistence, different medications have different persistence rates, Prim. Care Diab. (2017), http://dx.doi.org/10.1016/j.pcd.2017.03.006

PCD-596; No. of Pages 5

ARTICLE IN PRESS

4

p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 7 ) xxx–xxx

Table 4 – Odds ratio (OR) for good persistence to medications according to demographic characteristics.

Statin

ACEI

ARBs

Metformin

Glibenclamide

Acarbose

a b

Age (per year)

Gender (female vs. male)

Low SESa

Immigrantb BMI (per 1 BMI unit)

No of chronic disease

Insulin use

Endo crinologist

No of medication

1.02 (1.02–1.03) p < 0.001 1.02 (1.01–1.02) p < 0.001 1.01 (1.00–1.03) p = 0.04 1.02 (1.02–1.02) p < 0.001 1.02 (1.01–1.03) p < 0.001 1.01 (0.98–1.04) p = NS

0.97 (0.83–1.00) p = NS 0.99 (0.90–1.09) p = NS 1.07 (0.85–1.35) p = NS 1.08 (0.99–1.19) p = NS 1.04 (0.90–1.20) p = NS 1.78 (0.99–3.17) p = NS

0.97 (0.89–1.06) p = NS 0.89 (0.80–0.98) p = 0.02 0.94 (0.73–1.20) p = NS 0.92 (0.84–1.02) p = NS 0.93 (0.80–1.07) p = NS 1.05 (0.56–1.94) p = NS

0.85 (0.75–0.96) p = 0.01 0.68 (0.58–0.77) p < 0.001 0.80 (0.57–1.11) p = NS 0.91 (0.79–1.04) p = NS 0.86 (0.71–1.03) p = NS 0.35 (0.17–0.74) p = 0.006

1.13 (1.08–1.18) p < 0.001 1.08 (1.03–1.13) p = 0.001 0.97 (0.87–1.08 p = NS 1.00 (0.95–1.05) p = NS 0.98 (0.92–1.06) p = NS 1.06 (0.80–1.39) p = NS

1.08 (0.96–1.23) p = NS 1.25 (1.10–1.44) p = 0.001 1.04 (0.78–1.39) p = NS 0.77 (0.67–0.90) p = 0.001 1.20 (0.83–1.74) p = NS 1.40 (0.58–3.41) p = NS

1.15 (1.01–1.29) p = 0.03 1.18 (1.04–1.34) p = 0.01 1.12 (0.85–1.49) p = NS 1.03 (0.91–1.17) p = NS 1.05 (0.86–1.29) p = NS 0.68 (0.35–1.30) p = NS

1.11 (1.06–1.16) p < 0.001 1.33 (1.26–1.39) p < 0.001 1.29 (1.13–1.48) p < 0.001 1.27 (1.20–1.33) p < 0.001 1.29 (1.20–1.39) p < 0.001 0.78 (0.59–1.03) p = NS

1.01 (1.00–1.02) p = 0.02 1.02 (0.01–0.02) p < 0.001 1.00 (0.98–1.02) p = NS 1.01 (1.00–1.02) p = 0.01 1.00 (0.99–1.02) p = 0.01 1.00 (0.95–1.05) p = NS

Patients with low SES were defined as those exempt from payments based on their income by the national insurance. Patients who immigrated to Israel after 1990 were considered as immigrants.

Table 5 – Cross medication persistence for patients who used metformin, statins and ACEI (% of patients with good medication persistence for each medication).

Metformin Statins ACEI

Only 1 medication

Any combination of 2 medications

All 3 medication

44.2% 60.8% 55.8%

55.4% 65.5% 68.4%

63.6% 69.2% 72.2%

Also patients with higher BMI showed better persistence. This could be due to a greater appreciation of the need for treatment. Surprisingly, low socioeconomic status was not found to be associated with lower persistence for the medications tested, contrasting with findings of other studies [16,18]. We note that in Israel the co-payment for all the medications examined is fairly low (3–15 USD) and is reduced for patients with low SES, so in the Israeli health care system co-payment is a less important factor. An increased number of medications used by the patient was associated with higher persistence. Similarly in a study in general practices in the Netherland the number of drug prescriptions was not related with adherence to oral blood glucose or lipid-lowering medications [19] this may reflect acceptance of the need for chronic medications. As expected, good medication persistence was associated with better glycemic control and lower LDL-cholesterol levels, as was shown in previous studies [4,5,20]. Less expected was the better glycemic control and lower LDL-cholesterol levels observed for all the medications tested, regardless of whether they were hypoglycemic, antihypertensive, or lipid lowering medications. Good medication persistence for even one drug was found to be an indicator for better diabetes control. This is probably because good medication persistence is also related to better medication persistence to other drugs, and may be related to better adherence to suitable diet and physical activity. Our findings suggest that investing effort in improving persistence for one medication may influence persistence to other drugs as well. Moreover, low persistence to

any chronic medication may be an important indicator for low persistence with medical care. Focusing on patients with low persistence to chronic medications may help identify those who need particular attention.

4.1.

Study limitations

We used medication purchasing as a proxy for medication persistence. However, this does not necessarily reflect medication utilization. We had no information about medication side effects, the patient support system or patient–physician relations, all of which influence medication persistence. We had no direct information as to whether medications were discontinued for a period by a physician, which would result in an underestimation of persistence. On the other hand, the large population and the completeness of the acquisition data likely enable a good estimate of medication persistence. Since purchase of each medication the year before and the year after analysis was a study inclusion criterion, we assume that no major changes were made in treatment regimens. The data is referred to persistence information regarding 2009 and newer drugs for diabetes that were emerging in the recent years were not included in this study. However we chose to study common oral diabetes drugs since many newer drugs are taken by injection and not all of them are included in the treatment basket supplied by the government so other persistence issues may be involved.

Please cite this article in press as: M. Shani, et al., Diabetes medication persistence, different medications have different persistence rates, Prim. Care Diab. (2017), http://dx.doi.org/10.1016/j.pcd.2017.03.006

PCD-596; No. of Pages 5

ARTICLE IN PRESS p r i m a r y c a r e d i a b e t e s x x x ( 2 0 1 7 ) xxx–xxx

In conclusion, although different medications have different persistence rates, poor persistence to any single medication was found to be associated with both poorer glycemic control and lower persistence to other medications. Investment in enhancing medication persistence in persons with diabetes may improve persistence to other medications, as well as improve glycemic control.

Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of interests The authors state that they have no conflict of interest.

references

[1] L. Osterberg, T. Blaschke, Adherence to medication, N. Engl. J. Med. 353 (2005) 487–497. [2] J.A. Glombiewski, Y. Nestoriuc, W. Rief, et al., Medication adherence in the general population, PLoS One 7 (2012) e50537. [3] J.A. Cramer, A systematic review of adherence with medications for diabetes, Diabetes Care 27 (2004) 1218–1224. [4] L.E. Egede, C.P. Lynch, M. Gebregziabher, et al., Differential impact of longitudinal medication non-adherence on mortality by race/ethnicity among veterans with diabetes, J. Gen. Intern. Med. 28 (2013) 208–215. [5] C. Asche, J. LaFleur, C. Conner, A review of diabetes treatment adherence and the association with clinical and economic outcomes, Clin. Ther. 33 (2011) 74–109. [6] M. Salas, D. Hughes, A. Zuluaga, et al., Costs of medication nonadherence in patients with diabetes mellitus: a systematic review and critical analysis of the literature, Value Health 12 (2009) 915–922. [7] B.C. Stuart, L. Simoni-Wastila, L. Zhao, et al., Increased persistency in medication use by U.S. medicare beneficiaries with diabetes is associated with lower hospitalization rates and cost savings, Diabetes Care 32 (2009) 647–649.

5

[8] M.K. Rhee, W. Slocum, D.C. Ziemer, et al., Patient adherence improves glycemic control, Diabetes Educ. 31 (2005) 240–250. [9] J.A. Cramer, A. Roy, A. Burrell, et al., Medication compliance and persistence: terminology and definitions, Value Health 11 (January–February (1)) (2008) 44–47. [10] L. Osterberg, T. Blaschke, Adherence to medication, N. Engl. J. Med. 353 (2005) 487–497. [11] S. Vinker, M. Shani, T. Baevsky, A. Elhayany, Adherence with statins over 8 years in a usual care setting, Am. J. Manage. Care 14 (2008) 388–392. [12] R. Halpern, S. Agarwal, C. Dembek, et al., Comparison of adherence and persistence among multiple sclerosis patients treated with disease-modifying therapies: a retrospective administrative claims analysis, Patient Prefer Adher. 5 (2011) 73–84. [14] B. Zinman, S.B. Harris, J. Neuman, et al., Low-dose combination therapy with rosiglitazone and metformin to prevent type 2 diabetes mellitus (CANOE trial): a double-blind randomised controlled study, Lancet 376 (2010) 103–111. [15] S.M. Curkendall, N. Thomas, K.F. Bell, et al., Predictors of medication adherence in patients with type 2 diabetes mellitus, Curr. Med. Res. Opin. 29 (2013) 1275–1286. [16] M. Tiv, J.F. Viel, F. Mauny, et al., Medication adherence in type 2 diabetes: the ENTRED study 2007, a French population-based study, PLoS One 7 (2012) e32412. [17] J.M. Lopez, R.A. Bailey, M.F. Rupnow, K. Annunziata, Characterization of type 2 diabetes mellitus burden by age and ethnic groups based on a nationwide survey, Clin. Ther. 36 (2014) 494–506. [18] Sharon J. Rolnick, A. Pamala, et al., Patient characteristics associated with medication adherence, Clin. Med. Res. 11 (2013) 54–65. [19] R. van Bruggen, K. Gorter, R.P. Stolk, P. Zuithoff, O.H. Klungel, G.E. Rutten, Refill adherence and polypharmacy among patients with type 2 diabetes in general practice, Pharmacoepidemiol. Drug Saf. 18 (November (11)) (2009) 983–991. [20] B.S. Feldman, C.J. Cohen-Stavi, M. Leibowitz, et al., Defining the role of medication adherence in poor glycemic control among a general adult population with diabetes, PLoS One 9 (2014) e108145.

Please cite this article in press as: M. Shani, et al., Diabetes medication persistence, different medications have different persistence rates, Prim. Care Diab. (2017), http://dx.doi.org/10.1016/j.pcd.2017.03.006