Background
According to the World Health Organization (WHO), rational use of drugs means that patients receive medication appropriate to their individual clinical needs in the right dosage for an adequate period of time at the lowest possible cost [
1]. To improve rational drug prescribing, several countries have implemented a prescribing monitoring system. For example, the UK, Spain, and Sweden use national prescribing monitoring and feedback systems. Their prescribing indicators include the percentage of generic prescribing, cost of statin prescriptions, and antibiotic prescribing rates.
However, physicians’ behavior has not necessarily changed since the introduction of these systems. Worldwide, approximately two-thirds of all prescriptions for antibiotics prescribed by physicians are known to be for the treatment of respiratory tract infections (RTIs) [
2]. According to Gulliford et al. [
3] antibiotics are prescribed for 36.5 % of common colds in the UK, 48.7 % in France [
4], 39.7 % in Spain [
5], 16 % in Holland [
2], and 7 % in Sweden [
6]. When Akkerman et al. [
7] surveyed 146 physicians to find determinants of antibiotic overprescribing for sinusitis, tonsillitis, and bronchitis patients, only 50 % of them reported following national guidelines, which indicate that antibiotics should not be prescribed for common colds.
Antibiotic prescription abuse is also a serious issue in South Korea. In 2000, the total antibiotic prescribing rate was 57.9 % [
8]. In 2002, the antibiotic prescribing rate for the common cold was 73.33 %, and the total antibiotic prescribing rate was 42.39 % [
9]. Nonsteroidal anti-inflammatory drugs and corticosteroids also have some issues related to overuse [
10,
11]. As a response, in 2001, South Korea introduced the Prescribing Analysis System (PAS) to promote appropriate prescribing. It is not yet clear, however, that the program has produced sufficient changes in prescribing behavior. For example, in 2013, the average rate of antibiotic prescription for upper respiratory tract infections by private clinics was substantially higher than in tertiary hospitals (43.33/hundred visits vs. 23.99/hundred visits, respectively) [
12].
In this research, we aimed to understand the impact of the PAS on physician behavior, and whether it has the potential to reduce pharmaceutical misuse or abuse and manage pharmaceutical expenditure by reducing unnecessary and inappropriate prescriptions. We attempted to identify factors affecting physicians’ prescribing patterns, examined their perceptions of the PAS, and assessed the system’s strengths and limitations. Our research questions were: 1) What factors do physicians report as affecting prescribing behavior? 2) What are physicians’ perceptions of the PAS? 3) Do physicians perceive the program to be effective? If not, what are the main reasons? and 4) Are there ways to improve the current system to make it more effective?
Methods
Participants
We conducted focus group interviews on May 7–14, 2009. Four medical specialty areas were identified as having the most outpatient visits and prescriptions in the Seoul and Kyungki-do areas: internal medicine, otorhinolaryngology, pediatrics and primary care. We recruited 28 physicians across the specialties and geographical areas, and 27 agreed to participate in the study. They were divided into four focus groups by specialty, each consisting of six to seven physicians. In each focus group, participants’ ages ranged from 39 to 56 with an average age of 40. There were two females and 25 males. The gender distribution of primary care physicians in South Korea is 86 % male and 14 % female. Six participants worked in the Kyungki-do area, and the rest in Seoul (Table
1).
Table 1
Characteristics of participants
No. of participants | 27 | 7 | 7 | 7 | 6 |
Average age (years) | | 49.3 | 42.6 | 46.9 | 46.5 |
Age range (years) | | 44–51 | 39–50 | 41–56 | 44–48 |
Gender: male/female | 25/2 | 6/1 | 7/0 | 7/0 | 5/1 |
Location: Seoul/Kyungki-do | 21/6 | 7/0 | 6/1 | 4/3 | 4/2 |
Focus group interview questions were generated by literature review and discussion among the authors of this study. Participants were asked to reflect on their prescribing behavior and their experiences and perceptions of the PAS. In each focus group, a group discussion lasting one-and-a-half hours was conducted on the following topics: 1) determinants of prescribing behavior; 2) perceptions and attitudes toward the PAS; 3) PAS indicators; 4) feedback methods used by the PAS; and 5) overall opinions and comments about the system.
Interviews were conducted at a research center with videotaping facilities. Two of the authors, who are health policy researchers and familiar with government decision-making processes, moderated the focus group discussions as facilitators. Other researchers participated in the focus group as observers.
The study was approved by the Health Insurance Review and Assessment Service’s (HIRA) Institutional Review Board. Informed consent was obtained from the participants at the beginning of each session. Pseudonyms were used in data analysis to protect participants’ anonymity.
Coding and analysis
Focus group interviews were videotaped and transcribed verbatim by a professional transcriber. All four researchers monitored the interviews and, upon completion of each one, wrote reflective notes and memos. The researchers conducted hour-long sessions for debriefing and discussion, following the completion of each group interview.
The four resulting transcripts were coded by all four authors. We used open-coding of transcripts to identify key words, phrases, and statements. As far as possible, the codes and categories were created for consistency in reflecting emerging ideas, rather than merely describing topics. In the initial line-by-line coding process, researchers tried to observe data analytically without preconceived assumptions. Through iterative reading, we identified repeated and notable words, phrases, and patterns among the participants’ responses. Several common themes emerged from the first stage of ‘interpretive’ reading, while others emerged during, or as a result of, several additional rounds of reading. We then collapsed the codes and categorized them into themes by similarity of meaning within our research questions. Disagreements and differences in describing, synthesizing, and explaining the data were resolved by continuous discussions. Emerging codes, categories, and themes were entered into matrices by specialties to enable comparisons and contrasts within and among groups.
Prescribing Analysis System (PAS)
The South Korean government adopted the PAS in 2001 to reduce misuse and abuse of prescription medicines, and to promote proper usage, by improving each institution's autonomous management of medication. There are indicators to measure performance in three categories, including antibiotic prescription, injection frequencies, and the medication cost per day of use, which have been tracked and recorded since 2001. Later additions include the number of items per prescription (2003), the proportion of high-priced prescriptions (2003), and the duplication rate of NSAIDs (2005) (Table
2). The PAS reports prescribing tendencies of medical care institutions in a comparative format and provides feedback to each institution.
Table 2
Indicators used in the Prescribing Analysis System (PAS)
Injections | Injection prescription rate |
Antibiotics | Antibiotic prescription rate (all diseases) |
Prescription rate for acute upper respiratory infections |
No. of drugs per prescription | No. of drugs per prescription (for all diseases) |
No. of drugs per prescription (respiratory diseases) |
No. of drugs per prescription (musculoskeletal diseases) |
Prescription rate with 6 or more items |
Prescription rate of digestive medicines |
Medication cost per day of use | Medication cost per day of use |
Prescribing expensive medications | Proportion of prescribing of high-priced medicines |
Proportion of cost of high-priced medicines |
NSAIDsa and steroids for osteoarthritis | Duplicate prescription rate for NSAIDs |
Prescription rate for steroids |
Consumers can also see information about PAS results to help them make informed choices when selecting a medical care institution. Since 2006, indicator results, such as the number and the rate of Caesarean sections at each institution, have been openly reported.
Discussion
This study was conducted to understand how South Korean physicians perceive the PAS. A total of 27 participants from internal medicine, otorhinolaryngology, pediatrics, and primary care were interviewed to identify factors determining physicians’ prescribing behavior, and the effects of the PAS on this, together with their response to the program. Like previous studies on factors affecting prescribing behavior [
16‐
19], we found that South Korean physicians’ prescribing behavior is determined by both internal and external factors, including their training and experience [
20], patient expectations and demands [
21], competitive market forces [
22], and promotion and marketing [
23].
Most of the participants in this study had negative perceptions of the PAS, stemming from a belief that the reports were distorted by inadequate data collection or interpretation. This result is similar to findings in the study of Jones et al. [
18], which found that UK physicians seldom use the prescribing analysis and cost data from the Prescription Pricing Authority [
24]. However, Axelsson et al., who studied 603 physicians in Sweden, reported that attitudes to the prescribing guidelines were very positive; 42 % of physicians used the guidelines every day, and 34 % every week [
25]. Most of the participants in this study acknowledged that excessive antibiotic prescribing is problematic. They all agreed that some changes are necessary, but questioned whether the PAS is the right vehicle. Enforcement by government guidelines does not seem to be an effective way to change physician behavior.
Physicians’ negative perceptions of government policy seemed to come from various sources, including concerns about violations of their autonomy and expertise, suspicion about the program’s intention and purpose, distrust of HIRA, dissatisfaction with notification methods, and concerns about the publication of results. Physicians worried that the current PAS would not only promote conformity to a single standard but would also lower the quality of medical services. Moreover, as a highly autonomous and professional group, participants strongly objected to a system which results in them being evaluated and compared with others in terms of prescriptions written, or rates of injections of antibiotics [
26,
27].
Interestingly, we found that physicians’ resistance to perceived violations of their autonomy were often inconsistent and contradictory. While they were very positive about new information or printed materials provided by pharmaceutical representatives, they were less enthusiastic about government guidelines or standards. They appear willing to comply with patient demands as a means of surviving in a competitive market. Physicians’ inclinations to readily accept patient demands in making prescribing decisions has been reported in previous studies [
21]. In an interview with a focus group composed of 24 physicians [
19], it was found that decisions to prescribe antibiotics were made in the context of the physician–patient relationship to prevent potential tension and patients potentially finding other physicians. This implies that an education campaign for patients may improve the effectiveness of the PAS.
Care is necessary when interpreting physicians’ comments that they had no choice but to comply with patients’ requests for antibiotics, injections, or high-priced drugs. A qualitative study conducted with 8 physicians and 42 patients in North Rhine, Germany, showed that physicians tended to over-assume patient demand for antibiotics [
18]. In South Korea, Cho reported that while 73 % of pediatricians thought that patients wanted antibiotics, only 2 % of patients expected them [
28]. Petursson conducted qualitative interviews with 16 physicians in Iceland, and found that unstable physician–patient relationships, due to a lack of continuity of care, was the most important reason for prescribing of antibiotics in situations with low pharmacological indications [
17]. Although several studies have shown that elderly patients in South Korea prefer injections or antibiotics [
29], the claim that physicians had no choice other than to prescribe high-priced drugs, injections, or antibiotics due to patient demand may be somewhat exaggerated.
It is notable that physicians’ negative perceptions of the PAS seemed to be aggravated by suspicion and distrust about the purpose of this program. The most appropriate prescription would be a cost-effective one that provides high quality medical care at low cost. Participants, however, seemed to believe that improvement in the quality of care is only the stated purpose, with cost reduction being the real goal. Physicians expressed a deeply rooted distrust of HIRA, the body in charge of the system. They believed that HIRA’s role is to reduce medication costs by scrutinizing their claims and often rejecting their requests for reimbursements. In fact, the PAS is meant to encourage physicians to improve their practice autonomously, by providing information on their prescribing behavior. PAS is therefore entirely separated from HIRA’s reimbursement review process. Physicians continued to use negative terms such as “threat”, “daunted”, and “pressured” to characterize the PAS, associating the system with their distrust of HIRA.
Since this study was conducted in 2009, there have been several changes in government policy, including P4P (Pay for Performance), which was implemented in 2011; targeted acute myocardial infarction (AMI); and caesarean section delivery. In October 2010, the South Korean government introduced a prescribing incentive scheme to reduce physicians’ prescribing rates, and an online, computerized, prospective drug utilization review (pDUR) has been in operation since December 2010. This seems to have had some effect on overuse and misuse of medicines. Policies that target antibiotic prescription rates for upper respiratory tract infections, overuse of injections, and polypharmacy (the number of medications prescribed ≥6), were implemented in July 2013. Despite this ongoing introduction of new policies, any change in physicians’ prescribing behavior has yet to be shown. Findings from studies that have examined the effectiveness of these policies are not consistent [
30,
31]. Despite several interventions since the PAS program started, no research has shown significant effects on prescribing behavior.
Conclusions
Overall, the results from our study indicate that forming a social consensus on the purpose of the PAS is the most critical prerequisite. Physicians do not agree that they are responsible for cost containment. Participants in our study felt that maintaining the quality of service was the most important issue. They view the PAS as a system that forces physicians to focus on cost rather than quality. Without narrowing such gaps in perceptions and eliminating deeply rooted suspicions, any government efforts will continue to fail.
For HIRA to regain the trust of physicians, it must listen to them more proactively and have open channels for receiving comments and inquiries. Physicians place a higher priority on their professional autonomy and on market competition than on government surveillance. If change is to happen, this cannot be overlooked.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
DSK took part in the design of the study and was moderator of the focus group interviews. GB undertook the analysis of data, interpretation of results, and preparation of the manuscript. MK was actively involved in every step of the study, contributed to the interpretation of the results, and was moderator of the focus group interviews. SYY took part in the data analysis and contributed to the interpretation of the results. All authors read and approved the final manuscript. All authors read and approved the final manuscript.
Funding support
This work was supported by the Health Insurance Review and Assessment Service. This work was partially supported by the National Research Foundation of Korea Grant provided by the Korean Government (NRF-2012S1A3A2033416) to Minah Kang and Green Bae.