Trends in the multiple prescriptions of hypnotic drugs in a university outpatient in Japan

Abstract Aims In Japan, the daily dosage of hypnotic drugs for insomnia treatment is increasing year by year, and over‐dependence on treatment with hypnotic drugs is a major problem. This study aimed to examine the factors related to the elimination of prescriptions of three or more hypnotic drugs within 1 year in our clinic. Methods We conducted two surveys. Survey ① assessed the frequency of prescriptions of three or more hypnotic drugs by retrospectively reviewing the medical records of all patients who visited general and psychiatric outpatient clinics from January 2013 to March 2019. Survey ② assessed changes in prescriptions of hypnotic and psychotropic drugs within the subsequent year by retrospectively reviewing the medical records of all patients prescribed three or more hypnotic drugs who visited neuropsychiatric outpatient clinics multiple times between April 2013 and March 2019. Results The frequency of prescribing three or more hypnotic drugs was six to nine times higher in psychiatry than in other departments. Flunitrazepam and brotizolam were the most common drugs prescribed and had the second lowest discontinuation rate after zolpidem. Conversely, eszopiclone, zopiclone, and suvorexant had the highest discontinuation rates. The success factors for drug reduction were age (odds ratio [OR]: 0.97, p < 0.0037), trazodone addition (OR: 12.86, p < 0.0194) and number of years of psychiatric experience. Conclusions The characteristics and success factors in relation to drug reduction in patients with multiple prescriptions of hypnotic drugs identified in this study may contribute to solving the problem of multiple prescriptions of hypnotic drugs.


| INTRODUC TI ON
Insomnia is associated with generalized daytime malaise and difficulty concentrating, which can easily lead to a decline in psychomotor function and quality of life. 1,2In addition, insomnia has been shown to be a risk factor for not only physical diseases, such as hypertension and diabetes, 3,4 but also psychiatric disorders, such as depression. 5Therefore, the proper treatment of insomnia is extremely important.However, looking at the actual situation of insomnia treatment in Japan, the daily dosage of hypnotic drugs is increasing year by year, 6 and there is a major problem of over-dependence on drug treatment with hypnotic drugs.As a measure to improve this situation, the 2014 revision of medical reimbursements in Japan eliminated continuing psychiatric outpatient support and made it so that guidance fees would not be calculated when more than three hypnotic drugs were included in a single prescription.In addition, the prescription, prescribing, and drug fees were also reduced.Although these measures have resulted in a slight decrease in the rate of prescriptions of five or more hypnotic drugs, from 1.8% before to 0.9% at 6 months after the revision, no change in the downward trend of the rate of multiple prescriptions of two to four hypnotic drugs has been observed since the revision, indicating that the situation has not yet been sufficiently corrected. 7In addition, the characteristics of patients with multiple prescriptions of hypnotic drugs and the factors that determine whether they subsequently succeed in reducing their dosage have not yet been examined.Few previous studies have examined the factors that may have contributed to the resolution of poly-administration of three or more hypnotic drugs in patients on poly-administration therapy.
Therefore, we conducted a retrospective, noninterventional, observational study based on medical records to investigate the factors that help patients prescribed three or more hypnotic drugs to reduce their medications to fewer than three drugs within 1 year at our outpatient clinic.

| Participants
The hypnotic drugs investigated in this study were γ-aminobutyric acid (GABA) receptor agonist sleep medications (eszopiclone, zolpidem, zopiclone, triazolam, flunitrazepam, brotizolam, quazepam, estazolam, rilmazafone, and lormetazepam), ramelteon, and suvorexant, which were used in the hospital as of 2019.First, the change in the frequency of hypnotic prescriptions for three or more drugs in patients diagnosed with insomnia at our outpatient clinic was investigated from April 2013 to March 2019 among all those who visited general departments (all departments except pediatrics and neuropsychiatry) and psychiatric outpatients.
Second, prescriptions for the 1 year after a multidrug prescription were also investigated in psychiatric ambulatory patients who were prescribed a hypnotic drug multiple times over any 3-month period between April 2013 and March 2019 and continued medical examinations for at least 1 year afterwards.In this study, patients who were prescribed hypnotic drugs based on a diagnosis of insomnia in the attending physician's medical record were included in the analysis.

| Study methods
In this study, prescriptions for three or more sleep medications were defined as "multiple prescriptions," and the ratio of the number of multiple prescriptions for hypnotic drugs to the total number of prescriptions was defined as the "multiple prescription rate." Changes in the multiple prescription rate in general medicine and psychiatry between April 2013 and March 2019 were calculated by surveying all patients (excluding pediatric patients) who visited the outpatient department during this period and surveying the status of hypnotic drug prescriptions through the electronic medical record system.For those who had been prescribed multiple medications in psychiatry between April 2013 and March 2019, changes in prescriptions during the subsequent 1 year were investigated retrospectively.We also investigated the percentage of attending physicians who had reduced their patients' prescriptions to two or fewer hypnotic drugs within 1 year.To identify factors that would help eliminate the use of multiple hypnotic drugs, patients for whom the number of hypnotic drugs had been reduced to two or fewer within 1 year were classified into the reduction group, and those who had remained on three or more drugs were classified into the non-reduction group.

| Survey items
The primary survey items were as follows: (1) the multiple prescription rate of hypnotic drugs in general medicine and psychiatry, and (2) change in the use of sleeping pills 1 year later among those prescribed multiple sleeping pills in psychiatry.The secondary items were as follows: sex, age, psychiatric disorder, and psychotropic medications added within 1 year (i.e., anxiolytics, antidepressants, and antipsychotics), attending physician, and number of years of psychiatric experience.

| Statistical analysis
Statistical analysis was performed to compare items such as age, gender, psychiatric disorders, and psychotropic medications added within 1 year between the reduction and non-reduction groups.An unpaired t test was used to compare normal factors, the Mann-Whitney U test to compare non-normal factors, and the chi-square test to compare differences in proportions between the two groups.
In addition, we assessed the patients' clinical variables using univariate and multivariate logistic regression modeling to determine odds ratios (ORs) and 95% confidence intervals (CIs) to determine whether multiple prescriptions for hypnotic drugs had been eliminated.Variables included in the multivariate models had a p ≤ 0.2 in the univariate analysis. 8To test the success rate of reducing hypnotic drugs by the number of years of psychiatric experience, we used receiver operating characteristic (ROC) curve analyses.The chi-square test was used to test differences in the drug reduction ratio between two groups: "sleep specialist or physician specializing in psychopharmacology" and "other psychiatrists."All data analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 4.

| Background characteristics of patients with multiple prescriptions for hypnotic drugs in outpatient psychiatry
Table 1 shows the background characteristics (gender, age, and primary psychiatric illness) of patients with multiple prescriptions of hypnotic drugs in a psychiatric outpatient clinic (84 males [39.8%], 127 females [60.2%], mean age ± standard deviation [SD]: 47.9 ± 15.8 years).

| Psychotropic drugs added within 1 year
The most common psychotropic drugs added within 1 year from the day when three or more hypnotic drugs were prescribed for the first time were quetiapine in 23 patients (10.9%), followed by trazodone F I G U R E 1 Frequency of prescribing three or more hypnotic drugs inside and outside psychiatry. in 14 (6.6%), flunitrazepam in 14 (6.6%), chlorpromazine in 13 (6.2%),olanzapine and eszopiclone in 12 (5.7%),suvorexant in 11 (5.2%), aripiprazole in 10 (4.7%), and etizolam in 9 (4.3%).

| Reduced group and non-reduced drug groups
Table 2 shows the profiles of the patients in the reduction group, for whom the number of prescribed hypnotic drugs had been reduced to two or fewer within 1 year from the date of the first prescription of three or more hypnotic drugs, and in the non-reduction group, for whom three or more hypnotic drugs were still being prescribed.The proportion of the reduction group to the total number of patients (reduction rate of hypnotic drugs) was 54%.

| Age and gender
The mean ages of the patients in the reduction and non-reduction groups were 52.1 ± 15.8 and 44.3 ± 16.4 years, respectively, indicating a significant difference between the two groups, with younger patients less able to reduce their dose (p < 0.0009).On the other hand, no significant difference in the dose reduction rate was seen by gender (p < 0.34).

| Psychotropic drugs added within 1 year
Regarding sleep medications by psychotropic drug added within 1 year, the highest the reduction rate to two drugs or fewer was 86% (12/14 patients) in the trazodone addition group, followed by 77% (10/13 patients) in the chlorpromazine addition group, 70% (7/10 patients) in the aripiprazole addition group, and 61% (14/23 patients) in the quetiapine addition group.

| Factors involved in a successful hypnotic drug reduction
In a univariate logistic regression analysis conducted between the reduction and non-reduction groups on sleeping pills within 1 year, significantly associations were found with age (mean ± SD: reduction group = 52.1 ± 15.8 years vs. non-reduction group = 44.3± 16.4 years), the addition of trazodone [reduction group (n = 12) vs. nonreduction group (n = 2)], and number of years of psychiatric experience (mean ± SD: reduction group = 14.4 ± 6.2 years vs. nonreduction group = 12.3 ± 8.1 years).To adjust for confounding, multivariate logistic regression analysis was performed for "age," "bipolar disorder," "substance dependence," "addition of trazodone," "addition of chlorpromazine," and "number of years of psychiatric experience," the p values for all of which were <0.2 in the univariate logistic regression analysis. 8The results showed that the items with significant differences were age (every 1-year increase: OR: 0.97, p < 0.0037; 95%CI: 0.95-0.99), the addition of trazodone (OR: 12.86, TA B L E 1 Backgrounds of patients with multiple prescriptions of hypnotic drugs.

Age
1][12][13][14][15][16][17] The results were summarized in a Table S1 with adjusted for these factors.The results showed that "age," "the addition of trazodone," and "number of years of psychiatric experience" were still significantly associated factors after adjustment.

| Percentage of hypnotic drugs that had been discontinued after 1 year by drug
The percentage of hypnotic drugs that had been discontinued after 1 year was calculated for each of the multiple prescribed drugs (Figure 2).The hypnotic drug with the highest discontinuation rate after 1 year was eszopiclone (48.2%), followed by zopiclone (45.2%) and suvorexant (44.4%).By contrast, the hypnotic drug with the lowest rate of discontinuation after 1 year was zolpidem (16.7%), followed by flunitrazepam (20.5%) and brotizolam (26.2%).

| Hypnotic drug reduction rate by the attending physician and number of years of psychiatric experience
Figure 3 shows the hypnotic drug reduction rates after 1 year by the clinical physician who was in charge of the patient with multiple hypnotic drug prescriptions.The percentage of patients who had eliminated multiple prescriptions of hypnotic drugs after 1 year varied widely by attending physician, with a maximum of 100% (4/4 patients) and a minimum of 14.3% (1/7 patients).Figure 4 shows the hypnotic drug reduction rates after 1 year according to the attending physician by years of experience and specialty ("sleep specialist or physician TA B L E 2 Profiles of the reduction and non-reduction groups.specializing in psychopharmacology" and "other psychiatrists").The rates of eliminating multidrug prescriptions of hypnotic drugs after 1 year were 61% for "sleep specialist or physician specializing in psychopharmacology" and 45% for "other psychiatrists."The drug reduction success rate was significantly different between the two groups (p < 0.0316).
In addition, the cutoff values for the success rate of reducing hypnotic drugs by the number of years of psychiatric experience were explored using ROC analysis of Youden's Index.The results showed that the cutoff value was 7.5 years, with a sensitivity of 95.6% and a specificity of 39.2% (Figure 5).

| Frequency of prescription of three or more hypnotic drugs
In this study, the frequency of prescribing three or more hypnotic drugs in our psychiatry department was about 6 times higher than that in other departments (excluding the pediatric department) when compared with the number of prescriptions by month.This result was similar to that reported by Mishima et al. 6 who found that the frequency of use of multiple hypnotic drugs in psychiatry reported that rate of concomitant use of hypnotics was not affected after the 2014 revision (from 5% to 4.8% in 1 year), but a decreasing trend was observed after the 2016 revision (from 4.7% to 3.7%). 18However, few reports have discussed these subsequent changes.According to the present study, from 2013 to 2019, the multidrug prescription rate for hypnotic drugs decreased from about 6% to 3.6% in psychiatric outpatients and from about 0.7% to 0.4% in general departments, indicating that revisions to medical fees may have had a substantial impact on the prevention of multidrug prescriptions afterwards.

| Success factors for reducing multiple prescriptions of hypnotic drugs
Regarding the examination of factors related to the success or absence of a reduction in three or more drug prescriptions, the mean age ± SD of the reduction group was 52.1 ± 15.8 years, while that of the non-reduction group was 44.3 ± 16.4 years; this significant difference was also recognized in the factorial analysis.When considering the frequency of hypnotic drug prescriptions, it is important to look at the extent to which polypharmacy is continued, as well as the characteristics of the patient population prone to polypharmacy, because a longer period of prescribing three drugs leads to an increase in the rate of prescribing multiple drugs.In a previous study, Arakawa et al. 19 reported that the prescription rate of hypnotic drugs in psychiatry and psychosomatic medicine F I G U R E 5 ROC curve analysis of the success rate of reducing hypnotic drugs by number of years of psychiatric experience.The AUC was 0.65, sensitivity 95.6%, and specificity 39.2% based on Youden's Index (intersection of dashed lines).
was higher among young to middle-aged patients in their 20s to 50s than among those aged 65 years and older.The results of the present study also suggest that a reduction in the prescription rate for hypnotic drugs is less likely to be achieved in younger patients, which may be related to the high rate of multiple prescriptions among younger patients.Furthermore, the administration of multiple hypnotic drugs, such as benzodiazepines, to older adults has been shown to increase the risk of falls and fractures. 20,21erefore, older adults are often in a situation where hypnotic drugs must be reduced for adverse effects, whereas this is less likely to be the case in younger patients.In other words, it could be speculated that the difference in those incidences may have contributed to our finding that medications are less likely to be reduced in younger patients.
In addition, our results also indicated that the addition of trazodone was a significant success factor for drug reduction (p < 0.0194; 95%CI: 1.511-109.4).Trazodone is a sedating antidepressant widely used in clinical practice to treat insomnia and is the second most commonly prescribed drug for insomnia treatment in the United States after non-benzodiazepine drugs. 22Trazodone is also prescribed in many insomnia treatment settings in Japan.A systematic review 23 of trazodone in 2017 showed that the effect of improving total sleep duration for insomnia was greater than that of quetiapine. 24In addition, the efficacy of trazodone in reducing the dose of medication after 5 weeks among benzodiazepine-dependent patients and as an alternative to benzodiazepines has also been reported. 25,26In accordance with these previous studies, the present results also suggest that the addition of trazodone to patients with more than three prescriptions may promote success in reducing the number of drug prescriptions.However, even in previous studies, only reports of the effectiveness over the short term and in small doses are available; the effectiveness and safety of long-term administration have not yet been established.In addition, in the present study, the starting dose of trazodone, changes in trazodone oral administration, and duration of oral administration (e.g., long-term, short-term) were investigated in detail.These factors need to be reevaluated in additional investigations in the future.

| For discontinuation rates at 1 year by each hypnotic drug
In the present study, the lowest discontinuation rate among all hypnotic drugs after 1 year was 16.7% for zolpidem, an ultra-short-acting sleep medication.The Japanese guidelines for drug withdrawal recommend that patients taking multiple hypnotic drugs should reduce their dosages over a longer period of time compared with those taking a single hypnotic drug, taking rebound insomnia, tolerance formation, and withdrawal symptoms into account, and should start with hypnotic drugs with a shorter half-life.It has also been reported that benzodiazepines with short blood half-lives and high potencies tend to cause withdrawal symptoms and be difficult to discontinue. 25,26In addition, in terms of the pharmacological mechanisms of action, zolpidem is known to have a particularly higher affinity for α 1 subunits of the GABA A receptor α subunit and to be more closely associated with drug-dependent formation than non-benzodiazepine hypnotics, such as eszopiclone and zopiclone. 27,28Therefore, our results indicate that zolpidem may be a sleep medication that is clinically more difficult to discontinue, which suggests that zolpidem prescriptions are a factor in the difficulties associated with the amelioration of multiple drug prescribing situations in patients at risk of developing multiple drug use in the future, and as such, should be administered with caution when prescribed.

| Drug reduction rate by attending physician and number of years of psychiatric experience
The results of this study indicated that although the elimination of multiple prescriptions after 1 year varied by attending physicians, sleep and pharmacology specialists had the highest rate of elimination of multiple prescription.In addition, 7.5 years was identified as the most successful amount of psychiatric experience for reducing multiple prescriptions of hypnotic drugs.However, it is difficult to discuss these results because of the lack of previous studies.In one of the few previous studies, the "Survey of the Actual Conditions of Drug-related Psychiatric Disorders in Psychiatric Medical Facilities throughout Japan" 29 reported that the majority of prescriptions (60.0%) were made by "psychiatrists," 15.0% by "physicians in both psychiatry and physical departments," and 7.1% by "physicians" for patients with sleeping/anxiolytic drug-related disorders, and that many of these patients developed issues in relation to abuse and dependence during treatment for comorbid psychiatric disorders. 30A similar trend was found in a previous study by Matsumoto et al. 31 who pointed out the existence of problems such as "rambling prescribing" by some psychiatrists, as well as "multiple drug combination therapy and mass therapy," and "prescribing without consultation."Although we were unable to examine each attending physician's patient population in detail in the present survey, the fact that the rate of elimination of multiple drug use was higher among attending physicians who were specializing in psychopharmacology clinical research and sleep medicine suggests that a thorough understanding of drug characteristics and attitudes regarding proceeding with treatment with an eye toward a way out may also have a small influence on whether a patient would reduce his or her medication.
As previously mentioned, it is difficult to discuss the years of psychiatric experience because of the lack of studies.However, the curriculum for psychiatrists involves learning general psychiatry within 5 years.After that, the curriculum is designed to enable the mastery of subspecialties such as sleep research and pharmacology over the next 5 years.The results of the present study regarding the effectiveness of 7.5 years of psychiatric experience coincide with the period during which subspecialties are learned.In other words, experience as a psychiatrist with expertise in the field was thought to lead to a better understanding of drug characteristics and increase the likelihood of successful drug reduction.

2 . 1 . 9 3 | RE SULTS 3 . 1 |
Figure 1 shows the results of our primary study on the multiple prescription rate of hypnotic medications in psychiatry and other departments (excluding pediatrics) for the 6-year period from April 2013.The multiple prescription rate of hypnotic drugs in psychiatry trended downward from 5.1% to 5.4% in 2013, from 4.9% to 5.6% in 2014, from 4.8% to 5.4% in 2015, from 5.4% to 6% in 2016, from 4.2% to 4.9% in 2017, and from 3.6% to 4.1% in 2018, whereas the rate in other departments remained flat, at around 0.4-0.6% in all years.

F I G U R E 2 F I G U R E 3
Discontinuation rate after 1 year for each hypnotic drug in outpatient psychiatry.Rate of hypnotic drug reduction by each attending physician.(9.8%-11.3%)was about 6-10 times higher than that in general medicine (1.3%-2.1%)during 2005-2009, indicating a tendency among patients with psychiatric disorders to receive prescriptions for multiple drugs.A previous study of the deterrent effect of the revision of medical fees on the concomitant use of hypnotics

F I G U R E 4
Success rate of reducing hypnotic drugs by years of experience and specialty.*Groups differed significantly at p < 0.05.

Total (211 person) Univariate analysis Multivariate analysis with p < 0.2 as forced entry (excluding others) Reduction group (n = 114) Non-reduction group (n = 97) p-Value Odds ratio 95%CI p-Value
Abbreviations: CI, confidence interval; SD, standard deviation.a Groups differed significantly at p < 0.05.