The protocol was registered on the Open Science Framework platform on 12 December 2023 (https://osf.io/rykwt). This review’s report followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines checklist (Suppl. material 1) (Page et al. 2021).

The search for studies relevant to the research question was initiated in the PubMed scientific database between 01 December 2023 and 26 January 2024, regardless of the publication year or publication type. Boolean operators (AND, OR) and MeSH search strategies were used to identify relevant articles. The detailed search strategy can be found in Suppl. material 2. All identified articles were exported to the Rayyan systematic review application (https://rayyan.qcri.org/) and Mendeley Reference Management Software (https://www.mendeley.com/reference-management/reference-manager) to screen and remove duplicate papers. Two authors (IH, PP) reviewed the selected articles to determine their relevance to the inclusion and exclusion criteria and consolidated the evidence. In cases of disagreement, a third reviewer (MK) was consulted. A flow diagram was generated to systematize the data.

Articles were included based on the following criteria: 1) written in English; 2) patients with RDs who had received or were currently receiving treatment for their primary condition; and 3) studies addressing the research questions and reporting data on МА. Studies were excluded if they 1) did not report MA, 2) did not provide summarized results, 3) did not cover patients with RD, or 4) did not have full text available or sufficient information in the abstract.

Our search was conducted using routine practice approaches to the literature searches such as PICO (Population, Intervention, Comparison, Outcomes):

• Population: Patients diagnosed with a RD (adults and children), who have received or currently receive therapy prescribed for their illness;
• Intervention: MA issues and predictors among patients with RDs;
• Comparison: Watchful waiting without any interventions for the assessment and/or improvement of MA in the selected group of patients;
• Outcome(s): The primary outcome measures are the type of methods for assessment of MA and corresponding levels of MA among the selected population after diagnosing and prescribing therapy. The secondary outcomes include (a) type of risk factors, barriers, and predictors of MNA; (b) types of MAEIs; (c) outcomes associated with MNA such as hospital readmission rate, worsening of the condition, death, increasing direct and indirect costs, and other relevant consequences.

The methodological quality of all included studies was assessed using the Mixed Methods Appraisal Tool [MMAT] 2018, except for systematic reviews (SRs) (n = 4), for which the Joanna Briggs Institute (JBI) meta-aggregative approach was used (Aromataris et al. 2015; Quan et al. 2018). The MMAT is a checklist used to evaluate the quality of various study designs, including quantitative, qualitative, and mixed-methods studies, focusing on the latter. In our SR, we utilized the MMAT to assess how effectively the included studies addressed potential biases that could affect the interpretation of the results. IH evaluated all studies, while MK and PP assessed 20% of the studies each. The JBI critical appraisal tool, used to determine the quality of the included SRs, contains eleven questions with four rating choices (i.e., yes, no, unclear, not applicable) for each question. Two appraisers (IH and MK) subjected these studies to rigorous appraisal. Detailed quality appraisal information is presented in Suppl. materials 3, 4.

One investigator (IH) entered the study data into the extraction table, while a second investigator (MK) verified the information for accuracy and completeness. The following data were extracted: author, year, country of origin, type of study, RD, MA aspect (assessment, improvement, factors), and results and conclusions. The characteristics of the 49 included studies are presented in Suppl. material 5.

The initial search across PubMed identified 308 results. One duplicate sample was found and removed through the Rayyan and Mendeley platforms (n = 1). Then, 307 studies were screened by title and abstract, and 212 records were removed. Ninety-five records were retained for full-text review. Finally, 29 articles from the PubMed database were included for analysis. The 20 articles included from our previous review (a scoping review published in 2023 in the International Journal of Clinical Pharmacy (PCNE abstract number 576) (Kamusheva et al. 2023) remain highly relevant and provide foundational insights into MA. As rare disease research often evolves slowly due to limited patient populations and available treatment, these studies contribute critical data and context to the overall assessment of MA. Fig. 1 visualizes a PRISMA flowchart of the selection process.

Figure 1. 

PRISMA flow diagram.

The risk of bias assessment results for each of the 49 studies is reported in Suppl. materials 3, 4. All included papers, appraised with the MMAT tool, had clear research questions, and the collected data adequately addressed them. However, some studies indicated unclear reporting regarding sampling and bias. This was not considered a rational basis for their exclusion. The included SRs, appraised with the JBI tool, show clearly and explicitly stated review questions with appropriate inclusion criteria, search strategies, and sources used to search for studies.

The studies were published between 2001 and 2023, with most published in 2022 (n = 7) and 2021 (n = 8) (Fig. 2). Of the 49 included studies, 6 were conducted in the USA, 4 in Spain, 3 in France, Italy, Brazil, 2 in Japan, Turkey, Germany, and the United Kingdom (UK), and 1 each in China, Sweden, Denmark, Switzerland, and Bulgaria. Additionally, 6 studies were multicentered, and 11 were systematic/literature reviews or conducted via telephone or social media. The detailed study characteristics are presented in Suppl. material 5.

Figure 2. 

Number of publications by year 2001–2023.

Medication adherence rate was assessed using the Morisky Medication Adherence Scale (MMAS) in most of the selected studies (n = 6), as wide variability (21%–62.7%) was demonstrated among 2,165 patients with RD (Dwyer 2014; Dzemaili et al. 2017; Ivarsson et al. 2018; Vitturi et al. 2020; Karaca et al. 2022; García-Muñoz et al. 2023). Some studies (n = 5) used large databases and registries to assess MA in 4,244 patients (Farach et al. 2012; Johansen et al. 2016; Mehta et al. 2021; Oro-Ayude et al. 2022; Saito et al. 2022). Seven studies included questionnaires with various scales to determine MA levels. These studies involved 1462 patients, 223 physicians, and 22 caregivers (De Moerloose et al. 2008; Alexander et al. 2017; Vatier et al. 2019; Walburn et al. 2019; Aycan et al. 2021; Torregrosa Prats et al. 2021; Hollmen et al. 2023). Moreover, five studies used or described various approaches to measure MA (Uomo et al. 2001; Sultan et al. 2018; Kamusheva et al. 2020; González-Lamuño et al. 2021; Jacquelet et al. 2021). The MA rate of each study is reported in Table 1.

The WHO categorizes factors leading to MNA into five main categories (World Health Organization 2003) as factors for MNA among RD patients, as shown in Fig. 3.

Figure 3. 

Factors leading to medication nonadherence among patients with rare diseases.

Financial difficulties experienced by acromegaly patients were the reason for MNA in 89% of them (Kamusheva et al. 2021). A lack of health insurance was defined as a factor for refusing medication therapy, leading to multiple hospital readmissions (Shenoy et al. 2015).

One identified factor contributing to MNA was a lack of systemic education and information about the disease among patients (Ivarsson et al. 2018; Vitturi et al. 2020; Hollmen et al. 2023). In one study, 52 out of 55 women with congenital hypogonadotropic hypogonadism (CHH) reported using the internet for information, with 46 involved in online communities and social media; 24 expressed a need for further information and support (Dzemaili et al. 2017).

Late diagnosis, which causes a delay in finding expert help, was revealed as a reason for MNA in 13 out of 55 patients with CHH (Dzemaili et al. 2017). The time spent with the RD also affected MA; the longer the patients were treated, the less they adhered to therapy (Vitturi et al. 2020; Saito et al. 2022; Wang et al. 2022). The clinical phenotype can also lead to variations in the MA rate: patients suffering from Wilson’s disease with high MA mainly have a neurological form (75.9%), while 26.7% of patients with an asymptomatic form have low MA (Jacquelet et al. 2021).

Side effects seem to be a common barrier to MA. Local side effects (pain, redness, skin infections or irritation, skin induration at the site of the injection) (Cappellini and Taher 2009; Vatier et al. 2019; Kamusheva et al. 2020; Pogna et al. 2021; Orso et al. 2022), as well as photosensitivity (Hollmen et al. 2023) and gastrointestinal issues (Pogna et al. 2021; Acquadro et al. 2022; Hollmen et al. 2023) are frequently mentioned. Drug toxicity (Costa et al. 2012) and concerns about potential adverse effects (Ivarsson et al. 2018) are also significant problems impacting adherence. For example, patients often discontinue their prescribed therapy without consulting their physician after experiencing side effects (Morrow 2019; Ueda et al. 2019).

Additionally, symptom improvement or disappearance can lead to treatment abandonment (De Moerloose et al. 2008; Kamusheva et al. 2021). The need for multiple daily or weekly medications and the large number of pills required have been associated with low MA rates, which are often below the 50% observed in other chronic diseases (Cappellini and Taher 2009; Thomson et al. 2011; Vitturi et al. 2020; Torregrosa Prats et al. 2021; Hollmen et al. 2023).

Other therapy-related issues were drug administration difficulties (Vatier et al. 2019; Kamusheva et al. 2020), the unavailability of medications on the market due to export issues, cold storage requirements, and increased distance to treatment centers (Aycan et al. 2021; Jacquelet et al. 2021; Hollmen et al. 2023).

Forgetfulness regarding medication intake (De Moerloose et al. 2008; Jabbour et al. 2012; Orso et al. 2022; Hollmen et al. 2023), adolescence (De Moerloose et al. 2008; Jabbour et al. 2012; Malik et al. 2015; Sultan et al. 2018; Medina and Román 2020; González-Lamuño et al. 2021), older age and male gender (Jabbour et al. 2012; Fentiman 2018; Wang et al. 2022; Hollmen et al. 2023), a low level of education (Sultan et al. 2018; Orso et al. 2022; Wang et al. 2022), and unpleasant organoleptic characteristics of the therapy (Oro-Ayude et al. 2022) were factors responsible for low MA. Moreover, anxiety and depressive episodes are prevalent in nonadherent patients (Jabbour et al. 2012; Dwyer 2014; Malik et al. 2015; Dzemaili et al. 2017; Jacquelet et al. 2021).

Educating patients about what to expect from their prescribed therapy, possible side effects, and how to prevent or manage them has emerged as one of the most important MAEIs (De Moerloose et al. 2008; Jabbour et al. 2012; Malik et al. 2015; Ivarsson et al. 2018; Kamusheva et al. 2021; Nguyen et al. 2023). Two of the studies highlighted that implementing practical guidelines is essential for managing the information patients receive about their therapy (Ivarsson et al. 2018; Kamusheva et al. 2021). Several studies have shown the significant impact of interventions and policies implemented by the healthcare system, like collaborative care models, involving close coordination between healthcare professionals (HCPs) and patients (De Moerloose et al. 2008; Van Groenendael et al. 2015; Morrow 2019; Vaisbich et al. 2019; Nguyen et al. 2023). Additionally, technological solutions, such as medication reminder apps, devices that permit medications to be released at the right time, and electronic diaries, contributed to fostering adherence by leveraging advancements in digital health (Farach et al. 2012; Choi 2015; Kamusheva et al. 2020; Rüther et al. 2021; Orso et al. 2022). Behavioral interventions such as motivational interviewing with specialists from multidisciplinary teams seem effective in improving MA (Malik et al. 2015; Alexander et al. 2017; Vaisbich et al. 2019; González-Lamuño et al. 2021). Other MAEIs addressed patient preferences about the type of therapy, administration frequency, treatment duration, and reduction in patient cost burden (De Moerloose et al. 2008; Cappellini and Taher 2009; Thomson et al. 2011; Pogna et al. 2021; Acquadro et al. 2022; Orso et al. 2022). The different MAEIs are summarized in Fig. 4.

Figure 4. 

Factors for enhancing medication adherence among patients with rare diseases.

Non-adherence to therapy results in poor patient outcomes, increasing healthcare service utilization and overall costs. In our study, MNA was one of the factors contributing to frequent hospital readmissions or prolonged stays (Shenoy et al. 2015; Chaudhary et al. 2017; Ueda et al. 2019; Vitturi et al. 2020; Dong et al. 2022). The condition and quality of life of non-compliant with the treatment patients could worsen, and new previously unexperienced symptoms may appear (Gallin et al. 2003; Chaudhary et al. 2017; Westermann et al. 2019; Vitturi et al. 2020; Dong et al. 2022). Strict adherence to therapy can result in complete remission of RD or its complications, opposing previous claims (Westermann et al. 2019). MNA led to greater resource utilization and costs (Jabbour et al. 2012; Chen 2016; Lashilola et al. 2023). A recent study shows that improving MA in patients with kidney failure would result in savings of approximately £25,000 per patient in the UK (Lashilola et al. 2023).

The study highlights several critical aspects related to MA among patients with RDs, including the level of MA, assessment methods, and improvement strategies. The number of published studies inciting MA among patients with RDs increased in recent years. Notably, adherence levels showed significant variability, ranging from 0% (complete rejection of therapy) to 97.7% (excellent MA). Despite the importance of MA topics, there are no standardized methods for measuring and reporting MA among patients with RDs, which represents a significant gap in clinical practice. Logically, we revealed that the level of MA is influenced by the same factors identified in other patient populations, as well as by the consequences of MNA. The findings suggest that improving MA would enhance patient outcomes and reduce the economic burden of RDs by reducing hospitalizations and optimizing treatment efficacy.

In our study, the primary reasons for non-adherence were therapy-related and patient-related factors. The most commonly cited reasons in the first group were side effects (Cappellini and Taher 2009; Costa et al. 2012; Ivarsson et al. 2018; Vatier et al. 2019; Kamusheva et al. 2020; Pogna et al. 2021; Acquadro et al. 2022; Orso et al. 2022; Hollmen et al. 2023) and the need for multiple drug administrations daily or weekly (Cappellini and Taher 2009; Thomson et al. 2011; Vitturi et al. 2020; Torregrosa Prats et al. 2021; Hollmen et al. 2023). On the other hand, adolescence was one of the most cited patient-related reasons for MNA (De Moerloose et al. 2008; Jabbour et al. 2012; Malik et al. 2015; Medina and Román 2020; González-Lamuño et al. 2021), which is probably due to denial of the severity of the disease, concerns about side effects such as weight gain or other changes in appearance, the convenience of not taking the drug, fear of their friends seeing medication intake, and psychological aspects such as stress, anxiety, depression, and others (Taddeo et al. 2008; Anon 2016). Other reasons for MNA related to the patients were forgetfulness and low education level. The findings in this review underline again the need to improve the educational programs for patients, which can be pointed toward the whole family or the individual.

The consequences of MNA affect not only the individual but also the whole society and health system. For instance, MNA among patients with RD could lead to exacerbation of the patient’s condition, resulting in disability and/or death (Gallin et al. 2003; Jabbour et al. 2012; Shenoy et al. 2015; Chen 2016; Chaudhary et al. 2017; Ueda et al. 2019; Westermann et al. 2019; Vitturi et al. 2020; Dong et al. 2022; Lashilola et al. 2023). Considering the complexity of RD, it is unsurprising that patients with RD are more likely to visit a hospital (46.1%) than patients without RD (23.6%, p < 0.0001) (Bennett et al. 2019). In addition, numerous studies have concluded that non-adherence to therapy for chronic diseases is a major cause of frequent and/or longer hospital admissions (Iuga and McGuire 2013; Payero et al. 2014; Mongkhon et al. 2018). The societal and economic burdens associated with MNA are substantial, manifesting as increased direct and indirect costs (Jabbour et al. 2012; Chen 2016; Lashilola et al. 2023). These costs are projected to continue increasing, as concluded by a study from 2018, which also reported that the annual costs of MNA range from $949 to $44,190 per person (Cutler et al. 2018).

The clinical guideline “Medicines Adherence: involving patients in decisions about prescribed medicines and supporting adherence” (2009) focuses on patient-centered care. It explains the necessity for making informed decisions about therapy and active involvement in the treatment process. According to this guideline, no universal MAEI exists (NICE 2009). Health systems could implement diverse and effective strategies to identify nonadherent patients and the underlying reasons for MNA. However, due to the wide variety of RDs, it is important to individualize and select the most suitable approach for each patient. A multidisciplinary team is required to provide comprehensive care and information, answer questions that arise, and follow up with patients during therapy. This care includes timely addressing of adverse events and ensuring treatment adherence. One of the most important things when living with a lifelong condition is for the patient to not feel alone during the journey and to believe that the prescribed therapy is effective for their condition. For this reason, establishing a robust social support system to aid patients in MA can have a profound impact on MA levels (Malik et al. 2015; Walburn et al. 2019; Wang et al. 2022; Nguyen et al. 2023). Considering MAEIs included in our study, both patients and healthcare professionals believe mobile apps can significantly enhance adherence rates. While numerous apps are currently available that target common chronic diseases, very few apps address RDs. For instance, an application was developed to assist patients with rare inborn errors of metabolism in adhering to a strict diet that was essential for controlling the disease (Ho et al. 2016). Additionally, companies have created digital apps to support patients throughout the diagnosis, acceptance, and treatment phases (Atlantis Health 2023). These innovative solutions hold promise for improving adherence and enhancing the overall management of RDs.

To the best of our knowledge, this is one of the few comprehensive, systematic, and rigorous reviews that analyze MA issues related to patients with RDs. In line with our methodology, we reassessed the 20 articles from our prior search to ensure a consistent and thorough evaluation of all relevant data. Re-including these studies ensures that we capture all relevant data, enhancing the robustness of our findings and ensuring a more comprehensive review. It attempts to target all types of studies and RDs covered in the scientific literature. Thus, considering all specifics and nationally-based practices, a broad “picture” of the current situation regarding MA issues in this group of patients could be drawn and analyzed. The current study could be used as a basis for developing further policies and good practices for improving MA management, thus avoiding the negative consequences of nonadherence.

However, it is essential to acknowledge certain limitations of the study. First, the search was limited to studies published in English, potentially excluding valuable contributions in other languages. Additionally, the diversity in methodologies, designs, data collection methods, sample sizes, and evaluation criteria across the included publications posed challenges for direct comparisons and meta-analyses. The literature reviews (n = 6) included in our study were not assessed with the JBI tool or MMAT, as they failed to meet the criteria for evaluation. Another limitation is that this SR was restricted to articles published solely in the PubMed electronic database, unlike our previous scoping review, which covered two databases: PubMed and Google Scholar. This narrow focus may have inadvertently overlooked relevant studies in other databases or sources.

This review can be useful in future efforts to improve patient and public reporting systems for MA. Currently, there is an absence of developed and implemented methods for measuring, reporting, and improving MA in patients with RD. This reduces the potential for achieving optimal therapeutic outcomes. Therefore, care management programs are needed to understand the underlying causes of MNA. Such insights are critical for the development and implementation of innovative and effective solutions. A targeted approach is needed, combining different measures and focusing on overcoming the barriers that influence MA. Only when HCPs better and fully understand the reasons leading to MNA can they provide targeted assistance and effective interventions (Atlantis Health, 2023). It is also crucial to develop guidelines that indicate when and how information about the treatment should be given without causing concern to the patients. Given that frequent hospital admissions are the most common consequence of MNA, the implementation of government-supported home therapy programs is paramount. Problems with access to medicines (i.e., financial ones, lack of therapy on the market, etc.) must be eradicated as a prerequisite for MNA. Additionally, disparities in drug reimbursement policies across different countries contribute to unequal access to treatment, affecting patients’ ability to adhere to prescribed regimens (Pejcic et al. 2018).

Further research is crucial to fully understand how patient characteristics, such as age, gender, education, residence, and occupation, impact therapy adherence. Additionally, more research is needed to evaluate the preferences of patients with RDs regarding the use of various MAEIs. It is also important to examine how different methods to increase therapy adherence affect the overall adherence percentage. Research should also aim to develop standardized methods for measuring and reporting MA, as well as strategies that can be integrated into healthcare systems to identify non-adherent patients early and provide targeted support.