Do cost containment policies save money and influence physicians’ prescribing behavior? Lessons from South Korea’s drug policy for diabetes medication

Do cost containment policies save money and influence physicians’ prescribing behavior? Lessons... Abstract Objective We evaluate the effects of drug price reduction policy on pharmaceutical expenditure and prescription patterns in diabetes medication. Design An interrupted time series study design using generalized estimating equations. Setting This study used National Health Insurance claim data from 2010 to 2013. Participants A total of 68 127 diabetes patients and 12 465 hospitals. Intervention(s) The drug price reduction policy. Main Outcome Measures The primary outcome is pharmaceutical expenditure and prescription rate. To evaluate changes in prescription rate, we measured prescription rates such a brand-name drug and drug price reduction rate. Results Although the drug price reduction policy associated with decreased pharmaceutical expenditure (–13.22%, P < 0.0001), the trend (–0.01%, P = 0.9201) did not change significantly compared with the pre-intervention period. In addition, the trends in the monthly prescription rate of brand-name drugs decreased (–0.14%, P = 0.0091), while the immediate change was an increase (5.72%, P < 0.0001). Regardless of the drug reduction rate, the prescription rate after the introduction of the drug price reduction policy decreased compared with the pre-intervention period, and this decline was significant for reduction rates of 0% (–2.74%, P < 0.0001) and 10% (–0.13%, P = 0.0018). Conclusions Our results provide evidence of the effects of the drug price reduction policy on pharmaceutical expenditure and prescription patterns. This policy did not affect the prescribing behavior of healthcare providers and did not increase the use of drugs not subject to this policy. Although this study did not observe changes in the cost of pharmaceuticals after the introduction of the drug price reduction policy, further research is needed on the long-term changes in such costs. cost containment policy, prescription rate, pharmaceutical expenditure, prescription behavior Introduction Over the past few decades, health spending growth has exceeded economic growth, and despite declines in recent year, it remains a major concern for policymakers. According to the Organization for Economic Co-operation and Development (OECD), the recent decline in health expenditure has been caused by a decrease in pharmaceutical expenditure. Nevertheless, such spending still accounts for a large share of gross domestic product (GDP) among OECD countries (average of 9.0% of GDP) [1]. Further, health expenditure is expected to increase with the aging population and increasing prevalence of chronic diseases [2, 3]. Thus, cost containment policies are important to control the growth in health expenditure in many countries [4, 5]. In South Korea, healthcare spending accounts for 7.1% of GDP, below the average for OECD countries [1]. However, it is rising faster than the OECD average, with an annual growth rate for 2005–2010 of 6.8% (OECD average: 2.1%) [1]. In particular, the growth in pharmaceutical expenditure is one of the major problems of health expenditure. In 2014, pharmaceutical expenditure in South Korea accounted for 20.6% of health expenditure, above the OECD average of 16.3% [1]. One reason for this is the increase in the consumption of drugs for chronic diseases such as diabetes, which doubled between 2000 and 2013 [6]. These problems have led the government to introduce various drug policies to control pharmaceutical expenditure. One change to drug policy introduced in 2006 was a positive listing system that required pharmaceutical companies to reduce the costs of original drugs by 20% when their patents expired [7, 8]. Further, there was a re-evaluation of listed drugs; if cost-effectiveness was low, they were delisted or their costs were reduced. In addition, interventions to control drug costs for chronic diseases including hypertension and diabetes have been considered. As a result, clinical practice guidelines for diabetes medication reimbursement were introduced in 2011 [9]. Since then, physicians have had to consider patients’ clinical conditions when prescribing diabetes medication. Despite these efforts, however, the problem of high pharmaceutical expenditure has remained, and in April 2012, the largest drug price reduction policy ever seen in South Korea was introduced. Under this policy, approximately half of the covered drugs were reduced by an average of 21% [10]. Since the Korean healthcare system is a single-payer system, the reimbursement rate for healthcare providers depends on government policy; this means that healthcare providers are sensitive to government policy changes. Therefore, there is political interest in whether this policy has actually led to the intended results. Many studies have evaluated the effects of this drug price reduction policy on health expenditure [11–13]. These results have provided mixed evidence on whether cost containment policies are associated with a reduction in pharmaceutical expenditure [14–16]. This is because the cost savings from cost containment policies have been offset by other influences, particularly the prescribing behavior of health providers [17]. In other words, to measure the effect of cost containment policies, the changes have to be assessed in terms of costs as well as other aspects. However, there is lack of evidence as to whether changes in drug policy affect the prescription behavior of healthcare providers. This study evaluates changes in pharmaceutical costs and the prescription behavior of healthcare providers after the introduction of South Korea’s diabetes drug price reduction policy. First, we assumed that drug price would decrease after the drug price reduction policy, and measured changes in pharmaceutical expenditure. In order to examine the effect of the policy on pharmaceutical expenditure, it is necessary to examine the drug prescribing behavior which may affect the drug cost. This is because the introduction of a policy can lead to a change in physician’s choice of medication, which can offset the effect of the policy. Thus, this study evaluates changes in the prescription rate according to the provision of brand-name drugs and the drug price reduction rate. Materials and Methods Database and data collection This study used National Health Insurance (NHI) elderly cohort data from August 2010 to December 2013 to investigate changes in pharmaceutical expenditure and prescription patterns after the introduction of the drug price reduction policy. The baseline population was 558 147 participants who were randomly selected, accounting for 10% of the Korean population aged over 60 years old in 2002 [18]. These data were representative of population-based cohort data in South Korea. Among them, patients with type 2 diabetes were selected by using International Classification of Disease (ICD)-10 codes (E11–E14). A total of 101 648 patients used outpatient care during the study period. First, this study excluded outpatients not prescribed medication for diabetes. Because the use of oral hypoglycemic agents and changes in medication expenditure were taken into consideration, patients without oral diabetes medication were excluded. Next, patients who did not use outpatient care were excluded. Finally, those patients who had used a community health center or long-term care facilities were also excluded, as these facilities have specific functions and provide different healthcare services. Ultimately, 68 127 patients were included in our study (see Appendix A). Variables The outcome variables were pharmaceutical expenditure and the prescription rate in diabetes (brand-name drugs, drug price reduction rate). To evaluate the effect of introduction of drug policy, we measured changes in outcome variable before and after the introduction of policy. First, we selected oral hypoglycemic agents based on Anatomical Therapeutic Chemical Classification System codes (ATC code: A10B). Next, we selected a diabetes drug that was reimbursed by the NHI. NHI’s reimbursement criteria included single component and multiple components, and the drugs classified into 10 components are included in this study. Second, the drug prescription rate was measured by dividing the total diabetes prescription by the number of prescriptions of a specific drug. The prescription rate of brand-name drugs, one of our outcome variables, was measured on the basis of health insurance entry dates. If a drug is the first listed drug in each drug classification, we defined it as a brand-name drug. Next, we measured drug price reduction rate for each diabetes drugs. Drug reduction rates vary from drug to drug, and from an average of 0–23% depending on the drug ingredients. Then, we classified the drug price reduction rates into three categories: 0%, <10% and ≥10%. Third, pharmaceutical expenditures were calculated based on outpatient visits to diabetes medications included in our study. Finally, all outcome variables were aggregated by month (see Appendix A). The primary variable of interest in this study was the effects of drug price reduction policy. To evaluate the effects of drug policy, we measured the level and overall trend change in the outcome variables. Level change indicates the change at the moment of intervention. Trend change is the rate of change of an outcome variable, defined as an increase or decrease in the slope of the segment after the intervention compared with the segment preceding the intervention. During the study period, there were two interventions for diabetes medication, namely the July 2011 reimbursement restriction and the April 2012 price reduction policy. Thus, this study considers two-time points to evaluate the effects of the drug price reduction policy on the outcome variables. We used a dummy variable based on the index date (July 2011 or April 2012). The time before the intervention was defined as 0 and the time after the introduction of each policy was defined as 1. In addition, the linear changes after each intervention were measured as continuous variables based on the index date. The overall trends were stratified by month and included data from August 2010 to December 2013 (See Appendix A). We adjusted for patient and hospital characteristics when analyzing the changes in the outcome variables after the introduction of the drug policy. Patient characteristics such as sex, age (66–74, 75–84, ≥84), complications (none, single or multiple), Charlson comorbidity index (CCI), insurance type (medical aid, self-employed or employee), income status (low, moderate, high) and year of ambulatory care were included in our analysis. Complications were based on specific ICD-10 codes; if the code was 0.9 (without complications) or 0.7 (multiple complications), we classified the patient as ‘none’ or ‘multiple’, respectively. Other patients with specific ICD-10 codes were classified as a ‘single’ complication. Hospital characteristics included type of hospital (clinic, hospital or general hospital), hospital location (metropolitan, non-metropolitan), ownership status (public, private), bed for admission (yes, no) and number of doctors. Statistical analysis To evaluate the changes in the outcome variables, we used an interrupted time series study design, using generalized estimating equations [19–22]. We performed Poisson regression analysis with a log link function to evaluate the changes in the pattern of prescriptions for brand-name drugs and price reduction rates. A gamma generalized linear model based on the log link function was used to evaluate pharmaceutical expenditure. The unit of analysis is the patient case. All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Inc.; Cary, NC, USA). P-values <0.05 were considered to be statistically significant. Ethical consideration This study was approved by the Institutional Review Board, Yonsei University Health System (Y-2017-0058). Results The data used in our study were from 68 127 diabetes patients and 12 465 medical institutions. Most patients were over 85 years old (n = 37 519, 55.1%), and the smallest age group was 75–84 years old (n = 3423, 5.0%). Most patients with diabetes had multiple complications (n = 36 611, 53.7%), but 40.4% had no complications. Regarding insurance type, most patients were enrolled as employees, but there was a small proportion of medical aid patients (n = 6786, 10.0%). Regarding hospital type, clinics (n = 11 105) were the most common (Table 1). Table 1 Baseline characteristics of diabetes patient and hospital   N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)    N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)  CCI, Charlson comorbidity index. Table 1 Baseline characteristics of diabetes patient and hospital   N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)    N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)  CCI, Charlson comorbidity index. The average prescription rate of brand-name drugs rose from 56.93% before the introduction of the drug price reduction policy to 57.16% thereafter; however, this difference was not statistically significant. Pharmaceutical expenditure decreased significantly after the policy introduction. The prescription rate according to drug price reduction rate is also seen to decrease in drugs with a high drug price cut rate (Table 2). Table 2 Change in prescription rate and pharmaceutical expenditure since the introduction of Pharmaceutical price control policy   Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001    Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001  0%: DPP-IV inhibitor, Metformin+DPP-IV inhibitor, Metformin+Thiazolidinedione, Sulfonylurea+Thiazolidinedione; <10%: Metformin, Meglitinide, Sulfonlyurea+Metformin; ≤10%: Sulfonlyurea, α-glucosidase inhibitor, Thiazolidinedione. aAverage monthly prescription rate. Table 2 Change in prescription rate and pharmaceutical expenditure since the introduction of Pharmaceutical price control policy   Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001    Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001  0%: DPP-IV inhibitor, Metformin+DPP-IV inhibitor, Metformin+Thiazolidinedione, Sulfonylurea+Thiazolidinedione; <10%: Metformin, Meglitinide, Sulfonlyurea+Metformin; ≤10%: Sulfonlyurea, α-glucosidase inhibitor, Thiazolidinedione. aAverage monthly prescription rate. Figure 1 illustrates the results of the interrupted analysis, showing that monthly prescription rates changed after the introduction of the drug price reduction policy. An immediate change (5.72%) and a trend change (–0.14%) compared with the pre-intervention period were observed in the prescription rate of brand-name drugs, and this difference was significant (Fig. 1a). Pharmaceutical expenditure also decreased significantly (–13.22%) immediately after the policy introduction; however, the trend did not change significantly (Fig. 1b). Figure 2 shows the results of the prescription rate according to the drug reduction rate. Regardless of the reduction rate in drug costs, the prescription rates were decreased by 0.13% (reduction rate <10% or more) and 2.74% (reduction rate of 0%), respectively, compared to pre-policy rates. Figure 1 View largeDownload slide Changes in prescription rate of brand-name drugs and pharmaceutical expenditure. Figure 1 View largeDownload slide Changes in prescription rate of brand-name drugs and pharmaceutical expenditure. Figure 2 View largeDownload slide Changes in prescription rate by drug reduction rate. Figure 2 View largeDownload slide Changes in prescription rate by drug reduction rate. The subgroup analyses by type of hospital showed that the prescription rates of brand-name drugs decreased for all hospital types compared with the pre-intervention period. Regarding pharmaceutical expenditure, only clinics showed decreased trends compared with the pre-intervention period (–0.02%). The trends of pharmaceutical expenditure increased in both hospitals (0.04%) and general hospitals (0.06%). In most hospitals, the prescription rate decreased compared with the previous trend according to the drug reduction rate (Table 3). Table 3 Change in prescription patterns and costs by type of hospital (Unit: RR, 95% CI)   Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966    Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966  Table 3 Change in prescription patterns and costs by type of hospital (Unit: RR, 95% CI)   Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966    Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966  Discussion Drug price reduction policies are considered one way to reduce pharmaceutical expenditure, but unlike the initial intention, they can increase the use of drugs that are not covered by the policy. However, our study shows that drug price policies have impacted pharmaceutical expenditure, but have not changed the pattern of prescription, such as increased use of non-targeted drugs. Previous findings on the impact of cost reduction policies on cost savings are mixed. Some studies, similar to our results, show that cost containment policies are associated with cost savings [12, 23]. However, they do not always lead to the cost savings initially intended [24, 25] because the policy effects on cost savings can be offset by changes to the prescription patterns of healthcare prescribers. In particular, prescription decisions are often affected by illegal marketing by pharmaceutical wholesalers with under-the-table negotiations [26]. This has led to unintended consequences, such as an increase in the use of drugs not subject to the drug policy, in turn raising pharmaceutical expenditure. In this study, while immediate cost savings were noted upon the policy introduction, this cost trend did not change in the long run, even after controlling for the reimbursement restriction effect. One plausible explanation may be physicians’ prescribing decisions. Physicians’ prescribing behavior is affected by both extrinsic and intrinsic factors [17]. Intrinsic factors are associated with physician characteristics such as medical specialization and previous clinical experience. Extrinsic factors include patient factors, pharmaceutical industry factors (e.g. the pressure placed upon healthcare providers to prescribe various medications) and healthcare system-related factors. These factors influence the knowledge of the physician, and based on this, a variety of attitudes are formed and eventually the prescription is decided. This means that the prescribing behavior of physicians may not be changed by any one factor. Indeed, various factors have to be taken into consideration, leading to changes in prescribing behavior. In particular, prescription decisions are more likely to be related to physician factors such as the physician’s belief in the probability of a particular outcome and the personal value attached to these outcomes [27–29]. In our study, the prescription rate increased immediately for drugs in two categories, namely no cost change and a cost decrease of 10%. However, these effects were temporary and did not lead to an increase in monthly prescription rates. Similarly, there was no significant trend change in drug prescribing rates that declined more than 10% since the policy introduction. Further, the use of brand-name drugs did not increase compared with the pre-intervention period and was rather influenced by the introduction of reimbursement restrictions. If only the factors of the pharmaceutical company influence the physician’s prescription, we would have observed that the non-subject prescription rate increased in our study. However, physician prescribing behavior is affected by many factors, and the effects of other policies in this study can affect prescribing behavior. Previous studies on the prescribing behavior of hypertensive drugs since the introduction of the reimbursement restriction and drug price reduction policies showed a similar pattern to our study [10]. Both policies reduced drug costs, but the policy that has a greater impact on drug prescribing behavior was reimbursement restrictions rather than drug price reduction policies. Similarly, in our study, reimbursement policies were effective in prescribing behavior and no unintended effects such as an increased prescription volume of non-target drugs due to drug price reduction policy were observed. This phenomenon shows that drug price reduction policies are a strong policy that can directly affect saving costs, but physician prescribing behavior has not largely affected by drug price reduction policies because it is affected by a variety of factors. According to the subgroup analysis by type of hospital, pharmaceutical expenditure decreased significantly following the introduction of the drug price reduction policy, but the trend was different. In hospitals and general hospitals, the cost trend increased compared with the pre-intervention period. This phenomenon can be related to the different effects of the reimbursement restrictions depending on the type of hospital and patient severity. The introduction of reimbursement restrictions reduced the cost trend compared with previous periods, but these hospitals are used by patients with relatively high severity compared with clinics, meaning that the use of drugs according to the guidelines may be inappropriate. However, because drug use is reimbursed under the clinical guidelines, changes in prescribing behavior may occur later. Indeed, this effect may appear as an increase in pharmaceutical costs after the introduction of the drug price reduction policy. Therefore, a longer observation period would be needed to accurately determine drug cost changes by hospital type. Cost containment policies are important to control the growth in health expenditure and reduce the wastage of health resources. Our results provide evidence of the effects of drug price reduction policies on pharmaceutical expenditure and prescription patterns. First, such policies lead to pharmaceutical cost savings. This finding provides evidence that a drug price reduction policy is important to reduce health expenditure. Second, this study did not find any negative effects of such a policy. Unintended consequences of the cost policy could increase the use of drugs not covered by this policy, but this study did not observe changes in prescription patterns. These results indicate that the drug price reduction policy is effective at reducing costs but does not significantly affect prescribing behavior. Therefore, the drug price reduction policy can be an important strategy for controlling the growth in health expenditure. Our study has several strengths. First, although much research has been carried out on the cost-reducing effects of drug policy, few studies have examined the effect on prescribing patterns. Second, the interrupted time series analysis design is a powerful method for evaluating political change that provides immediate and trend changes. Third, we used NHI cohort data with representative samples, and thus these results provide meaningful evidence for policymakers. Finally, our results are also meaningful to countries considering introducing drug policies to reduce health expenditure. Despite these strengths, our study does have some limitations. First, a patient’s clinical condition can affect prescribing changes, which can lead to cost changes. Second, short-term evaluations can be confusing because the introduction of reimbursement restrictions and of drug price reduction policies are close in time. Therefore, to accurately assess the effect of this policy on cost savings, further studies must account for its long-term effects. Third, the effects of the policy may be different in other drugs because we evaluated only diabetes drugs. Finally, because costs may be affected by other factors, unmeasured factors may affect our results. Conclusion This study provides evidence to policymakers that drug price reduction policies are an effective way in which to control the growth in health expenditure while not negatively affecting changes in prescription patterns. South Korea’s drug price reduction policy did not affect the prescribing behavior of healthcare providers and did not increase the use of drugs not subject to this policy. Although this study did not observe changes in the cost of pharmaceuticals after the policy introduction, further research is needed on the long-term changes in such costs. Supplementary material Supplementary material is available at International Journal for Quality in Health Care online. Acknowledgements None. Funding This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors. References 1 Organizations for Economic Co-operation and Development (OECD). OECD Health Statistics 2016. OECD; 2017. 2 Mongan JJ, Ferris TG, Lee TH. Options for slowing the growth of health care costs. N Engl J Med  2008; 358: 1509. Google Scholar CrossRef Search ADS PubMed  3 Bodenheimer T, Chen E, Bennett HD. Confronting the growing burden of chronic disease: can the US health care workforce do the job? Health Aff  2009; 28: 64– 74. Google Scholar CrossRef Search ADS   4 Ess SM, Schneeweiss S, Szucs TD. European healthcare policies for controlling drug expenditure. Pharmacoeconomics  2003; 21: 89– 103. Google Scholar CrossRef Search ADS PubMed  5 Pita Barros P. Pharmaceutical policies in European countries. In: Pharmaceutical Markets and Insurance Worldwide . Bingley: Emerald Group Publishing Limited, 2010: 3– 27. Google Scholar CrossRef Search ADS   6 Organizations for Economic Co-operation and Development (OECD). Health at a Glance 2015: OECD indicators . Paris: OECD Publishing, 2015. 7 Bae EY, Lee EK. Pharmacoeconomic guidelines and their implementation in the positive list system in South Korea. Value Health  2009; 12: S36– 41. Google Scholar CrossRef Search ADS PubMed  8 Park SE, Lim SH, Choi HW et al.  . Evaluation on the first 2 years of the positive list system in South Korea. Health Policy  2012; 104: 32– 9. Google Scholar CrossRef Search ADS PubMed  9 Ko SH, Oh SJ, Lee HJ et al.  . 2011 Clinica Practice Guideline for Type 2 Diabetes in Korea. J Korean Diabetes  2011; 12: 183– 9. Google Scholar CrossRef Search ADS   10 Yoo K-B, Lee SG, Park S et al.  . Effects of drug price reduction and prescribing restrictions on expenditures and utilisation of antihypertensive drugs in Korea. BMJ open  2015; 5: e006940. Google Scholar CrossRef Search ADS PubMed  11 Carone G, Schwierz C, Xavier A Cost-containment policies in public pharmaceutical spending in the EU. 2012. 12 Lee Y-C, Yang M-C, Huang Y-T et al.  . Impacts of cost containment strategies on pharmaceutical expenditures of the National Health Insurance in Taiwan, 1996–2003. Pharmacoeconomics  2006; 24: 891– 902. Google Scholar CrossRef Search ADS PubMed  13 Moreno-Serra R. The impact of cost-containment policies on health expenditure: evidence from recent OECD experiences. OECD J Budget  2014; 13: 1B. Google Scholar CrossRef Search ADS   14 Moreno-Torres I, Puig-Junoy J, Raya JM. The impact of repeated cost containment policies on pharmaceutical expenditure: experience in Spain. Eur J Health Econ  2011; 12: 563– 73. Google Scholar CrossRef Search ADS PubMed  15 Lambrelli D, O’Donnell O. The impotence of price controls: failed attempts to constrain pharmaceutical expenditures in Greece. Health Policy  2011; 101: 162– 71. Google Scholar CrossRef Search ADS PubMed  16 Kenneally M, Walshe V. Pharmaceutical cost-containment policies and sustainability: recent Irish experience. Value Health  2012; 15: 389– 93. Google Scholar CrossRef Search ADS PubMed  17 Rodrigues AT, Roque F, Falcão A et al.  . Understanding physician antibiotic prescribing behaviour: a systematic review of qualitative studies. Int J Antimicrob Agents  2013; 41: 203– 12. Google Scholar CrossRef Search ADS PubMed  18 Lee J, Lee JS, Park S-H et al.  . Cohort profile: the National Health Insurance Service–National Sample Cohort (NHIS-NSC), South Korea. Int J Epidemiol  2016; 46: e15. doi:10.1093/ije/dyv319. 19 Zeger SL, Liang K-Y, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics  1988; 44: 1049– 60. Google Scholar CrossRef Search ADS PubMed  20 Hanley JA, Negassa A, Forrester JE. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol  2003; 157: 364– 75. Google Scholar CrossRef Search ADS PubMed  21 Wagner AK, Soumerai SB, Zhang F et al.  . Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther  2002; 27: 299– 309. Google Scholar CrossRef Search ADS PubMed  22 Zhang F, Wagner AK, Soumerai SB et al.  . Methods for estimating confidence intervals in interrupted time series analyses of health interventions. J Clin Epidemiol  2009; 62: 143– 8. Google Scholar CrossRef Search ADS PubMed  23 Brekke KR, Grasdal AL, Holmås TH. Regulation and pricing of pharmaceuticals: reference pricing or price cap regulation? Eur Econ Rev  2009; 53: 170– 85. Google Scholar CrossRef Search ADS   24 Chu H-L, Liu S-Z, Romeis JC. Assessing the effects of drug price reduction policies on older people in Taiwan. Health Serv Manag Res  2011; 24: 1– 7. Google Scholar CrossRef Search ADS   25 Kwon H-Y, Hong J-M, Godman B et al.  . Price cuts and drug spending in South Korea: the case of antihyperlipidemic agents. Health Policy  2013; 112: 217– 26. Google Scholar CrossRef Search ADS PubMed  26 Kwon S. Pharmaceutical reform and physician strikes in Korea: separation of drug prescribing and dispensing. Soc Sci Med  2003; 57: 529– 38. Google Scholar CrossRef Search ADS PubMed  27 Lilja J. How physicians choose their drugs. Soc Sci Med  1976; 10: 363– 5. Google Scholar CrossRef Search ADS PubMed  28 Segal R, Hepler CD. Drug choice as a problem-solving process. Med Care  1985; 23: 967– 76. Google Scholar CrossRef Search ADS PubMed  29 Coscelli A. The importance of doctors’ and patients’ preferences in the prescription decision. J Ind Econ  2000; 48: 349– 69. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal for Quality in Health Care Oxford University Press

Do cost containment policies save money and influence physicians’ prescribing behavior? Lessons from South Korea’s drug policy for diabetes medication

Loading next page...
 
/lp/ou_press/do-cost-containment-policies-save-money-and-influence-physicians-lfB1UGljLU
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1353-4505
eISSN
1464-3677
D.O.I.
10.1093/intqhc/mzy114
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective We evaluate the effects of drug price reduction policy on pharmaceutical expenditure and prescription patterns in diabetes medication. Design An interrupted time series study design using generalized estimating equations. Setting This study used National Health Insurance claim data from 2010 to 2013. Participants A total of 68 127 diabetes patients and 12 465 hospitals. Intervention(s) The drug price reduction policy. Main Outcome Measures The primary outcome is pharmaceutical expenditure and prescription rate. To evaluate changes in prescription rate, we measured prescription rates such a brand-name drug and drug price reduction rate. Results Although the drug price reduction policy associated with decreased pharmaceutical expenditure (–13.22%, P < 0.0001), the trend (–0.01%, P = 0.9201) did not change significantly compared with the pre-intervention period. In addition, the trends in the monthly prescription rate of brand-name drugs decreased (–0.14%, P = 0.0091), while the immediate change was an increase (5.72%, P < 0.0001). Regardless of the drug reduction rate, the prescription rate after the introduction of the drug price reduction policy decreased compared with the pre-intervention period, and this decline was significant for reduction rates of 0% (–2.74%, P < 0.0001) and 10% (–0.13%, P = 0.0018). Conclusions Our results provide evidence of the effects of the drug price reduction policy on pharmaceutical expenditure and prescription patterns. This policy did not affect the prescribing behavior of healthcare providers and did not increase the use of drugs not subject to this policy. Although this study did not observe changes in the cost of pharmaceuticals after the introduction of the drug price reduction policy, further research is needed on the long-term changes in such costs. cost containment policy, prescription rate, pharmaceutical expenditure, prescription behavior Introduction Over the past few decades, health spending growth has exceeded economic growth, and despite declines in recent year, it remains a major concern for policymakers. According to the Organization for Economic Co-operation and Development (OECD), the recent decline in health expenditure has been caused by a decrease in pharmaceutical expenditure. Nevertheless, such spending still accounts for a large share of gross domestic product (GDP) among OECD countries (average of 9.0% of GDP) [1]. Further, health expenditure is expected to increase with the aging population and increasing prevalence of chronic diseases [2, 3]. Thus, cost containment policies are important to control the growth in health expenditure in many countries [4, 5]. In South Korea, healthcare spending accounts for 7.1% of GDP, below the average for OECD countries [1]. However, it is rising faster than the OECD average, with an annual growth rate for 2005–2010 of 6.8% (OECD average: 2.1%) [1]. In particular, the growth in pharmaceutical expenditure is one of the major problems of health expenditure. In 2014, pharmaceutical expenditure in South Korea accounted for 20.6% of health expenditure, above the OECD average of 16.3% [1]. One reason for this is the increase in the consumption of drugs for chronic diseases such as diabetes, which doubled between 2000 and 2013 [6]. These problems have led the government to introduce various drug policies to control pharmaceutical expenditure. One change to drug policy introduced in 2006 was a positive listing system that required pharmaceutical companies to reduce the costs of original drugs by 20% when their patents expired [7, 8]. Further, there was a re-evaluation of listed drugs; if cost-effectiveness was low, they were delisted or their costs were reduced. In addition, interventions to control drug costs for chronic diseases including hypertension and diabetes have been considered. As a result, clinical practice guidelines for diabetes medication reimbursement were introduced in 2011 [9]. Since then, physicians have had to consider patients’ clinical conditions when prescribing diabetes medication. Despite these efforts, however, the problem of high pharmaceutical expenditure has remained, and in April 2012, the largest drug price reduction policy ever seen in South Korea was introduced. Under this policy, approximately half of the covered drugs were reduced by an average of 21% [10]. Since the Korean healthcare system is a single-payer system, the reimbursement rate for healthcare providers depends on government policy; this means that healthcare providers are sensitive to government policy changes. Therefore, there is political interest in whether this policy has actually led to the intended results. Many studies have evaluated the effects of this drug price reduction policy on health expenditure [11–13]. These results have provided mixed evidence on whether cost containment policies are associated with a reduction in pharmaceutical expenditure [14–16]. This is because the cost savings from cost containment policies have been offset by other influences, particularly the prescribing behavior of health providers [17]. In other words, to measure the effect of cost containment policies, the changes have to be assessed in terms of costs as well as other aspects. However, there is lack of evidence as to whether changes in drug policy affect the prescription behavior of healthcare providers. This study evaluates changes in pharmaceutical costs and the prescription behavior of healthcare providers after the introduction of South Korea’s diabetes drug price reduction policy. First, we assumed that drug price would decrease after the drug price reduction policy, and measured changes in pharmaceutical expenditure. In order to examine the effect of the policy on pharmaceutical expenditure, it is necessary to examine the drug prescribing behavior which may affect the drug cost. This is because the introduction of a policy can lead to a change in physician’s choice of medication, which can offset the effect of the policy. Thus, this study evaluates changes in the prescription rate according to the provision of brand-name drugs and the drug price reduction rate. Materials and Methods Database and data collection This study used National Health Insurance (NHI) elderly cohort data from August 2010 to December 2013 to investigate changes in pharmaceutical expenditure and prescription patterns after the introduction of the drug price reduction policy. The baseline population was 558 147 participants who were randomly selected, accounting for 10% of the Korean population aged over 60 years old in 2002 [18]. These data were representative of population-based cohort data in South Korea. Among them, patients with type 2 diabetes were selected by using International Classification of Disease (ICD)-10 codes (E11–E14). A total of 101 648 patients used outpatient care during the study period. First, this study excluded outpatients not prescribed medication for diabetes. Because the use of oral hypoglycemic agents and changes in medication expenditure were taken into consideration, patients without oral diabetes medication were excluded. Next, patients who did not use outpatient care were excluded. Finally, those patients who had used a community health center or long-term care facilities were also excluded, as these facilities have specific functions and provide different healthcare services. Ultimately, 68 127 patients were included in our study (see Appendix A). Variables The outcome variables were pharmaceutical expenditure and the prescription rate in diabetes (brand-name drugs, drug price reduction rate). To evaluate the effect of introduction of drug policy, we measured changes in outcome variable before and after the introduction of policy. First, we selected oral hypoglycemic agents based on Anatomical Therapeutic Chemical Classification System codes (ATC code: A10B). Next, we selected a diabetes drug that was reimbursed by the NHI. NHI’s reimbursement criteria included single component and multiple components, and the drugs classified into 10 components are included in this study. Second, the drug prescription rate was measured by dividing the total diabetes prescription by the number of prescriptions of a specific drug. The prescription rate of brand-name drugs, one of our outcome variables, was measured on the basis of health insurance entry dates. If a drug is the first listed drug in each drug classification, we defined it as a brand-name drug. Next, we measured drug price reduction rate for each diabetes drugs. Drug reduction rates vary from drug to drug, and from an average of 0–23% depending on the drug ingredients. Then, we classified the drug price reduction rates into three categories: 0%, <10% and ≥10%. Third, pharmaceutical expenditures were calculated based on outpatient visits to diabetes medications included in our study. Finally, all outcome variables were aggregated by month (see Appendix A). The primary variable of interest in this study was the effects of drug price reduction policy. To evaluate the effects of drug policy, we measured the level and overall trend change in the outcome variables. Level change indicates the change at the moment of intervention. Trend change is the rate of change of an outcome variable, defined as an increase or decrease in the slope of the segment after the intervention compared with the segment preceding the intervention. During the study period, there were two interventions for diabetes medication, namely the July 2011 reimbursement restriction and the April 2012 price reduction policy. Thus, this study considers two-time points to evaluate the effects of the drug price reduction policy on the outcome variables. We used a dummy variable based on the index date (July 2011 or April 2012). The time before the intervention was defined as 0 and the time after the introduction of each policy was defined as 1. In addition, the linear changes after each intervention were measured as continuous variables based on the index date. The overall trends were stratified by month and included data from August 2010 to December 2013 (See Appendix A). We adjusted for patient and hospital characteristics when analyzing the changes in the outcome variables after the introduction of the drug policy. Patient characteristics such as sex, age (66–74, 75–84, ≥84), complications (none, single or multiple), Charlson comorbidity index (CCI), insurance type (medical aid, self-employed or employee), income status (low, moderate, high) and year of ambulatory care were included in our analysis. Complications were based on specific ICD-10 codes; if the code was 0.9 (without complications) or 0.7 (multiple complications), we classified the patient as ‘none’ or ‘multiple’, respectively. Other patients with specific ICD-10 codes were classified as a ‘single’ complication. Hospital characteristics included type of hospital (clinic, hospital or general hospital), hospital location (metropolitan, non-metropolitan), ownership status (public, private), bed for admission (yes, no) and number of doctors. Statistical analysis To evaluate the changes in the outcome variables, we used an interrupted time series study design, using generalized estimating equations [19–22]. We performed Poisson regression analysis with a log link function to evaluate the changes in the pattern of prescriptions for brand-name drugs and price reduction rates. A gamma generalized linear model based on the log link function was used to evaluate pharmaceutical expenditure. The unit of analysis is the patient case. All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Inc.; Cary, NC, USA). P-values <0.05 were considered to be statistically significant. Ethical consideration This study was approved by the Institutional Review Board, Yonsei University Health System (Y-2017-0058). Results The data used in our study were from 68 127 diabetes patients and 12 465 medical institutions. Most patients were over 85 years old (n = 37 519, 55.1%), and the smallest age group was 75–84 years old (n = 3423, 5.0%). Most patients with diabetes had multiple complications (n = 36 611, 53.7%), but 40.4% had no complications. Regarding insurance type, most patients were enrolled as employees, but there was a small proportion of medical aid patients (n = 6786, 10.0%). Regarding hospital type, clinics (n = 11 105) were the most common (Table 1). Table 1 Baseline characteristics of diabetes patient and hospital   N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)    N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)  CCI, Charlson comorbidity index. Table 1 Baseline characteristics of diabetes patient and hospital   N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)    N/M  %/SD  Sex   Male  26 448  (38.8)   Female  41 679  (61.2)  Age   66–74  37 519  (55.1)   75–84  27 185  (39.9)   ≥85  3423  (5.0)  Complication   None  36 611  (53.7)   Single  27 487  (40.4)   Multiple  4029  (5.9)  CCI  0.81  ±0.75  Duration of diabetes (year)   ≤5  26 955  (39.6)   ≥6  41 172  (60.4)  Insurance type   Medical aid  6786  (10.0)   Self-employed  19 818  (29.1)   Employees  41 523  (61.0)  Income   Low  16 239  (23.8)   Moderate  26 316  (38.6)   High  25 572  (37.5)  Year of ambulatory care   2010  41 157  (60.4)   2011  13 337  (19.6)   2012  7701  (11.3)   2013  5932  (8.7)  Hospital characteristics (n = 12 465)   Type of hospital    Clinic (n = 11 105)  44 743  (65.7)    Hospital (n = 1010)  4955  (7.3)    General hospital (n = 350)  18 429  (27.1)   Hospital location    Metropolitan (n = 5657)  31 916  (46.9)    Non-metropolitan (n = 6808)  36 211  (53.2)   Ownership status    Public (n = 43)  1444  (2.1)    Private (n = 12 422)  66 683  (97.9)   Bed for admission    Yes (n = 7379)  44 645  (65.5)    No (n = 5086)  23 482  (34.5)   Number of doctors  85.67  ±234.65  Total  68 127  (100.0)  CCI, Charlson comorbidity index. The average prescription rate of brand-name drugs rose from 56.93% before the introduction of the drug price reduction policy to 57.16% thereafter; however, this difference was not statistically significant. Pharmaceutical expenditure decreased significantly after the policy introduction. The prescription rate according to drug price reduction rate is also seen to decrease in drugs with a high drug price cut rate (Table 2). Table 2 Change in prescription rate and pharmaceutical expenditure since the introduction of Pharmaceutical price control policy   Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001    Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001  0%: DPP-IV inhibitor, Metformin+DPP-IV inhibitor, Metformin+Thiazolidinedione, Sulfonylurea+Thiazolidinedione; <10%: Metformin, Meglitinide, Sulfonlyurea+Metformin; ≤10%: Sulfonlyurea, α-glucosidase inhibitor, Thiazolidinedione. aAverage monthly prescription rate. Table 2 Change in prescription rate and pharmaceutical expenditure since the introduction of Pharmaceutical price control policy   Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001    Introduction of pharmaceutical price control policy  P-value  Before  After  Outcome variable   Brand-name drug (%)  56.93 ± 1.22  57.16 ± 1.00  0.5157   Pharmaceutical expenditure (KRW)  20 974 ± 691  16 918 ± 307  <0.0001   Reduction ratea    0%  11.77 ± 2.50  17.59 ± 0.26  <0.0001    <10%  68.63 ± 0.99  70.40 ± 0.59  <0.0001    ≥10%  68.69 ± 2.82  59.55 ± 2.03  <0.0001  0%: DPP-IV inhibitor, Metformin+DPP-IV inhibitor, Metformin+Thiazolidinedione, Sulfonylurea+Thiazolidinedione; <10%: Metformin, Meglitinide, Sulfonlyurea+Metformin; ≤10%: Sulfonlyurea, α-glucosidase inhibitor, Thiazolidinedione. aAverage monthly prescription rate. Figure 1 illustrates the results of the interrupted analysis, showing that monthly prescription rates changed after the introduction of the drug price reduction policy. An immediate change (5.72%) and a trend change (–0.14%) compared with the pre-intervention period were observed in the prescription rate of brand-name drugs, and this difference was significant (Fig. 1a). Pharmaceutical expenditure also decreased significantly (–13.22%) immediately after the policy introduction; however, the trend did not change significantly (Fig. 1b). Figure 2 shows the results of the prescription rate according to the drug reduction rate. Regardless of the reduction rate in drug costs, the prescription rates were decreased by 0.13% (reduction rate <10% or more) and 2.74% (reduction rate of 0%), respectively, compared to pre-policy rates. Figure 1 View largeDownload slide Changes in prescription rate of brand-name drugs and pharmaceutical expenditure. Figure 1 View largeDownload slide Changes in prescription rate of brand-name drugs and pharmaceutical expenditure. Figure 2 View largeDownload slide Changes in prescription rate by drug reduction rate. Figure 2 View largeDownload slide Changes in prescription rate by drug reduction rate. The subgroup analyses by type of hospital showed that the prescription rates of brand-name drugs decreased for all hospital types compared with the pre-intervention period. Regarding pharmaceutical expenditure, only clinics showed decreased trends compared with the pre-intervention period (–0.02%). The trends of pharmaceutical expenditure increased in both hospitals (0.04%) and general hospitals (0.06%). In most hospitals, the prescription rate decreased compared with the previous trend according to the drug reduction rate (Table 3). Table 3 Change in prescription patterns and costs by type of hospital (Unit: RR, 95% CI)   Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966    Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966  Table 3 Change in prescription patterns and costs by type of hospital (Unit: RR, 95% CI)   Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966    Level change  Trend change  RR  95%CI  P-value  RR  95% CI  P-value  Clinic   Brand-name drug (%)  1.061  1.053  1.069  <0.0001  0.999  0.998  1.000  0.0848   Pharmaceutical expenditure  0.861  0.855  0.866  <0.0001  0.998  0.996  0.999  <0.0001   Reduction rate    0%  1.086  1.064  1.108  <0.0001  0.972  0.968  0.976  <0.0001    <10%  1.011  1.006  1.016  <0.0001  0.998  0.997  0.999  <0.0001    ≥10%  0.984  0.979  0.989  <0.0001  0.999  0.998  1.000  0.1620  Hospital   Brand-name drug (%)  1.026  1.001  1.052  0.0419  0.999  0.994  1.003  0.5627   Pharmaceutical expenditure  0.883  0.860  0.907  <0.0001  1.013  1.008  1.018  <0.0001   Reduction rate                    0%  1.089  1.019  1.163  0.0113  0.969  0.956  0.981  <0.0001    <10%  1.012  0.990  1.034  0.2916  1.004  1.000  1.008  0.0385    ≥10%  0.983  0.960  1.006  0.1477  1.000  0.996  1.004  0.9499  General hospital   Brand-name drug (%)  1.052  1.042  1.062  <0.0001  0.998  0.997  0.999  0.0064   Pharmaceutical expenditure  0.901  0.889  0.914  <0.0001  1.008  1.006  1.011  <0.0001   Reduction rate    0%  1.099  1.070  1.129  <0.0001  0.975  0.969  0.980  <0.0001    <10%  1.009  0.998  1.020  0.1175  1.000  0.998  1.002  0.8934    ≥10%  0.987  0.976  0.998  0.0240  1.000  0.998  1.003  0.6966  Discussion Drug price reduction policies are considered one way to reduce pharmaceutical expenditure, but unlike the initial intention, they can increase the use of drugs that are not covered by the policy. However, our study shows that drug price policies have impacted pharmaceutical expenditure, but have not changed the pattern of prescription, such as increased use of non-targeted drugs. Previous findings on the impact of cost reduction policies on cost savings are mixed. Some studies, similar to our results, show that cost containment policies are associated with cost savings [12, 23]. However, they do not always lead to the cost savings initially intended [24, 25] because the policy effects on cost savings can be offset by changes to the prescription patterns of healthcare prescribers. In particular, prescription decisions are often affected by illegal marketing by pharmaceutical wholesalers with under-the-table negotiations [26]. This has led to unintended consequences, such as an increase in the use of drugs not subject to the drug policy, in turn raising pharmaceutical expenditure. In this study, while immediate cost savings were noted upon the policy introduction, this cost trend did not change in the long run, even after controlling for the reimbursement restriction effect. One plausible explanation may be physicians’ prescribing decisions. Physicians’ prescribing behavior is affected by both extrinsic and intrinsic factors [17]. Intrinsic factors are associated with physician characteristics such as medical specialization and previous clinical experience. Extrinsic factors include patient factors, pharmaceutical industry factors (e.g. the pressure placed upon healthcare providers to prescribe various medications) and healthcare system-related factors. These factors influence the knowledge of the physician, and based on this, a variety of attitudes are formed and eventually the prescription is decided. This means that the prescribing behavior of physicians may not be changed by any one factor. Indeed, various factors have to be taken into consideration, leading to changes in prescribing behavior. In particular, prescription decisions are more likely to be related to physician factors such as the physician’s belief in the probability of a particular outcome and the personal value attached to these outcomes [27–29]. In our study, the prescription rate increased immediately for drugs in two categories, namely no cost change and a cost decrease of 10%. However, these effects were temporary and did not lead to an increase in monthly prescription rates. Similarly, there was no significant trend change in drug prescribing rates that declined more than 10% since the policy introduction. Further, the use of brand-name drugs did not increase compared with the pre-intervention period and was rather influenced by the introduction of reimbursement restrictions. If only the factors of the pharmaceutical company influence the physician’s prescription, we would have observed that the non-subject prescription rate increased in our study. However, physician prescribing behavior is affected by many factors, and the effects of other policies in this study can affect prescribing behavior. Previous studies on the prescribing behavior of hypertensive drugs since the introduction of the reimbursement restriction and drug price reduction policies showed a similar pattern to our study [10]. Both policies reduced drug costs, but the policy that has a greater impact on drug prescribing behavior was reimbursement restrictions rather than drug price reduction policies. Similarly, in our study, reimbursement policies were effective in prescribing behavior and no unintended effects such as an increased prescription volume of non-target drugs due to drug price reduction policy were observed. This phenomenon shows that drug price reduction policies are a strong policy that can directly affect saving costs, but physician prescribing behavior has not largely affected by drug price reduction policies because it is affected by a variety of factors. According to the subgroup analysis by type of hospital, pharmaceutical expenditure decreased significantly following the introduction of the drug price reduction policy, but the trend was different. In hospitals and general hospitals, the cost trend increased compared with the pre-intervention period. This phenomenon can be related to the different effects of the reimbursement restrictions depending on the type of hospital and patient severity. The introduction of reimbursement restrictions reduced the cost trend compared with previous periods, but these hospitals are used by patients with relatively high severity compared with clinics, meaning that the use of drugs according to the guidelines may be inappropriate. However, because drug use is reimbursed under the clinical guidelines, changes in prescribing behavior may occur later. Indeed, this effect may appear as an increase in pharmaceutical costs after the introduction of the drug price reduction policy. Therefore, a longer observation period would be needed to accurately determine drug cost changes by hospital type. Cost containment policies are important to control the growth in health expenditure and reduce the wastage of health resources. Our results provide evidence of the effects of drug price reduction policies on pharmaceutical expenditure and prescription patterns. First, such policies lead to pharmaceutical cost savings. This finding provides evidence that a drug price reduction policy is important to reduce health expenditure. Second, this study did not find any negative effects of such a policy. Unintended consequences of the cost policy could increase the use of drugs not covered by this policy, but this study did not observe changes in prescription patterns. These results indicate that the drug price reduction policy is effective at reducing costs but does not significantly affect prescribing behavior. Therefore, the drug price reduction policy can be an important strategy for controlling the growth in health expenditure. Our study has several strengths. First, although much research has been carried out on the cost-reducing effects of drug policy, few studies have examined the effect on prescribing patterns. Second, the interrupted time series analysis design is a powerful method for evaluating political change that provides immediate and trend changes. Third, we used NHI cohort data with representative samples, and thus these results provide meaningful evidence for policymakers. Finally, our results are also meaningful to countries considering introducing drug policies to reduce health expenditure. Despite these strengths, our study does have some limitations. First, a patient’s clinical condition can affect prescribing changes, which can lead to cost changes. Second, short-term evaluations can be confusing because the introduction of reimbursement restrictions and of drug price reduction policies are close in time. Therefore, to accurately assess the effect of this policy on cost savings, further studies must account for its long-term effects. Third, the effects of the policy may be different in other drugs because we evaluated only diabetes drugs. Finally, because costs may be affected by other factors, unmeasured factors may affect our results. Conclusion This study provides evidence to policymakers that drug price reduction policies are an effective way in which to control the growth in health expenditure while not negatively affecting changes in prescription patterns. South Korea’s drug price reduction policy did not affect the prescribing behavior of healthcare providers and did not increase the use of drugs not subject to this policy. Although this study did not observe changes in the cost of pharmaceuticals after the policy introduction, further research is needed on the long-term changes in such costs. Supplementary material Supplementary material is available at International Journal for Quality in Health Care online. Acknowledgements None. Funding This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors. References 1 Organizations for Economic Co-operation and Development (OECD). OECD Health Statistics 2016. OECD; 2017. 2 Mongan JJ, Ferris TG, Lee TH. Options for slowing the growth of health care costs. N Engl J Med  2008; 358: 1509. Google Scholar CrossRef Search ADS PubMed  3 Bodenheimer T, Chen E, Bennett HD. Confronting the growing burden of chronic disease: can the US health care workforce do the job? Health Aff  2009; 28: 64– 74. Google Scholar CrossRef Search ADS   4 Ess SM, Schneeweiss S, Szucs TD. European healthcare policies for controlling drug expenditure. Pharmacoeconomics  2003; 21: 89– 103. Google Scholar CrossRef Search ADS PubMed  5 Pita Barros P. Pharmaceutical policies in European countries. In: Pharmaceutical Markets and Insurance Worldwide . Bingley: Emerald Group Publishing Limited, 2010: 3– 27. Google Scholar CrossRef Search ADS   6 Organizations for Economic Co-operation and Development (OECD). Health at a Glance 2015: OECD indicators . Paris: OECD Publishing, 2015. 7 Bae EY, Lee EK. Pharmacoeconomic guidelines and their implementation in the positive list system in South Korea. Value Health  2009; 12: S36– 41. Google Scholar CrossRef Search ADS PubMed  8 Park SE, Lim SH, Choi HW et al.  . Evaluation on the first 2 years of the positive list system in South Korea. Health Policy  2012; 104: 32– 9. Google Scholar CrossRef Search ADS PubMed  9 Ko SH, Oh SJ, Lee HJ et al.  . 2011 Clinica Practice Guideline for Type 2 Diabetes in Korea. J Korean Diabetes  2011; 12: 183– 9. Google Scholar CrossRef Search ADS   10 Yoo K-B, Lee SG, Park S et al.  . Effects of drug price reduction and prescribing restrictions on expenditures and utilisation of antihypertensive drugs in Korea. BMJ open  2015; 5: e006940. Google Scholar CrossRef Search ADS PubMed  11 Carone G, Schwierz C, Xavier A Cost-containment policies in public pharmaceutical spending in the EU. 2012. 12 Lee Y-C, Yang M-C, Huang Y-T et al.  . Impacts of cost containment strategies on pharmaceutical expenditures of the National Health Insurance in Taiwan, 1996–2003. Pharmacoeconomics  2006; 24: 891– 902. Google Scholar CrossRef Search ADS PubMed  13 Moreno-Serra R. The impact of cost-containment policies on health expenditure: evidence from recent OECD experiences. OECD J Budget  2014; 13: 1B. Google Scholar CrossRef Search ADS   14 Moreno-Torres I, Puig-Junoy J, Raya JM. The impact of repeated cost containment policies on pharmaceutical expenditure: experience in Spain. Eur J Health Econ  2011; 12: 563– 73. Google Scholar CrossRef Search ADS PubMed  15 Lambrelli D, O’Donnell O. The impotence of price controls: failed attempts to constrain pharmaceutical expenditures in Greece. Health Policy  2011; 101: 162– 71. Google Scholar CrossRef Search ADS PubMed  16 Kenneally M, Walshe V. Pharmaceutical cost-containment policies and sustainability: recent Irish experience. Value Health  2012; 15: 389– 93. Google Scholar CrossRef Search ADS PubMed  17 Rodrigues AT, Roque F, Falcão A et al.  . Understanding physician antibiotic prescribing behaviour: a systematic review of qualitative studies. Int J Antimicrob Agents  2013; 41: 203– 12. Google Scholar CrossRef Search ADS PubMed  18 Lee J, Lee JS, Park S-H et al.  . Cohort profile: the National Health Insurance Service–National Sample Cohort (NHIS-NSC), South Korea. Int J Epidemiol  2016; 46: e15. doi:10.1093/ije/dyv319. 19 Zeger SL, Liang K-Y, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics  1988; 44: 1049– 60. Google Scholar CrossRef Search ADS PubMed  20 Hanley JA, Negassa A, Forrester JE. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol  2003; 157: 364– 75. Google Scholar CrossRef Search ADS PubMed  21 Wagner AK, Soumerai SB, Zhang F et al.  . Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther  2002; 27: 299– 309. Google Scholar CrossRef Search ADS PubMed  22 Zhang F, Wagner AK, Soumerai SB et al.  . Methods for estimating confidence intervals in interrupted time series analyses of health interventions. J Clin Epidemiol  2009; 62: 143– 8. Google Scholar CrossRef Search ADS PubMed  23 Brekke KR, Grasdal AL, Holmås TH. Regulation and pricing of pharmaceuticals: reference pricing or price cap regulation? Eur Econ Rev  2009; 53: 170– 85. Google Scholar CrossRef Search ADS   24 Chu H-L, Liu S-Z, Romeis JC. Assessing the effects of drug price reduction policies on older people in Taiwan. Health Serv Manag Res  2011; 24: 1– 7. Google Scholar CrossRef Search ADS   25 Kwon H-Y, Hong J-M, Godman B et al.  . Price cuts and drug spending in South Korea: the case of antihyperlipidemic agents. Health Policy  2013; 112: 217– 26. Google Scholar CrossRef Search ADS PubMed  26 Kwon S. Pharmaceutical reform and physician strikes in Korea: separation of drug prescribing and dispensing. Soc Sci Med  2003; 57: 529– 38. Google Scholar CrossRef Search ADS PubMed  27 Lilja J. How physicians choose their drugs. Soc Sci Med  1976; 10: 363– 5. Google Scholar CrossRef Search ADS PubMed  28 Segal R, Hepler CD. Drug choice as a problem-solving process. Med Care  1985; 23: 967– 76. Google Scholar CrossRef Search ADS PubMed  29 Coscelli A. The importance of doctors’ and patients’ preferences in the prescription decision. J Ind Econ  2000; 48: 349– 69. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

International Journal for Quality in Health CareOxford University Press

Published: May 17, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off