1. Abstract
Decreased Inappropriate Antibiotic Use Following a Korean National Policy to Prohibit Medication Dispensing by Physicians
Sylvia Park, PhD; Stephen B. Soumerai, ScD; Alyce S. Adams, PhD; Jonathan A. Finkelstein, MD,MPH; Sunmee Jang, PhD*; Dennis Ross-Degnan, ScD
Department of Ambulatory Care and Prevention, Harvard Medical School, USA;Health Insurance Review Agency, Korea*
Problem Statement: Korean government introduced a new policy in July 2000 that prohibited physicians from dispensing and pharmacists from prescribing medication.
Objectives: To evaluate the impact of the new policy on antibiotic prescribing for cases of viral disease, in which antibiotic prescribing was likely inappropriate, compared to bacterial disease, where antibiotic prescribing could be appropriate; to determine provider factors associated with reductions in inappropriate antibiotic prescribing for viral illness following the policy change.
Design: Retrospective, before/after study.
Setting and Population: National health insurance claims data on monthly episodes of care for patients with viral or bacterial illness, collected for January 2000 and January Nationally representative sample consisted of 50,999 cases from 1,372 primary care clinics.
Intervention: As of July 2000, physicians were prohibited from dispensing and pharmacists were prohibited from prescribing.
Outcome Measures: Rate of antibiotic prescribing; average number of different antibiotics per case.
Results: After the dispensing restriction, antibiotic prescribing declined for both patients with viral illness (from 80.8% to 72.8%, adjusted relative risk (RR)= 0.89, [95% confidence interval: 0.86, 0.91], p<0.0001) and patients with bacterial illness (from 91.6% to 89.7%, adjusted RR= 0.98, [0.97, 0.99], p=0.0171). Reductions in antibiotic prescribing were significantly larger (adjusted RR=0.90, [0.87, 0.93], p<0.0001) for patients with viral illness. The number of different antibiotics prescribed per episode also decreased significantly after the policy, but there were no significant differences in these reductions between viral and bacterial illness. The dispensing restriction also reduced prescribing of non-antibiotic drugs, with no difference by diagnosis. Provider factors found to be associated with reduced inappropriate antibiotic prescribing were young age and practice location in an urban area.
Conclusions: Prohibiting doctors from dispensing drugs reduced prescribing overall and selectively reduced inappropriate antibiotic prescribing for patients with viral diagnoses. Since our findings were based on single observations before and after the policy intervention, further study using longitudinal data is needed to evaluate the long-term effects of such policies.
I am happy to be here to share my research with you. I’ll welcome and appreciate any of your comments about this research. This research is about the impacts of new pharmaceutical policies in Korea on physicians’ antibiotics prescriptions. The two new policies are prohibiting physicians from dispensing drugs and antibiotics utilization monitoring, both of these are newly introduced in Korea recently.

2. Introduction Research on Dispensing Doctors
Dispensing Doctors were found to
- prescribe greater numbers of drugs
- prescribe more antibiotics and injections
- have higher prescribing costs.
Little is known about whether incentives related to dispensing affect the quality of prescribing.
Most of previous research has been cross-sectional.
New Policy in Korea (July, 2000)
Prohibiting doctors from dispensing drugs and pharmacists from prescribing drugs
Antibiotics in Korea
Antibiotics accounted for 20% of ambulatory drug expenditures in 2000.
Korea has very high resistance rates, with 86% of Streptococcus pneumoniae resistant to penicillin in 2001.

3. Objectives
To evaluate the impact of the dispensing restriction policy in Korea on the quantity and quality of physician prescribing
- selectivity in the decrease of antibiotic prescribing in cases with viral illness, in which antibiotic prescribing was likely inappropriate, compared to bacterial illness, where antibiotic prescribing could be appropriate
To investigate provider characteristics related to the decrease of inappropriate antibiotic prescribing in viral illness
The object of this research is to analyze the impact of these two policies on drug prescription.
So the first object is to….. and the second object is to…..

4. Methods Data Collection Jan. 2000 Jan. 2001
NHI monthly claims data(patient level)
monthly episodes
diagnosis, prescription, patient information
Jan. 2000
(6 months before policy)
Jan. 2001
(6 months after policy)
Viral Illness
Common Cold / Upper respiratory tract infection / Bronchiolitis
Bacterial Illness
Penumonia/ Otitis media/ Tonsilitis/ Strep. Sore throat/ Sinusitis/ Urinary tract infection/ Skin and soft tissue infection
This is the research flow.
We collected data and analyzed for these three time points. The first policy that prohibited doctors from dispensing drugs started here.
And the second policy of antibiotics monitoring started here. For these three points, we collected claims data of these two disease groups, viral infectious diseases and bacterial infectious diseases. And we analyzed prescription variables on antibiotics and other drugs. We compared the prescription values for these two disease groups.
To evaluate the impacts of the first policy on prescriptions, we compared the change of prescription values between Jan, 2000 and Jan, 2001 for these two disease groups. And then, to evaluate the impacts of the second policy on prescriptions, we compared the change of prescription values between Jan, 2001 and Jan, 2002 for these two disease groups.
All the analyses were done in clinic level.
Sampling: 10% of all clinics (1476 clinics)  20% of claims with above diagnoses in sampled clinics
Including cases with no commorbidity
From 1372 clinics,
Viral: 18,656(pre), 16,736(post) cases
Bacterial: 7758(pre), 7849(post) cases

5. : Impact of the Policy on Prescribing
Analysis
: Impact of the Policy on Prescribing
Prescription Variables (patient level)
Antibiotics
- antibiotic prescribing
- number of different antibiotics
Non-antibiotics
- gastrointestinal drug prescribing
- number of different non-antibiotic drugs
Generalized Estimating Equations
Y = ß0 + ß1×Policy + ß2×Illness
+ ß3×Policy × Illness
(+ ß4 Patient or provider char. + ß5 … ) + 
- Y: Prescription Variables (patient level)
- X : Policy: after policy=1 / before policy=0
: Illness: viral=1 / bacterial=0
: Policy × illness: Interaction
(different policy effect between illnesses)
: Patient or provider characteristics
: gender, age, location, size, type
- Cluster effect : clinic

6. Analysis
: Provider Characteristics Related to Decrease of Inappropriate Antibiotic Prescribing in Viral Illness
Data Rearrangement
Only clinics having >= 10 cases in each period in viral illness
Aggregating data to clinic level
Adjusting to patient gender, age distribution and diagnosis mix of total sample in Jan. 2001
Multiple Regression
Y= ß0 + ß1 Location + ß2 Type + ß3 Size
(+ ß4 Age + ß5 Gender) + 
Y: Antibiotic prescribing rate
: Average # of different antibiotics per case
(baseline & pre/post change) (clinic level)
X: Location : Urban / Rural
Type : Group / Solo
Size : <= 150 pt / pt / >= 251 pt
Age : <= 39 / / >= 50
Gender : Male / Female

7. Characteristics of Cases
Results
Characteristics of Cases

8. Impact of the Policy on Antibiotic Prescribing
Antibiotic prescribing decreased after the policy in both illness groups.
Reduction of antibiotic prescribing was significantly larger in viral illness than in bacterial illness.

9. Impact of the Policy on the Number of Different Antibiotics
Number of different antibiotics decreased after the policy, with no difference between illness groups.

10. Impact of the Policy on Gastrointestinal Drug Prescribing
Gastrointestinal drug prescribing decreased after the policy, with no difference between illness groups.
Impact of the Policy on the Number of Different Non-antibiotic Drugs
Number of different non-antibiotic drugs decreased after the policy, with no difference between illness groups.

11. Provider Characteristics Related to Decrease of Inappropriate Antibiotic Prescribing in Viral Illness
At baseline, physicians in group practice were less likely to prescribing antibiotics for viral illness than those in solo practice by 14.3 percentage points.
After the policy, younger physicians were more likely to decrease antibiotic polypharmacy than those age 50 and older.
Physicians in urban area were more likely to reduced prescribing per patient by 0.14 antibiotics than those in rural area.

12. Conclusion
Prohibiting doctors from dispensing drugs reduced prescribing overall, both antibiotics and other drugs, and selectively reduced inappropriate antibiotic prescribing for patients with viral diagnosis.
There was no evidence of diagnosis shift or change in the proportion with bacterial or possibly bacterial diagnoses as either primary or secondary diagnosis.
Data on pre-intervention trends supports the validity of our findings: antibiotic prescribing had not decreased before intervention in
Still high rate of antibiotic prescribing for viral illness after policy indicates the need for further targeted interventions.
Further study using longitudinal data is needed to evaluate whether these reductions in prescribing and improvements in quality are maintained.
Removing the financial motivation to prescribe can contribute significantly to quality use of medicines.
In this study, two new policies and their effects were investigated. The first policy is preventing dr from dispensing drugs.
This removed doctors’ economic gains of dispensing drugs. This policy provide the incentive not to increase drug px for economic reason. And it may reduce unnecessary drug prescription.
The result of this research showed decreases of drug prescriptions including anti and other drugs. For these drugs, both of px rate and the average # of different drugs per case decreased.
This result is common with the existing research about prescription behavior of dispensing doctors and non-dispensing doctors.
The new policy also selectively reduced inapp antibiotics px rather than appro anti px. So there were greater decrease of anti px rate in viral disease dropped more than in bacterial disease.
We tested the possibility of diagnosis shift and adding secondary diagnosis. In terms of diagnosis shift, the proportion of bacterial or potentially bacterial disease did not increase and the proportion of viral disease did not decrease after the policy. In terms of commorbidity, the proportion of cases with bacterial secondary disease did not increase after the policy.