Browsing by Title

Today, many of the new drugs are poorly soluble in water, which can be a problem in the drug development. Solid dispersion is a formulation technique, which improves the dissolution rate of the drug. However solid dispersions, where the drug is in amorphous form, are often unstable. Because of that, solid dispersions, where the drug is in crystalline form, have been developed. Drug crystallization and factors affecting to the crystallization, such as amount of the polymer, are important to examine to be able to develop better drug products. Different kinds of mathematical models, which describe the kinetics of crystallization, has been developed to help to understand the crystallization event more comprehensively. In this study, the crystallization of the amorphous drug, in the absence of polymer and with a low polymer concentration, was investigated. The crystallization was also examined using a mathematical model designed to determine the kinetics of crystallization in order to find out does it work in this case. A model drug was felodipine and polymers used in this study were HPMCAS-LF and PVP K30. The concentration of polymers in the solid dispersions was 10% and 20%. It was found that a small amount of polymer has a very significant effect on crystallization rate of felodipine. Mathematically defined crystallization rate constant k increased by 13 times, when the amount of PVP was decreased 20 % to 10 %. The polymer concentration also had an effect on nucleation time which is the time before crystallization occurs. For example in the solid dispersion, where PVP concentration was 10 %, the nucleation time was five times slower and 20 % PVP consentration ten times slower than felodipine alone. The work also showed that HPMCAS stabilizes the amorphous state of felodipine better than PVP at 40 ° C / 75% RH conditions. This was observed in both MTDSC-measurements and the polarizing light microscopy. The difference between polymers was thought to be due to weakening of the interactions between PVP and felodipine by the influence of water in humid conditions. However, the different formulations had no significant effect on dissolution characteristics of felodipine. There is a possibility that felodipine crystallizes at the beginning of dissolution. It should be noted that mathematical method tested was not able to model crystallization kinetics properly in this study. So care should be given, when using a mathematical model in the product development.

Parkinson’s disease (PD) is a progressive neurodegenerative disease characterized by the death of nigrostriatal dopaminergic neurons and formation of intraneuronal protein aggregates called Lewy bodies and Lewy neurites. These inclusions consist of a protein called α-synuclein (aSyn) but also of other proteins, lipids and cell organelles. Progressive cell death leads to nonmotor and motor symptoms. Current therapies for PD are symptomatic and do not modify the disease progression. Therefore, there is a need for the development of therapies attenuating the neurodegeneration. The pre-formed fibrils (PFF) model enables studying of aSyn aggregation and mechanisms behind inclusion formation. The PFF model is based on the exogenous aSyn fibrils’ tendency to result in formation of Lewy body -like inclusions when added in cell culture or in animals. Primary neuronal cultures of mice and rats have typically been used to model aSyn aggregation in vitro with the PFF model. Primary neuronal cultures provide practicality and are able to depict relevant features of dopaminergic neurons. To gain insight about the composition of E13.5 primary embryonic mouse midbrain culture and to enable adaptation of an existing protocol to study other cell types, this study identified and quantified several relevant cellular phenotypes in the micro island culture. The cells were fixed on day in vitro (DIV) 8 or DIV 22 and analysis was conducted using fluorescent immunocytochemistry combined with automated image analysis software, CellProfiler. On DIV 8, tyrosine hydroxylase -positive dopaminergic neurons represented 5 % of the total cells in the culture. Neuronal nuclear antigen -positive neurons resulted representing 30 % of the total cells. Gabaergic neurons were identified to be abundant in the culture and certain dopaminergic neurons were identified as immunoreactive for GABA. Choline acetyltransferase -positive cholinergic neurons were also identified to be present in the culture. The number of oligodendrocyte precursors (OPCs) was observed to be significantly smaller than the number of dopaminergic neurons. OPCs represented around 1 % of the culture on DIV 8. Glutaminergic neurons, parvalbumin-positive interneurons, microglia or astrocytes were not identified in the culture on DIV 8. The number of astrocytes was observed to increase as the incubation time was prolonged to DIV 22. Overall these findings provide valuable insights of the composition of cell phenotypes in E13.5 mouse midbrain culture. The results also provide additional validation for suitability of the original protocol to robustly produce midbrain dopaminergic cultures with minimal number of glial cells. Understanding more about the relevance and interplay of different cell phenotypes in PD pathophysiology can provide valuable insight for the development of potential therapeutic strategies.

Diseases of the posterior eye segment, such as age-related macular degeneration (AMD), diabetic retinopathy, diabetic macular edema and glaucoma are the leading cause of blindness worldwide. Current therapy to treat these vision-threatening diseases relies on intravitreal injections to maintain a desired therapeutic drug concentration in the back of the eye. Frequent intravitreal injections are uncomfortable with poor patient compliance and causes major burden to the healthcare systems as well as to the patients. Small molecule drugs have shorter half-life in the vitreous and are eliminated rapidly. This requires frequent intravitreal dosing intervals that are not feasible in the clinical settings. Also, intravitreally injected small molecule drugs are often poorly and non-specifically distributed to the ocular tissues causing adverse effects. To address these issues, controlled and sustained drug delivery systems in the form of drug conjugates are desirable. Conjugating small molecule drugs with enzymatically cleavable peptide linkers increases the residence time in the vitreous. The peptide linker gets cleaved by vitreal enzyme and the released drug reaches the target in retina and choroid. Aim of this thesis was to screen a library of 25 peptide linkers for cleavage in the presence of porcine vitreal enzymes. The peptide linkers were chemically synthesized and the in vitro stability of the peptide linkers were studied in freshly isolated porcine vitreous. Ten time point samples were collected over a period of 45 days and the peptide cleavage in porcine vitreous was assessed by LC-MS method. A TQ-S liquid chromatography-mass spectrometer was used to study the linker cleavage. LC-MS method development for the peptide library was carried out using IntelliStart wizard function. Out of the 25 peptide linker in the library, stability of eight linkers were not included in the LC-MS analysis as a mass method could not be developed. Out of 17 peptide linkers studied, 14 were categorized as fast cleaving linkers (>90% of the linker cleaved in porcine vitreous after 5 h). Three linker peptides; P4, P5 and P25 were categorized as slow cleaving linkers. Conjugating slow cleaving peptide linkers to small molecule drugs will increase the half-life and enhance the duration of drug action upon intravitreal injection. In this study, linkers that are hydrolyzed by specific enzymes present in vitreous or ocular tissues are exploited to investigate their potential for delivering small molecule drugs.

The solubility of a poorly water-soluble drug can be improved by converting the crystalline drug into an amorphous form. However, the amorphous form is metastable due to the higher energy state and recrystallization may occur during storage and dissolution. The amorphous form can be stabilized by forming an amorphous solid dispersion (ASD), where the drug molecules are dispersed to the solid medium, e.g. hydrophilic polymer. One preparation method for amorphous solid dispersions is spray drying, where a solution containing a drug and polymer is converted into small droplets in a drying chamber, in which the solvent evaporates in a hot gas stream and solid particles are formed. The aim of this study was to investigate whether an ASD of a poorly water-soluble drug can be prepared by spray drying using 20:80 (V/V) ethanol-water mixture as a solvent in a feed solution. Indomethacin (γ-polymorph) was used as a model drug and polyvinylpyrrolidone vinyl acetate (PVPVA) as a polymer. The aim was to find a suitable formulation where the drug is in the amorphous form after spray drying and remains in the amorphous form during storage. The ratios of the drug to polymer in the spray-dried formulations were 1:4, 1:6, 1:8, 1:10, 1:12 and 1:16. The study also examined whether a change in one process parameter, pump feed rate, affects the amorphous nature and stability of the resulting spray-dried solid dispersions. Two different pump feed rates, a higher 30% and a lower 15%, were used in the study. X-ray powder diffraction (XRPD) was used to characterize the solid-state of the spray-dried formulations. XRPD measurements were performed immediately after spray drying and on selected time points during storage. Formulations 1:10 at 30% feed rate, 1:12 at both feed rates and 1:16 at 30% feed rate were amorphous after spray drying. In 1:12 (30%) and 1:16 (30%) formulations indomethacin remained in amorphous form over the study periods (22 and 56 days, respectively). In other formulations, indomethacin was found to be in crystalline α-form immediately after spray drying or recrystallization to the α-form occurred during storage. The interaction between indomethacin and PVPVA was studied by surface plasmon resonance spectroscopy (SPR). The aim of the SPR measurements was to understand the interaction between these substances in the feed solution used in spray drying. PVPVA solutions of various concentrations (1%, 0.5%, 0.1% and 0.01%) were injected to the surface of the gold sensor coated with crystalline γ-indomethacin, and the changes in the SPR signal responses were monitored during the interaction. The same measurements were also performed on a pure gold sensor without indomethacin. An interaction between indomethacin and PVPVA can be observed, and based on the measurements, a polymer layer with a thickness of about 1 nm was formed on the surface of the indomethacin sensor regardless of the concentration of the polymer solution. Thus, even a small amount of polymer in solution is sufficient to cover the indomethacin crystals. This may also occur in the feed solution during spray drying, but further studies with SPR are still needed, especially with amorphous indomethacin. This study showed that an ASD of indomethacin and PVPVA can be successfully prepared by spray drying using an aqueous feed solution. Spray-dried 1:12 and 1:16 formulations at a higher pump feed rate were found to be stable enough for further studies. If the spray-dried material is further formulated into a pharmaceutical product, indomethacin must remain in amorphous form throughout the shelf-life of the product to maintain the improved solubility.

Introduction: When people age, the composition of sleep changes and sleep becomes more sensitive to external disturbances, making insomnia also more common. Medication is not the first-line treatment option for insomnia. Benzodiazepines or benzodiazepine receptor agonists for the treatment of insomnia have been in the focus of past studies. The content of the dosing instructions for the supplied medicines has not been studied. The dosing instructions should provide clear instructions on how to dose the medicine prescribed to the patient. The aim of this study was to investigate the content and quality of dosing instructions prescribed for the treatment of insomnia for Finns aged ≥75 years in 2020 based on the prospective reimbursement register data by the Social Insurance Institution of Finland (Kela). Materials and methods: The reimbursed purchases of all medicines by persons aged ≥75 years from 1.1.2020 to 31.12.2020 were selected by ATC code from the medicines data according to the Insomnia: Current Care guidelines. The data was gathered from Kela’s dispensations reimbursed under the National Health Insurance scheme. The dataset consisted of 1,080,843 delivery lines, which were screened, and 328,285 lines were included in the analyses. Dosage instructions were reviewed according to the following predetermined five categories: frequency of use, dose, timing, warnings or remarks, and inappropriate instructions. In addition, 1000 dosing instructions were randomly derived to study the phrasing and appropriateness of the dosing instructions in more detail. Results: In 2020, an average of 3.8 reimbursed hypnotic drug deliveries were made per elderly person. Of the deliveries, 68% were for women. 52% of drug deliveries were partially made later and not by original prescription. In the hypnotic dataset, the three most administered drugs were zopiclone (41%), mirtazapine (34%) and zolpidem (12%). The dosage was prescribed in 98–99% of the dosage instructions. Dosing schedule was reported in 83% of dosing instructions and regularity of use was reported in 57% of them. Only 3-6% of the dosing instructions had comments or warnings. 1–2% of the dosing instructions were vague. The duration or regularity of use was clearly indicated in 5% of the dosing instructions. Only 0.1% of dosage instructions contained instructions for discontinuation or reduction. Discussion: The dose and timing of administration were well reported, but the frequency of use was reported in only about half of the dosing instructions. Only few dosing instructions contained remarks or warnings even though hypnotics are at risk for the elderly. Among the three most administered drugs for the treatment of insomnia were two benzodiazepine receptor agonists, zopiclone and zolpidem. However, they may not be suitable for the elderly according to Beers criteria and their use should be avoided. During 2020, an average of four drug deliveries were made per elderly person for the treatment of insomnia, which may indicate prolonged hypnotics use. In addition, more than half of the deliveries were partially made later and not by original prescription. Thus, several drug packages are prescribed for prescriptions, although the drug-based treatment of insomnia should only be short-lived. Conclusions: There are significant deficiencies in the contents and quality in the dosing instructions for drugs delivered to the elderly for insomnia. Minimum information on dose, timing and duration of use was not found in all dosing instructions in this study. Understandable dosing instructions and the reduction in the amount of medication in the prescription could have a further effect on reducing the long-term use of hypnotics, also increasing the safety of medicine use in the elderly.

Anticholinergic medicines are commonly used to treat e.g. incontinence. These medicines have side effects, which may cause and also exacerbate e.g. dryness of the mouth, increased heart rate, and even cognitive impairment. Older people may be more at risk for these side effects as they may be experiencing similar symptoms as a natural effect of aging, and because they may be using several medicines causing these effects. Older people often have a high medicine burden and also a high disease burden. Measuring anticholinergic effects to change medicine regimens and to reduce the symptoms is difficult as there is no golden standard method. This thesis investigated the published methods available for estimating anticholinergic burden in the literature review part, and used one anticholinergic scoring system, the Anticholinergic Risk Scale, in a cross-sectional study to test the effects of anticholinergics on mortality in 1004 older institutionalised patients from Helsinki area public hospitals. Cross-tabulations and Kruskal-Wallis or Chi square methods were used to detect differences between variables such as nutritional status or certain diagnoses when the patients were stratified according to their anticholinergic use. Cox Proportional Hazard regression, the logrank test and Kaplan-Meier curve were used to investigate the effects of anticholinergics on 5-year all-cause mortality. An in vitro serum assay and seven anticholinergic scoring systems were identified in the literature search. Also, 17 anticholinergic lists were identified, which covered 278 medicines, of which 21 appeared on at least eight of the lists. In the empirical study, the women's (n = 745) mean (± SD) age was 83.35 (± 9.99) years, and they were older than the men (n = 241, mean age ± SD 75.11 ± 11.48, p < 0.001). The 1004 patients (response rate 70 %) were using a mean (± SD) number of 7.1 ± 3.4 regular medicines (range 0-20). 455 patients used no anticholinergics, 363 had some anticholinergic burden (score 1 or 2), and 186 had a high burden, with anticholinergic scores of 3 or more. The mean ARS score (± SD) was 1.2 ± 1.5 (range 0-10). When three anticholinergic lists were compared, all three lists identified only 280/791 of patients who were anticholinergic users according to at least one list. No association was found between anticholinergic medicine use and mortality. There are several methods available for measuring anticholinergic burden, but there is a need for a consensus method. This was highlighted by the lack of agreement on medicines on different lists and when three anticholinergic lists tested identified different patients when compared to each other. Anticholinergic use was common in this frail, older patient sample, but no effect on mortality was shown in this study setting. The cross-sectional nature of the data limits the reliability of the study, and any conclusions beyond older patients in Helsinki area must be done very cautiously. Future research should define anticholinergics better and investigate their possible effect on mortality in a prospective, randomised, and controlled setting.

Multiple Sclerosis (MS) is an incurable autoimmune demyelinating disease affecting the central nervous system (CNS). Although the detailed pathogenesis remains unclear, recent research has highlighted the involvement of B cells. For decades, however, MS research was based on T cell-focused animal models of autoimmune encephalomyelitis (EAE), which do not reflect the involvement of B cells in the pathogenesis. Therefore, B cell-dependent EAE models are hypothesized to allow a better understanding of MS immunohistopathology and may therefore lead to the development of efficient treatments. In our spontaneous relapsing-remitting (RR) EAE model, B cells are recruited from the endogenous repertoire by transgenic myelin oligodendrocyte glycoprotein (MOG) -reactive T cells, causing the development of EAE in 3–4-month-old mice. Interestingly, MOG-specific antibodies are present long before actual onset of clinical disease and can be detected already in 5-week-old RR mice and disease development in RR mice is dependent both on the presence of (presumably MOG-specific) B cells as well as on stimuli provided by intestinal microbiota. Firstly, we evaluated the broader usability of induced germinal center cell (iGB) culture as a model for B cell repertoire studies. Then, by using iGB culture, we studied whether MOG-specific B cells are present in secondary lymphoid organs of younger than 4-week-old and germ-free RR mice. Finally, this study aimed to investigate whether the repertoire of MOG-specific B cells undergoes significant qualitative changes from young healthy mice to older acutely sick RR mice, and whether at the time of disease onset the recruited MOG-specific B cells expand and mature in the cervical nodes (cLN) or in the CNS. To do so, following the hosting-lab’s previous single-cell RNA sequencing (scRNA-seq) of B cells derived from cLN of 5-week-old RR mice, we performed the scRNA-seq of B cells from CNS, spleen, and cLN of acutely sick RR EAE mice. We demonstrated that iGB culture is an unsuitable tool to expand pre-activated B cells, and hence, in our hands it was inappropriate for repertoire studies. However, iGB culture proved to be useful for screening different organs for MOG-specific B lymphocytes, and we found that anti-MOG antibodies were firstly detected in 3-4-week-old RR mice, and MOG-specific B cells were present also in germ-free RR mice. Our scRNA-seq results revealed many highly expanded MOG-specific B cell clonotypes in acutely sick RR mice. Moreover, the B cell repertoire of sick RR mice was more diverse, including IgG1, IgM, IgG2b, IgG2c, and IgG3 isotypes, compared to healthy 5-week-old RR mice that had only IgG1 or IgM isotypes. Two-thirds of the expanded clonotypes were primarily detected in the CNS in sick RR mice, indicating that clonotypes develop further and continue isotype switching within the CNS. We also detected more somatic mutation in the variable region of expanded clones of sick RR mice compared to 5-week-old RR mice. The results of this study clearly show an antigen-driven evolution of the MOG-specific B cell repertoire from healthy young to acutely sick RR mice, which seems to occur mainly in CNS itself. In contrast, cLN are the major initial priming site of MOG-specific B cells in healthy RR mice, even under germ-free conditions. This suggests that commensal microbiota is not required for initial recruitment of MOG-specific B cells, but for the development of EAE. To further validate our encouraging scRNA-Seq results, it is necessary, in future experiments, to confirm the MOG-specificity of expanded clonotypes.

Medication safety is an important target of development in health and social services systems internationally. Medication errors are one of the biggest risk factors in medication safety. Majority of the medication incidents could be avoided by improving the medication treatment process. Patient safety incident reporting systems enable health and social services to collect systematic data from risk factors within the medication treatment process. This study was conducted as a retrospective registry-based study where medication incidents that occurred in health and social care units reported by community pharmacies to the incident reporting system HaiPro from 21st of September 2021 to 31st of October 2022 were analysed. Cases that did not meet the criteria for this study (n=55) were removed from the original data (n=3841). If needed, the nature and type of the reported error were corrected. A descriptive quantitative analysis was conducted for the final data (n=3786) using Microsoft Excel. The number, natures, types, observers, and prescription types of medication errors were investigated from the data. In addition, the most common groups of medicinal substance and high risk medicines were identified. A qualitive content analysis was performed to near miss cases involving high-risk medications (n=446) using the Atlas.ti program. Interventions, measures following the interventions and risks prevented by the measures were identified from the open description in the incident reports. The qualitative analysis was performed as an abductive content analysis. Of the medication errors included in the study (n=3786) 91% were detected by community pharmacies and the majority (68%) of the reported incidents were near misses. Most (96%) of the safety incidents (n=3786) were associated with the patient’s medication treatment and had occurred mostly during the prescribing process (92%). As a result from the prescribing errors, patients were most commonly prescribed wrong dose or strength of the medicine (26%) or the prescription lacked SIC marking (26%). High-risk medications occurred in 16% (n=591) of the incidents (n=3786). Most frequently detected high-risk medications were opioids (35 %). Three quarters (76 %) of safety incidents associated with high-risk medications were near misses (n=446). The majority (92 %) of interventions (n=471) made to prevent safety incidents associated with high-risk medications were made by community pharmacies. The most frequent intervention was community pharmacies contacting the doctor. Based on the HaiPro incident reports made about medication errors in health and social care units reported by community pharmacies, it can be concluded that community pharmacies are a central barrier in primary care medication treatment process. Community pharmacies detect and report medication errors that have occurred in other health and social care units. Safety incidents reported by pharmacies systematically accumulate important information that can be used in the development of medication safety in primary care at a unit, wellbeing services county and national levels.

The pharmacy operations are strictly regulated in Finland and the operation of a pharmacy business requires a licence. Number of community pharmacies has stayed quite steady for the past 10 years. At the end of the year 2019 there were 817 pharmacies or their subsidiaries in Finland. The number of pharmacies is expected to increase, since 29 new pharmacies has been established since 2016. The inspection of pharmacies is a part of the legal duties of the Finnish Medicines Agency. In Finland, pharmacies are inspected based on a risk assessment, as often as it is necessary to ensure appropriate operations of a pharmacy. During a pharmacy inspection, the focus is on operations that are critical to drug safety and medication safety. The aim of an inspection is to make sure that pharmacy operations comply with the regulations. There are only few studies made on remote inspection of the pharmacies. The Finnish Medicines Agency hasn’t made remote pharmacy inspections before. The aim of this study was to develop, validate and test a questionnaire, which could be used to inspect community pharmacies remotely and to develop and test a process for remote inspections. The study was done in two parts. In the first part, a draft of the questionnaire was developed by studying the regulations and laws regulating the operations of a pharmacy and by using a content analysis. The material for the content analysis was a pre-inspection questionnaire form, answers to the pre-inspection questionnaires and defect lists of the inspection reports of those pharmacies (n=37), which had answered to the pre-inspection questionnaire before pharmacy inspection in 2019. Content of the pre-inspection questionnaire and the answers of pre-inspection questionnaire were compared to the content of the defect lists of inspection reports. The aim of the comparison was to find out how the existing pre-inspection questionnaire could be utilized when developing the questionnaire for the remote pharmacy inspections. In addition, the listed defects of the inspection reports were categorized to explore what were the most common defects observed during pharmacy inspections. In the second part of the study, the content of the developed questionnaire was validated by using a three round modified Delphi survey. Seven experts with good knowledge of the pharmacy inspections were chosen to the Delphi panel. The aim of the Delphi rounds was to achieve full consensus among the experts about the content of the questionnaire. Alongside the Delphi rounds, a process to remote inspect a pharmacy was developed. The remote inspection questionnaire and the process were tested internally in the Finnish Medicines Agency at the end of the second stage of study. The draft of the questionnaire included 15 sections and 164 questions. Based on the comments received during the Delphi rounds, the content of the questionnaire was modified. On the third Delphi round a full consensus of the content of the questionnaire was achieved among the experts. The final questionnaire for the pharmacy remote inspection included 14 sections and 184 questions. The process of the remote pharmacy inspection follows the procedure of an on-site pharmacy inspection. In the internal test, the process of the remote inspection was found to be a good way to inspect pharmacies remotely. The remote inspection process is a new way to inspect pharmacies. With the remote inspection, it is possible to find out the most common defects on the pharmacy operations by using the questionnaire and contact calls. The remote inspection questionnaire and the process need to be further tested to ensure that the process is optimal from the perspective of the authority and the pharmacies.

Background and objectives: Documenting and processing of dispensing errors at both organizational and national levels is one of the basic preconditions for effective medication risk management. Since the most recently accomplished national register research of dispensing errors in Finland, there have been several changes in the medication dispensing process that advance medication safety. Thus, the previous study does not provide an up-to-date picture of the current situation. The primary objective of this study was to find out the trends in dispensing errors that were reported to the Finnish Pharmacy Association's registry of dispensing errors in 2015–2019. The secondary objective was to identify risk factors expository to dispensing errors in the dispensing process and to review the measures utilized by community pharmacies to prevent dispensing errors. Materials and methods: The retrospective registry study, in which the register of dispensing errors maintained by the Finnish Association of Pharmacists for the period from 1 January 2015 to 31 December 2019, was analyzed. Cases that did not fulfil the definition of a dispensing error (n=829) were removed from the original data (n=17763). In addition, clear errors (n=2130) were corrected in the data and cases (n=499) that were initially insufficiently entered in the register were added. 17433 dispensing errors were included in the study. The data was analyzed using Microsoft Excel. The number, qualities, prescription types, observers, therapeutic harms and contributory factors of the dispensing errors were investigated in the data. The most common groups of medicinal substance, high-alert medications and risk factors in the medication dispensing process were identified in the data. In addition, interventions reported by community pharmacies to prevent dispensing errors were collected from the data. Results: The number of cases reported to the dispensing error register has decreased annually (2015 n=3913, 2016 n=3795, 2017 n=3708, 2018 n=3578, 2019 n=2439). The most common types of dispensing errors are incorrect strength (51 % of all the reported dispensing errors) and incorrect quantity or package size (14 %). Slightly more than a half (51 %) of the reported dispensing errors were noticed by medicine users. The percentage of electronic prescriptions in dispensing errors has increased and is clearly the most common prescription type in dispensing errors (2015: 79 %, 2016: 84 %, 2017: 93 %, 2018: 96 %, 2019: 95 %). The majority of dispensing errors occurred with cardiovascular medicines (29 %) and medicines affecting the nervous system (26 %). 7 % of dispensing errors caused therapeutic harm to the medicine user. As a result of dispensing errors, 21 medicine users were hospitalized. 13 % of dispensing errors occurred with high-alert medications (n=2244). The high-alert medications were involved in one-third (n=7) of dispensing errors that led to hospitalization. Factors related to the employee (25 %), similar packaging (19 %), and similar medicine name (15%) were most commonly considered to be the main contributory factors for the occurrence of the dispensing errors. The risk factors identified in the medicine dispensing process were related to the pharmacy system, the characteristics of the prescription, the storage method of the medicine and the characteristics of the medicine packaging. In the automated dose dispensing process, the risk of dispensing error increased if changes had to be made to the dose dispensing order. The risk factors for automated dose dispensing were related to the pharmacy system and the characteristics of the prescription. The community pharmacies had mentioned taking measures to prevent dispensing errors in one-fifth (21 %) of the reported cases. In addition to developing their own operations, community pharmacies saw cooperation with other healthcare professionals as an important factor in preventing medication errors. In addition, community pharmacies reported exposing properties for dispensing errors of pharmaceutical products and systems to pharmaceutical companies and providers of pharmacy systems and automated dose dispensing. Conclusions: Trends, risk factors of the dispensing process and interventions to prevent dispensing errors can be identified in the dispensing errors reported to the Finnish Association of Pharmacists’ dispensing error registry. The dispensing error register provides valuable information on dispensing errors at the national level, but it is no longer able to fulfil completely the current medication safety needs. In the future, the role of the pharmacy as a promoter of medication safety should be perceived as more comprehensive. In the development of medication safety, special attention should be paid to the risk factors of the dispensing process, the high-alert medications and to new risks arising from the increase of electronic prescriptions and automated dose dispensing. In addition, cooperation between pharmacies and other healthcare professionals and the medication safety culture of pharmacies should be further strengthened.