Pain is defined as an unpleasant sensory and emotional experience associated with actual or potential tissue damage or described in terms of such damage.¹⁾ It is commonly experienced by patients with cancer, which has negative impact on their quality of life.²⁾ The prevalence of pain among patients with cancer ranges from 20% to 50%,³⁾ and approximately 80% of patients with advanced cancer experience moderate to severe cancer pain.⁴⁾ In Korea, more than 50% to 60% of patients with cancer suffer from undertreated cancer pain.⁵⁾ According to the World Health Organization (WHO)’s three-step analgesic ladder for cancer pain relief, opioids play an important role in treating cancer pain.⁶⁾ Recently, these pain management guidelines have been revised to allow for the more liberal prescription of opioids at all stages of pain.^7,8)

The use of opioids is known to be an important indicator of national progress in relieving cancer pain.^9,10) Since the 1990s, the consumption of opioids to treat non-cancer chronic pain and cancer pain has increased,¹¹⁾ resulting in a parallel increase in safety concerns, such as opioid-related disorders and mortalities.¹²⁾ Although the consumption of opioids has increased globally, the average rate of consumption in Korea was lower than that of other developed countries.^13,14) Some reasons for inadequate pain management include ineffective assessment of pain, opioid access and regulations, and the social and moral environment associated with using opioids.⁵⁾ Thus, it is necessary to use opioids safely and properly to avoid 0 pain.

Therefore, this study aimed to provide an update on the trends of the most commonly prescribed strong opioids in Korea from 2002 to 2015 using data from the Korean National Health Insurance Service (NHIS)-National Health Screening Cohort (NHIS-HEALS), stratified by cancer and non-cancer groups.

1. Study population and total prescription

Data on opioid consumption from 2002 to 2015 were extracted from the NHIS-HEALS cohort. The data comprised 514,866 individuals, taken as a random sample from 5.15 million health insurance holders at the end of December 2002. All study participants underwent at least one health screening between January 2002 and December 2003. The claims data include the patients’ medical history records, such as diagnosis codes, prescription details, and health screening results from January 2002 to December 2015. In addition, the data included sociodemographic information, such as age, sex, death, past medical history (malignant neoplasm, hypertension, and diabetes), health behaviors (physical activity, cigarette smoking, and alcohol intake), household income level, and healthcare unit information that the subjects visited. Because this cohort was a closed system and no new individuals were registered after the initial registration, the number of registrants decreased each year due to death or censoring.

This study was approved by the Institutional Review Board of Chungbuk National University Hospital (CBNUH-2019-12-034) and followed the guidelines of the Declaration of Helsinki (1975). The Ethics Committee of NHIS waived the need for informed consent because the data from the NHIS-HEALS were anonymized at all stages, including during data cleaning and statistical analysis. All experiments were conducted in accordance with the relevant guidelines and regulations.

2. Definition and variables

Four commonly prescribed opioids were chosen for analysis, namely morphine, oxycodone, fentanyl, and hydromorphone. We examined the number of opioid prescriptions per 10,000 registrants among patients with and without cancer, which were calculated annually. Data are presented as the number per 10,000 registrants, including individuals receiving more than one opioid prescription in a given year. To compare the trends of opioid prescriptions in patients according to the presence of cancer diagnosis, individuals were divided into cancer and non-cancer groups, based on the main sick code of the tenth edition of the International Classification of Disease (C00–C97). The definition of the cancer group included patients whose main diagnosis code was cancer at least once during the study period.

3. Statistical analysis

The statistical package SAS enterprise guide version 7.1 (SAS Inc., Cary, NC, USA) and R studio version 3.3.3 (R Foundation for Statistical Computing, Vienna, Austria) were used to perform the analyses in this study. To test whether there was a change in opioid prescriptions, a simple linear regression was fitted after the logarithmic transformation, considering the non-zero slopes of the years. Furthermore, a joinpoint regression analysis (version 4.7.0.0; National Cancer Institute, Bethesda, MD, USA) was performed to detect any significant trend changes.¹⁵⁾ The overall trend from 2002 to 2015 was calculated as the average annual percentage change (AAPC). When the trend changed significantly, the trends of shorter time segments were calculated as the annual percentage change (APC). APC and AAPC are marked with 95% confidence intervals (CIs). All P-values were two-sided, and statistical significance was set at P<0.05. All the results were calculated as weight per 10,000 registrants, considering the number of registrants each year. All zero values were replaced with 0.00001 to ensure an accurate calculation.

Table 1 presents the baseline characteristics of the study population. Between 2002 and 2015, there were 18,150 prescriptions of the four aforementioned opioid types. During the study period, the number of newly prescribed opioids increased from 5 to 990 (β coefficient=1.514, P for trend<0.001).

Table 1

Baseline characteristics of study population

Year	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	Exp (β)	P-value
Total number of prescription	514,866	514,531	512,802	509,900	506,189	503,007	498,566	494,192	490,255	487,835	483,421	478,740	472,214	467,132
Newly prescribed	5	14	23	25	16	16	60	295	237	313	448	673	834	990	1.514	<0.001

The annual opioid prescriptions per 10,000 registrants, depending on whether they had a cancer diagnosis, are shown in Supplement 1. In the cancer group, opioid prescriptions increased from 0.18 in 2002 to 62.27 in 2015 (β coefficient=1.623, P for trend<0.001). On the other hand, in the non-cancer group, these increased from 0.25 in 2002 to 45.30 in 2015 (β coefficient=1.648, P for trend<0.001).

Table 2 shows the trends in the rate of annual opioid prescriptions per 10,000 registrants using joinpoint regression analysis. The proportion of the cancer group to the total number of opioid prescriptions increased significantly, with an AAPC of 62.3% (95% CI, 44.7–82.1; P<0.05). The rate of annual opioid prescriptions also increased by 49.9% (95% CI, 22.6–83.2; P<0.05) in the non-cancer group between 2002 and 2015. In particular, there was a dramatic increase in the prescriptions from 2006 to 2009, with an APC of 230.2% (95% CI, 27.0–758.6; P<0.05). Additionally, we found that the number of each opioid prescription per 10,000 registrants increased over 14 years between 2002 and 2015. Of the four types of opioids, oxycodone was the most prescribed in 2015, followed by fentanyl, hydromorphone, and morphine. The greatest increases in the numbers of prescriptions were seen in fentanyl (0.18 to 11.62) and oxycodone (0.34 to 5.83) between 2007 and 2009, respectively. In 2010, the number of prescriptions of oxycodone surpassed that of fentanyl (11.00 vs 10.44 per 10,000 registrants, respectively). The number of hydromorphone prescriptions has steadily increased since its introduction in Korea in 2006, surpassing that of morphine in 2009 (data not shown).

Table 2

Trends in rate of opioid prescriptions per 10,000 registrants during 2002–2015

	Overall trend(2002–2015)		Joinpoint analysis
			Trend 1			Trend 2			Trend 3
	AAPC, % (95% CIs)	Year	APC, % (95% CIs)	Year		APC, % (95% CIs)	Year		APC, % (95% CIs)
Cancer	62.3a (44.7–82.1)		N.A.	N.A.		N.A.	N.A.		N.A.	N.A.
Non-cancer	49.9a (22.6–83.2)		2002–2006	-3.9 (-28.9–30.0)		2006–2009	230.2a (27.0–758.6)		2009–2015	35.8a (15.5–59.6)

AAPC, average annual percentage change; APC, annual percentage change; CIs, confidence intervals; N.A., not available.

^aThe APC or AAPC is significantly different from zero (P<0.05).

In this study, we observed that the proportion of patients with cancer who were prescribed opioids increased in Korea. Furthermore, there was an increase in the number of prescriptions regardless of the presence of cancer diagnosis, implying a shift to more aggressive pain control in recent years.

Based on our findings, there might be some possible factors related to the increasing number of opioid prescriptions in Korea. First, the prevalence of cancer and cancer-related government support has increased. Every year, 10,000–20,000 patients are diagnosed with cancer, and the number of cancer survivors and their 5-year relative survival rate have both increased due to advances in cancer diagnostic tools and pharmaceutical developments.^16,17) In addition, there is improved awareness of cancer pain management and increased government support through hospices and palliative care services.¹⁸⁾ For example, patients with cancer pay only 5% of their total hospital bills due to insurance coverage being recorded with a special code, leading to improvements in active pain management. Second, the number of elderly people has increased as Korea is one of the fastest aging countries. Currently, 13% of the population is aged over 65 years.¹⁶⁾ According to previous studies, age was associated with opioid consumption.^9,19,20) Third, various formulations and dosages of various opioids have been introduced in Korea. Oxycodone and hydromorphone were introduced in Korea in 2001 and 2005, respectively. In addition, the easy administration of transdermal and transmucosal fentanyl and the various formulations of opioids, such as sustained-release oral and parenteral types, might have promoted the increase in opioid prescriptions for pain control, including fentanyl, oxycodone, and hydromorphone (data not shown). These results were consistent with previous studies that reported higher consumption rates for fentanyl, oxycodone, and hydromorphone.^21-23) These results also support our findings of increased opioid prescriptions in the non-cancer group, which may mean that aggressive treatment is also prescribed for non-cancer chronic pain. Opioids are known to be effective against chronic non-cancer diseases, such as osteoarthritis, herpes zoster, and rheumatoid arthritis. However, national guidelines for the use of opioids for non-cancer pain have yet to be established.^17,24) Therefore, further studies may be needed to prevent misuse, addiction, and death caused by excessive use of opioids in non-cancer patients.

This study had several limitations. First, we assumed that the prescription of opioids increased in proportion to the intensity of pain. That is, it was not possible to directly explain whether there was an adequate prescription of opioids for individuals because individual pain was not evaluated. However, a recent study reported that the use of strong opioids increased proportionally with increasing pain severity,²⁵⁾ which may validate our study. Second, the data did not include clinical information on the reason for prescription or compliance with treatment using opioids. Thus, we might have overestimated the number of prescriptions for cancer pain because a patient with cancer may have been prescribed opioids for the relief of non-cancer pain, which we could not rule out. Additionally, it was difficult to discern the number of opioids consumed. Third, because the Korean NHIS-HEALS cohort data did not enroll new registrants during the study period after baseline enrollment, the accumulation of double-counted registrants in each year may have influenced the observed trends of opioid prescriptions. As for this study’s strengths, it was able to investigate long-term trends in the prescription of opioids for cancer pain using a nationally representative cohort.

In summary, the prescription of opioids increased between 2002 and 2015, indicating an advancement in cancer and non-cancer pain control in Korea. Further studies on the prescriptions and misuse of opioids at the individual and population level are needed.

This research was supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI19C0526). Hee-Taik Kang has received research grants from Ministry of Health and Welfare in Korea.

No potential conflict of interest relevant to this article was reported.