Potentially Inappropriate Medication: A Pilot Study in Institutionalized Older Adults

1. Introduction

Advancements in healthcare and medicine have significantly improved quality of life and life expectancy, contributing to an increase in the global population ageing. Europe is notably the continent with the oldest population, inhabited by around 25% of the world’s population aged over 60 [1]. Portugal reflects this demographic trend, showing an rate of ageing of 183.5% in 2022 [2], with 29.7% of the population aged 60 or over and projected to rise to around 40.3% by 2050 [3]. Globally, according to the United Nations, the number of elderly people aged 60 or over could double by 2050, from 962 million in 2017 to 2.1 billion in 2050 [1].

As physical and cognitive capacities decline with age, the elderly face diminished self-care abilities [4,5]. This problem, further complicated by the fragility of the family environment and social, economic, and cultural challenges, jeopardizes legally guaranteed basic rights. Residential Structures for the Elderly (ERPI) play a crucial role in upholding these rights by providing health care, psychosocial support, and promoting autonomy and protection for the elderly [6].

Ageing is responsible for numerous limitations and difficulties. Most elders have chronic illnesses which contribute to a decrease in autonomy and self-care ability, which results in the prescription of various pharmacological therapies and, consequently, polymedication [7].

According to the World Health Organization, there is no universal definition of polymedication. However, it is commonly understood to involve the concurrent use of five or more medications [8]. Even if there is no consensus on what is considered polymedication, taking five or more medications increases the risk of adverse reactions and drug interactions and can lead to negative and unexpected clinical results [9]. In elderly patients, polymedication is often categorized into two large groups: i) polymedication, when five to nine drugs are taken simultaneously; and ii) excessive polymedication, when 10 or more drugs are taken [10,11,12]. Polypharmacy, a term encompassing both categories, is a growing problem, whether due to an ageing population, multimorbidity, lack of communication between different prescribing doctors or the discovery of new drugs [8,12,13]. In the elderly population, the number of polymedicated and excessively polymedicated patients is 31.5% and 20.1% respectively [11]. In addition to drug interactions and the consequences they entail, polymedication leads to significant economic and clinical problems, including increased risk of falls, hospitalizations, and healthcare costs [13,14].

Promoting the appropriate use of medications trough multidisciplinary teams, reviewing pharmacotherapy, and tailoring medication to individual needs could help decrease mortality and the risks associated with polypharmacy [8,15]. In environments where elderly individuals receive care from specialized institutions or qualified health professionals, the risks of medication-related problems are notably decreased [16,17]. The elderly population is diverse, with varying medical conditions and degrees of physical, cognitive, and metabolic decline, making it challenging to identify and prevent adverse drug reactions [18]. Thus, individual needs, pharmacokinetic and pharmacodynamic changes, polymedication, and multimorbidity are risk factors that should be considered in pharmacotherapeutic analyses in this population [19].

The use of Potentially Inappropriate Medications (PIM) represents an increase in the complexity of pharmacotherapy in the elderly. In addition to anatomophysiological and metabolic changes, interactions, and risk for adverse reactions associated with any medication, research shows that there are medications that are considered inappropriate for elders [20]. PIM are drugs where the associated risks outweigh the clinical benefits, putting the patient at risk and prone to serious health risks [21,22]. Analyses of routine prescriptions and medication use in hospitals and nursing homes have shown that the prevalence of PIM use by elderly patients is high [23,24].

The use of PIM is associated with polymedication and the with the incidence of adverse effects [23,25,26,27]. Although PIM should be avoided in the elderly, sometimes their use is necessary, due to the lack of effective alternatives or when their discontinuation could cause more harm. When used inappropriately, without following medical recommendations and without proper monitoring, PIM can lead to increased hospitalizations, falls, adverse reactions and mortality among the elderly [28,29].

The pharmacotherapeutic regime for the elderly is complex, influenced by multimorbidity and age-related factors that contribute to polymedication and associated medication problems [17]. In addition to the extensive medical care that is already necessary at this stage of life, pharmaceutical care and medication review can favor quality of life and patient safety [30]. According to the Pharmaceutical Care Network Europe (PCNE), medication review “…is a structured assessment of a patient’s medicines with the aim of optimizing the use of medicines and improving health outcomes. This involves detecting medication-related problems and recommending interventions” [31]. Medication review leads to better clinical results from therapeutic practice and is therefore an extremely important procedure, especially for elderly, polymedicated patients, with multiple health issues [32].

This pilot study aimed to characterize the sociodemographic, clinical and pharmacotherapeutic profile and the use of PIM of institutionalized elders residing in two ERPI in the Faro municipality, in the Portuguese region of the Algarve.

2. Materials and Methods

We conducted a cross-sectional descriptive study in two ERPI located in the municipality of Faro, Portugal. ERPI selection and invitation were decided by convenience, without additional criteria apart from distance. The study included a total of 96 elderly people aged 65 or over, taking at least 2 medications and who consented the collection of their clinical and medication data, and summary clinical and demographic characteristics.

Analysis of the existing pharmacotherapeutic profile for each participant allowed to retrieve the names of the medicines (prescription and nonprescription) and food supplements used, pharmaceutical form, dosage and dosage regimen. A Type 2B medication review was carried out, according to the Pharmaceutical Care Network Europe criteria [31].

All medications were organized according to the Anatomical Therapeutic Chemical Code (ATC), level 3 [33]. The use of ≥5 medications was considered “polymedication” and the use of ≥10 medications was considered “excessive polymedication”. Individuals with less than five medications were considered “not polymedicated” [11].

Potential medication interactions were classified using the Lexidrug^® Drug Interaction Module tool from the UpToDate platform. This tool classifies interactions into five different categories, according to the associated risk: type A (no potential pharmacokinetic or pharmacodynamic interactions; no known interaction), type B (little to no evidence of clinical concern; no action needed), type C (clinically significant interactions; monitor therapy), type D (clinically significant interactions and risk-benefit must be assessed; consider therapy modification), and type X (clinically significant interactions where the risks outweigh the benefits; avoid combination) [34].

Potentially inappropriate medicines for the elderly were considered according to the European EU(7)- PIM List [35], and the complexity of the pharmacotherapeutic regime was assessed using the criteria established by the Medication Regimen Complexity Index tool (MRCI) [36]. The MRCI score is calculated from three main subtotals, each contributing to the overall complexity. Subtotal A refers to the pharmaceutical form of the medication (e.g., tablet, injection, inhaler), as different forms may have different complexities. Subtotal B refers to the dosage frequency; more frequent dosing schedules increase the complexity. Subtotal C refers to the instructions and additional precautions for each medication, such as dietary restrictions or to avoid certain activities. Points are added with increased complexity added by the form, with increased frequency of dosing, and with additional specific instructions. The MRCI score is the sum of all subgroup scores and a total score >16.5 can be interpreted as high pharmacotherapeutic complexity [37].

Statistical analysis was carried out using IBM SPSS Statistics software for Macintosh, version 29.0. Armonk, NY: IBM Corp.

Data were described using absolute and relative frequencies, mean (M), median (Mdn), standard deviation (SD) and interquartile range (IQR). We used the Kolmogorov-Smirnov test to assess adherence to the Normal distribution. All variables in our study were either categorical or with a non-Normal distribution. Thus, group comparisons were made using Kruskal-Wallis, chi-square, and Mann-Whitney’s tests. When the results of the chi-square test were considered non-valid due to low expected frequency count, Fisher’s or Fisher-Freeman-Halton’s exact tests were computed. Paired comparisons were analyzed with Friedman’s test. Correlations were computed using Spearman’s correlation coefficient. Statistical significance for all procedures was set at 0.05.

3. Results

3.1. Sociodemographic and Pharmacotherapeutic Characterization

There were 96 patients included in this study, all aged 65 or over and residing in care institutions for the elderly. Mean age was 86.6±7.86 years, ranging from 68-109 years, with only 9.4% aged between 65-74 and 66.7% aged 85 or above. Most participants were women (68.8%). We found significant gender differences in age (p=0.044), with women (87.9±6.9) being older than men (83.7±9.04).

Mean number of medications used was 9.1±4.15, ranging from two to 24, and 90.6% of participants (n=87) were considered polymedicated (≥5 medications). Of the polymedicated, 47.1% (n=41) were excessively polymedicated (≥10 drugs). Only 9.4% of patients (n=9) had a pharmacotherapeutic profile with fewer than five drugs. Of the patients identified as polymedicated, 60.4% were aged 85 or over and 62.5% were female. The mean number of medications in the participants without polypharmacy was 3±0.87, while those with polypharmacy showed a significantly higher (p<0.001) mean of 9.8±3.83 medications.

According to the MRCI score, pharmacotherapeutic complexity, showed a mean score of 19.2±9.90, ranging between 5-55 points. Subtotals A, B and C had means of 3.2±2.39, 10.3±5.26 and 5.7±3.95 points, respectively. Table 1 presents age and selected data on medication use. No statistically significant differences in medication use were found between male and female participants (p>0.05).

MRCI score showed a strong positive correlation with the total number of medications used (R=0.873; p<0.001) and a negative correlation with age (R= -0.219, p=0.032), suggesting that younger patients and those that use more medications have a higher pharmacotherapeutic complexity. The statistical significant correlation of MRCI with age does not hold up when considering patients aged 65-74 years (R= 0.156, p=0.688; n=9) or ≥85 years (R= -0.005, p=0.967; n=64). This correlation is still significant in patients aged 75-84 years (n=23), but suggests a positive association, i.e., higher complexity in older patients (R=0.550, p=0.007). After adjusting for the number of medications, the correlation between MRCI and age was not significant (R=0.038, p=0.867) in this demographic. This suggests that the number of medications may have confounded the initial association, and that factors other than age, such as clinical condition, could be contributing to pharmacological complexity. Table 2 presents the correlation matrix between MRCI and other variables.

Of the participants, 91.7% (n=88) had at least one potential drug interaction, with a mean of 10.7±10.55 interactions per participant, ranging between 0-57. At least one potential interaction of each of the five types considered by the Lexidrug^® tool was detected. Type C interactions were the most prevalent (86.5%; n=83) and had the highest mean occurrence per patient (7.93±7.68). Type A interactions, those with no potential pharmacokinetic or pharmacodynamic interactions, were seen in only two patients (2.0%), with a mean of 0.1±0.92 interactions. The most serious and clinically significant interactions (type X), where the risks outweigh the benefits and that should be avoided, were identified in 13 patients (13.5%), with a mean of 0.19±0.53. One of the patients presented those three interactions of this type simultaneously. Our data also shows that 59.4% (n=57) of the sample had at least one type B interaction, with an average of 1.34±1.89 of these type of interactions per participant. Type D interactions were identified in 40.6% (n=39) of patients, with an average of 1.10±2.02 interactions. The number of interactions is positively correlated with pharmacologic complexity (R=0.692, p<0.001). All interaction types apart from type A were also correlated with pharmacologic complexity (p<0.001).

3.2. Potentially Inappropriate Medication Characterization

Our data show that 86.5% (n=83) of the patients used at least one PIM (with a mean of 2.2±1.56 PIM per participant, ranging between 0-6), 67.5% (n=56) were women, 63.9% (n=53) were aged 85 or over, and 92.8% (n=77) had some kind of potential drug interaction and were polymedicated (≥5 medications). More than half of the patients who used any PIM (54.2%, n=45) had high pharmacotherapeutic complexity scores (MRCI>16.5 points). PIM was positively correlated with MRCI score (R=0.412, p<0.001), but no statistically significant difference in total MRCI score (p=0.368) was found when comparing patients who used PIM and those who do not. Nevertheless, subtotal B from the MCRI score was different in these patients (p=0.047).

No statistically significant differences in age, gender, polymedication, and interactions, were observed when comparing patients who used PIM and those who do not (p>0.05). Table 3 presents the comparisons between patients using PIM and those who do not use PIM.

PIM from the EU(7)-PIM list were classified according to their ATC. 209 PIM were identified and distributed among seven ATC groups (A, B, C, G, M, N and R). Of the total number of PIM, 49.8% (n=104) belong to central nervous system medications (ATC N) and 29.2% (n=61) to alimentary tract and metabolism medications (ATC A). Within these two groups, the use of anxiolytics (ATC N05B) and drugs to treat peptic ulcers and gastroesophageal reflux (ATC A02B) stood out, accounting for 23.4% (n=49) and 23.9% (n=50), respectively, of the total number of PIM.

Mean PIM in ATC G is significantly lower (p=0.024) in patients ≥85 years old (M=0.19±0.5) than in younger patients (M=0.44±0.67). Similarly, patients ≥85 years old also have lower (p<0.001) PIM in ATC N (M=3±1.79) than younger patients (M=4.8±2.55). Polymedicated patients have a higher (p=0.009) mean number of PIM in ATC A (M=1.6±1.25) than non-polymedicated patients (M=0.6±0.53), and patients with a high pharmacotherapeutic complexity have a higher (p<0.001) PIM in ATC C (M=2.5±1.7) than all other patients (M=1.3±1.1).

Of the patients who use PIM, 53.1% (n=51) use at least one from ATC A and 62.5% (n=60) use at least one from ATC N.

4. Discussion

In this cross-sectional study, we reviewed the medication of elders institutionalized in two ERPI. Our sample primarily consisted of individuals over 80, predominantly women, aligning with demographics reported in similar studies from the United States, Spain, Portugal, France, and Colombia [38,39,40,41]. This demographic trend underscores the critical need for targeted pharmacotherapy management, particularly given the elderly’s propensity for chronic diseases and the consequent polymedication [19,42].

Our data show the use of 9.13±4.15 medicines, ranging from two to 24, in accordance with previous research in Portugal and France suggesting high medication usage among the institutionalized elderly. The Portuguese study reports a mean use of 10.1±3.9 medications, ranging from 5 to 28 [43], and the French study reports a mean of 8.1±4.0, ranging from 1 to 20 medications [44]. However, the tools used to analyze the PIM were different, which may reflect differences in the results. Other studies reflect slight variances in medication usage, emphasizing the complexity of pharmacotherapy in this demographic [23,43].

Pharmacotherapeutic complexity can be assessed considering numerous factors, but the complexity index is one of the few methodologies capable of numerically quantifying this criterion [36,45,46] and is validated for Portuguese [36]. Studies from Brazil and Portugal reported similar MRCI means to the one in our data, while Australian studies found higher complexity [47,48,49,50]. The MRCI is instrumental in gauging the challenge each pharmacotherapeutic profile presents, and our findings corroborate a direct correlation between MRCI values, the number of medications, and patient age [49,51].

One of the decisive points for a good therapeutic review is the identification of potential drug interactions that could jeopardize clinical conditions and the treatment of pathological conditions [30,52]. Studies carried out in this area show different results in terms of the percentage of institutionalized elderly people with potential drug interactions. Research carried out in Brazil, Greenland and Spain revealed 25.4%, 61.0% and 81.1% of patients, respectively, with potential interactions [53,54,55], although our data showed more than 90% of the sample with potential interactions underscoring the critical need for vigilant drug management in this population. The differences found in potential drug interactions may be due to the assessment tool used. The literature shows that, depending on the tool used to assess potential drug interactions, the results can be different [56,57]. However, Lexidrug^® is considered a tool with excellent performance and comprehensiveness in identifying potential drug interactions, according to the scientific literature [58].

Like the potential drug interactions, the various existing tools for assessing PIM make it difficult to compare our data with results from the literature. For instance, a study comparing three PIM identification tools in the same sample found variability between 35.3% and 96.3% in the detection rates of participants with at least one inappropriate medication. PRISCUS list had the lowest identification rate (35.5%), followed by the Beers criteria (90.8%) and finally, the tool with the highest identification rate (96.3%) was the STOPP Criteria [59]. This variability points to a generally high tendency toward prescribing PIM in institutionalized settings, confirmed by our findings (86.5%) and literature suggesting prevalence rates from 79.1% to 96.8% [23,60,61].

Population-based studies involving both institutionalized and non-institutionalized elders reported lower PIM usage (24.1% and 40%) likely due to reduced PIM consumption among the non-institutionalized [62,63,64,65]. The organization and care provided in institutions, the pathologies diagnosed, the greater number of drugs and the classes of drugs prescribed in institutions are pointed to as possible reasons for the differences found in the number of potentially inappropriate drugs between institutionalized and non-institutionalized elderly people [64,65]. A study comparing institutionalized elders and elderly outpatients showed that 82.4% of institutionalized participants where prescribed PIM, while this prevalence was 34.4% in outpatients [66]. This trend was also reported by a Swedish study, which suggested institutionalization as a potential risk factor for the use of PIM and proposing greater attention and care when prescribing drugs for institutionalized elderly people [67]. Contrary results were described in a study carried out in Ontario, Canada, where the number of potentially inappropriate medications in institutionalized elderly people was lower than that found in elderly people in the community. However, clinical pharmaceutical services, such as medication review, are mandatory in institutions for the elderly, which may justify the results found [68].

Moreover, our study aligns with other research indicating that female and polymedicated elderly patients are more likely to receive at least one PIM [60,63,65].

In Portugal, a national study suggested a significant consumption of PIM, with 1232 billion defined daily doses prescribed, resulting in an overall prevalence of 9.20%. This study showed a high consumption of PIMs, particularly anxiolytics and antidepressants, despite the implementation of national measures to prescribe these medicines. In the Algarve region, there seems to be an upward trend in the number of potentially inappropriate medicines [69]. This reality shows that it will be important to implement different measures to improve the safe use of medicines in institutionalized elderly people. Medication reviews could be an important service to implement [64].

Regarding ATC classification of PIM, our data are in accordance with the literature. A study on institutionalized elderly with dementia across eight countries identified major PIM within psycholeptics and drugs treating acid-related disorders (ATC N05 and ATC A02, respectively) [70]. A Portuguese study with institutionalized patients, which also used the EU(7)-PIM List, revealed a similar trend: proton pump inhibitors (A02BC) were the most prescribed PIM (25.8%), followed by anxiolytics (N05BA, 20.5%) [23]. Another study carried out using three different tools to identify PIM reported significant prescription of proton pump inhibitors and drugs associated with the nervous system among the PIM detected [59]. As mentioned above, it is to be expected that drugs from the ATC A and N groups are the most prescribed PIM, since they also correspond to the therapeutic groups more often used by institutionalized elderly people.

Regarding the number of medications used and the prescription of PIM, our results are also in accordance with the literature, such as a population-based study suggesting a significant association between these two variables (p<0.001) and higher mean medications in use by patients with PIM (7.7 vs. 5.6) [61]. Furthermore, regarding pharmacotherapeutic complexity, the literature suggests that the presence of PIM is also associated with a likelihood of higher MCI in institutionalized elderly (p<0.001) [48].

Our findings suggest that elderly patients below 85 years old use more PIM, as a study reporting a similar trend for patients under 70 [71]. This may reflect a tendency to prescribe fewer medications to older individuals [72,73].

We identify some limitations in our study. Even if this was not an objective in our study, our data does not allow us to conduct an in-depth analysis of the clinical variables and of the assumptions and reasons for the prescription of some medications. The non-random nature of the sample implies limited external validity for our results and our findings should be considered according to context. We cannot extrapolate for the institutionalized elderly population in the Algarve region, but this pilot study allowed us to assess the applicability of the data collection tools on this population and to carry out a type 2B medication review.

5. Conclusions

This study highlighted several key challenges associated with pharmacotherapeutics in the elderly, such as the excessive number of medicines prescribed, high therapeutic complexity, the frequent use of PIM, and a high prevalence of potential drug interactions.

Medication review is a critical intervention for preventing and correcting issues associated with medication use, particularly among the elderly, who are often subjected to polymedication. This clinical practice is essential to prevent the exacerbation of chronic illnesses and the overall decline in health status among older adults.

Medication review is an accessible and advantageous tool aimed at therapeutic success, and its benefits extend beyond the identification and correction of drug-related problems; they also contribute significantly to improving the physical condition and overall well-being of patients. Despite its clear benefits, our results suggest that medication review may remain underutilized within care institutions. This underscores a critical gap in practice that needs addressing. Enhancing awareness and understanding of medication review’s benefits among healthcare providers and institutional managers is essential. A broader dissemination of this practice should be considered, in order to contributing to integrate it as a standard component of elderly care, ensuring that elderly patients receive the most appropriate, effective, and safe pharmacotherapy possible. It is therefore important to sensitize health students at undergraduate level to the im-portance of medication review and interdisciplinarity in this process.