Objectives: The study aimed to assess the relationship between the Fit fOR The Aged (FORTA) score – a classification system designed to evaluate medication appropriateness in older adults – and several negative outcomes, including impaired cognitive performance, functional status, adverse clinical events, and all-cause mortality at 3, 6, and 12 months after hospital discharge. Methods: This retrospective study utilized data from the ELICADHE cohort, a cluster-randomized trial conducted across 20 Italian internal medicine and geriatric wards. The study included patients aged 75 and older with complete FORTA score assessments. Demographics, medication history, and comorbidities were collected. The FORTA classification system assessed medication appropriateness. FORTA scores were calculated and FORTA score cut-offs (3 and 5) were applied. Statistical analyses included descriptive statistics, survival analysis with Cox regression, logistic regression, and negative-binomial regression using SAS 9.4 and RStudio 12.1. Ethical approval was obtained. Results: Of the 506 patients included, 171 (33.8%) were fully assessable with complete FORTA scores. The study found no significant association between higher FORTA scores and impaired cognitive performance, functional status, or mortality. Additionally, no clear relationship was observed between FORTA scores and adverse clinical events or mortality. The analysis indicated that age was a significant factor associated with mortality and adverse clinical events. Conclusion: The study did not find a significant relationship between the FORTA score and negative outcomes in older patients discharged from internal medicine and geriatric wards. Further research is needed to define specific FORTA score cut-off values and expand the criteria to improve medication assessment in this population.

Highlights of the Study

  • The study aimed to associate the Fit fOR The Aged score with negative outcomes in older patients.

  • No significant association was found between the Fit fOR The Aged score and negative outcomes like impaired cognition, adverse events or mortality.

  • The Fit fOR The Aged score did not predict negative outcomes, more research is needed to define specific cut-offs for better evaluation.

The older population is growing worldwide [1] and the progressive increase in the age diversity of the population implies a rise in the number of the most “fragile,” as they are more likely to need hospitalization. Older people frequently have multiple chronic conditions [2, 3] and are likely to receive multiple drug treatments: polypharmacy may precipitate adverse drug reactions (ADRs), which may lead to a prescribing cascade, drug-drug or drug-disease interactions, dosing and medication errors and even death [4, 5].

Strategies for prescribing medications more safely in older adults and pharmaco-epidemiological studies to assess the relation between potentially inappropriate medication (PIM) and adverse outcomes are mainly based on negative lists, such as the Beers or Screening Tool of Older Persons' Prescriptions STOPP criteria [6]. In 2008, Wehling et al. [7] introduced the Fit fOR The Aged (FORTA) classification system containing negative and positive labelling of treatments or drugs thus supporting the screening for unnecessary, inappropriate or harmful medications and for the omission of individual drugs. The FORTA system has undergone subsequent updates, with the latest revision dating from 2021 [8]. Pilot intervention studies have shown improvement of medication quality and reduction of clinical endpoints (such as falls) with FORTA compared to standard treatment [9]. To further validate the FORTA concept regarding its impact on medication quality and relevant clinical end-points, and on its practical teachability and implementation, the VALFORTA study, a randomized controlled clinical trial was carried out comparing FORTA-guided versus standard care in older patients recruited in two geriatric clinics and using the FORTA score [10]. In the VALFORTA trial, using the FORTA score improved medication quality and reduced clinical endpoints like falls. The study also established a relationship between the FORTA score and mortality, as well as cognitive and physical function outcomes among geriatric in-hospital patients.

However, only few studies have been conducted to validate the association between the FORTA score and its relationship with mortality and other adverse outcomes [11]. This study aimed to assess the relation between FORTA score and negative outcomes (impaired cognitive performance and functional status, ADR, and all-cause mortality 3, 6, and 12 months after hospital discharge) in a sample of older adult patients discharged by a sample of Italian internal medicine and geriatric wards.

Study Design and Population

This retrospective study was conducted on the ELICADHE cohort, a cluster randomized, single-blind controlled was run in 20 Italian internal medicine and geriatric wards from July 2014 to July 2015 [12], including all patients aged 75 years or over consecutively admitted to the participating wards. The ELICADHE cohort was chosen for this analysis due to its comprehensive data collection, which included detailed information on drug prescriptions, comorbidities, cognitive and physical functions, and follow-up outcomes. This made it particularly suitable for evaluating the impact of the FORTA score on various clinical endpoints and the multicenter nature of the cohort, covering a diverse range of Italian hospitals, reduces the risk of bias. Exclusion criteria were refusal of consent or estimated life expectancy <6 months. In addition to the criteria applied by the ELICADHE cohort, we excluded patients who died during hospitalization and those with incomplete follow-up data.

Data Collection

Sociodemographic details were recorded (age, sex, alcohol consumption, smoking), drug therapy (at discharge and during follow-ups), comorbidities (acute and chronic), Cumulative Illness Rating Scale (CIRS) index [13], Barthel Index (BI) [14], Mini Mental State Examination (MMSE) [15], and date of death (when recorded). Data were collected via face-to-face structured interviews of the monitors in the wards. Patients who agreed to participate gave written consent to participate in the study. The attending physicians were asked to provide all necessary information regarding the patients’ chronic diseases and prescribed medications through accessing the patients’ medical electronic and nonelectronic records. Patients assuming a missing value for a specific variable were excluded only from the analysis including that variable in the process.

FORTA Score

The FORTA classification, developed in Germany, is a patient-centered approach for evaluating the appropriateness of medications in older adults. It incorporates both negative and positive labeling for individual drugs or drug classes [8]. The FORTA criteria classify medications with regard to their overall age-appropriateness in four categories from A to D and facilitates (i) detection of therapeutic gaps (undertreatment); (ii) nonoptimal therapy, and (iii) treatment without indication (overtreatment). FORTA A (A-bsolutely) drugs have clear-cut benefits in terms of the efficacy/safety ratio; an example is β-blockers for atrial fibrillation. FORTA B (B-eneficial) drugs are beneficial but have limitations with regard to safety and efficacy: an example is sertraline prescribed for depression. FORTA C (C-areful) drugs have a questionable safety/efficacy profile, require monitoring and should be avoided. An example is amantadine for Parkinson’s disease. FORTA D (D-on’t) drugs should generally be avoided. Every long-acting benzodiazepine falls into this category.

The FORTA score is obtained by summing over-, under-, and mistreatments [10]. For each drug, one point is assigned if a symptom or condition is left untreated when beneficial pharmacological options exist (undertreatment, i.e., the absence of drugs in the FORTA categories A and B despite medical necessity) or a drug is prescribed without an appropriate medical indication (overtreatment). Two points are assigned if a symptom or condition is treated with a drug that does not belong to the best available FORTA category (mistreatment), as it represents both overtreatment and undertreatment. The total score is the sum of all single scores. Some examples are available in online supplementary Appendix S1 (for all online suppl. material, see https://doi.org/10.1159/000542109). The FORTA score was calculated at discharge and at follow-ups. Drugs not included in the FORTA score calculation were: (i) medications that were not assigned a classification by the FORTA criteria; (ii) medications prescribed for a medical condition not included in FORTA; (iii) drugs used for conditions not covered by the FORTA criteria. The alignments between FORTA diagnoses and the patients’ diagnosis are reported in online supplementary Appendix S2.

To determine the patients at risk of an unfavorable outcome, we considered a FORTA score cut-off of 3 in relation to the MMSE score and the BI score since a FORTA score greater than 3 was associated with an increased risk of dementia [16]; and a cut-off value of 5 for the relationship between the score and adverse clinical events, readmission, and all-cause mortality at follow-up since a FORTA score higher than 6 is associated with increased mortality [17]. The cut-off reported in the cited studies was chosen on the basis of the median FORTA score, analyses were repeated using the median value of our sample. A Sankey plot is employed to depict the dynamic alterations in the proportion of patients across distinct FORTA score categories (0–3, 4–5, 6+) at discharge and throughout the three follow-up: individuals who were either lost to follow-up or for whom information is unattainable are classified in the “censored” class.

Assessment of Cognitive and Daily Activities Performances

Cognitive patients’ status was evaluated by the MMSE. It includes 30 items and assesses temporal and spatial orientation, working memory, recall, attention, arithmetic capacity, and linguistic and visual motor skills. The maximum score is 30 points (one point per correct item). Any score ≥24 indicates a normal cognition. Lower scores can indicate severe (≤8 points), moderate (9–18 points) or mild (19–23 points) cognitive impairment [15]. The BI evaluates the functional ability in 10 activities of daily living. The BI total score spans from 0 to 100 points and indicates the person’s degree of dependence as follows: a score below 24 indicates total dependence, a score between 25 and 49 signifies a high level of dependence, 50–74 indicates partial dependence, 75 to 90 suggests minimal dependence, and a score between 91 and 100 reflects the ability to live independently [14].

Outcomes

The primary outcomes of this study were the association of the FORTA score with impaired cognitive performance, functional status, ADR, and all-cause mortality at 3, 6, and 12 months after hospital discharge. The relation between FORTA score and cognitive and physical functions was evaluated using MMSE and BI scores. Cognitive performance and physical function were evaluated at baseline, as well as at 3, 6, and 12 months, post-discharge. A decline was defined as a decrease in points on the assessment tool compared to baseline. Adverse clinical events were defined as any new hospitalization or acute clinical problem, including ADR that occurred from discharge to the follow-up date. For these purposes, we only considered patients eligible for this analysis who had at least one complete follow-up.

Statistical Analysis

Sociodemographic characteristics of patients were described using standard descriptive statistics. We tabulated the percentages for discrete variables and differences were evaluated with Pearson’s chi-squared test. Differences in groups were analyzed with a T test or Wilcoxon test according to their distribution.

A survival analysis was developed to assess the relationship between increasing FORTA score and 1-year mortality. First, we estimated the survival function using Kaplan-Meier curves then, having checked that proportional hazard assumption was not violated using zph tests, we used a Cox’s regression model to estimate the risk of death in the first year from discharge. This model was conducted first univariately and then adjusted by age, sex, and comorbidity burden using the CIRS.

For detection of events, we used date of death reported in our database: subjects not found during the follow-up calls after discharge were considered right-censored at the first missing time-point according to survival analysis theory. Risk of death was checked from the resulting hazard ratios (HRs) with two-sided p values and 95% confidence intervals. We then examined the relationship between FORTA score and adverse clinical events arising in the first year from discharge. Due to the lack of dates for these events in most cases, we employed a logit regression to avoid the problem. Also, for this model we showed first a univariate model and then the version corrected by age, sex, and CIRS comorbidity index. The risk of occurrences of these kinds of event according to FORTA score was evaluated using odds ratios (ORs) with two-sided p values and 95% confidence intervals.

To check the relationship between FORTA criteria and cognitive or physical functions of our sample, we assumed, respectively, MMSE and BI as the count of points scored by patients in these tests. According to generalized linear models’ theory, we conducted a negative-binomial model regression univariately, then corrected by age, sex, and CIRS, to assess how the results of these tests were related to the increasing of FORTA scores, using rate ratios (RRs) and estimated mean scores with two-sided p values and 95% confidence intervals.

For all these analyses, we considered the FORTA score at discharge, using a continuous model or different cut-offs: (0–3 vs. 4+, 0–5 vs. 6+ and a mixed three-way level variable 0–3 vs. 4–5 vs. 6+), and using our FORTA score median (0–4 vs. 5+). We performed ROC curves to detect optimal cut-offs for mortality and for the combined outcome. We used Youden’s J to identify the best threshold in our sample. Results are shown in online supplementary Appendix S3.

The significance criterion (alpha) was set at 0.05 for all tests. Analyses were using SAS 9.4 (SAS Institute Inc., Cary, NC, USA) and RStudio 12.1 (RStudio Inc., Boston, MA, USA).

Of the 700 in-patients initially considered for this study, 194 were excluded: of these, 71 patients (10.1%) were excluded because they died during hospitalization, and 123 patients (17.6%) were excluded due to incomplete follow-up data. A total of 506 patients were included in the study. Of these, 171 (33.8%) patients had prescribed drugs fully meeting the FORTA criteria and were therefore completely assessable and included in the following analysis. Supplementary analysis, also conducted on patients who are not fully assessable according to FORTA criteria, is reported in online supplementary Appendix S4.

The main sociodemographic details and characteristics of these patients, such as the mean FORTA score and adverse outcomes are provided in Table 1. Over 85% of the patients had at least one adverse clinical event and about 10% had died at 1-year follow-up.

Table 1.

Main sociodemographic details and characteristics of patients

Sample
Patients, n 171 
Age (mean±SD), years 84.4±6.0 
Women, n (%) 116 (68.2) 
FORTA score at discharge 
 Mean±SD 4.6±3.2 
 Median (25–75 percentile) 4 (2–7) 
MMSE 22.7±5.7 
 Normal 78 (53.4) 
 Mild 39 (26.7) 
 Moderate 29 (19.9) 
 Severe 0 (0) 
 Missing 15 
Barthel Index 71.1±29.1 
 Minimal 52 (33.5) 
 Mild 35 (22.6) 
 Moderate 31 (20.0) 
 Severe 22 (14.2) 
 Complete 15 (9.7) 
 Missing 16 
CIRS, comorbidity index 3.2±1.9 
Observation time, days (mean±SD) 305.9±95.1 
Δ FORTA score (discharge – last follow-up) −1.1±2.0 
Adverse clinical events, n (%) 146 (85.4) 
Mortality, n (%) 20 (11.7) 
Sample
Patients, n 171 
Age (mean±SD), years 84.4±6.0 
Women, n (%) 116 (68.2) 
FORTA score at discharge 
 Mean±SD 4.6±3.2 
 Median (25–75 percentile) 4 (2–7) 
MMSE 22.7±5.7 
 Normal 78 (53.4) 
 Mild 39 (26.7) 
 Moderate 29 (19.9) 
 Severe 0 (0) 
 Missing 15 
Barthel Index 71.1±29.1 
 Minimal 52 (33.5) 
 Mild 35 (22.6) 
 Moderate 31 (20.0) 
 Severe 22 (14.2) 
 Complete 15 (9.7) 
 Missing 16 
CIRS, comorbidity index 3.2±1.9 
Observation time, days (mean±SD) 305.9±95.1 
Δ FORTA score (discharge – last follow-up) −1.1±2.0 
Adverse clinical events, n (%) 146 (85.4) 
Mortality, n (%) 20 (11.7) 

CIRS, Cumulative Illness Rating Scale; MMSE, Mini Mental State Exam; SD, standard deviation.

The characteristics of patients at discharge and follow-ups are provided in Table 2. Most of the patients are in the group with the lowest FORTA score (lower than 3), while the mean number of drugs in the group with the highest FORTA score (higher than 6) is definitely larger than in the other groups at baseline and at each follow-up. The rate of mortality, rehospitalization, and adverse clinical outcome was not different in the three groups. Figure 1 provides the Sankey plot for patients with FORTA scores 0–3, 4–5, and 6+ and shows that most patients’ FORTA scores did not change during the follow-up.

Table 2.

Characteristics of patients at discharge and follow-ups

At dischargeFORTA scorep value
01–34–56+
Patients, n 21 (12.3) 48 (28.8) 36 (21.0) 66 (38.6) <0.0001 
Age (mean±SD), years 80.5±7.8 85.5±4.9 85.1±6.4 84.6±5.5 0.0278 
Women, n (%) 11 (47.6) 30 (63.8) 23 (63.9) 52 (78.8) 0.09 
Daily drugs (mean±SD) 6.0±2.2 6.9±3.4 8.7±2.8 11.4±3.7 <0.0001 
Three-month follow-up 
 Patients, n 25 (17.7) 48 (34.0) 31 (22.0) 37 (26.2) 0.04 
 Daily drugs (mean±SD) 4.9±1.8 5.6±1.9 6.6±2.5 8.5±2.0 <0.0001 
Six-month follow-up 
 Patients, n 27 (18.6) 52 (35.9) 31 (21.4) 35 (24.1) 0.0185 
 Daily drugs (mean±SD) 4.7±1.8 5.6±1.7 6.8±2.5 7.9±1.9 <0.0001 
Twelve-month follow-up 
 Patients, n 19 (16.7) 41 (36.0) 27 (23.7) 27 (23.7) 0.032 
 Daily drugs (mean±SD) 4.5±1.9 5.2±1.7 7.8±2.5 8.1±2.0 <0.0001 
At dischargeFORTA scorep value
01–34–56+
Patients, n 21 (12.3) 48 (28.8) 36 (21.0) 66 (38.6) <0.0001 
Age (mean±SD), years 80.5±7.8 85.5±4.9 85.1±6.4 84.6±5.5 0.0278 
Women, n (%) 11 (47.6) 30 (63.8) 23 (63.9) 52 (78.8) 0.09 
Daily drugs (mean±SD) 6.0±2.2 6.9±3.4 8.7±2.8 11.4±3.7 <0.0001 
Three-month follow-up 
 Patients, n 25 (17.7) 48 (34.0) 31 (22.0) 37 (26.2) 0.04 
 Daily drugs (mean±SD) 4.9±1.8 5.6±1.9 6.6±2.5 8.5±2.0 <0.0001 
Six-month follow-up 
 Patients, n 27 (18.6) 52 (35.9) 31 (21.4) 35 (24.1) 0.0185 
 Daily drugs (mean±SD) 4.7±1.8 5.6±1.7 6.8±2.5 7.9±1.9 <0.0001 
Twelve-month follow-up 
 Patients, n 19 (16.7) 41 (36.0) 27 (23.7) 27 (23.7) 0.032 
 Daily drugs (mean±SD) 4.5±1.9 5.2±1.7 7.8±2.5 8.1±2.0 <0.0001 

SD, standard deviation.

Fig. 1.

Sankey plot for patients with Fit fOr The Aged (FORTA) scores 0–3 (green), 4–5 (yellow), and 6+ (red) at discharge and follow-ups.

Fig. 1.

Sankey plot for patients with Fit fOr The Aged (FORTA) scores 0–3 (green), 4–5 (yellow), and 6+ (red) at discharge and follow-ups.

Close modal

Cognitive and Physical Function Outcomes

No significant relation was found between impaired cognitive performance at MMSE or impaired physical function at BI and a higher FORTA score. We used a negative-binomial regression to analyze scores in either unadjusted or adjusted models (RR (95% CI) 0.98 (0.88–1.09) p = 0.051, 0.97 (0.79–1.20) p = 0.784, FORTA class 6+ for MMSE or BI (Table 3). The analysis of the cohort of 506 patients revealed a modest association between the FORTA score and cognitive performance in the univariate model (RR [95% CI]: 0.94 [0.98–0.99)]p = 0.04), with higher FORTA scores (6+) corresponding to lower MMSE scores. However, this association lost significance after adjustment in the multivariate model (RR [95% CI]: 0.96 [0.91–1.01], p = 0.15) (online suppl. Appendix S4; Table S3).

Table 3.

Relation between FORTA score and adverse outcome

Univariate modelAdjusted modela,b
estimated meanRR – p valueestimated meanRR – p value
Physical functionc 
 FORTA Score 0–3 76.0 (66.0–87.4) 73.4 (63.8–84.4) 
 4–5 67.7 (56.1–81.6) 0.89 (0.71–1.13) – 0.33 65.9 (54.9–79.1) 0.90 (0.72–1.13) – 0.36 
 6+ 68.2 (59.3–78.4) 0.90 (0.74–1.09) – 0.28 69.4 (60.0–80.2) 0.95 (0.78–1.15) – 0.58 
Cognitive functiond 
 FORTA Score 0–3 23.6 (22.1–25.2) 23.7 (22.1–25.3) 
 4–5 22.4 (20.4–24.5) 0.95 (0.85–1.14) – 0.34 22.3 (20.3–24.4) 0.94 (0.84–1.05) – 0.29 
 6+ 21.9 (20.5–23.4) 0.93 (0.85–1.02) – 0.10 22.2 (20.7–23.9) 0.94 (0.85–1.03) – 0.20 
Univariate modelAdjusted modela,b
estimated meanRR – p valueestimated meanRR – p value
Physical functionc 
 FORTA Score 0–3 76.0 (66.0–87.4) 73.4 (63.8–84.4) 
 4–5 67.7 (56.1–81.6) 0.89 (0.71–1.13) – 0.33 65.9 (54.9–79.1) 0.90 (0.72–1.13) – 0.36 
 6+ 68.2 (59.3–78.4) 0.90 (0.74–1.09) – 0.28 69.4 (60.0–80.2) 0.95 (0.78–1.15) – 0.58 
Cognitive functiond 
 FORTA Score 0–3 23.6 (22.1–25.2) 23.7 (22.1–25.3) 
 4–5 22.4 (20.4–24.5) 0.95 (0.85–1.14) – 0.34 22.3 (20.3–24.4) 0.94 (0.84–1.05) – 0.29 
 6+ 21.9 (20.5–23.4) 0.93 (0.85–1.02) – 0.10 22.2 (20.7–23.9) 0.94 (0.85–1.03) – 0.20 
events, N (%)OR – p valueevents, N (%)OR – p value
Adverse clinical events 
 FORTA Score 0–3 58 (84.1) 56 (83.6) 
 4–5 34 (94.4) 3.22 (0.67–15.42) – 0.14 33 (94.3) 3.88 (0.78–19.2) – 0.10 
 6+ 54 (81.8) 0.85 (0.35–2.10) – 0.73 54 (81.8) 0.95 (0.36–2.52) – 0.92 
events, N (%)OR – p valueevents, N (%)OR – p value
Adverse clinical events 
 FORTA Score 0–3 58 (84.1) 56 (83.6) 
 4–5 34 (94.4) 3.22 (0.67–15.42) – 0.14 33 (94.3) 3.88 (0.78–19.2) – 0.10 
 6+ 54 (81.8) 0.85 (0.35–2.10) – 0.73 54 (81.8) 0.95 (0.36–2.52) – 0.92 
deaths, N (%)HR – p valuedeaths, N (%)HR – p value
1-year mortality 
 FORTA Score 0–3 6 (8.7) 6 (9.0) 
 4–5 6 (16.7) 2.01 (0.65–6.24) – 0.22 8 (17.1) 1.77 (0.56–5.55) – 0.32 
 6+ 8 (12.1) 1.55 (0.54–4.47) – 0.42 8 (12.1) 1.14 (0.38–3.42) – 0.82 
deaths, N (%)HR – p valuedeaths, N (%)HR – p value
1-year mortality 
 FORTA Score 0–3 6 (8.7) 6 (9.0) 
 4–5 6 (16.7) 2.01 (0.65–6.24) – 0.22 8 (17.1) 1.77 (0.56–5.55) – 0.32 
 6+ 8 (12.1) 1.55 (0.54–4.47) – 0.42 8 (12.1) 1.14 (0.38–3.42) – 0.82 

BI, Barthel Index; CIRS, Cumulative Illness Rating Scale; FORTA, Fit fOr The Aged; HR, hazard ratio; MMSE, Mini Mental State Exam; OR, odds ratio; RR, rate ratio. aCorrected by age, sex, CIRS.

bThree missing data due to sex and CIRS.

cSixteen missing data due to BI + 3 missing data due to sex, CIRS.

dTwenty-five missing data due to MMSE +3 missing data due to sex, CIRS.

Reference level: FORTA = 0–3.

Adverse Clinical Events

Of the 171 patients, 146 (85.4%) were readmitted to hospital in the 12 months after discharge or had at least one adverse clinical event. New acute clinical problems, including ADRs and hospital readmissions, do not appear to be related to a higher FORTA score, in a logit regression for the occurrence of clinical events during the follow-up period, neither in univariate nor in the multivariate models (OR [95% CI] 4.99 [0.99–25.2] p = 0.627 class FORTA 4–5; 1.50 [0.57–3.91], p = 0.441 class FORTA 6+). The odds ratios are reported in Table 3.

The analysis was repeated utilizing the median FORTA score of our sample, yielding comparable outcomes OR (95% CI) 1.98 (0.81–4.85), p = 0.134 for univariate analysis; 1.48 (0.57–3.87), p = 0.644 for multivariate analysis. Age was significantly related with an increased risk of adverse clinical events (OR [95% CI] 1.12 [1.04–1.21], p = 0.002). Similar results emerged from the analyses conducted on the cohort of 506 patients (online suppl. Appendix S4; Table S3), and no relation was found between FORTA score and adverse clinical events.

Mortality

In total, 20 (11.7%) of the 171 patients died during the observation period. No relationship was found between a higher FORTA score at discharge and mortality in a proportional hazard regression for 1-year mortality, in either univariate analysis or in a model adjusted for age, sex, and CIRS (Table 3). The Cox model showed that age was significantly associated with a higher risk of mortality HR (95% CI) 1.12 (1.04–1.21) p = 0.002.

The analysis, repeated using the median FORTA score of our sample, yielded similar results (HR (95% CI) 1.98 (0.81–4.85), p = 0.134 for univariate analysis; 1.48 (0.57–3.87), p = 0.644 for multivariate analysis and only age was significantly associated with an increased risk of mortality (HR (95% CI) 1.12 (1.04–1.21), p = 0.002. Similar results emerged from the analyses conducted on the cohort of 506 patients (online suppl. Appendix S4; Table S3), and no relation was found between FORTA score and mortality.

This study found no significant relation with any negative outcomes (impaired cognitive performance, functional status, adverse clinical events, and all-cause mortality) among older adult patients discharged from Italian internal medicine and geriatric wards. Furthermore, no correlation was found when applying specific cut-offs to distinguish between high and low FORTA scores, based on previous literature or using the median FORTA score of our sample, following the method employed by the authors of the FORTA to define cut-offs. The analysis including those patients receiving medications not covered by the FORTA criteria confirmed the main findings of the study, showing no significant relationship between FORTA scores and clinical outcomes. Importantly, this further analysis underscores the limitations of the FORTA tool in covering the full spectrum of medications used in clinical practice, which could partly explain the lack of association with adverse outcomes.

This investigation builds on the foundation laid by the VALFORTA study, which demonstrated the efficacy of FORTA-guided care in improving medication quality and clinical endpoints, and found relationships between the FORTA score, mortality and cognitive and physical function outcomes among geriatric in-hospital patients (mean age 81.5 years) [10]. However, we found no relation with clinical outcome despite a sample of patients with similar characteristics to those included in the VALFORTA study. This discrepancy may be attributed to difficulties in the FORTA assessment and differences in the follow-up.

First, about 67% of the patients were excluded from the analysis because of the lack of a complete FORTA score, as some medications are not considered by FORTA criteria (e.g., paroxetine and promazine) or for the lack of inclusion of numerous chronic conditions with their corresponding medications, such as benign prostatic hyperplasia or gout. This issue has been extensively discussed in other studies where, due to the scarcity of drugs and classified conditions, FORTA identified a smaller percentage of potentially inappropriate psychotropic medications in nursing home residents compared to the Beers and STOPP criteria [18] and compared to the EU(7)-PIM list in eight different study centers in Germany [19].

Again, the different observation periods we considered might explain our findings. While Pazan et al.’s [17] study involved six follow-up assessments at an average interval of 1 and a half years, our observation period was only 1 year with follow-up assessments at 3, 6, and 12 months. In a prior association study by one of the creators of the FORTA criteria, higher FORTA scores, indicating more frequent medication errors, were linked to impaired cognitive and physical function tests in older hospital patients [16]. The VALFORTA trial demonstrated a significant improvement in activities of daily living through the application of the FORTA intervention [10]. So, in the smaller number of patients evaluated at 1 year the relation between inappropriate prescribing measured with the FORTA score and certain adverse clinical outcomes, such as mortality, may not be detectable.

Differences in results may also be due to the unclear definition of the cut-offs for the FORTA score. Because of the low dimension of our sample, our ROC analysis may suffer from overfitting bias, so we used author’s cut-offs. Few studies propose specific cut-offs, mainly established by the criteria authors, and the figurers vary.

One aspect that warrants discussion is the source of our data, the ELICADHE study – a randomized controlled trial aimed at optimizing prescribing quality. The intervention’s focus on optimizing prescribing quality might have mitigated some of the risks associated with polypharmacy and PIM use, potentially blurring the relationship between FORTA scores and adverse outcomes but the ELCIADHE study failed to improve clinician drug prescription for hospitalized older patients and found no significant differences in the prevalence of PIM, drug-disease interaction or mortality between the intervention and control groups [12].

One of the primary strengths of this observational study lies in its assessment of the FORTA score in a sample of older adults and its correlation with functional status, adverse clinical events, and all-cause mortality. This study stands out as one of the few, beyond the VALFORTA and those conducted by the authors of the criteria, that assesses the applicability of the FORTA score in a setting where patient data are largely comprehensive, both pharmacologically and clinically: the data used in the study were in fact collected during the ELICADHE study, approved and financially supported by the Italian Medicines Agency (AIFA), ensuring their accuracy and comprehensiveness.

Our study has several potential biases that may affect the interpretation of the results. First, selection bias could have been introduced by including only patients with a complete FORTA score assessment, potentially limiting the representativeness of the sample, reducing the statistical power of the study, and increasing the uncertainty of the findings, as reflected in the wide confidence intervals. The exclusion of several patients, due to the strict application of the FORTA criteria, further contributed to influence the small sample size. Although our study focused on a precise and strict application of these criteria, this selection process impacted the generalizability of our findings: the study sample may not fully reflect the broader population of older adults in hospital settings. Attrition bias is also a concern due to the loss of patients during follow-up, particularly in the 12-month follow-up, which may have resulted in an unbalanced analysis. The short observation period might not have been sufficient to capture the full impact of PIM, especially long-term effects. Furthermore, we did not evaluate in-hospital mortality, which may have introduced additional limitations in assessing the overall impact of medication appropriateness on patient outcomes. The large number of exclusions due to incomplete FORTA assessments, coupled with the relatively short duration of patient observation and significant follow-up losses introduce uncertainty in our results, and broader generalizations should be made with caution. The lack of statistical significance and broad confidence intervals further underscore the need for careful interpretation of our findings.

Further research is necessary to clarify the relationship between the FORTA score with a specific cut-off and different outcomes. We hope that future versions of the FORTA criteria will include more pathologies, for a fuller evaluation of patients and a list of ICD codes for FORTA diagnosis to reduce the risk of wrong alignment.

Our findings suggest that the FORTA score may not have a clear or consistent relationship with impaired cognitive and physical function, adverse clinical events, or mortality among older adult discharged from internal medicine and geriatric wards. However, these results are uncertain and should be interpreted with caution due to potential biases and relevant limitations, such as the small sample size, exclusion of patients with incomplete FORTA scores, and the relatively short observation period. Further research is needed to define specific cut-off values for different clinical outcomes and the number of drugs and clinical conditions considered in the FORTA criteria should be expanded to improve the accuracy and comprehensiveness of the FORTA score assessment.

The authors are very grateful to the investigators for data collection and to J.D. Baggott for language editing. See Electronic online supplementary material Appendix S5 for a list of investigators and co-authors.

This was a retrospective study and data collection complied fully with Italian law on personal data protection. All data were anonymous and informed consent was not required for the purpose of the study.

The study was first approved by the Ethical Committee of the coordinating clinical unit (IRCCS Cà Granda Maggiore Hospital Foundation, Milan, Italy). The ELICADHE study was approved and financially supported by the Italian Medicines Agency (AIFA) according to the 2008 Italian Program for Independent Research (Project no. FARM87SA2B).

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

All authors participated in drafting or critical revision for important intellectual content. Individual contributions are as follows: Marina Azab and Luca Pasina designed the study, interpreted data, and wrote the manuscript; Alessio Novella did and interpreted statistical analyses. All authors read, approved the final version of the paper, and agreed to be accountable for the work.

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The codes that support the findings of this study are available on request from the corresponding author.

1.
Statistics | eurostat [Internet]. [cited 2023 Apr 18]. Available from: https://ec.europa.eu/eurostat/databrowser/view/demo_pjanind/default/table?lang=en
2.
Fried
LP
,
Tangen
CM
,
Walston
J
,
Newman
AB
,
Hirsch
C
,
Gottdiener
J
, et al
.
Frailty in older adults: evidence for a phenotype
.
J Gerontol A Biol Sci Med Sci
.
2001
;
56
(
3
):
M146
156
.
3.
Rockwood
K
,
Stadnyk
K
,
MacKnight
C
,
McDowell
I
,
Hébert
R
,
Hogan
DB
.
A brief clinical instrument to classify frailty in elderly people
.
Lancet
.
1999
;
353
(
9148
):
205
6
.
4.
Eckerblad
J
,
Theander
K
,
Ekdahl
A
,
Unosson
M
,
Wirehn
A-B
,
Milberg
A
, et al
.
Symptom burden in community-dwelling older people with multimorbidity: a cross-sectional study
.
BMC Geriatr
.
2015
;
15
(
1
):
1
.
5.
Viktil
KK
,
Blix
HS
,
Moger
TA
,
Reikvam
A
.
Polypharmacy as commonly defined is an indicator of limited value in the assessment of drug-related problems
.
Br J Clin Pharmacol
.
2007
;
63
(
2
):
187
95
.
6.
Curtin
D
,
Gallagher
PF
,
O’Mahony
D
.
Explicit criteria as clinical tools to minimize inappropriate medication use and its consequences
.
Ther Adv Drug Saf
.
2019
;
10
:
2042098619829431
.
7.
Wehling
M
.
[Drug therapy in the elderly: too much or too little, what to do? A new assessment system: fit for the aged (FORTA)]
.
Dtsch Med Wochenschr
.
2008
;
133
(
44
):
2289
91
.
8.
Pazan
F
,
Weiss
C
,
Wehling
M
,
Bauer
JM
,
Berthold
HK
.
The FORTA (fit fOR the aged) list 2021: fourth version of a validated clinical aid for improved pharmacotherapy in older adults
.
Drugs Aging
.
2022
;
39
(
3
):
245
7
.
9.
Michalek
C
,
Wehling
M
,
Schlitzer
J
,
Frohnhofen
H
.
Effects of “Fit fOR the Aged” (FORTA) on pharmacotherapy and clinical endpoints: a pilot randomized controlled study
.
Eur J Clin Pharmacol
.
2014
;
70
(
10
):
1261
7
.
10.
Wehling
M
,
Burkhardt
H
,
Kuhn-Thiel
A
,
Pazan
F
,
Throm
C
,
Weiss
C
, et al
.
VALFORTA: a randomised trial to validate the FORTA (Fit fOR the Aged) classification
.
Age Ageing
.
2016
;
45
(
2
):
262
7
.
11.
Chen
L-J
,
Nguyen
TNM
,
Laetsch
DC
,
Chang-Claude
J
,
Hoffmeister
M
,
Brenner
H
, et al
.
Association of comedication quality with chemotherapy-related adverse drug reactions and survival in older colorectal cancer patients
.
J Gerontol A Biol Sci Med Sci
.
2022
;
77
(
5
):
1009
19
.
12.
Franchi
C
,
Tettamanti
M
,
Djade
CD
,
Pasina
L
,
Mannucci
PM
,
Onder
G
, et al
.
E-learning in order to improve drug prescription for hospitalized older patients: a cluster-randomized controlled study
.
Br J Clin Pharmacol
.
2016
;
82
(
1
):
53
63
.
13.
Salvi
F
,
Miller
MD
,
Grilli
A
,
Giorgi
R
,
Towers
AL
,
Morichi
V
, et al
.
A manual of guidelines to score the modified cumulative illness rating scale and its validation in acute hospitalized elderly patients
.
J Am Geriatr Soc
.
2008
;
56
(
10
):
1926
31
.
14.
Stone
SP
,
Ali
B
,
Auberleek
I
,
Thompsell
A
,
Young
A
.
The Barthel index in clinical practice: use on a rehabilitation ward for elderly people
.
J R Coll Physicians Lond
.
1994
;
28
(
5
):
419
23
.
15.
Folstein
MF
,
Folstein
SE
,
McHugh
PR
.
“Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician
.
J Psychiatr Res
.
1975
;
12
(
3
):
189
98
.
16.
Pazan
F
,
Burkhardt
H
,
Frohnhofen
H
,
Weiss
C
,
Throm
C
,
Kuhn-Thiel
A
, et al
.
Higher Fit-fOR-the-Aged (FORTA) scores comprising medication errors are associated with impaired cognitive and physical function tests in the VALFORTA trial
.
Drugs Aging
.
2019
;
36
(
3
):
269
77
.
17.
Pazan
F
,
Breunig
H
,
Weiss
C
,
Röhr
S
,
Luppa
M
,
Pentzek
M
, et al
.
Higher FORTA (Fit fOR the Aged) scores are associated with poor functional outcomes, dementia, and mortality in older people
.
Eur J Clin Pharmacol
.
2022
;
78
(
11
):
1851
9
.
18.
Azab
M
,
Novella
A
,
Ianes
A
,
Pasina
L
.
Potentially inappropriate psychotropic drugs in nursing homes: an Italian observational study
.
Drugs Aging
.
2024
;
41
(
2
):
187
97
.
19.
Krüger
C
,
Schäfer
I
,
van den Bussche
H
,
Bickel
H
,
Dreischulte
T
,
Fuchs
A
, et al
.
Comparison of FORTA, PRISCUS and EU(7)-PIM lists on identifying potentially inappropriate medication and its impact on cognitive function in multimorbid elderly German people in primary care: a multicentre observational study
.
BMJ Open
.
2021
;
11
(
9
):
e050344
.