1 Introduction

The term prescribing cascades was first introduced by Rochon and Gurwitz in 1995 [1]. It describes the event in which adverse drug reactions (ADRs) to medications lead to the prescription of additional medications [1]. Besides prescription medications, also over-the-counter medications, supplements, medical devices and diagnostic interventions have been reported as healthcare interventions used to treat ADRs [2]. Prescribing cascades can be intentional when an ADR is recognized and treated or when a healthcare provider (HCP) aims to prevent an ADR (e.g., prescribing a laxative alongside an opioid) [3]. However, regardless of intent, prescribing cascades have been associated with various negative outcomes for both patients and the healthcare system in general [4-6].

For the identification of prescribing cascades, multiple studies have published lists of prescribing cascades that may occur in practice [4, 7-10]. For instance, a systematic review by Adrien et al. presented 115 individual prescribing cascades identified in previous studies [7]. These lists generally focus on a singular concept involving individual prescribing cascades, i.e., a single primary medication leading to one ADR and subsequently one additional medication. Examples mentioned across the different lists include the prescription of diuretics for peripheral oedema induced by calcium channel blockers or the prescription of cough suppressants for ACE inhibitor–induced cough [4, 7-10]. These examples involve well-known ADRs. A recent case vignette study showed that most (28/30) physicians recognized peripheral oedema induced by a calcium channel blocker and preferred to substitute this medication rather than prescribing diuretics [11]. In clinical practice, however, this potential prescribing cascade seems unrecognized as it has been repeatedly observed, involving large numbers of patients (in 2.4%–11.8% of calcium channel blocker users) [5, 6, 12-14]. Apparently, the assessment and management of ADRs and prescribing cascades in clinical practice can be difficult, particularly in patients with polypharmacy and multimorbidity [15, 16]. Moreover, of great concern are cases in which patients are exposed to multiple prescribing cascades. For instance, a recent scoping review by Chen et al. identified several publications, including case reports, describing prescribing cascades involving more than two additional medications to treat sequential ADRs—for example, peripheral oedema induced by calcium channel blockers followed by diuretics, which then result in urinary incontinence followed by medication for an overactive bladder [17]. This indicates that real-world situations may be more complex to implement deprescribing, with multiple contributing factors, making it less straightforward to find potential strategies to identify, mitigate or ideally prevent prescribing cascades [15, 16, 18].

This study aims to illustrate the real-world complexity of prescribing cascades, using four real-world exemplary cases from literature, and to provide insight into strategies to identify, mitigate and prevent these prescribing cascades [19-22].

2 Case Studies

2.1 Case 1: Repeated Misinterpretation

A 68-year-old woman consulted her general practitioner for the first time because of a persistent cough [19]. She presented a medication list including amitriptyline as a sleeping aid, a nasal steroid, a proton pump inhibitor and two asthma inhalers (albuterol and an inhaled corticosteroid). However, the patient reported not to have any history of hay fever, heartburn or asthma. According to the patient, these medications were prescribed to treat a chronic cough. Nevertheless, the cough did not resolve.

In the year prior to this consultation, the patient had consulted a gastroenterologist, an allergist and an otolaryngologist for her cough, each of whom prescribed medications based on their suspected diagnoses (Figure 1). The consulting general practitioner suspected that the anticholinergic properties of amitriptyline might be contributing to a dry mouth which then could result in the patient's cough. The general practitioner advised the patient not to use amitriptyline for 1 week. This deprescribing resulted in the cough being resolved and eventually deprescribing of the four additional medications described above.

2.2 Case 2: Consecutive Cascades

A 71-year-old woman presented to the emergency department following a fall resulting in multiple fractures [20]. At the time of admission, she used 17 different medications, including medication for hypertension, asthma, hypothyroidism, depression, osteoarthritis and Ménière's disease.

A few months prior to presentation, amlodipine had been prescribed by a general practitioner to control the patient's blood pressure (Figure 2). Three weeks later, the patient developed lower limb oedema, and a cardiologist prescribed furosemide and spironolactone. As a result of the use of the diuretics, the patient experienced symptoms of urinary incontinence in the subsequent weeks. Therefore, the antimuscarinic agent fesoterodine was initiated by a urologist. One month later, the patient developed a dry mouth, for which her general practitioner prescribed anetholtrithion. Ultimately, possibly due to the ADRs of the different medications and her comorbidities, the patient experienced gait and balance problems, which led to a fall in her bathroom, and she was admitted to the hospital.

A clinical pharmacist identified the consecutive prescribing cascades after which the amlodipine was replaced by an ACE inhibitor to manage the patient's hypertension. Thereafter, the four additional medications spironolactone, furosemide, fesoterodine and anetholtrithion were deprescribed. The oedema, urinary symptoms and dry mouth resolved within 14 days, and the patient was discharged from the hospital.

2.3 Case 3: The Increased Dose

An 80-year-old woman presented to the hospital with acute delirium, a 7-day history of dry cough and 2 days of severe diarrhoea [21]. The patient's medical history included hypertension, osteoporosis, osteoarthritis and macular degeneration. At the time of presentation, the patient's medications consisted of hydrochlorothiazide, enalapril, alendronate, calcium, paracetamol, guaifenesin/codeine and levofloxacin.

One week prior to presentation to the hospital, the patient's enalapril dose was increased by her general practitioner from 5 to 10 mg once daily to control blood pressure (Figure 3). Five days after dose increase, the patient developed a non-productive cough, for which a cough syrup containing guaifenesin and codeine was prescribed. Two days later, the patient called her general practitioner by phone to report increasing cough and lethargy. These symptoms were misinterpreted as pneumonia symptoms. The general practitioner subsequently prescribed levofloxacin. Three days after initiating levofloxacin, the patient developed watery diarrhoea and delirium and was admitted. In the hospital, the patient's stool tested positive for Clostridium difficile toxin, and therefore, metronidazole was started. The next day, the patient's condition improved. However, non-productive cough still persisted. Seven days after hospital admission, a consulting geriatrician deprescribed enalapril and cough syrup. As a result, the patient's cough improved, and the patient was discharged after 10 days.

2.4 Case 4: Cumulative Effects

In 2023, a 58-year-old man presented to the emergency department after 3 days of vomiting and reduced oral intake [22]. Notably, since 2022, he had been receiving metoclopramide, and over the preceding 4 months, two additional antiemetics had been added for refractory nausea. During admission, starvation, euglycaemic ketoacidosis, dehydration, a mild acute kidney injury and constipation were also noted. His extensive medical history included the use of more than 20 medications. Several complex cascades were identified as possible contributors to his current presentation (Figure 4).

The first cascade began in 2018 when an attempt was made to rationalize his psychotropic medication for bipolar affective disorder because of concerns about drug-induced parkinsonism. Although several changes to his psychotropic regimen were made in the following period, a neurologist prescribed levodopa in combination with benserazide in 2020 with the intention of discontinuing this treatment once the patient's mood stabilized. However, the patient continued the levodopa/benserazide treatment, which may have contributed to persistent nausea. In 2022, metoclopramide was introduced to alleviate the nausea. However, its known potential to induce parkinsonism likely compounded his clinical issues. That same year, lactulose was initiated to treat the emerging constipation. Furthermore, 7 months prior to presentation, his general practitioner prescribed pre-exposure prophylaxis (PrEP) with tenofovir and emtricitabine for human immunodeficiency virus (HIV) prevention. As PrEP can also cause nausea, the patient's nausea worsened, prompting the addition of prochlorperazine and ondansetron—both of which are associated with constipation—to his regimen. Dehydration secondary to nausea may have further aggravated his constipation. The patient was already using dapagliflozin and telmisartan long term. Dapagliflozin is associated with ketoacidosis, particularly in fasting and dehydrated individuals. The acute kidney injury may have resulted from a combination of telmisartan, PrEP and dehydration.

Another cascade was noted. Limited mobility and an increased risk of falls prompted a rehabilitation physician to recommend the use of a four-wheeled walker in 2020. At that time, the patient was on long-term therapy with two benzodiazepines, and the antipsychotic-induced parkinsonism may have further contributed to his impaired mobility and risk of falls.

During hospitalization, a medical doctor and pharmacist conducted a comprehensive review of the patient's medication history, including the timing of each prescription. In response, several adjustments were made. To reduce nausea, the levodopa dose was reduced, and PrEP was deprescribed (the patient was no longer sexually active). Dapagliflozin was deprescribed to resolve euglycaemic ketoacidosis. Metoclopramide and prochlorperazine were deprescribed to reduce the risk of drug-induced parkinsonism and replaced with domperidone. Ondansetron was discontinued to reduce constipation. To further manage his constipation, the lactulose dose was increased, and additional laxatives—docusate combined with senna and macrogol—were initiated. Lastly, temazepam was deprescribed to reduce the risk of falls, and telmisartan was deprescribed to resolve the acute kidney injury. Within 1 month, the patient's nausea improved, and domperidone was deprescribed, and his renal function normalized.

3 Underlying Factors

The first case describes a patient who presented with a misinterpreted ADR, involving multiple clinicians, ultimately resulting in the prescription of four unnecessary additional medications [19]. The second case highlights the sequential misinterpretation of different ADRs, resulting in multiple consecutive prescribing cascades [20]. The third case demonstrates a pattern of consecutive prescribing cascades following a dose increase [21]. Finally, the fourth case illustrates the cumulative effects of multiple medications and comorbidities, contributing to the same symptoms [22]. Of note, three cases (Cases 2–4) resulted in hospital admissions, and in all cases, multiple HCPs were involved. In Case 3, complex prescribing cascades developed within a couple of weeks, whereas in Case 4, cascades unfolded over several years. These exemplary cases illustrate the complexity and multifaceted nature of prescribing cascades in clinical practice with major consequences. It should be noted that these specific cases may represent extreme examples. It is unknown what proportion of prescribing cascades is complex in clinical practice.

ADRs can manifest in delayed or unexpected ways, making identification challenging. They may present as rare or atypical symptoms, increasing the likelihood of misinterpretation as a new medical condition rather than an ADR. Depending on the medical specialty, different interpretations can be made, known as cognitive bias (Case 1) [23]. Furthermore, variations in how symptoms are experienced or reported by patients can complicate recognition, potentially leading to misattribution and misdiagnosis (possibly Case 1). Alternatively, identification of an ADR may be limited by incomplete examination of the patient, for instance, due to consultations by telephone (possibly Case 3). ADRs resulting from a dose increase of previously well-tolerated medication can also be overlooked, particularly when there is no information or evidence of the ADR being dose related (Case 3) [15]. In addition, ADRs may evolve gradually over time rather than appearing immediately after the initiation of a medication (Cases 3 and 4). Consequently, the ADR appears unrelated to the medication causing it [15]. The further the cascade progresses, the more difficult it becomes to recognize the original cause [22].

Fragmented healthcare systems across the care continuum can make it challenging to recognize ADRs and prevent prescribing cascades, as HCPs may not have access to the same patient information [16]. When multiple providers are involved in a patient's care (all cases), medications may be prescribed without full awareness of the existing treatment regimen, increasing the risk of unrecognized ADRs and subsequent prescribing cascades [15, 16]. Inadequate availability of medication and medical history or incomplete documentation of treatment changes can further complicate ADR identification [15, 22]. The presented cases illustrate that having one (team of) HCP(s) taking responsibility to review all medication is needed to end prescribing cascades and to deprescribe multiple medications.

The accumulation of pharmacological effects and comorbidities within a patient increases the likelihood of ADRs to occur and complicates the determination of causality and consequently management of the patient's condition (Case 4). For Case 4, it was not clear whether changes and adverse effects had been communicated between the patient's various prescribers [22]. Identifying the root causes and managing the prescribing cascade become particularly challenging when medications and a patient's health condition exert similar effects or when escalating loops are triggered. Polypharmacy, and particularly the use of multiple medications with similar effects, can also contribute to the development of prescribing cascades (Case 4) [15]. Previous work has shown that each new medication increases the risk of an ADR up to 10% [24].

4 Strategies to Identify, Mitigate and Prevent Prescribing Cascades

Several strategies have been suggested to improve the identification, mitigation and prevention of prescribing cascades (Figure 5) [10, 18, 25-28]. Key components include (1) educating both HCPs and patients to increase awareness and facilitate identification; (2) providing tools and training to assist HCPs in managing ADRs and addressing prescribing cascades; (3) enhancing communication and information exchange among HCPs and with patients regarding ADRs and the rationale for medication changes; (4) conducting comprehensive medication reviews; and (5) monitoring patients proactively for ADRs. Furthermore, integrating risk-based strategies and patient-centred approaches may strengthen these efforts.

Although basic knowledge about ADRs and prescribing cascades is a prerequisite for addressing them, relying solely on education of HCPs appears insufficient to change clinical practice. In fact, education is one of the least effective intervention strategies to reduce medication errors and inappropriate prescribing [29, 30]. Therefore, augmenting educational efforts with practical knowledge and recommendations is necessary. A systematic review showed that for many cascades, there are no explicit or well-substantiated recommendations on how to reverse or prevent them [7]. Furthermore, community pharmacists have mentioned they would like more explicit information on how best to resolve specific prescribing cascades [28, 31]. Evidence-based knowledge documents and tools can support the management of prescribing cascades. For example, readily accessible information is needed to decide whether a dose reduction could resolve an ADR and which alternative medication could be prescribed, either within the same pharmacological group or across different groups. Also, information on when an ADR typically occurs and what the expected resolution time is after deprescribing could be helpful.

The complex cases presented in this study illustrate that additional supportive strategies are needed. Given the large number of potential ADRs and related prescribing cascades, an HCP may not be able to identify all issues during a routine care visit [32]. Therefore, web-based or e-tools with algorithms to identify potential prescribing cascades have been recommended [27]. These tools should include warnings when initiating medications with similar ADR profiles. However, alerts running on pharmacy information systems are often limited by lacking clinical patient information. This makes such approaches inefficient for identifying prescribing cascades and could result in alert fatigue [28]. Therefore, there is a need to develop clinical rules, taking all patient-relevant information available in the electronic medical records into account, which can be integrated into computerized physician order entry systems (CPOEs) [33]. There is moderate-certainty evidence that CPOE systems can reduce medication errors and inappropriate prescribing in hospital settings [29]. In outpatient settings, however, computer alerts and decision support have not always been successful in reducing inappropriate prescribing, which may be explained by difficulties in integrating the computerized rules in existing systems in outpatient care [30]. Besides tools, training is needed for HCPs to address prescribing cascades in complex cases going beyond the simplistic singular concept of individual prescribing cascades. Interactive case demonstrations and discussions and learning modules demonstrating good practice may be helpful. These modules should include consideration of alternative non-pharmacological treatments. For instance, lifestyle modifications, physical therapy or dietary interventions can sometimes replace medications that contribute to prescribing cascades. For some complex cases, a consult service with a physician specialist or pharmacist to help identify and manage potential prescribing cascades might be helpful [27].

Enhanced communication and information exchange between HCPs and with the patient is another strategy to mitigate or prevent prescribing cascades. Ndai et al. found that multiple HCPs were involved in almost two-thirds of patients who were prescribed diuretics after initiating a calcium channel blocker [13]. Another study found that multiple HCPs were involved in nearly half of the patients who were prescribed anticholinergic medication after initiating dementia treatment [34]. This could be caused by a lack of acknowledgment of the ADR in relation to medication started by others or of incomplete patient records. Therefore, extensive medication reconciliation is essential, not only to record current medications but also to understand the timeline of medication changes. This is especially of importance when patient records are fragmented among multiple HCPs [22, 35]. Piggott et al. highlight the need for a clinical process map to gather and structure the necessary information to assess prescribing cascades [25]. Furthermore, when additional medication is intentionally initiated as the best treatment option, it is critical to document the rationale. If eventually the primary medication causing the ADR is deprescribed, this documentation can also guide subsequent HCPs to understand the context and safely deprescribe the additional medication as well. Finally, to empower patients, they need to be informed about their medication and the rationale for prescribing (additional) medication.

Regular medication reviews, including medication reconciliation, particularly during care transitions, can help identify when a prescription might be initiated to manage an ADR rather than an independent condition. In countries where such reviews are integrated into routine clinical practice, it has been shown they can reduce inappropriate medication use [36-40]. The effect of medication reviews on reducing prescribing cascades is yet to be established. Nevertheless, evidence-based frameworks and guidelines providing structured approaches to safely remove, reduce or replace medications that are no longer beneficial or suboptimal are available and can be incorporated into clinical practice [41, 42]. Systematic deprescribing interventions have been shown to improve patient outcomes by reducing medication burden and minimizing ADRs [43, 44]. However, to optimally mitigate the effects of prescribing cascades, strategies should aim beyond deprescribing alone.

Deprescribing is inherently reactive. By the time deprescribing occurs, harm may already be done. Thus, exploring proactive approaches to prevent the initiation of unnecessary additional medications is critical. A proactive strategy involves informing patients about common ADRs and instructing them to promptly report any symptoms. Furthermore, targeted monitoring of high-risk groups including older adults who have a lower socio-economic status, have polypharmacy and are female is another proactive strategy [4, 6, 45]. This approach includes follow-up calls or automated online questionnaires after starting high-risk medications or following care transitions [16, 28, 46]. So far, most studies have primarily included adults or older patients, resulting in a lack of data on other potential high-risk groups. Chen et al. reported a case involving a young adolescent, and a scoping review of Khalil and Huang found that only a limited number of studies have evaluated ADRs in the paediatric population [17, 47]. Patient-centred approaches, including supportive tools, enabling and empowering patients to participate actively in their medication management and report ADRs may be helpful [26, 48].

When looking at complex cases, however, both HCPs and patients struggle with balancing benefits and harms, have beliefs that ADRs and prescribing cascades are unavoidable and can be unwilling to change complex treatment regimes, particularly when multiple prescribers are involved [49]. Nonetheless, initial intentional prescribing cascades can still be problematic [3]. Furthermore, HCPs and patients experience difficulties recognizing which symptom could be related to which medication [15, 16]. Also, patients themselves find it difficult to provide a reliable chronology of symptoms and medications [15, 16]. This may be especially true in cases of fragmented care. This calls for strategies where patients can report possible ADRs with minimal effort to an HCP who can take responsibility for managing complex medication regimes [50]. Shared decision-making between HCPs and patients can help identify the most appropriate action while minimizing medication-related harm. The inclusion of the patient in any assessment of the appropriateness of the cascade is critical, with particular attention given to whether the cascade's initiation matches with the patient's goals and their awareness of the cascade's possible long-term risks [35].

5 Conclusion

Prescribing cascades are complex processes that extend beyond the simplistic singular concept of one medication leading to one ADR leading to a second medication. They involve misinterpreted symptoms, consecutive prescribing cascades and additive pharmacological and clinical effects. Given this complexity, a holistic approach to medication management is essential to identify, mitigate and prevent prescribing cascades at an early stage. Implementation of supportive tools to deprescribe, such as alerting systems, clinical medication reviews and practical guidance with explicit recommendations, along with enhanced communication among HCPs and patients regarding ADRs and the rationale for medication changes and proactive monitoring of patients, are promising strategies for preventing unnecessary medication use and associated harms. Additionally, increasing awareness among HCPs and patients through education and patient-centred approaches play a role in minimizing prescribing cascades. As the global population continues to age and more patients experience polypharmacy, the importance of these strategies will grow in the coming years. Future research should aim at development of these tools and recommendations targeting both HCPs and patients.

Conflicts of Interest

The authors declare no conflicts of interest.