Impact of delisting melatonin from a health-system formulary on psychotropic utilization

Abstract

Objective

To characterize the trends of psychotropic utilization before and after the formulary removal of melatonin in Alberta, Canada.

Methods

We conducted a retrospective drug utilization review of patients admitted to inpatient, long-term care, and pediatric units who received psychotropics for insomnia treatment. Utilization patterns were compared between two periods; pre-removal: January to December 2023 and post-removal: October 2024 to March 2025. Data were analyzed using defined daily dose per 1000 patient-days (DDD/1000PD).

Results

A total of 101 pre-removal and 115 post-removal sites were included. There was a 13.5% relative decrease in psychotropic use, from 731.83 to 632.78 DDD/1000PD. Melatonin use decreased from 95.60 to 13.94 DDD/1000PD. In the post-removal period, the psychotropic classes with the highest utilization rates were benzodiazepines (203.26 DDD/1000PD), Z-drugs (167.78 DDD/1000PD), and antidepressants (148.75 DDD/1000PD). The overall use of antidepressants increased by 17.3% while Z-drugs and benzodiazepines decreased by 8.8% and 8.4%, respectively.

Conclusions

The temporal trends of psychotropics varied among zones, sites, and years within the same site. Overall, there was a decrease in psychotropic utilization, largely driven by decreases in the use of Z-drugs and benzodiazepines. The findings should be interpreted with caution, as results may not be generalizable to other institutions.

Hospital-acquired sleeplessness is a complication of inpatient care and can lead to significant morbidity and mortality.^1,2 Sleep duration in hospitalized patients is variable, with children sleeping 1 to 4 hours less than recommended, while adults sleep 1 to 3 hours less. ³ These disturbances can be influenced by environmental factors such as noisy and disruptive environments, patient care interventions, bright lights, and unfamiliar surroundings, as well as non-environmental factors such as acute illness, medication-related adverse effects, and pre-admission insomnia. ⁴ Inadequate sleep can lead to short-term consequences, including increased pain perception, risk of falls, and delirium; while long-term effects include anxiety, cardiovascular disease, and type 2 diabetes mellitus. ⁵ Given the importance of sleep to one’s quality of life, finding safe and effective management strategies for insomnia are valuable. ⁶

While non-pharmacological management strategies are the first-line treatment for insomnia, the evidence for their effectiveness in the hospital setting is conflicting, with a multicomponent approach recommended as the most effective intervention.^7,8 Benzodiazepines and Z-drugs are commonly prescribed pharmacotherapy to inpatients; however, current guidelines do not recommend these agents for older adults due to increased risk of falls, physical dependence, and functional impairment.^9,10 Furthermore, clinicians may consider prescribing alternatives such as low doses of mirtazapine, quetiapine, or trazodone, all of which come with their own safety concerns. ¹¹

In recent years, melatonin has seen increased utilization as a potential alternative for insomnia despite limited evidence supporting its efficacy and safety.^12,13 A growing body of research exists on melatonin prescribing in various practice settings; however, there is currently no literature published on the impact of removing melatonin from a health-system formulary.^14,15 In September 2024, Alberta Health Services (AHS) delisted melatonin to non-formulary status following the Drugs and Therapeutics Committee’s decision based on clinical and operational concerns. As a result, AHS can no longer provide melatonin, but patients may use their home supply while hospitalized.

The trends in and characteristics of psychotropic utilization in Alberta, Canada, are unknown. This study aimed to characterize the trends of psychotropic utilization before and after the formulary removal of melatonin in Alberta, Canada.

Methods

Study design and setting

This retrospective drug utilization review was conducted across all sites and zones within AHS. The pre-removal period, during which melatonin was supplied by AHS, occurred between January 1 to December 31, 2023. The post-removal period, during which melatonin was no longer supplied by AHS, covered October 1, 2024, to March 31, 2025, to characterize the initial temporal trends following the formulary removal of melatonin.

Data were collected from the electronic health record (EHR; Epic Systems Corporation) and the Drug Optimization, Sustainability, and Evaluation (DOSE) platform. Research ethics approval was obtained from the University of Alberta Health Research Ethics Board (Pro00146476) and Administrative Approval from the Northern Alberta Clinical Trials and Research Center (RR19271).

Patient population

All patients admitted to inpatient, long-term care, and pediatric units were identified using the DOSE platform to run a drug utilization report for all patients who received a psychotropic. The medications considered psychotropics were melatonin, amitriptyline, doxepin, mirtazapine, trazodone, diphenhydramine, methotrimeprazine, quetiapine, alprazolam, bromazepam, clobazam, clonazepam, diazepam, lorazepam, midazolam, nitrazepam, oxazepam, temazepam, triazolam, zopiclone, chloral hydrate, and clonidine at a dose that would treat insomnia. Patients were excluded if they were managed in ambulatory settings, including the emergency department or primary care.

Data collection

Aggregated data from each site and zone were collected, including the generic and brand name for each psychotropic, transaction type (i.e., medication administration, inventory transfer), strength, unit of measurement, dosage form, route of administration (i.e., oral, buccal/sublingual, parenteral), quantity dispensed, and the total number of patient-days. This information is available in the form of monthly and annual datasets; however, for the present study, the aggregate annual defined daily dose (DDD) value of each psychotropic is considered.

Data analysis

Data were analyzed using R Statistical Software (v4.5.0: R Core Team 2025) to report descriptive statistics. The unit of measurement used for standardizing psychotropic doses was the World Health Organization (WHO) Collaborating Center for Drug Statistics Methodology system of DDD per 1000 patient-days (DDD/1000PD). DDD/1000PD is the average number of individuals per 1000 patients on medication treatment each day. All DDDs were based on the 2025 version of the WHO Anatomical Therapeutic Chemical Classification System and the DDD Index. ¹⁶ DDD/1000PD were calculated separately by psychotropic, site, and zone, by dividing the total amount of utilization in milligrams by the total number of patient-days.

Results

Summary of site demographic characteristics

A total of 101 pre-removal and 115 post-removal sites were included in this study (Table 1). Most sites were acute care institutions (87.0%). There was a total of 4,969,096 patient-days over the 18-month study period. Sites in the post-removal period were primarily located in the North Zone (31 sites) and Central Zone (28 sites). Additionally, there were 24 sites in the Calgary Zone, 20 in the Edmonton Zone, and 12 in the South Zone.

Table 1.

Demographic characteristics of health-system institutions

Period	Number of sites	Type of health-system institutions
		Teaching hospital	Community and regional hospital	Psychiatric center	Long-term care	Rural hospital with long-term care
Pre-removal	101	9	8	5	2	77
Post-removal	115	9	8	5	13	80

Psychotropic utilization

There was a 13.5% relative decrease in overall psychotropic use between 2023 and 2025, declining from 731.83 to 632.78 DDD/1000PD (Table 2). In the post-removal period, the psychotropic classes with the highest utilization rates were benzodiazepines (203.26 DDD/1000PD), Z-drugs (167.78 DDD/1000PD), and antidepressants (148.75 DDD/1000PD), which accounted for 82.1% of the total psychotropics used.

Table 2.

Utilization of Psychotropics expressed in Defined Daily Dose per 1000 patient-days

Psychotropic	ATC code	Pre-removal		Post-removal
		Number of sites	DDD/1000PD	Number of sites	DDD/1000PD
Antidepressants
Amitriptyline	N06AA09	62	2.28	103	14.26
Doxepin	N06AA12	64	0.68	59	0.77
Mirtazapine	N06AX11	101	81.67	103	85.96
Trazodone	N06AX05	101	42.15	103	47.76
Antihistamines
Diphenhydramine	R06AA02	100	5.80	103	5.48
Antipsychotics
Methotrimeprazine	N05AA02	97	2.40	96	3.58
Quetiapine	N05AH04	101	43.86	103	42.30
Benzodiazepines
Alprazolam	N05BA12	62	4.09	62	3.67
Bromazepam	N05BA08	3	0.09	3	0.14
Clobazam	N05BA09	78	10.60	76	11.50
Clonazepam	N03AE01	100	8.71	99	8.16
Diazepam	N05BA01	97	41.37	96	38.00
Lorazepam	N05BA06	100	76.78	102	76.18
Midazolam	N05CD08	97	57.43	96	47.34
Nitrazepam	N05CD02	29	3.21	28	2.00
Oxazepam	N05BA04	59	1.35	45	1.25
Temazepam	N05CD07	90	17.55	89	14.53
Triazolam	N05CD05	25	0.65	15	0.49
Centrally acting alpha-agonists
Clonidine	N02CX02	95	51.18	96	46.12
Z-drugs
Zopiclone	N05CF01	101	183.92	101	167.78
Other
Chloral hydrate	N05CC01	13	0.46	11	1.57
Melatonin	N05CH01	32	95.60	10	13.94

Abbreviations: DDD/1000PD = defined daily dose per 1000 patient-days; ATC = Anatomical Therapeutic Chemical.

The rates of the five most used psychotropics varied (Figure 1). Zopiclone use decreased from 183.92 to 167.78 DDD/1000PD. Overall use of benzodiazepines also declined, including lorazepam from 76.78 to 76.18 DDD/1000PD and midazolam from 57.43 to 47.34 DDD/1000PD. In contrast, the use of antidepressants increased, including mirtazapine from 81.67 to 85.96 DDD/1000PD and trazodone from 42.15 to 47.76 DDD/1000PD. These five agents represented 67.2% of the total psychotropics used during the post-removal period.

Figure 1.

Utilization rates of the most frequently used psychotropics during the pre- and post-removal periods. Abbreviation: DDD/1000PD = defined daily dose per 1000 patient-days.

Most notably, melatonin use decreased from 95.60 to 13.94 DDD/1000PD (85.4%), resulting in a 68.8% decline in the number of sites using melatonin (n = 32 vs n = 10). At the hospitals where patients supplied their own melatonin, four sites were psychiatric centers (40.0%), three were teaching hospitals (30.0%), and three were rural hospitals with long-term care (30.0%).

There were specific psychotropics which were more likely to be utilized in certain zones (Table 3). Among all psychotropic classes, antidepressants showed the largest increase in percentage change. The use of mirtazapine increased by 32.6% in the Edmonton Zone, while trazodone use increased by 32.7% and 15.1% in the Edmonton and Calgary Zones, respectively. Additionally, the use of lorazepam increased by 15.8% in the Edmonton Zone, while all other psychotropics declined in utilization. Most zones with the highest overall utilization rates were in the North, Central, and South, which comprised of rural facilities with long-term care (91.5%), regional hospitals (7.1%), and one psychiatric center (1.4%).

Table 3.

Utilization of Psychotropics by Zones expressed in Defined Daily Dose per 1000 patient-days

Psychotropic	Pre-removal		Post-removal		Percentage change a
	Zone	DDD/1000PD	Zone	DDD/1000PD
Antidepressants
Mirtazapine	North	132.33	North	112.92	−14.7%
	Edmonton	56.40	Edmonton	74.78	+32.6%
	Central	164.41	Central	143.68	−12.6%
	Calgary	63.41	Calgary	64.62	+1.9%
	South	124.34	South	127.95	+2.9%
Trazodone	North	58.52	North	63.94	+9.3%
	Edmonton	42.12	Edmonton	55.91	+32.7%
	Central	53.90	Central	47.72	−11.5%
	Calgary	27.27	Calgary	31.38	+15.1%
	South	83.70	South	67.88	−18.9%
Benzodiazepines
Lorazepam	North	91.92	North	76.53	−16.7%
	Edmonton	85.00	Edmonton	98.44	+15.8%
	Central	93.28	Central	67.26	−27.9%
	Calgary	57.48	Calgary	52.85	−8.1%
	South	83.61	South	83.34	0.0%
Midazolam	North	79.97	North	57.91	−27.6%
	Edmonton	52.47	Edmonton	48.20	−8.1%
	Central	38.10	Central	37.15	−2.5%
	Calgary	54.83	Calgary	46.61	−15.0%
	South	113.23	South	53.02	−53.2%
Z-drugs
Zopiclone	North	245.12	North	211.86	−13.6%
	Edmonton	134.36	Edmonton	130.61	−2.8%
	Central	334.42	Central	277.88	−16.9%
	Calgary	155.64	Calgary	145.04	−6.8%
	South	293.08	South	257.09	−12.3%
Other
Melatonin	North	0.29	North	0.00	0.0%
	Edmonton	193.13	Edmonton	22.31	−88.4%
	Central	164.03	Central	30.03	−81.7%
	Calgary	0.07	Calgary	4.35	+62.1%
	South	0.00	South	0.00	0.00

Abbreviation: DDD/1000PD = defined daily dose per 1000 patient-days.

^aPercentage Change = (Post-Removal DDD/1000PD/Pre-Removal DDD/1000PD) × 100.0%.

Discussion

This study is among the first conducted in Canada that investigated psychotropic utilization before and after the formulary removal of melatonin. From 2023 to 2025, there was a 13.5% relative decrease in overall psychotropic use, from 731.83 to 632.78 DDD/1000PD, with observed decreases in the use of Z-drugs and benzodiazepines. Additionally, use of antidepressants increased, while melatonin use continued despite its formulary removal.

Traditional psychotropics, including benzodiazepines and Z-drugs, had the highest overall use in our study, despite the presence of deprescribing resources designed to limit their use, particularly among older adults.^10,17 Older adults make up nearly 20.0% of Alberta’s total population, especially in the rural Central and South Zones, which utilized the most psychotropics. ¹⁸ Although there were no differences in medical comorbidities, medications, or socioeconomic status among older adults, a correlational study identified that individuals living in rural areas were more likely to be prescribed psychotropics as compared with those living in urban areas (25.0% vs 15.0%, p = .02). ¹⁹ These findings are consistent with previous research, indicating that rural residents may be at increased risk of inappropriate prescribing, inadequate deprescribing, and inequitable access to health services.^20,21

Antidepressants were among the most utilized psychotropics, with current guidelines recommending their use for managing comorbid psychiatric disorders. ²² Over 20.0% of Canadians suffer from mental illness, with depression and anxiety being among the top concerns for patients seeking care in Canada. ²³ We did not analyze comorbid psychiatric disorders as a factor related to psychotropic utilization which differs from the findings of a Danish drug utilization study that examined psychotropic use over a 20-year period and reported a higher prevalence of depression, anxiety, and substance use disorders among those prescribed psychotropics. ¹³ In our study, antidepressants emerged as a potential alternative to melatonin, irrespective of any comorbid psychiatric disorders.

A key finding of our study is that nearly 15.0% of inpatients supplied their own melatonin in the post-removal period. Melatonin use in hospital has been associated with the management of insomnia, delirium, and as an adjunct to facilitate the discontinuation of benzodiazepines. ²⁴ Research from Europe and the United States have shown limited and inconsistent evidence for melatonin use, especially in pediatric and adult inpatients suffering from insomnia.^24–26 However, despite the formulary removal of melatonin, its continued use in hospital suggests that clinicians may prescribe non-formulary drugs based on factors such as the patient’s history of treatment resistance to previously trialled medications, patient preference, and the clinician’s professional identity as a specialist. ²⁷

We observed an overall decline in psychotropic use with variation among zones, sites, and years within the same site. This may be explained by several factors. In 2019, AHS implemented an integrated EHR system which was launched in nine phases over 5 years, with each site transitioning at different intervals. As these transitions took place, the corresponding data within the DOSE platform for each site updated accordingly; however, it became evident that pre- and post-DOSE data were not directly comparable. This was apparent in the addition of 11 new long-term care sites during the post-removal period. Thus, our reported temporal trends should be interpreted with caution due to the rollout of the integrated EHR, which may limit the generalizability and precision of the results.

The limitations of this study are related to the retrospective study design. Data collection was only as thorough as the information documented within the EHR and DOSE platforms, which is unclear if this is a true reflection of the reported temporal trends. The pre-removal period spanned 12 months, while the post-removal period was limited to 6 months, rather than 1 year. This approach was taken to characterize the initial trends in psychotropic utilization following the formulary removal of melatonin; a process that may have inadvertently been impacted by changes in clinician prescribing preferences, site- and zone-specific treatment practices, and variation in zone populations. Variability among psychotropic prescribing was not characterized; further limiting our ability to determine if the differences between pre- and post-removal periods are within standard variability. Notably, we did not analyze the indication for use as a factor related to psychotropic utilization, and it is possible that psychotropics were used for indications other than insomnia. This is especially the case for antidepressants, antihistamines, antipsychotics, and centrally acting alpha-agonists. Regardless of indication, all psychotropics included are known to be sedating and have associated safety concerns. Additionally, our analyses were based on administrative data; thus, clinical characteristics, including demographic information, treatment duration, and patient outcomes, that are expected to impact psychotropic use were not available. Finally, statistical significance testing was not performed due to variability in temporal trends associated with the implementation of the EHR.

Conclusion

This study revealed variable temporal trends of psychotropic utilization before and after the formulary removal of melatonin. There was a 13.5% relative decrease in overall psychotropic use between 2023 and 2025, largely driven by decreases in the use of Z-drugs and benzodiazepines. An increase in antidepressant use was observed, which accounted for nearly 25.0% of post-removal utilization. These findings should be interpreted with caution due to the rollout of the integrated EHR, which may limit the generalizability and precision of the results. Further investigation is warranted with a longer period and multistate study design, as a clearer understanding of psychotropic utilization following the formulary removal of melatonin in acute care settings may contribute to the improvement of patient care. These findings underscore the need for tailored psychotropic prescribing tools adapted for all inpatient settings.