Key summary points
Aim
To assess the prevalence of the latest version of STOPP/START criteria (Version 3) in a nationally representative sample of older community-dwelling adults and explore any association with healthcare utilisation and functional decline over time.
Findings
In a sample of 3619 older Irish adults, a total of 31% (n = 1123) experienced STOPP potentially inappropriate medications (PIMs) and exposure to any STOPP PIM was associated with increased hospital admissions and functional decline over time. Just over one-third of participants (36.2%; n = 1309) experienced START potential prescribing omissions (PPOs), with increasing age and number of chronic conditions associated with START PPOs.
Message
Over one-third of participants experienced a subset of STOPP PIMs, another one-third experienced a subset of START PPOs. Balancing the relative benefits and risks of medication in older adults with multimorbidity continues to present challenges for both prescribers and older adults.
Supplementary Information
The online version contains supplementary material available at 10.1007/s41999-025-01201-3.
Keywords: Potentially inappropriate prescribing, STOPP, START, Healthcare utilisation, Activities of daily living
Abstract
Purpose
Potentially inappropriate prescribing includes prescribing potentially inappropriate medicines (PIMs), where risk of medication-related harm may outweigh the clinical benefit(s), and potential prescribing omissions (PPOs), whereby clinically indicated medications are unprescribed without good reason. This study aimed to assess prevalence of PIMs and PPOs (subset of STOPP/START version 3) in older community-dwelling adults and any association with healthcare utilisation and functional decline over time.
Methods
Retrospective cohort study of a nationally representative longitudinal study of ageing in Ireland (n = 3619) (2016–2018). Logistic regressions examined association of patient characteristics with PIMs/PPOs and between prevalent PIMs/PPOs and functional decline. Negative binomial regressions examined association between PIM/PPO with healthcare utilisation over time.
Results
Participants’ mean age was 74.2 years (SD 6.99), 53.9% were female and were prescribed a mean of 4.02 (SD 3.16) medications. A total of 1123 (31.0%) participants experienced STOPP PIMs and 1309 (36.2%) START PPOs. STOPP PIMs were associated with increased hospital admissions (adjusted incident rate ratio (aIRR) 1.38, 95% confidence interval (CI) 1.08, 1.75), and functional decline (adjusted odds ratio (aOR) 1.46, 95% CI 1.11, 1.91) at follow-up. Age ≥ 75 years (aOR 1.32, 95% CI 1.10, 1.57) and three or more chronic conditions (aOR 5.19, 95% CI 3.69, 7.31) were significantly associated with START PPOs.
Conclusion
Approximately one-third of study participants experienced STOPP PIMs, associated with an increased risk of hospital admissions and functional decline. START PPOs also occurred in over one-third, associated with increasing age and degree of multimorbidity. Balancing the risk: benefit of medications for older people with multimorbidity remains challenging.
Supplementary Information
The online version contains supplementary material available at 10.1007/s41999-025-01201-3.
Introduction
Potentially inappropriate medications (PIMs) are medications where the potential for medication-related harm may outweigh the clinical benefit(s) to the patient. Potentially inappropriate prescribing (PIP) includes PIMs as well as potential prescribing omissions (PPOs), whereby clinically indicated medications are not prescribed [1].
Older adults are particularly vulnerable to PIP as they are more likely to live with multiple long-term health conditions. More than half of the global population aged ≥ 60 years are estimated to have multimorbidity, defined as the presence of two or more chronic conditions [2]. Consequently, older adults are more likely to experience polypharmacy (often defined as the use of five or more daily long-term prescribed medications), with international prevalence rates of 45%, and higher rates in frail older adults (59%) [3]. Age-related changes in physiology (e.g. renal function, liver function, body mass etc.) mean that often both the metabolism of and response to medication is altered in older adults, resulting in an underlying biological vulnerability to adverse drug reactions (ADRs) [4]. The prevalence of ADRs is higher amongst older adults (≥ 65 years), with more than one-quarter shown to experience an ADR and 1 in 10 hospital admissions associated with serious ADRs [5–7]. PIP is also associated with a range of other adverse outcomes including falls and functional decline [8]. Nevertheless, comparatively fewer studies have explored associations between PIP and functional decline, with limited exploration of the relationship for PPOs specifically [8]. Direct healthcare costs for those patients exposed to PIP are more than twice the total cost of those without PIP [9].
Explicit criteria to identify PIP are lists of medications or medication/disease combinations considered harmful when used in older adults. Numerous PIP explicit criteria for older adults have been published internationally including the US Beers’ list, French Laroche list and Norwegian General Practice (NORGEP) criteria [10–12]. In addition, the European Screening Tool of Older Person’s potentially inappropriate Prescriptions (STOPP) and the Screening Tool to Alert to Right Treatment (START) criteria assess PIMs and PPOs, respectively [13, 14]. Originally developed by expert consensus in Ireland, the 2008 STOPP/START criteria have undergone considerable revision since the first iteration. STOPP/START Version 3, published in 2023 represents an approximate 200% increase in the original number of prescribing indicators in version 1, reflecting Europe-wide consensus on newer medications and the expanding pharmacotherapy evidence base for older people [15].
Longitudinal cohort studies of ageing allow examination of temporal changes in PIP through the application of explicit criteria such as STOPP/START. In Ireland, The Irish Longitudinal Study on Ageing (TILDA), has recently been designated as the World Health Organisation Collaborating Centre for Longitudinal Studies on Ageing and the Life Course [16, 17]. Previous application of STOPP/START version 1 criteria to TILDA self-reported medication data identified prevalence rates of PIMs and PPOs of 14.6% and 30%, respectively [18]. When later examined in a subset of TILDA participants with linked administrative pharmacy claims data, the overall prevalence was higher; 57% and 41.8% for PIM and PPO, respectively [19]. This observable increase likely reflects the application of more indicators due to additional data provided by the linked pharmacy claims data. The aim of this study was to examine the prevalence of PIM and PPO, identified by STOPP/START Version 3, in older Irish adults and to examine any association with healthcare utilisation and functional decline over time.
Methods
The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement has been used in reporting this study (see Additional File 1) [20].
Study design, population and data collection
A retrospective cohort study was conducted using a nationally representative longitudinal study of ageing in Ireland (TILDA). TILDA respondents are community-dwelling adults aged ≥ 50 years. Baseline data collection was conducted between 2009 and 2011, with respondents followed up every two years thereafter. Data collection procedures have been detailed elsewhere [16, 17], but in summary involved the completion of a computer-assisted personal interview (CAPI) and self-completion questionnaire (SCQ) at each wave, that includes questions regarding current medications and diagnoses. Participants complete the CAPI and SCQ at each wave (every 2 years) and the health assessment at alternate waves (e.g. Waves 1 and 3). Medication data, collected by CAPI, demonstrates both internal and external validity [21]. Morbidity self-reports from TILDA participants have shown moderate to very good agreement with medication-based morbidity assessments for many chronic conditions included in the present study [22]. This study utilised data from the CAPI and SCQ for waves 1–5. For this study, Wave 4 (2016) was used as the baseline assessment point with demographical characteristics at Wave 4 used to describe the sample. Follow-up outcomes (healthcare utilisation and functional decline) were examined at Wave 5 (2018). Data from Waves 1–3 were examined for some indicators where prior use of a medication or reporting of a disease was relevant e.g. primary prevention of cardiovascular disease required an examination as to whether cardiovascular disease was ever reported at an earlier wave. At Wave 4, a total of 5973 respondents completed data collection. Data pertaining to those who were aged ≥ 65 years at Wave 4 (N = 3619) were included in the analysis.
Variables
Primary outcome variables
The primary outcome of interest was exposure to PIMs, assessed using STOPP criteria, and PPOs, assessed using START criteria (Version 3) at Wave 4. A subset of STOPP/START Version 3 were applied in line with previous applications of Version 1 to earlier TILDA data [18, 19]. The selected subset of indicators was based on the availability of medication and disease information within the database. Sixty STOPP criteria (45% of 133 STOPP criteria) and 16 START criteria (28% of 57 START criteria) were applied. In the absence of disease information, prescription medications for the treatment of certain conditions were used as proxies e.g. the use of allopurinol or colchicine was used as a marker of history of gout. A detailed description of how the various indicators were defined is provided in Additional File 2.
Secondary outcome variables
Healthcare utilization (self-reported use over the previous 12 months) at Wave 5 (2018) was operationalized as the number of general practitioner (GP) visits; Emergency Department (ED) visits; number of outpatient visits; and number of hospital admissions. Self-reported limitations in activities of daily living (ADLs) and instrumental activities of daily living (IADLs) were reported during the CAPI. This is measured in TILDA as limitations in ADLs (help with walking across a room, dressing, bathing, eating, getting in and out of bed, and using the toilet) and IADLs (preparing meals, shopping for groceries, making telephone calls, taking medications and managing money) [23]. The total number of both ADL and IADL limitations was calculated at waves 4 and 5. To operationalise functional decline, a binary variable was computed to capture those participants who acquired at least one additional ADL or IADL impairment between waves 4 and 5.
Explanatory variables
Explanatory variables included age group (65–74 years; ≥ 75 years) and sex (male/female). To examine socioeconomic status three variables were included, i) educational attainment; ii) employment status; and iii) type of healthcare coverage. Educational attainment was categorised as primary level/none; secondary level; third level/higher education. Employment status was categorised as employed; retired; or other. Healthcare coverage was categorised as no cover; public General Medical Services (GMS) scheme cover only; private health insurance cover only; and dual GMS and private coverage. The GMS scheme provides access to free healthcare, as well as access to eligible prescriptions which are available via a low co-payment that is capped monthly per household. GMS scheme eligibility is based upon income-related means testing, with a higher income threshold set for those ≥ 70 years. The scheme provides coverage for more than 60% of those aged ≥ 65 years and approximately 80% of those aged ≥ 75 years in Ireland [24].
Multimorbidity was defined as the number of chronic conditions, from a list of 21 conditions captured during data collection and previously applied by Larkin et al. [25]. Further detail regarding the list of conditions included is reported in Additional File 3. Multimorbidity was further categorised as none, one, two, or ≥ 3 chronic conditions, in line with previous TILDA studies [18, 19]. Polypharmacy was categorised as 0–4 medications (no polypharmacy); 5–9 medications (polypharmacy); ≥ 10 medications (hyperpolypharmacy), based upon self-reported regular medication use reported during the CAPI. Participants were asked to report all medications taken regularly, including prescription and non-prescription medication, over-the-counter medications and supplements.
Analytical plan
Descriptive statistics were used to summarise the population. The overall prevalence of PIM/PPO (any STOPP/START indicator) and the prevalence per individual criteria were calculated as the proportion of all eligible participants aged ≥ 65 years. The prevalence of PIM/PPO was also calculated as the proportion of overall disease or medication prevalence (e.g. drugs with potent anticholinergic/antimuscarinic effects in patients with dementia as a proportion of dementia prevalence). Further detail on the drug/disease denominator used for each criterion is provided in Additional File 2.
Univariate associations between age, sex, education, employment status, healthcare coverage, multimorbidity, and polypharmacy and presence of PIM/PPO (with none as referent) were explored using logistic regression analyses, reported as odds ratios (ORs) with 95% confidence intervals (CI). Multivariable associations were explored using multivariable logistic regression analysis.
Separate negative binomial regressions were conducted to examine the relationship between experiencing PIM/PPO (any PIM/PPO (model 1) and 0, 1 and ≥ 2 STOPP/START criteria (model 2)) at Wave 4 and healthcare utilisation at Wave 5. The models were adjusted for relevant confounders (age, sex, education, employment status, healthcare coverage, multimorbidity, polypharmacy, and healthcare utilisation at Wave 4). Results are reported as incident rate ratios (IRRs) with 95% CIs. Logistic regression analyses examined the relationship between experiencing PIM/PPO and functional decline from Wave 4 to Wave 5, adjusted for relevant confounders for each outcome (age, sex, education, employment status, healthcare coverage, multimorbidity, polypharmacy, and the presence of any ADL/IADL impairment at Wave 4). Results are reported as odds ratios with 95% CIs. Data analyses were performed using Stata version 18 (StataCorp) with an alpha level of p < 0.05.
Results
Descriptive statistics for the overall sample (n = 3619) and by experience of any PIM and any PPO are presented in Table 1. The mean age of study participants was 74.20 years (SD 6.99) and just over half (53.94%) were female. The mean number of regular medications was 4.02 (SD 3.16). Approximately two-fifths of participants were prescribed five or more medications, with 6.02% of participants prescribed ≥ 10 medications.
Table 1.
Descriptive statistics for study participants
| Overall N = 3619 n (%) |
No STOPP N = 2496 n (%) |
Any STOPP N = 1123 n (%) |
No START N = 2310 n (%) |
Any START N = 1309 n (%) |
|
|---|---|---|---|---|---|
| Age in years, M (SD) | 74.20 (6.99) | 73.67 (6.88) | 75.37 (7.09) | 73.68 (6.89) | 75.12 (7.06) |
| 65–74 years | 2114 (58.41) | 1532 (61.38) | 582 (51.83) | 1423 (61.60) | 691 (52.79) |
| ≥ 75 years | 1505 (41.59) | 964 (38.62) | 541 (48.17) | 887 (38.40) | 618 (47.21) |
| Sex | |||||
| Male | 1667 (46.06) | 1191 (47.72) | 476 (42.39) | 1165 (50.43) | 502 (38.35) |
| Female | 1952 (53.94) | 1305 (52.28) | 647 (57.61) | 1145 (49.57) | 807 (61.65) |
| Education | |||||
| Primary/none | 1194 (32.99) | 763 (30.57) | 431 (38.38) | 761 (32.94) | 433 (33.08) |
| Secondary | 1316 (36.36) | 919 (36.82) | 397 (35.35) | 818 (35.41) | 498 (38.04) |
| Third/higher | 1109 (30.64) | 814 (32.61) | 295 (26.27) | 731 (31.65) | 378 (28.88) |
| Employment statusa | |||||
| Employed | 456 (12.61) | 360 (14.43) | 96 (8.55) | 354 (15.34) | 102 (7.79) |
| Retired | 2559 (70.75) | 1741 (69.81) | 818 (72.84) | 1610 (69.76) | 949 (72.50) |
| Other | 602 (16.64) | 393 (15.76) | 209 (18.61) | 344 (14.90) | 258 (19.71) |
| Health insuranceb | |||||
| Not covered | 122 (4.04) | 100 (4.82) | 22 (2.34) | 87 (4.50) | 35 (3.23) |
| Private insurance only | 734 (24.32) | 585 (28.18) | 149 (15.82) | 518 (26.80) | 216 (19.91) |
| GMS cover only | 1319 (43.70) | 838 (40.37) | 481 (51.06) | 804 (41.59) | 515 (47.47) |
| Dual cover | 843 (27.93) | 553 (26.64) | 290 (30.79) | 524 (27.11) | 319 (29.40) |
| Number of morbidities, M (SD) | 2.36 (1.62) | 2.09 (1.53) | 2.97 (1.66) | 2.00 (1.48) | 2.98 (1.67) |
| Multimorbidity category | |||||
| None | 419 (11.58) | 367 (14.70) | 52 (4.63) | 356 (15.41) | 63 (4.81) |
| 1 chronic condition | 738 (20.39) | 578 (23.16) | 160 (14.25) | 560 (24.24) | 178 (13.60) |
| 2 chronic conditions | 940 (25.97) | 693 (27.76) | 247 (21.99) | 628 (27.19) | 312 (23.83) |
| ≥ 3 chronic conditions | 1522 (42.06) | 858 (34.38) | 664 (59.13) | 766 (33.16) | 756 (57.75) |
| Number of prescribed medications, M (SD) | 4.02 (3.16) | 3.17 (2.73) | 6.09 (3.18) | 3.65 (3.06) | 4.69 (3.21) |
| Polypharmacy category | |||||
| 0–4 medications | 2213 (61.15) | 1822 (73.00) | 391 (34.82) | 1523 (65.93) | 690 (52.71) |
| 5–9 medications | 1188 (32.83) | 614 (24.60) | 574 (51.11) | 679 (29.39) | 509 (38.88) |
| ≥ 10 medications | 218 (6.02) | 60 (2.40) | 158 (14.07) | 108 (4.68) | 110 (8.40) |
aData missing for two participants; b data missing for 601 participants
Just under one-third of participants (n = 1123, 31.03%) experienced one or more PIMs, with 814 participants (22.49%) having one STOPP PIM indicator and 309 (8.54%) with two or more STOPP PIM indicators (Table 2). Of the total study population (n = 3619) the most prevalent STOPP indicator was the use of aspirin for primary prevention of cardiovascular disease (14.78%). Across 15 different drug classes a total of 125 participants (3.45%) reported duplicate prescribing (Table 2). Other prevalent indicators included the use of NSAIDs and z-drugs beyond their recommended duration, representing 2.76% and 2.57%, respectively. When both indicators were expressed as the proportion of drug class prescribing the prevalence was found to be 68.03% and 59.62%, respectively. A total of 45.45% of TILDA participants aged ≥ 85 years with established frailty were found to be using statins for primary cardiovascular prevention.
Table 2.
STOPP Version 3 criteria applied to TILDA data for study participants aged ≥ 65 years (N = 3619)
| STOPP criteria description | PIM (n) | PIM (%) | Prescribing per indication (%)a |
|---|---|---|---|
| Indication of medication | |||
| Any drug prescribed beyond the recommended duration | |||
| Z-drugs | 93 | 2.57 | 59.62 |
| NSAIDs | 100 | 2.76 | 68.03 |
| Any duplicate drug class prescription | 125 | 3.45 | 4.73 |
| Duplicate NSAIDs | 11 | 0.3 | 4.82 |
| Duplicate SSRIs | 7 | 0.19 | 3.52 |
| Duplicate loop diuretics | 4 | 0.11 | 2 |
| Duplicate ACEIs | 12 | 0.33 | 1.88 |
| Duplicate anticoagulants | 6 | 0.17 | 1.81 |
| Duplicate antipsychotics | 6 | 0.17 | 8.45 |
| Duplicate opioids | 11 | 0.3 | 6.04 |
| Duplicate stimulant laxatives | 0 | – | – |
| Duplicate thiazides | 0 | – | – |
| Duplicate beta blockers | 16 | 0.44 | 1.94 |
| Duplicate ARBs | 14 | 0.39 | 2.01 |
| Duplicate CCBs | 9 | 0.25 | 1.44 |
| Duplicate statins | 25 | 0.69 | 1.59 |
| Duplicate benzodiazepines | 8 | 0.22 | 5.63 |
| Duplicate TCAs | 0 | – | – |
| Cardiovascular system | |||
| Beta-blocker in combination with verapamil or diltiazem | 11 | 0.3 | 17.19 |
| Beta-blocker as monotherapy for uncomplicated hypertension | 42 | 1.16 | 4.12 |
| Loop diuretic as first-line treatment for hypertension unless there is concurrent heart failure | 24 | 0.66 | 1.64 |
| Thiazide diuretic with a history of gout (defined as prior use of anti-gout medication) | 4 | 0.11 | 2.53 |
| Loop diuretic for treatment of hypertension with concurrent urinary incontinence | 21 | 0.58 | 6.46 |
| Centrally acting antihypertensives | 0 | – | – |
| Phosphodiesterase type-5 inhibitors with concurrent nitrate therapy for angina | 1 | 0.03 | 1.3 |
| Statins for primary cardiovascular prevention in persons aged ≥ 85 and established frailty | 10 | 0.28 | 45.45 |
| Long-term systemic i.e. non-topical NSAIDs with known history of coronary, cerebral or peripheral vascular disease | 14 | 0.39 | 1.43 |
| Long-term antipsychotics with known history of coronary, cerebral or peripheral vascular disease | 11 | 0.30 | 1.12 |
| NSAIDs or systemic corticosteroids with heart failure requiring loop diuretic therapy | 1 | 0.03 | 7.69 |
| Coagulation system | |||
| Ticlopidine in any circumstances | 0 | – | – |
| Antiplatelet agents as alternatives to vitamin K antagonists, direct thrombin inhibitors or factor Xa inhibitors for stroke prevention in patients with chronic atrial fibrillation | 30 | 0.83 | 19.48 |
| NSAIDs and vitamin K antagonist, direct thrombin inhibitor or factor Xa inhibitors in combination | 6 | 0.17 | 2.48 |
| Direct thrombin inhibitor (e.g. dabigatran) and diltiazem or verapamil | 7 | 0.19 | 10.94 |
| Apixaban, dabigatran, edoxaban, rivaroxaban and P-glycoprotein (P-gp) drug efflux pump inhibitors | 15 | 0.41 | 11.72 |
| Aspirin for primary prevention in cardiovascular disease | 535 | 14.78 | 21.80 |
| Central nervous system | |||
| Tricyclic antidepressants (TCAs) in patients with dementia, glaucoma or recent falls | 35 | 0.97 | 3.49 |
| Benzodiazepines for ≥ 4 weeks | 64 | 1.77 | 55.65 |
| Benzodiazepines for agitated behaviour or psychotic symptoms of dementia | 0 | – | – |
| Benzodiazepines for insomnia for ≥ 2 weeks | 27 | 0.75 | 23.48 |
| Z-drugs (zolpidem, zopiclone, zaleplon) for insomnia for ≥ 2 weeks | 39 | 1.08 | 25.00 |
| Antipsychotics (i.e. other than clozapine or quetiapine) in those with parkinsonism | 1 | 0.03 | 2.5 |
| Anticholinergic/antimuscarinic drugs to treat extra-pyramidal side-effects of antipsychotic medications | 0 | – | – |
| Drugs with potent anticholinergics/antimuscarinic effects in patients with dementia | 7 | 0.19 | 8.14 |
| Antipsychotics as hypnotics, unless sleep disorder is due to psychosis or BPSD effects of dementia | 10 | 0.28 | 8.4 |
| Acetylcholinesterase inhibitors with concurrent treatment with drugs that induce persistent bradycardia such as beta-blockers, digoxin, diltiazem, verapamil | 15 | 0.41 | 1.68 |
| Memantine with known current or previous seizure disorder | 0 | – | – |
| Nootropics in dementia | 0 | – | – |
| Levodopa or dopamine agonists for treatment of extra-pyramidal side-effects of antipsychotics (inappropriate prescribing cascade) | 0 | – | – |
| Gastrointestinal system | |||
| Prochlorperazine or metoclopramide with Parkinsonism | 0 | – | – |
| Corticosteroids with a history of peptic ulcer disease | 2 | 0.06 | 0.62 |
| Respiratory system | |||
| Theophylline as monotherapy for COPD | 0 | – | – |
| Systemic corticosteroids instead of inhaled corticosteroids for maintenance therapy in moderate-severe COPD | 5 | 0.14 | 9.43 |
| Long-acting muscarinic antagonists with a history of glaucoma | 4 | 0.11 | 3.05 |
| Musculoskeletal system | |||
| Non-steroidal anti-inflammatory drugs (NSAIDs) other than COX-2 selective agents with history of peptic ulcer disease or gastrointestinal bleeding, unless with concurrent PPI or H2 antagonist | 14 | 0.39 | 4.32 |
| Long-term corticosteroids (> 3 months) as monotherapy for rheumatoid arthritis | 4 | 0.11 | 1.30 |
| Corticosteroids (other than periodic intra-articular injections for mono-articular pain) for osteoarthritis | 20 | 0.55 | 2.77 |
| NSAID with concurrent corticosteroids for treatment of arthritis/rheumatism of any kind | 3 | 0.08 | 0.22 |
| Oral bisphosphonates in patients with a current or recent history of upper gastrointestinal | 18 | 0.5 | 5.56 |
| Long-term opioids for osteoarthritis | 32 | 0.88 | 6.13 |
| Urogenital system | |||
| Systemic antimuscarinic drugs with dementia or chronic cognitive impairment | 8 | 0.22 | 9.3 |
| Systemic antimuscarinic drugs with glaucoma | 11 | 0.30 | 8.4 |
| Endocrine system | |||
| Sulphonylureas with a long half-life with type 2 diabetes mellitus | 0 | – | – |
| Thiazolidenediones with heart failure | 0 | – | – |
| Systemic oestrogens with a history of breast cancer | 1 | 0.03 | 0.89 |
| Vasopressin analogues for urinary incontinence or urinary frequency | 1 | 0.03 | 0.14 |
| Drug classes that predictably increase falls risk | |||
| Benzodiazepines in patients with recurrent falls | 18 | 0.50 | 4.83 |
| Antipsychotic drugs in patients with recurrent falls | 10 | 0.28 | 2.68 |
| Hypnotic Z-drugs in patients with recurrent falls | 30 | 0.83 | 8.04 |
| Anti-epileptic drugs in patients with recurrent falls | 47 | 1.30 | 12.60 |
| First generation antihistamines in patients with recurrent falls | 2 | 0.06 | 0.54 |
| Opioids in patients with recurrent falls | 39 | 1.08 | 10.46 |
| Antidepressants in patients with recurrent falls | 97 | 2.68 | 26.01 |
| Centrally acting antihypertensives | 0 | – | – |
| Antimuscarinics for treatment of overactive bladder or urge incontinence | 44 | 1.22 | 6.33 |
| Analgesic drugs | |||
| Topical lidocaine patch for treatment of chronic osteoarthritis pain | 13 | 0.36 | 1.54 |
| Antimuscarinic/anticholinergic drug burden | |||
| Concomitant use of two or more drugs with antimuscarinic/anticholinergic properties | 19 | 0.53 | 8.15 |
| Any STOPP indicator | 1123 | 31.03 | |
| 1 indicator | 814 | 22.49 | – |
| ≥ 2 indicators | 309 | 8.54 | – |
a Calculated as the proportion of overall disease or medication prevalence (e.g. use of medication with potent anticholinergic/antimuscarinic effects in participants with dementia as a proportion of dementia prevalence). Further detail on the disease/medication denominator used for each criterion is provided in Additional File 2
Just over one-third of participants (36.17%, n = 1309) experienced one or more PPOs, with 788 (21.77%) with one START indicator and 521 (14.40%) with two or more START indicators (Additional File 4). The most prevalent START indicator was the absence of an angiotensin converting enzyme inhibitor (ACEI) in those with coronary artery disease (n = 601; 16.61%), which represented 74.11% of the prescribing per indication. Other prevalent START indicators include the absence of vitamin D in patients with known osteoporosis (n = 439; 12.13%), no statin therapy in those a documented history of coronary, cerebral or peripheral vascular disease and no established frailty (n = 341; 9.42%), no antiplatelet therapy in those with a documented history of coronary, cerebral or peripheral vascular disease (7.79%), and the absence of a non-tricyclic antidepressant in those who met the Center for Epidemiologic Studies Depression 8-item scale (CESD-8) threshold for depression (7.38%) (Additional File 4, Table S5).
Of the 3619 TILDA participants aged ≥ 65 years at Wave 4, 449 (12.41%) had both STOPP and START indicators. A total of 674 (18.62%) had a STOPP indicator only and a further 860 (23.76%) had a START indicator only. Approximately two-fifths of the sample (45.21%, n = 1636) did not have either a STOPP or START criterion.
Unadjusted and adjusted associations between participant characteristics and experience of any STOPP and any START indicators are presented in Table 3. In the adjusted model, significant associations for any STOPP indicator were observed for those aged ≥ 75 years (adjusted (a) OR 1.43, 95% CI 1.25, 1.64), females (aOR 1.21, 95% CI 1.00, 1.46), those with one or more chronic conditions (Table 3), and for those with 5–9 daily medications (aOR 3.70, 95% CI 3.05, 4.47) and 10 or more daily medications (aOR 9.03, 95% CI 6.27, 13.00). Those with university or higher level educational attainment were less likely to receive any STOPP indicator (aOR 0.78, 95% CI 0.61, 0.99).
Table 3.
Unadjusted (n = 3619) and adjusted (n = 3016) logistic regression for study participants with at least one STOPP or at least one START indicator
| Characteristic | Any STOPP Unadjusted |
Any STOPP Adjusted |
Any START Unadjusted |
Any START Adjusted |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95%CI | p | OR | 95%CI | p | OR | 95%CI | p | OR | 95%CI | p | |
| Age ≥ 75 yearsa | 1.48 | 1.28, 1.70 | < 0.001h | 0.97 | 0.80, 1.17 | 0.728 | 1.43 | 1.25, 1.64 | < 0.001 | 1.32 | 1.10, 1.57 | 0.002 |
| Female sexb | 1.24 | 1.08, 1.43 | 0.003 | 1.21 | 1.00, 1.46 | 0.048 | 1.64 | 1.42, 1.88 | < 0.001 | 1.30 | 1.09, 1.55 | 0.003 |
| Educationc | ||||||||||||
| Secondary level | 0.76 | 0.65, 0.90 | 0.002 | 0.84 | 0.68, 1.03 | 0.091 | 1.07 | 0.91, 1.26 | 0.414 | 1.17 | 0.97, 1.42 | 0.105 |
| Third/higher level | 0.64 | 0.54, 0.77 | < 0.001 | 0.78 | 0.61, 0.99 | 0.043 | 0.91 | 0.77, 1.08 | 0.274 | 1.09 | 0.87, 1.37 | 0.447 |
| Employmentd | ||||||||||||
| Retired | 1.76 | 1.39, 2.24 | < 0.001 | 0.96 | 0.71, 1.30 | 0.793 | 2.04 | 1.62, 2.59 | < 0.001 | 1.89 | 1.41, 2.54 | < 0.001 |
| Other | 1.99 | 1.51, 2.64 | < 0.001 | 0.98 | 0.69, 1.41 | 0.932 | 2.60 | 1.98, 3.42 | < 0.001 | 1.98 | 1.40, 2.80 | < 0.001 |
| Health insurancee | ||||||||||||
| Private insurance only | 1.16 | 0.70, 1.90 | 0.562 | 1.19 | 0.69, 2.03 | 0.527 | 1.04 | 0.68, 1.58 | 0.868 | 0.85 | 0.54, 1.34 | 0.490 |
| GMS cover only | 2.61 | 1.62, 4.19 | < 0.001 | 1.38 | 0.81, 2.34 | 0.237 | 1.59 | 1.06, 2.39 | 0.025 | 0.79 | 0.50, 1.25 | 0.324 |
| Dual cover | 2.38 | 1.47, 3.86 | < 0.001 | 1.48 | 0.86, 2.53 | 0.156 | 1.51 | 0.99, 2.29 | 0.051 | 0.75 | 0.47, 1.18 | 0.213 |
| Multimorbidityf | ||||||||||||
| 1 condition | 1.95 | 1.39, 2.74 | < 0.001 | 1.62 | 1.10, 2.39 | 0.014 | 1.90 | 1.31, 2.46 | < 0.001 | 1.83 | 1.28, 2.63 | 0.001 |
| 2 conditions | 2.51 | 1.82, 3.48 | < 0.001 | 1.65 | 1.13, 2.40 | 0.009 | 2.81 | 2.08, 3.79 | < 0.001 | 2.76 | 1.96, 3.89 | < 0.001 |
| ≥ 3 conditions | 5.46 | 4.02, 7.43 | < 0.001 | 2.40 | 1.66, 3.48 | < 0.001 | 5.58 | 1.49, 7.42 | < 0.001 | 5.19 | 3.69, 7.31 | < 0.001 |
| Polypharmacyg | ||||||||||||
| 5–9 medications | 4.36 | 3.72, 5.10 | < 0.001 | 3.70 | 3.05, 4.47 | < 0.001 | 1.65 | 1.13, 1.91 | < 0.001 | 1.08 | 0.90, 1.30 | 0.390 |
| ≥ 10 medications | 12.27 | 8.94, 16.84 | < 0.001 | 9.03 | 6.27, 13.00 | < 0.001 | 2.25 | 1.70, 2.97 | < 0.001 | 1.32 | 0.95, 1.84 | 0.098 |
aReferent is 65–74 years
bReferent is male sex
cReferent is primary level/none
dReferent is employed; missing data for two participants
eReferent is no cover; missing data for 601 participants
fReferent is no chronic conditions
gReferent is 0–4 regular medications
h Associations with p <0.05 are in bold
When likelihood for any START indicator was examined in the adjusted model (Table 3), significant associations were identified for those aged ≥ 75 years (aOR 1.32, 95% CI 1.10, 1.57), females (aOR 1.30, 95% CI 1.09, 1.55), those who were retired (aOR 1.89, 95% CI 1.41, 2.54), or in an employment status other than employed (aOR1.98, 95% CI 1.40, 2.80), and those with one or more chronic conditions (Table 3). Compared to those with no chronic conditions, those with three or more chronic conditions were more than five times more likely to experience a START criterion PPO (aOR 5.19, 95% CI 3.69, 7.31).
In the 12 months prior to the Wave 5 TILDA interview, the average number of GP visits was 4.18 (SD 3.87, N = 2991). A total of 2.19% experienced one or more ED visits (M = 0.30, SD = 0.74, N = 3006) and 47.58% experienced one or more outpatient visits (M = 1.81, SD = 6.45, N = 2999). Approximately one fifth of the sample (19.65%) experienced one or more hospital admissions (M = 0.35, SD = 1.31, N = 3007). In total, 434 (14.41%, N = 3012) participants manifested functional decline between Wave 4 and Wave 5. Results of the secondary outcome analysis are summarised in Tables 4 and 5. In the adjusted models, significant associations were observed between any STOPP indicator at Wave 4 and the number of hospital admissions in the 12 months prior to the Wave 5 interview (adjusted IRR 1.38, 95% CI 1.08, 1.75; Table 4, model 1), as well as functional decline between Wave 4 and Wave 5 (aOR 1.46, 95% CI 1.11, 1.91; Table 5, model 1). When examined as a categorical exposure, those with two or more STOPP criteria had significantly more hospital admissions (Table 4, model 2) and were more likely to experience functional decline (Table 5, model 2). In contrast, those who experienced any START indicator had more GP visits (adjusted IRR 1.07, 95% CI 1.01, 1.13; Table 4, model 1), more ED visits (adjusted IRR 1.23, 95% CI 1.03, 1.48; Table 4, model 1) and more outpatient visits (adjusted IRR 1.21, 95% CI 1.02, 1.42; Table 4, model 1), but no significant associations were observed for hospital admissions nor functional decline.
Table 4.
Association between STOPP/START and healthcare utilization outcomes using negative binomial regression (incident rate ratios (IRR), 95% confidence intervals (CIs))
| M (SD)a | GP visits (n = 2487) Adjusted IRR (95% CI)b |
||
|---|---|---|---|
| Model 1c | Model 2d | ||
| Any STOPPe (n = 909) | 4.98 (4.79) | 1.06 (1.00, 1.13), p = 0.051 | – |
| Number of STOPP PIMS | |||
| 0 STOPP (n = 2082) | 3.83 (3.33) | – | (Referent) |
| 1 STOPP (n = 658) | 4.69 (4.28) | – | 1.05 (1.00, 1.13), p = 0.064 |
| ≥ 2 STOPP (n = 251) | 5.72 (5.87) | – | 1.06 (0.96, 1.17), p = 0.267 |
| Any STARTe (n = 1062) | 2.09 (6.77) | 1.07 (1.01, 1.13), p = 0.018f | – |
| Number of START PPOs | |||
| 0 START (n = 1933) | 3.89 (3.57) | – | (Referent) |
| 1 START (n = 651) | 4.39 (3.49) | – | 1.07 (1.00, 1.15), p = 0.038 |
| ≥ 2 START (n = 407) | 5.19 (5.35) | – | 1.07 (0.99, 1.16), p = 0.097 |
| M (SD)a | ED visit (n = 2500) Adjusted IRR (95% CI)b |
||
|---|---|---|---|
| Model 1c | Model 2d | ||
| Any STOPPe (n = 915) | 0.37 (0.85) | 0.91 (0.74, 1.11), p = 0.335 | – |
| Number of STOPP PIMS | |||
| 0 STOPP (n = 2091) | 0.27 (0.69) | – | (Referent) |
| 1 STOPP (n = 662) | 0.33 (0.81) | – | 0.88 (0.71, 1.10), p = 0.260 |
| ≥ 2 STOPP (n = 253) | 0.46 (0.93) | – | 0.97 (0.72, 1.31), p = 0.847 |
| Any STARTe (n = 1066) | 0.39 (0.80) | 1.23 (1.03, 1.48), p = 0.025 | – |
| Number of START PPOs | |||
| 0 START (n = 1940) | 0.22 (0.70) | – | (Referent) |
| 1 START (n = 653) | 0.36 (0.74) | – | 1.20 (0.97, 1.49), p = 0.089 |
| ≥ 2 START (n = 413) | 0.43 (0.87) | – | 1.28 (1.00, 1.63), p = 0.050 |
| M (SD)a | Outpatient visit (n = 2493) Adjusted IRR (95% CI)b |
||
|---|---|---|---|
| Model 1c | Model 2d | ||
| Any STOPPe (n = 914) | 2.34 (8.50) | 1.09 (0.92, 1.29), p = 0.304 | – |
| Number of STOPP PIMS | |||
| 0 STOPP (n = 2085) | 1.58 (5.30) | – | (Referent) |
| 1 STOPP (n = 660) | 2.21 (6.14) | – | 1.08 (0.90, 1.30), p = 0.381 |
| ≥ 2 STOPP (n = 254) | 2.67 (12.74) | – | 1.11 (0.84, 1.46), p = 0.452 |
| Any STARTe (n = 1062) | 2.09 (6.77) | 1.21 (1.02, 1.42), p = 0.024 | – |
| Number of START PPOs | |||
| 0 START (n = 1937) | 1.66 (6.27) | – | (Referent) |
| 1 START (n = 650) | 2.11 (8.07) | – | 1.26 (1.04, 1.52), p = 0.018 |
| ≥ 2 START (n = 412) | 2.06 (3.93) | – | 1.13 (0.91, 1.41), p = 0.271 |
| M (SD)a | Hospital admission (n = 2501) Adjusted IRR (95% CI)b |
||
|---|---|---|---|
| Model 1c | Model 2d | ||
| Any STOPPe (n = 916) | 0.49 (1.79) | 1.38 (1.08, 1.75), p = 0.009 | – |
| Number of STOPP PIMS | |||
| 0 STOPP (n = 2091) | 0.28 (1.03) | – | (Referent) |
| 1 STOPP (n = 662) | 0.46 (1.92) | – | 1.28 (0.99, 1.66), p = 0.062 |
| ≥ 2 STOPP (n = 254) | 0.59 (1.40) | – | 1.70 (1.17, 2.47), p = 0.006 |
| Any STARTe (n = 1066) | 0.44 (1.55) | 1.24 (0.99, 1.54), p = 0.063 | – |
| Number of START PPOs | |||
| 0 START (n = 1941) | 0.29 (1.16) | – | (Referent) |
| 1 START (n = 653) | 0.40 (1.66) | – | 1.21 (0.93, 1.57), p = 0.158 |
| ≥ 2 START (n = 413) | 0.52 (1.37) | – | 1.28 (0.95, 1.73), p = 0.108 |
aMean and standard deviation of healthcare utilisation variables at Wave 5
bAdjusted for age group, sex, education, employment status, insurance coverage, multimorbidity, polypharmacy, and wave 4 healthcare utilisation
cModel 1: PIP exposure assessed using binary variables for presence or absence of STOPP/START
dModel 2: PIP exposure assessed using categorical variables for presence of 0, 1 and ≥ 2 STOPP/START criteria
eReferent: none
fAssociations with p <0.05 are in bold
Table 5.
Association between STOPP/START and the outcome of functional decline using logistic regression (odds ratio (OR), 95% confidence intervals (CIs))
| n (%)a | Functional decline (n = 2507) Adjusted OR (95% CI)b |
||
|---|---|---|---|
| Model 1c | Model 2d | ||
| Any STOPPe (n = 917) | 204 (22.25) | 1.46 (1.11, 1.91), p = 0.007f | – |
| Number of STOPP PIMS | |||
| 0 STOPP (n = 2095) | 230 (10.98) | – | (Referent) |
| 1 STOPP (n = 663) | 116 (17.50) | – | 1.27 (0.94, 1.72), p = 0.116 |
| ≥ 2 STOPP (n = 254) | 88 (34.65) | – | 2.02 (1.36, 2.99), p < 0.001 |
| Any STARTe (n = 1067) | 212 (19.87) | 1.23 (0.95, 1.60), p = 0.114 | – |
| Number of START PPOs | |||
| 0 START (n = 1945) | 222 (11.41) | – | (Referent) |
| 1 START (n = 653) | 120 (18.38) | – | 1.28 (0.94, 1.73), p = 0.110 |
| ≥ 2 START (n = 414) | 92 (22.22) | – | 1.17 (0.83, 1.64), p = 0.364 |
aNumber and percentage who experienced functional decline at Wave 5, referent: none
bAdjusted for age group, sex, education, employment, insurance coverage, multimorbidity, polypharmacy, and functional impairment at Wave 4
cModel 1: PIP exposure assessed using binary variables for presence or absence of STOPP/START
dModel 2: PIP exposure assessed using categorical variables for presence of 0, 1 and ≥ 2 STOPP/START criteria
eReferent: none
fAssociations with p <0.05 are in bold
Discussion
Principal findings
Approximately one-third of study participants experienced one or more STOPP v3-defined PIMs. The majority of these had 1 STOPP PIM, with 8.5% of the total experiencing 2 STOPP PIMs. The most common examples of STOPP PIMs in this cohort were aspirin prescribed for the primary prevention of cardiovascular disease and duplicate drug class prescriptions.
More than one-third of study participants met START PPO criteria, reflecting an omission of clinically indicated medications. The most common PPOs were ACEI omitted in the context of known cardiovascular disease and vitamin D omission for patients with known osteoporosis. Of note, 7.4% of study participants met the criteria for depression using a validated scale but were not prescribed an antidepressant medication, indicating potentially unmet clinical need in this area.
Increasing age, female sex and polypharmacy were associated with an increased risk of STOPP PIMs. Hyperpolypharmacy was associated with a ninefold increased risk of STOPP PIMs compared to participants not experiencing polypharmacy. Over the follow-up period, participants with STOPP PIMs were significantly more likely to be admitted to hospital and to experience functional decline. Patients with START-defined prescribing omissions reported more GP, outpatient and ED visits over time, but there was no evidence of an association with emergency admissions or functional decline.
Comparison with other studies
Previous international longitudinal studies examining the prevalence of earlier versions of STOPP/START criteria found generally higher prevalence rates of PIMs and PPOs than the current study. In a systematic review (including 30 non-randomised studies) the prevalence of STOPP PIMs (version 2) was 42% in community settings and 51% in hospital settings [26]. Another systematic review reported prevalence rates of STOPP PIMs (version 1) as ranging between 21 and 79% due to significant heterogeneity in study design and study populations [27]. In a previous study using the same longitudinal cohort as the current study, with linked pharmacy claims data, rates of STOPP v2 PIMs and START v2 PPOs were 57% and 41% respectively [19]. The current study reported rates of STOPP v3 PIMs at 31% and START v3 PPOs at 36% and therefore suggest an improvement in prescribing appropriateness over time. It is important to note when comparing the two studies that there was overlap in less than 50% of the STOPP criteria applied and in one-third of START criteria applied. However, there has been a substantial expansion of the numbers of STOPP/START version 3 criteria reflecting the growth in the pharmacotherapy evidence base and medication availability over time. It must also be borne in mind that in the current study only 45% of STOPP criteria and 28% of START criteria could be applied due to the limited available data.
The current study found that older patient age, female sex and increasing number of medications are factors associated with an increased risk of STOPP PIMs, in line with the results of previous studies [19, 28]. Balancing the risks and benefits of multiple medications to treat multimorbidity with increasing age-related physiological changes requires constant trade-offs. Patients can experience high levels of treatment burden adhering to complex medication regimes from attending several healthcare professionals across different settings [29]. For clinicians, balancing the risk-to-benefit tipping point for individual patients is often very challenging and requires integration of available evidence, clinical experience and patient preferences [30].
The consistent emergence of female sex as a predictor of PIM occurrence may have implications for developing targeted interventions to address PIM use [28, 31, 32]. Both sex and gender may influence likelihood for PIMs via several mechanisms including biological differences in disease manifestation, differing pharmacokinetic and pharmacodynamic responses, as well as differences in both healthcare seeking behaviour and prescriber behaviour [33]. Women are more likely to be prescribed psychotropic medication and in Ireland rates of benzodiazepine and z-drug hypnotic prescribing was highest for older (≥ 65 years) Irish women over a ten-year period [34, 35]. Disaggregation of data to examine how sex and gender intersect with PIP may reveal further insights from which targeted interventions may be developed [33, 36].
In this study START PPOs were associated with increasing age (≥ 75 years) and patients living with three or more long-term conditions were most at risk. PPOs in this cohort may reflect some of the difficulties of applying clinical guideline recommendations that focus on single diseases to this group. Single disease focussed clinical guidelines are underpinned by evidence that often excludes older people with multimorbidity. Thus, applying these guidelines for this group can drive polypharmacy, in the absence of guidance on how best to prioritise treatment recommendations [29, 37]. Given these difficulties, clinicians should carefully consider the balance of pharmacotherapy recommendations to avoid the ‘risk/treatment paradox’ (higher‐risk older patients denied safe medications likely to improve survival or health-related quality of life) whilst avoiding potentially inappropriate use of medications for which risks are likely to outweigh benefits [38].
This study also found that patients in receipt of one or more STOPP PIMs were more likely to experience functional decline over time and more likely to experience an emergency hospital admission compared to participants not exposed to STOPP PIMs. This aligns with the findings of a 2019 systematic review and meta-analysis (n = 4 studies) which reported that the presence of STOPP PIMs was associated with a range of adverse outcomes including increased healthcare utilisation, adverse drug events and functional decline [39].
There is limited existing research examining START PPOs and subsequent health outcomes, however [8]. In the present study, START PPOs were associated with an increase in GP, outpatient and ED visits, indicating potentially unmet clinical need in this group. There was no impact on emergency admission or functional decline.
Implications for clinical practise, policy and research
A considerable proportion of NSAID and z-drug users at Wave 4 also reported use of these drug classes at Wave 3, suggesting continuous use. This is despite the known gastrointestinal, renal, cardiovascular and cerebrovascular risks associated with chronic NSAID use and prescribing advice to use the lowest effective dose for the shortest possible time [40]. Similarly, chronic z-drug use (≥ 2 weeks) increases the risk of falls and fractures, which can have potentially devastating consequences for older adults [41]. This finding aligns with a previous Irish study-in 2019, 27% of all patients dispensed z-drugs received these for three months or more, an increase from 22% in 2015 [42]. Deprescribing opportunities were identified in participants aged ≥ 85 years with established frailty who used statins for primary prevention. Evidence for statin efficacy in the ‘old old’ age cohort is uncertain and therefore statins should only be considered on an individual basis considering level of cardiovascular risk and life expectancy [43, 44].
Addressing PIP may also confer healthcare system benefits through a reduction in healthcare attendances at a time when global healthcare systems face increasing demand. Reducing the number of unplanned hospital admissions for older patients may in turn reduce the likelihood for further PIP, as transitions of care are a key risk factor for medication-related harm [45, 46]. One challenge with the application of the STOPP/START version 3 is the large number of criteria to be applied (n = 190). Automating prescribing indicators through integration with patient electronic record systems to alert prescribers to PIMs and PPOs is one way of addressing this issue [47].
Highlighting international concerns regarding medication safety the World Health Organisation launched a global patient safety challenge, Medication Without Harm, in March 2017 [46]. However, to date, multi-centre randomised controlled trials with interventions focussed on STOPP/START criteria and other prescribing indicators have not resulted in improvements in primary end-points such as ADRs, emergency admission and re-admission [48–51]. Recently, polypharmacy stewardship has been suggested as a novel way to approach this issue. This is defined as a “coordinated intervention designed to assess, monitor, improve, and measure the pharmacotherapeutic treatment of multimorbidity………with the aim of aligning treatment regimens with the overall condition, prognosis, and preferences of the individual patient” [52]. Core principles of polypharmacy stewardship include optimisation of prescribing (e.g. dose and frequency), setting of realistic treatment goals, ADR avoidance and provision of support and follow-up to patients with multimorbidity. Explicit prescribing indicator sets, such as STOPP/START criteria, have a role in identifying medicines that merit clinical review in the context of the individual patient’s clinical history and preferences, but broader healthcare system issues must also be addressed to optimise prescribing. This includes timely sharing of medication information and medicines reconciliation across fragmented healthcare settings and prioritising continuity of care [29, 30].
Strengths and limitations
This study has several strengths. It is a nationally representative study of the Irish population and longitudinal in design allowing for follow-up over time. Comprehensive information was collected on a range of demographic, socioeconomic and clinical characteristics through rigorous standardised protocols, allowing for baseline differences to be accounted for. Medication data, collected by in-home computer-assisted participant interview (CAPI), demonstrates both internal and external validity [21]. The study has some limitations also. Similar to previous studies, only a subset of the STOPP/START v3 criteria could be applied to the TILDA dataset due to the absence of certain clinical data required for application [19, 27]. The outcomes of healthcare utilization and functional decline were self-reported which may have affected accuracy. However, adjustment for the range of confounders should have addressed any systematic reporting bias amongst participant subgroups [53]. The proportion of study participants with complex polypharmacy (10 or more medications) was low at 6%. Finally, as this is a retrospective cohort study, results of association and outcome analyses should not be regarded as causal as there are potential confounders that could not be controlled for, that could affect these associations.
Conclusion
In this nationally representative longitudinal study, approximately one-third of participants were found to experience STOPP-defined PIMs and over one-third START-defined PPOs, when assessed using the recently updated STOPP/START prescribing indicator set. The presence of STOPP PIMs was associated with an increased risk of functional decline and hospital admission at follow-up. Increasing age and number of chronic health conditions were associated with START PPOs, highlighting the challenges in balancing the risk-to-benefit ratio of medications in older people with multimorbidity.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
The authors would like to acknowledge those older adults who participated in TILDA.
Author contributions
Study concept and design: EW, ASD, FM, RAK, DOM, FB, BC and TF. Acquisition of data: EW, RAK, FM and ASD. Analyses and interpretation: ASD, FM, EW, FB, TF, RAK, BC and DOM. Drafting and revising the manuscript: ASD and EW. All authors have critically reviewed and edited the manuscript and have approved the final version.
Funding
Open Access funding provided by the IReL Consortium. This study was funded by the Health Research Board Ireland Emerging Clinician Scientist Award (HRB-ECSA-2020-002) awarded to Prof Emma Wallace. The funder had no role in the design and conduct of the study; data analysis and interpretation; preparation and review of the manuscript; or the decision to submit for publication. Funding for the TILDA study is provided by the Irish Government, the Atlantic Philanthropies and Irish Life plc.
Data availability
Access to data from the TILDA study is granted upon approval of the request by the TILDA Data Access Committee. More information can be obtained at https://www.tilda.tcd.ie/. The publicly accessible dataset files are available via the Irish Social Science Data Archive (ISSDA) based in University College Dublin. Access to the publicly available dataset files is granted following application to the ISSDA. More information can be obtained at https://www.ucd.ie/issda/data/tilda/.
Declarations
Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Ethical approval
Ethical approval for this study was obtained from the Clinical Research Ethics Committee (CREC) of University College Cork (ECM 4 (g) 01/08/2023 & ECM 3 (vv) 24/10/2023 & ECM 3 (kkkkk) 02/07/2024). Ethical approval for each wave of TILDA data collection was provided by the Faculty of Health Sciences Research Ethics Committee at Trinity College Dublin. For each wave and each component of data collection (e.g. CAPI, health assessment etc.) participants are provided with an information leaflet and informed written consent completed prior to data collection.
Informed consent
All participants provided written informed consent.
Rights retention
This publication has emanated from research conducted with the financial support of the Health Research Board under Grant number HRB-ECSA-2020–002. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.O’Connor MN, Gallagher P, O’Mahony D (2012) Inappropriate prescribing. Drugs Aging 29:437–452. 10.2165/11632610-000000000-00000 [DOI] [PubMed] [Google Scholar]
- 2.Chowdhury SR, Das DC, Sunna TC, Beyene J, Hossain A (2023) Global and regional prevalence of multimorbidity in the adult population in community settings: a systematic review and meta-analysis. eClinicalMedicine 57:101860. 10.1016/j.eclinm.2023.101860 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kim S, Lee H, Park J, Kang J, Rahmati M, Rhee SY et al (2024) Global and regional prevalence of polypharmacy and related factors, 1997–2022: an umbrella review. Arch Gerontol Geriatr 124:105465. 10.1016/j.archger.2024.105465 [DOI] [PubMed] [Google Scholar]
- 4.Lavan AH, Gallagher P (2016) Predicting risk of adverse drug reactions in older adults. Ther Adv Drug Saf 7:11–22. 10.1177/2042098615615472 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Insani WN, Whittlesea C, Alwafi H, Man KKC, Chapman S, Wei L (2021) Prevalence of adverse drug reactions in the primary care setting: a systematic review and meta-analysis. PLoS ONE 16:e0252161. 10.1371/journal.pone.0252161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Doherty AS, Boland F, Moriarty F, Fahey T, Wallace E (2023) Adverse drug reactions and associated patient characteristics in older community-dwelling adults: a 6-year prospective cohort study. Br J Gen Pract 73:e211–e219. 10.3399/BJGP.2022.0181 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Oscanoa TJ, Lizaraso F, Carvajal A (2017) Hospital admissions due to adverse drug reactions in the elderly. A meta-analysis. Eur J Clin Pharmacol 73:759–770. 10.1007/s00228-017-2225-3 [DOI] [PubMed] [Google Scholar]
- 8.Mekonnen AB, Redley B, de Courten B, Manias E (2021) Potentially inappropriate prescribing and its associations with health-related and system-related outcomes in hospitalised older adults: a systematic review and meta-analysis. Br J Clin Pharmacol 87:4150–4172. 10.1111/bcp.14870 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Robinson EG, Hedna K, Hakkarainen KM, Gyllensten H (2022) Healthcare costs of adverse drug reactions and potentially inappropriate prescribing in older adults: a population-based study. BMJ Open 12:e062589. 10.1136/bmjopen-2022-062589 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Beers MH, Ouslander JG, Rollingher I, Reuben DB, Brooks J, Beck JC (1991) Explicit criteria for determining inappropriate medication use in nursing home residents. Arch Intern Med 151:1825–1832. 10.1001/archinte.1991.00400090107019 [PubMed] [Google Scholar]
- 11.Laroche M-L, Charmes J-P, Merle L (2007) Potentially inappropriate medications in the elderly: a French consensus panel list. Eur J Clin Pharmacol 63:725–731. 10.1007/s00228-007-0324-2 [DOI] [PubMed] [Google Scholar]
- 12.Rognstad S, Brekke M, Fetveit A, Spigset O, Wyller TB, Straand J (2009) The Norwegian General Practice (NORGEP) criteria for assessing potentially inappropriate prescriptions to elderly patients: a modified Delphi study. Scand J Prim Health Care 27:153–159. 10.1080/02813430902992215 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gallagher P, Ryan C, Byrne S, Kennedy J, O’Mahony D (2008) STOPP (Screening Tool of Older Persons Prescriptions) and START (Screening Tool to Alert doctors to Right Treatment). Consensus validation. CP 46:72–83. 10.5414/CPP46072 [DOI] [PubMed] [Google Scholar]
- 14.O’Mahony D, O’Sullivan D, Byrne S, O’Connor MN, Ryan C, Gallagher P (2015) STOPP/START criteria for potentially inappropriate prescribing in older people: version 2. Age Ageing 44:213–218. 10.1093/ageing/afu145 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.O’Mahony D, Cherubini A, Guiteras AR, Denkinger M, Beuscart J-B, Onder G et al (2023) STOPP/START criteria for potentially inappropriate prescribing in older people: version 3. Eur Geriatr Med 14:625–632. 10.1007/s41999-023-00777-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Donoghue OA, McGarrigle CA, Foley M, Fagan A, Meaney J, Kenny RA (2018) Cohort profile update: The Irish Longitudinal Study on Ageing (TILDA). Int J Epidemiol 47:1398–1398l. 10.1093/ije/dyy163 [DOI] [PubMed] [Google Scholar]
- 17.Kearney PM, Cronin H, O’Regan C, Kamiya Y, Savva GM, Whelan B et al (2011) Cohort profile: The Irish Longitudinal Study on Ageing. Int J Epidemiol 40:877–884. 10.1093/ije/dyr116 [DOI] [PubMed] [Google Scholar]
- 18.Galvin R, Moriarty F, Cousins G, Cahir C, Motterlini N, Bradley M et al (2014) Prevalence of potentially inappropriate prescribing and prescribing omissions in older Irish adults: findings from The Irish LongituDinal Study on Ageing study (TILDA). Eur J Clin Pharmacol 70:599–606. 10.1007/s00228-014-1651-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Moriarty F, Bennett K, Cahir C, Kenny RA, Fahey T (2016) Potentially inappropriate prescribing according to STOPP and START and adverse outcomes in community-dwelling older people: a prospective cohort study. Br J Clin Pharmacol 82:849–857. 10.1111/bcp.12995 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP (2007) Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ 335:806–808. 10.1136/bmj.39335.541782.AD [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Qato DM, Schumm LP, Johnson M, Mihai A, Lindau ST (2009) Medication data collection and coding in a home-based survey of older adults. J Gerontol B Psychol Sci Soc Sci 64B:i86-93. 10.1093/geronb/gbp036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Mannion C, Hughes J, Moriarty F, Bennett K, Cahir C (2020) Agreement between self-reported morbidity and pharmacy claims data for prescribed medications in an older community based population. BMC Geriatr 20:1–14. 10.1186/s12877-020-01684-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Turner N, Donoghue O, Kenny RA (2018) Wellbeing and Health in Ireland’s Over 50s 2009–2016. The Irish Longitudinal Study on Ageing, Dublin. 10.38018/TildaRe.2018-00 [Google Scholar]
- 24.Mattsson M, Flood M, Wallace E, Boland F, Moriarty F (2023) Eligibility rates and representativeness of the General Medical Services scheme population in Ireland 2016–2021: a methodological report. HRB Open Res. 10.12688/hrbopenres.13622.2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Larkin J, Walsh B, Moriarty F, Clyne B, Harrington P, Smith SM (2022) What is the impact of multimorbidity on out-of-pocket healthcare expenditure among community-dwelling older adults in Ireland? A cross-sectional study. BMJ Open 12:e060502. 10.1136/bmjopen-2021-060502 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Thomas RE, Thomas BC (2019) A systematic review of studies of the STOPP/START 2015 and American Geriatric Society Beers 2015 criteria in patients ≥ 65 years. Curr Aging Sci 12:121–154. 10.2174/1874609812666190516093742 [DOI] [PubMed] [Google Scholar]
- 27.Hill-Taylor B, Sketris I, Hayden J, Byrne S, O’Sullivan D, Christie R (2013) Application of the STOPP/START criteria: a systematic review of the prevalence of potentially inappropriate prescribing in older adults, and evidence of clinical, humanistic and economic impact. J Clin Pharm Ther 38:360–372. 10.1111/jcpt.12059 [DOI] [PubMed] [Google Scholar]
- 28.Moriarty F, Hardy C, Bennett K, Smith SM, Fahey T (2015) Trends and interaction of polypharmacy and potentially inappropriate prescribing in primary care over 15 years in Ireland: a repeated cross-sectional study. BMJ Open 5:e008656. 10.1136/bmjopen-2015-008656 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wallace E, Salisbury C, Guthrie B, Lewis C, Fahey T, Smith SM (2015) Managing patients with multimorbidity in primary care. BMJ 350:h176. 10.1136/bmj.h176 [DOI] [PubMed] [Google Scholar]
- 30.Jennings AA, Doherty AS, Clyne B, Boland F, Moriarty F, Fahey T et al (2024) Stakeholder perceptions of and attitudes towards problematic polypharmacy and prescribing cascades: a qualitative study. Age Ageing 53:afae116. 10.1093/ageing/afae116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cahir C, Fahey T, Teeling M, Teljeur C, Feely J, Bennett K (2010) Potentially inappropriate prescribing and cost outcomes for older people: a national population study. Br J Clin Pharmacol 69:543–552. 10.1111/j.1365-2125.2010.03628.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ukhanova M, Markwardt S, Furuno JP, Davis L, Noble BN, Quiñones AR (2021) Are there sex differences in potentially inappropriate prescribing in adults with multimorbidity? J Am Geriatr Soc 69:2163–2175. 10.1111/jgs.17194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sternberg SA, Petrovic M, Onder G, Cherubini A, O’Mahony D, Gurwitz JH et al (2021) Identifying key prescribing cascades in older people (iKASCADE): a transnational initiative on drug safety through a sex and gender lens-rationale and design. Eur Geriatr Med 12:475–483. 10.1007/s41999-021-00480-w [DOI] [PubMed] [Google Scholar]
- 34.Cadogan CA, Ryan C, Cahir C, Bradley CP, Bennett K (2018) Benzodiazepine and Z-drug prescribing in Ireland: analysis of national prescribing trends from 2005 to 2015. Br J Clin Pharmacol 84:1354–1363. 10.1111/bcp.13570 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Boyd A, Van de Velde S, Pivette M, Ten Have M, Florescu S, O’Neill S et al (2015) Gender differences in psychotropic use across Europe: results from a large cross-sectional, population-based study. Eur Psychiatry 30:778–788. 10.1016/j.eurpsy.2015.05.001 [DOI] [PubMed] [Google Scholar]
- 36.Shapiro JR, Klein SL, Morgan R (2021) Stop ‘controlling’ for sex and gender in global health research. BMJ Glob Health. 10.1136/bmjgh-2021-005714 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hughes LD, McMurdo MET, Guthrie B (2013) Guidelines for people not for diseases: the challenges of applying UK clinical guidelines to people with multimorbidity. Age Ageing 42:62–69. 10.1093/ageing/afs100 [DOI] [PubMed] [Google Scholar]
- 38.Scott IA, Guyatt GH (2010) Cautionary tales in the interpretation of clinical studies involving older persons. Arch Intern Med 170:587–595. 10.1001/archinternmed.2010.18 [DOI] [PubMed] [Google Scholar]
- 39.Liew TM, Lee CS, Goh Shawn KL, Chang ZY (2019) Potentially inappropriate prescribing among older persons: a meta-analysis of observational studies. Ann Fam Med 17:257–266. 10.1370/afm.2373 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Marcum ZA, Hanlon JT (2010) Recognizing the risks of chronic nonsteroidal anti-inflammatory drug use in older adults. Ann Longterm Care 18:24–27 [PMC free article] [PubMed] [Google Scholar]
- 41.Scharner V, Hasieber L, Sönnichsen A, Mann E (2022) Efficacy and safety of Z-substances in the management of insomnia in older adults: a systematic review for the development of recommendations to reduce potentially inappropriate prescribing. BMC Geriatr 22:87. 10.1186/s12877-022-02757-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Health Service Executive Medicines Management Programme. Guidance on appropriate prescribing of benzodiazepines and z-drugs (BZRA) for the treatment of anxiety and insomnia. 2021.
- 43.Marcellaud E, Jost J, Tchalla A, Magne J, Aboyans V (2023) Statins in primary prevention in people over 80 years. Am J Cardiol 187:62–73. 10.1016/j.amjcard.2022.10.015 [DOI] [PubMed] [Google Scholar]
- 44.Mondal A, Li A, Edusa S, Gogineni A, Karipineni S, Abdelhafez S et al (2024) Does statin use in frail patients provide survival benefits? Insights from a meta-analysis. Curr Probl Cardiol 49:102038. 10.1016/j.cpcardiol.2023.102038 [DOI] [PubMed] [Google Scholar]
- 45.Pérez T, Moriarty F, Wallace E, McDowell R, Redmond P, Fahey T (2018) Prevalence of potentially inappropriate prescribing in older people in primary care and its association with hospital admission: longitudinal study. BMJ 363:k4524. 10.1136/bmj.k4524 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Donaldson LJ, Kelley ET, Dhingra-Kumar N, Kieny M-P, Sheikh A (2017) Medication without harm: WHO’s third global patient safety challenge. Lancet 389:1680–1681. 10.1016/S0140-6736(17)31047-4 [DOI] [PubMed] [Google Scholar]
- 47.Rochon PA, Stall NM, Reppas-Rindlisbacher C, Gurwitz JH (2023) STOPP/START version 3: even better with age. Eur Geriatr Med. 10.1007/s41999-023-00784-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.O’Mahony D, Gudmundsson A, Soiza RL, Petrovic M, Cruz-Jentoft AJ, Cherubini A et al (2020) Prevention of adverse drug reactions in hospitalized older patients with multi-morbidity and polypharmacy: the SENATOR* randomized controlled clinical trial. Age Ageing 49:605–614. 10.1093/ageing/afaa072 [DOI] [PubMed] [Google Scholar]
- 49.Blum MR, Sallevelt BTGM, Spinewine A, O’Mahony D, Moutzouri E, Feller M et al (2021) Optimizing therapy to prevent avoidable hospital admissions in multimorbid older adults (OPERAM): cluster randomised controlled trial. BMJ 374:n1585. 10.1136/bmj.n1585 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.McCarthy C, Clyne B, Boland F, Moriarty F, Flood M, Wallace E et al (2022) GP-delivered medication review of polypharmacy, deprescribing, and patient priorities in older people with multimorbidity in Irish primary care (SPPiRE Study): a cluster randomised controlled trial. PLoS Med 19:e1003862. 10.1371/journal.pmed.1003862 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Hill-Taylor B, Walsh KA, Stewart S, Hayden J, Byrne S, Sketris IS (2016) Effectiveness of the STOPP/START (Screening Tool of Older Persons’ potentially inappropriate Prescriptions/Screening Tool to Alert doctors to the Right Treatment) criteria: systematic review and meta-analysis of randomized controlled studies. J Clin Pharm Ther 41:158–169. 10.1111/jcpt.12372 [DOI] [PubMed] [Google Scholar]
- 52.Daunt R, Curtin D, O’Mahony D (2023) Polypharmacy stewardship: a novel approach to tackle a major public health crisis. Lancet Healthy Longev 4:e228–e235. 10.1016/S2666-7568(23)00036-3 [DOI] [PubMed] [Google Scholar]
- 53.Short ME, Goetzel RZ, Pei X, Tabrizi MJ, Ozminkowski RJ, Gibson TB et al (2009) How accurate are self-reports? Analysis of self-reported health care utilization and absence when compared with administrative data. J Occup Environ Med 51:786–796. 10.1097/JOM.0b013e3181a86671 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Access to data from the TILDA study is granted upon approval of the request by the TILDA Data Access Committee. More information can be obtained at https://www.tilda.tcd.ie/. The publicly accessible dataset files are available via the Irish Social Science Data Archive (ISSDA) based in University College Dublin. Access to the publicly available dataset files is granted following application to the ISSDA. More information can be obtained at https://www.ucd.ie/issda/data/tilda/.