UEF//eRepository
DSpace https://erepo.uef.fi
Rinnakkaistallenteet Yhteiskuntatieteiden ja kauppatieteiden tiedekunta
2018
Patient- and health care-related factors associated with initiation of potentially inappropriate medication in
community-dwelling older persons.
Hyttinen, Viva
Wiley
Tieteelliset aikakauslehtiartikkelit
© Nordic Association for the Publication of BCPT All rights reserved
http://dx.doi.org/10.1111/bcpt.13096
https://erepo.uef.fi/handle/123456789/6861
Downloaded from University of Eastern Finland's eRepository
Accepted Article
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as Article Type: Original Article
Patient- and health care-related factors associated with initiation of potentially inappropriate medication in community-dwelling older persons
Running head: Factors associated with PIM initiation in older persons
Virva Hyttinen1, Johanna Jyrkkä2, Leena K. Saastamoinen3, Anna-Kaisa Vartiainen1 and Hannu Valtonen1
1Department of Health and Social Management, University of Eastern Finland, Kuopio, Finland
2Assessment of Pharmacotherapies, Finnish Medicines Agency, Kuopio, Finland
3Kela Research, The Social Insurance Institution, Helsinki, Finland
(Received 20 March 2018; Accepted 8 July 2018)
Correspondence: Virva Hyttinen, Department of Health and Social Management, University of Eastern Finland, Kuopio, Finland.
Email: virva.hyttinen@uef.fi
Conflict of Interest
Authors declare no conflicts of interest.
Accepted Article
Abstract
Potentially inappropriate medications (PIMs) in older persons are defined as medications of which the potential harms outweigh their benefits. The purpose of this study was determine how initiation of PIMs accumulate in community-dwelling persons aged 65–74 and ≥75 years, and which patient- and health care-related factors are associated with PIM initiation over time.
Data of this study were gathered from population-based registers by a 10% random sample of persons (n = 28,497) aged ≥65 years with no prior PIMs within a 2-year period preceding the index date (1 January 2002) and the study subjects were followed until 2013. The Finnish Prescription Register was linked by using a personal identity code to registers on inpatient care and causes of deaths and socioeconomic data.
In this study, 10,698 (37.5%) persons initiated PIMs during the study period. Female gender was associated with PIM initiation in 65–74-year-olds, but not in ≥75-year-olds. In 65–74-year-olds, the risk of PIM initiation increased with the higher income, whereas in ≥75-year-olds, the association between PIM initiation and the high income was not significant. The prescribing physician explained 9% to 16% of the variation in the probability of PIM initiation.
In conclusion, there were age-related differences in the factors associated with PIM initiation in relation to gender and socioeconomic status. Overall, patient-related factors explained a large proportion of variation of PIM initiation, but there were also differences in PIM prescribing among physicians. However, physician-related variance explained why PIM prescribing decreased during the 12-year follow-up.
Keywords: potentially inappropriate medications, older adults, register-based study, gerontopharmacology, rational pharmacotherapy
Accepted Article
Introduction
Pharmacotherapy in older adults is challenging due to age-related physiological changes and the use of multiple medications. In Finland, almost half of all medication costs accumulate to only five per cent of the whole population [1]. Of them, almost half are aged over 65 years and use more than ten medications [1]. Similar results have also been reported in Sweden and Canada; aged persons using multiple medications accounted for between 40% [2] and 75% of total pharmaceutical costs [3].
Polypharmacy increases the risk of use of potentially inappropriate medications (PIMs) [4] that are defined as medications of which the potential harms outweigh their benefits [5]. It is known that PIMs can lead to adverse drug events [6], such as increased risk of hospitalization [7] and thus higher health care costs [8].
Several criteria have been created to define PIM use in order to improve pharmacotherapy in older adults in various countries. The first ones are the well-known Beers Criteria, which were developed in the United States at the beginning of the 1990s [5]. The latest update of these criteria was published in 2015 [9]. However, many national criteria have also been developed to support country-specific clinical practices and treatment guidelines, because generalizations of criteria developed in other countries may be problematic to some extent [10]. In Finland, the Database of Medication for the Elderly (Meds75+) maintained by the Finnish Medicines Agency (FIMEA) was published in 2010 [11]. The database supports clinical decision-making by offering an easy and free tool for caregivers, which is intended to improve the safety of medication of those aged 75 years and over. However, these criteria can also be applied in persons aged <75 years.
Based on systematic reviews, the prevalence of PIM use ranges from 12%–21% to 63%–79% in older adults depending on criteria used and study setting [12,13]. A recent review estimated that overall PIM prevalence is more than 20% in older adults aged ≥65 years in Europe [14]. There are only a few studies about time trend, which have shown that the prevalence of PIM has been decreasing slightly despite the increasing use of medications [15-17]. However, previous studies are mainly conducted on a cross-sectional basis, and the longitudinal evidence about PIM use is lacking.
Accepted Article
The factors most commonly associated with PIM use are female gender, advanced age and polypharmacy [12]. In addition, a high comorbidity score was often found to be associated with PIM use in a recent review [14]. The results of this review also pointed out that older age is an important risk factor for PIM use, but there was a positive association in only about a half of the studies included. Other studies have also found opposite or mixed association of PIM use and age [18,19].
Even though female gender has been strongly associated with PIM use [12], some studies have reported that males have a greater risk to be prescribed PIMs than females [18]. In addition, previous studies have indicated that lower socioeconomic status is associated with PIM use [15,19]. However, there is still a need to identify patients with high-PIM-risk conditions [19]. A previous study found that persons aged over 80 years are more likely to receive PIMs than the younger counterparts [20].
However, this previous study did not analyse factors associated with PIM use between age groups.
Another study investigated PIM use between age groups among older persons [21]. The study found that female gender is not associated with PIM use in the oldest old (>85 years) persons.
Previous research on the association between PIM use and factors related to the health care system is scarce. Some studies have found that the risk of PIM prescription is higher in visits to family and general practitioners compared to specialized doctors [22]. Additionally, hospital characteristics were found to be associated with PIM use [23], but results on regional differences are mixed [23-26]
Based on previous literature, the prescribing decision is the result of the interaction of a variety of factors between patient and a prescribing physician [27]. Patient and physician characteristics are in a major part of the nature of the interaction. Reducing PIM use needs knowledge on the underlying factors associated with PIM use. Evidence is needed on the role of patient characteristics and prescribing practices (i.e., the supply of services, especially the prescribing physician) for the PIM use. Study is also needed whether these associations with PIM use vary in different age groups when targeting interventions, such as medication reviews.
Accepted Article
The aims of this study were to determine 1) how PIM initiation accumulate in the Finnish nationwide sample of community-dwelling older adults aged 65–74 and ≥75 years according to the Meds75+
database during the 12-year follow-up (years 2002–2013) and 2) what is the role of patient characteristics and health care–related variables in PIM initiation.
Materials and Methods
Data sources. Data of this study were extracted from nationwide registers in Finland. The Prescription Register was linked by using a unique personal identification code to the Care Register for Health Care and registers on causes of deaths and socioeconomic factors.
The Prescription Register is maintained by the Social Insurance Institution and it includes all Finnish citizens who have received reimbursement of their medication purchases. This means practically the whole population, because in Finland all residents are covered by National Health Insurance, which reimburses medications in ambulatory care. Medication purchases for persons living in long-term institutionalized settings are not included in the Prescription Register. Information gathered from the register included each person’s date of birth, gender and residential area (municipality, hospital district), and information on medication purchases (Anatomical Therapeutic Chemical (ATC) code, purchase date and physician’s sickness insurance number = identification code). All medications were classified based on the World Health Organization ATC classification system [28].
The Care Register for Health Care is maintained by the National Institute for Health and Welfare, and it includes information on the use of inpatient care: date of hospital admission and date of hospital discharge.
Causes of deaths and socioeconomic information were extracted from the registers maintained by Statistics Finland. Socioeconomic information included the spending money of household unit and the number of persons living in a household.
Accepted Article
Study population. The study population was constituted by drawing a 10% random sample of persons aged ≥65 years (N=64,250, 1 January 2000) from the Prescription Register, and they were followed until 31 December 2013. We excluded those persons who used at least one PIM within the 24-month period preceding the index date (1 January 2002, n=32,087) in order to identify new PIM-initiators. In addition, we excluded those persons who stayed in hospital or another institution for over 90 days during the preceding two years (years 2000–2001 = washout period) (n=431) or were hospitalized at the start of the follow-up (n=318). The persons who died (n=2,460) during the washout period were dropped from the sample, as well as altogether 457 persons who had no medication information in any subsequent year after the washout period. This resulted in a final study population of 28,497 community-dwelling older persons and their PIM initiation was followed from 2002 to 2013.
Meds75+ database. PIM use was defined according to the Meds75+ database [11]. In the database, medications most commonly used in older persons are divided into four categories: A (appropriate for older adults), B (limited research evidence of appropriateness or practical experience or efficacy in older adults), C (appropriate for older adults only under certain conditions) and D (potentially inappropriate for older adults; see Appendix 1). The database includes descriptive legend for each pharmaceutical substance, which gives information on the most common adverse reactions and interactions, the effects and dosing of the substance. The categories and legends were made by consensus in multidisciplinary working groups, which included experts in geriatrics, clinical pharmacology and physicians who worked with older persons. The database does not give information on indications or contra-indications [11].
PIM use. A new PIM initiator was defined as a person who had purchased at least one PIM (prescription) during the follow-up period from the beginning of the year 2002, but had no PIM purchases during 2000–2001. In this study, to assess PIMs in older adults, we considered only medications in the D category (which included approx. 90 pharmaceutical substances in 2010). In addition, the dichotomous variable of PIM use was assessed for each year separately. Accumulation of PIMs was measured by calculating the number of different PIMs per participant annually and for
Accepted Article
the whole study period. The follow-up ended on 31 December 2013, death or ≥90 days hospitalization. The long hospitalization was the reason for the end of the follow-up, because the Prescription Register does not include information on medications given to patients in hospital.
Measures and definitions. Patient characteristics included sociodemographic (age and gender) and socioeconomic factors and information on morbidity based on prescribed medications. Prescribing physician and hospital district of the patient refers to health care–related characteristics. Information about individual physicians was based only on their professional identification code, which is obligatory for every Finnish physician in order to practice as a licensed physician.
Socioeconomic position was described by dividing spending money of household-dwelling unit by the equivalent number of persons (number of equivalent consumers) living in a household, and this was coded into four different income classes. The information on the number of persons living in a household was used to formulate the variable recording whether a person was living alone or not.
Baseline morbidity was defined during the washout period (years 2000–2001) and was based on ATC groups (excluding PIMs in the list of Appendix 1), which were obtained as a proxy for metabolic syndrome (medications used in diabetes, A10), psychiatric disorders (psycholeptics, N05; and antidepressants, N06A), cardiovascular disease (cardiovascular system, C01–04 and C07–C10) and dementia (anti-dementia medications, N06D). Medications were chosen, because their indications are clear and these diseases are common among older persons. Excessive polypharmacy was defined as the use of at least 10 different medications (ATC-codes) annually during the washout period.
Statistical analysis. Baseline characteristics of PIM initiators and PIM non-initiators were analysed using cross tabulation and a chi-square test. Cox proportional hazards regression was used for investigating the time to the first PIM purchase (failure event) and calculating hazards for different factors associated with PIM initiation. In the Cox model, survival time was censored at the first PIM purchase or for any reason for end of follow-up, whichever came first. Models were adjusted for
Accepted Article
gender, socioeconomic position, the use of anti-diabetics, psychotropics, cardiovascular or anti- dementia medications, presence of excessive polypharmacy, living situation and hospital district.
Analyses were performed separately for those aged 65–74 and ≥75 years. The proportional hazard assumption was tested by Kaplan-Meier curves.
Physician effect on PIM initiation was investigated with multilevel mixed-effects logistic regression in both the first (year 2002) and the last year (2013) of the study period. All medication purchases in study years were gathered, and the prescribing physician was defined in each medication purchase with a physicians’ identification code. In these models, the dependent variable was dichotomous, a person’s new PIM purchase, which means that the models considered only different “new PIM initiations”. We formulated two models; the first one was the unconditional (constant-only) model, which estimates the overall probability of a PIM event and the variance between physicians. The second model (the random-intercept model) also includes the patient-related fixed predictor variables:
gender, socioeconomic position, the use of antidiabetics, psychotropics, cardiovascular or anti- dementia medications, excessive polypharmacy and living situation. Intraclass correlation (ICC) was determined to find out the proportion of the total variation in PIM purchases explained by the physician effect.
Only persons without any missing values of the covariates were included in the models. Results were reported as hazard ratios (HR) and odds ratios (OR). The significance level was set at 0.05. All analyses were performed by using the Stata statistical package (STATA IC14.1.). The study was conducted in accordance with ethical guidelines and approved by the Research Ethics Committee of the Northern Savo Hospital District.
Accepted Article
Results
The mean age of the study subjects was 75.6 years (median 74.5, SD 6.2), and 60.6% of them were women (Table 1). Nearly 38% lived alone. PIM initiators were more likely women and younger compared with non-initiators. In addition, PIM-initiators belonged less often to the lowest income class and were more likely to experience excessive polypharmacy than non-initiators. PIM-initiators lived alone less often than non-initiators. There were minor differences between the groups with the use of psychotropics and anti-dementia medications, but no difference was found for anti-diabetics or cardiovascular medications. The mean follow-up time of the study population was 2,447 days (median 2,408 days).
“Table 1 here”
Initiation of PIMs. Overall, 10,698 (37.5%) persons started using PIMs during the 12-year study period. The mean time for the first PIM purchase was 1,400 days (median 1,015 days). The study population started using 67 different PIMs (see Appendix 1). The most commonly initiated PIM was tizanidine (14.2% of the initiated PIMs; Fig. 1). When comparing the ten most frequently initiated PIMs by age group, the ten most commonly initiated PIMs included orphenadrine instead of oxubutynin in persons aged 65–74 years, whereas in persons over 75 years, there were nitrofurantoin and theophylline instead of solifenacin and propranolol. Among women, the ten most commonly initiated PIMs were similar than in the whole study population (medications listed in Fig. 1). The ten most commonly initiated PIMs among men included theophylline and indomethacin instead of propranolol and oxubutynin.
The mean number of different PIMs per participant was 1.1 per year. During the 12-year study period, PIM-initiators accumulated on average 2.8 different PIMs. Most of the PIM initiators (45.1%) purchased the same PIM during the study period, but over 17% of the PIM initiators purchased five or more different PIMs (see Table 1). The highest number of different PIMs taken was 26 over the
Accepted Article
course of the study period. As a whole, PIMs (at least one) were used for an average of 2.5 years. The years were not necessarily consecutive.
Factors associated with PIM initiation. The Cox models showed that female gender was significantly associated (HR 1.12, 95% CI 1.06–1.18, p<0.001) with PIM initiation in persons aged 65–74 years (Table 2). However, this association with gender was not statistically significant in the older age group (≥75 years). Income was also differently associated in different age groups. In persons aged 65–
74 years, the risk of PIM initiation was the largest at the highest income, whereas in the older age (≥75 years) group, the highest income group was not associated with PIM initiation.
Excessive polypharmacy was strongly associated with the higher risk of PIM initiation (p<0.001) in both age groups. In addition, the use of psychotropics increased the risk of PIM initiation in both age groups. There were also differences in PIM initiation between hospital districts.
Physician and patient effects on PIM initiation. The results of the multilevel mixed-effects logistic regression models indicated that 16% of the total variance of the PIM purchases in persons aged 65–
74 years is accounted for physicians in the year 2002. In older age group, the physician-related variance of PIM initiation was 11%. Thus, the physician effect was systematic but relatively small and patient factors explained a larger proportion of variation in the data. The results remained essentially the same after including the patient-related predictors in the second (random-intercept) model. In 2013, all persons were aged over 75 years and the physician-related variance had decreased to 9%.
Discussion
Based on this study, a high proportion (37.5%) of those aged 65 years or older initiated the use of PIM during the 12-year study period. To our knowledge, this is the first study to determine the factors associated with PIM initiation in two age groups. Our study indicated that there were differences in factors associated with PIM initiation between these two age groups. We found that female gender and higher income level were associated with higher probability of PIM initiation in persons aged 65–
Accepted Article
74 years, whereas in persons aged ≥75 years, gender and high income were not significantly associated with PIM initiation. In both age groups, excessive polypharmacy and the use of psychotropics were associated with PIM initiation. Patient factors explained a large proportion of variation of PIM initiation in the data, but our study also found a physician effect, which indicated that PIM prescribing is not randomly distributed among the physicians.
One reason for the remarkable number of PIM initiators in this study might be a long follow-up.
Previous studies have primarily concentrated on point prevalence instead of initiation of PIMs. The most frequently initiated PIMs were tizanidine, metoclopramide and cough medication with opioids.
A recent Finnish study on PIM initiation in persons with and without Alzheimer’s disease (AD) also found that tizanidine and metoclopramide were the most commonly initiated PIMs in persons without AD [29]. On average, the PIM initiators accumulated 2.8 different PIMs over the entire 12-year study period. That finding is in line with an Australian study that found the mean number of different PIMs taken was 2.2 during a 13-year study period [17].
The decision of a PIM prescription is the result of the interaction of a variety of factors between patient and a prescribing physician. This was shown, when this study indicated that patient characteristics were differently associated with PIM initiation in different age groups. For example, the gender was differently associated with PIM initiation between age groups. PIMs were initiated more often among women in the younger age group, but in the older ones the gender was not significant. This finding is in concordance with a previous study, which found that women had higher risk of PIMs in the younger age group and not in the oldest-old group, but the earlier study included only persons aged 75 years and older [21]. A recent study from Canada also reported that increasing age had a slightly greater protective effect among women than men [30]. The finding of the higher risk of PIMs among women is in concordance with previous studies [12,26,31]. A potential explanation for this result may be more frequent health care service use among women [32].
Accepted Article
Our new approach with separate age groups yields another interesting finding of the socioeconomic position being associated differently between age groups. The risk of PIM initiation increased with higher income among persons aged 65–74 years. That is contrary to previous studies which have mainly suggested that PIM use is associated with lower socioeconomic position [15,19,25]. Our results may partly be explained by a previous finding that older persons with higher income are using more private doctor care [33]. In both age groups, the risk of PIM initiation was also higher in the second lowest income compared to the lowest income group. As expected, excessive polypharmacy was also strongly associated with PIM initiation in our study, which is consistent with previous studies [12,14,17,31]. In addition, our finding of the higher risk of PIM initiation with the use of other psychotropics (than were listed in Meds75+ database) was in line with previous studies [34].
Our results indicated that the majority of the total variation in the probability of PIM initiation was explained by patient case-mix, but it is interesting that the physician effect remained essentially the same after controlling patient characteristics. It means that the probability of prescribing PIM is not the same for all physicians, and some of them prescribed PIMs more likely than other physicians. The physician effect was smaller in the older age group in 2002. This could suggest that physicians avoid PIM prescribing when patients are growing older, which is supported by findings from previous studies that physicians attempt to be more cautious to prescribe PIMs to the oldest old persons [19].
The physician effect was decreased in 2013, but it is noteworthy that the study population is also more selected. In addition, the increased patient heterogeneity would have a reducing effect on the differences in PIM prescribing between physicians. Previous studies have also found that PIM use by patients varied among prescribing physicians [31,35], but in an Irish study by Cahir et al. [35], the variation between physicians was not significant after controlling patient-related factors.
The results of this study also showed that there were regional differences in PIM initiation between hospital districts. Previous studies provide mixed findings related to regional differences with PIM use [23-26]. For example, studies have reported lower rates of PIMs in smaller hospitals or rural areas and regions in smaller populations [23,25]. A study by Zhan et al. [24] did not find any differences
Accepted Article
between rural or urban locations or regions after controlling for patient sociodemographic factors and health status. However, these studies were not comparable to other countries because of different health care systems and populations. A previous study has reported that some health care system- related factors affecting a given population’s PIM use could be physician characteristics such as limited knowledge of PIMs or experiences, or organizational characteristics such as lack of time, limited applicability of PIM lists in daily practice, or lack of alternatives for PIMs [36].
The main strength of this study is the large, nationally representative longitudinal 12-year data. The Finnish registers, as well as other health registers in the Nordic countries, offer a quite unique and valid opportunity to study medication use in the long term [37]. We restricted our analyses to new PIM users to avoid prevalent user bias by implementing a two-year washout period. However, the study population might be healthier and more wealthy after the exclusions of prevalent PIM users.
Our study has also a few limitations which should be considered. Firstly, the Prescription Register includes data on reimbursed medication purchases, so information on over-the-counter medications is not available. Medication purchases may not necessarily reflect the real use of medications. However, register-based information has been found to provide more valid information on medication exposures compared to self-reported data [38]. Secondly, the Finnish Meds75+ database was not published until the year 2010 and our follow-up already started in 2002, so in the data, there might have been changes in the prescribing practices and the availability of medications [39]. However, the first PIM criteria, the Beers, were already published in the 1990s [5] and were taken into account in addition to other commonly used criteria (Laroche, STOPP/START) when the database was composed [11]. Thirdly, our study included persons aged 65 years and over, although the Finnish criteria were developed to improve pharmacotherapy among persons aged 75 years and over. Nevertheless, the age limit of 65 years is commonly used also in other studies assessing PIMs, and the mean age of our study population was relatively high. Lastly, registers do not include information on indications, dosage or physicians’ specialities.
Accepted Article
It should be noted that explicit PIM criteria do not take into account the heterogeneity in older persons. In addition, in some cases PIMs might be needed, for example in hospice and palliative care.
PIM initiation should always be carefully considered, and the decision of PIM prescribing requires always the clinical reasoning based on the indications and the assessing risks and benefits.
Conclusion
A remarkable number of older adults started using PIMs during the 12-year study period. There were differences between two age groups in the factors associated with PIM initiation. For example, gender differences occurred in the younger age group (aged 65–74 years). PIM initiation was more dependent on patient characteristics, but also physicians’ prescription choices and possibly differences in regional prescribing practices. However, physician-related variance explained PIM prescribing was decreased during the follow-up. In future studies, it appears important to investigate which physician- related factors are behind this variation, and what are the effects of PIM criteria in the reduction of PIM prescribing.
Acknowledgements
The authors would like to thank The Social Insurance Institution (SII) of Finland for funding this study, and Statistics Finland's Research Services for offering the data via remote access service.
Accepted Article
References
1. Saastamoinen LK, Verho J. Drug expenditure of high-cost patients and their characteristics in Finland. Eur J Health Econ 2013;14:495-502.
2. Kozyrskyj A, Lix L, Dahl M, Soodeen R. High-cost users of pharmaceuticals: who are they?
Manitoba Centre for Health Policy. Winnipeg. 2005; Available at: http://mchp- appserv.cpe.umanitoba.ca/reference/high-cost.pdf. Accessed 6/27, 2017.
3. Hovstadius B, Åstrand B, Persson U, Petersson G. Acquisition cost of dispensed drugs in individuals with multiple medications—A register-based study in Sweden. Health Policy 2011;101:153-161.
4. Saastamoinen LK, Verho J. Register-based indicators for potentially inappropriate medication in high-cost patients with excessive polypharmacy. Pharmacoepidemiol Drug Saf 2015;24:610-618.
5. Beers MH, Ouslander JG, Rollingher I, Reuben DB, Brooks J, Beck JC. Explicit criteria for determining inappropriate medication use in nursing home residents. UCLA Division of Geriatric Medicine. Arch Intern Med 1991;151:1825-1832.
6. Lund BC, Carnahan RM, Egge JA, Chrischilles EA, Kaboli PJ. Inappropriate prescribing predicts adverse drug events in older adults. Ann Pharmacother 2010;44:957-963.
7. Reich O, Rosemann T, Rapold R, Blozik E, Senn O. Potentially inappropriate medication use in older patients in Swiss managed care plans: prevalence, determinants and association with hospitalization. PLoS One 2014;9:e105425.
8. Fu AZ, Jiang JZ, Reeves JH, Fincham JE, Liu GG, Perri M. Potentially inappropriate medication use and healthcare expenditures in the US community-dwelling elderly. Med Care 2007;45:472-476.
Accepted Article
9. By the American Geriatrics Society 2015 Beers Criteria Update Expert Panel. American Geriatrics Society 2015 Updated Beers Criteria for Potentially Inappropriate Medication Use in Older Adults. J Am Geriatr Soc 2015;63:2227-2246.
10. Dimitrow MS, Airaksinen MS, Kivela SL, Lyles A, Leikola SN. Comparison of prescribing criteria to evaluate the appropriateness of drug treatment in individuals aged 65 and older: a systematic review. J Am Geriatr Soc 2011;59:1521-1530.
11. The Finnish Medicines Agency. Meds75+. 2015; Available at:
http://www.fimea.fi/web/en/databases_and_registeries/medicines_information/database_of_medicatio n_for_the_elderly. Accessed 9/5, 2017.
12. Guaraldo L, Cano FG, Damasceno GS, Rozenfeld S. Inappropriate medication use among the elderly: a systematic review of administrative databases. BMC Geriatr 2011;11:79-2318-11-79.
13. Hill-Taylor B, Sketris I, Hayden J, Byrne S, O'Sullivan D, Christie R. 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 2013;38:360-372.
14. Tommelein E, Mehuys E, Petrovic M, Somers A, Colin P, Boussery K. Potentially inappropriate prescribing in community-dwelling older people across Europe: a systematic literature review. Eur J Clin Pharmacol 2015;71:1415-1427.
15. Bongue B, Naudin F, Laroche ML, Galteau MM, Guy C, Gueguen R, et al. Trends of the potentially inappropriate medication consumption over 10 years in older adults in the East of France.
Pharmacoepidemiol Drug Saf 2009;18:1125-1133.
16. Hovstadius B, Petersson G, Hellstrom L, Ericson L. Trends in inappropriate drug therapy prescription in the elderly in Sweden from 2006 to 2013: assessment using national indicators. Drugs Aging 2014;31:379-386.
Accepted Article
17. Price SD, Holman CD, Sanfilippo FM, Emery JD. Are older Western Australians exposed to potentially inappropriate medications according to the Beers Criteria? A 13-year prevalence study.
Australas J Ageing 2014;33:E39-48.
18. Chang CB, Lai HY, Yang SY, Wu RS, Liu HC, Hsu HY, et al. Patient- and clinic visit-related factors associated with potentially inappropriate medication use among older home healthcare service recipients. PLoS One 2014;9:e94350.
19. Miller GE, Sarpong EM, Davidoff AJ, Yang EY, Brandt NJ, Fick DM. Determinants of Potentially Inappropriate Medication Use among Community-Dwelling Older Adults. Health Serv Res 2017;52:1534-1549.
20. Mo L, Ding D, Pu S, Liu Q, Li H, Dong B, et al. Patients Aged 80 Years or Older are Encountered More Potentially Inappropriate Medication Use. Chin Med J 2015;129:22-27.
21. San-Jose A, Agusti A, Vidal X, Formiga F, Gomez-Hernandez M, Garcia J, et al. Inappropriate prescribing to the oldest old patients admitted to hospital: prevalence, most frequently used medicines, and associated factors. BMC Geriatr 2015;15:42-015-0038-8.
22. Goulding MR. Inappropriate medication prescribing for elderly ambulatory care patients. Arch Intern Med 2004;164:305-312.
23. Rothberg MB, Pekow PS, Liu F, Korc-Grodzicki B, Brennan MJ, Bellantonio S, et al. Potentially inappropriate medication use in hospitalized elders. Journal of Hospital Medicine 2008;3:91-102.
24. Zhan C, Sangl J, Bierman AS, Miller MR, Friedman B, Wickizer SW, et al. Potentially inappropriate medication use in the community-dwelling elderly: findings from the 1996 Medical Expenditure Panel Survey. JAMA 2001;286:2823-2829.
Accepted Article
25. Beuscart JB, Genin M, Dupont C, Verloop D, Duhamel A, Defebvre MM, et al. Potentially inappropriate medication prescribing is associated with socioeconomic factors: a spatial analysis in the French Nord-Pas-de-Calais Region. Age Ageing 2017;46:607-613.
26. Jiron M, Pate V, Hanson LC, Lund JL, Jonsson Funk M, Sturmer T. Trends in Prevalence and Determinants of Potentially Inappropriate Prescribing in the United States: 2007 to 2012. J Am Geriatr Soc 2016;64:788-797.
27. Lublóy Á. Factors affecting the uptake of new medicines: a systematic literature review. BMC Health Services Research 2014;14:469. Doi: 10.1186/1472-6963-14-469.
28. WHO Collaborating Centre for Drug Statistics Methodology. The Anatomical Therapeutic Chemical Classification System. Structure and principles. 2011; Available at:
http://www.whocc.no/atc/structure_and_principles/. Accessed 1/20, 2017.
29. Hyttinen V, Taipale H, Tanskanen A, Tiihonen J, Tolppanen A, Hartikainen S, et al. Risk Factors for Initiation of Potentially Inappropriate Medications in Community-Dwelling Older Adults with and without Alzheimer's Disease. Drugs Aging 2017;34:67-77.
30. Morgan SG, Weymann D, Pratt B, Smolina K, Gladstone EJ, Raymond C, et al. Sex differences in the risk of receiving potentially inappropriate prescriptions among older adults. Age Ageing 2016;45:535-542.
31. Holmes HM, Luo R, Kuo YF, Baillargeon J, Goodwin JS. Association of potentially inappropriate medication use with patient and prescriber characteristics in Medicare Part D. Pharmacoepidemiol Drug Saf 2013;22:728-734.
32. Suominen-Taipale A, Martelin T, Koskinen S, Holmen J, Johnsen R. Gender differences in health care use among the elderly population in areas of Norway and Finland. A cross-sectional analysis based on the HUNT study and the FINRISK Senior Survey. BMC Health Services Research 2006;6:110-110.
Accepted Article
33. Blomgren J. Transition to Retirement and Use of Private Health Care. Int J Health Serv 2017;47:312-332.
34. Nishtala PS, Bagge ML, Campbell AJ, Tordoff JM. Potentially inappropriate medicines in a cohort of community-dwelling older people in New Zealand. Geriatr Gerontol Int 2014;14:89-93.
35. 1. Cahir C, Fahey T, Teljeur C, Bennett K. Prescriber variation in potentially inappropriate prescribing in older populations in Ireland. BMC Fam Pract 2014;15:59. Doi: 10.1186/1471-2296-15- 59
36. Voigt K, Gottschall M, Köberlein-Neu J, Schübel J, Quint N, Bergmann A. Why do family doctors prescribe potentially inappropriate medication to elderly patients? BMC Family Practice 2016;17:93. Doi: 10.1186/s12875-016-0482-3
37. Furu K, Wettermark B, Andersen M, Martikainen JE, Almarsdottir AB, Sorensen HT. The Nordic countries as a cohort for pharmacoepidemiological research. Basic Clin Pharmacol Toxicol 2010;106:86-94.
38. Rikala M, Hartikainen S, Sulkava R, Korhonen MJ. Validity of the Finnish Prescription Register for measuring psychotropic drug exposures among elderly finns: a population-based intervention study. Drugs Aging 2010;27:337-349.
39. Bell JS, Ahonen J, Lavikainen P, Hartikainen S. Potentially inappropriate drug use among older persons in Finland: application of a new national categorization. Eur J Clin Pharmacol 2013;69:657- 664.
Accepted Article
Figure 1. The 10 most frequently initiateyears (n=10,698)
ed PIMs in community-dwelling persons aged 65-74 a in 2002
and ≥75 2–2013.
Accepted Article
Table 1. Baseline characteristics of the study population.
Total (n=28,497) n (%)
PIM initiators (n=10,698) n (%)
PIM-non- initiators (n=17,799)
n (%) p-value
Age1 <0.001
65–74 years 15,144 (53.1) 6,403 (59.9) 8,741 (49.1) ≥75 years 13,353 (46.9) 4,295 (40.1) 9,058 (50.9)
Gender <0.001
Male 11,221 (39.4) 3,988 (37.3) 7,233 (40.6) Female 17,276 (60.6) 6,710 (62.7) 10,566 (59.4) Socioeconomic position
(income)2 <0.001
<9999 8,871 (31.1) 2,995 (28.0) 5,876 (33.0) 10 000–19 999 16,219 (56.9) 6,322 (59.1) 9,897 (55.6) 20 000–29 999 2,129 (7.5) 893 (8.4) 1,236 (7.0)
>30 000 931 (3.3) 409 (3.8) 522 (2.9) N/A 347 (1.2) 79 (0.7) 268 (1.5)
Living alone2 10,706 (37.6) 3,776 (35.3) 6,930 (38.9) <0.001 Medication use3
Anti-diabetics 2,156 (7.6) 815 (7.6) 1,341 (7.5) 0.795 Psychotropics 5,896 (20.7) 2,375 (22.2) 3,521 (19.8) <0.001 Cardiovascular medications 19,963 (70.1) 7,426 (69.4) 12,537 (70.4) 0.068 Anti-dementia medications 283 (1.0) 63 (0.6) 220 (1.2) <0.001 Excessive polypharmacy3 (≥10
medications) 3,199 (11.2) 1,455 (13.6) 1,744 (9.8) <0.001
University hospital district1,4 0.026
Helsinki 8,034 (28.4) 3,109 (29.0) 4,925 (27.6) Turku 5,184 (18.4) 1,893 (17.7) 3,291 (18.5)
Accepted Article
Tampere 6,171 (21.8) 2,316 (21.7) 3,855 (21.7) Kuopio 5,191 (18.3) 1,885 (17.6) 3,306 (18.6) Oulu 3,718 (13.1) 1,422 (13.3) 2,296 (12.9) Åland5 / NA 199 (0.7) 73 (0.7) 126 (0.7) Number of different PIMs6
1 4,829 (45.1) N/A
2 2,123 (19.8) N/A
3 1,174 (11.0) N/A
4 685 (6.4) N/A
5 and over 1,887 (17.6) N/A
PIM indicates potentially inappropriate medication; NA, not available
1At the start of follow-up (1 January 2002)
2Year 2000
3At the washout period (years 2000–2001)
4Including 20 hospital districts
5Not included in the analyses
6Frequency of different PIMs accumulated during the study period
Accepted Article
Table 2. Associated factors for PIM initiation by age group.
Model 1: PIM initiation in persons aged1 65–74 years
Model 2: PIM initiation in persons aged1 ≥75 years
HR 95% CI p-value HR 95% CI p-value Gender
Male 1.00 (reference) 1.00 (reference)
Female 1.12 (1.06–1.18) <0.001 1.04 (0.97–1.12) 0.245
Socioeconomic position (income)2
<9,999 1.00 (reference) 1.00 (reference)
10,000–19,999 1.08 (1.01–1.15) 0.025 1.14 (1.06–1.22) <0.001 20,000–29,999 1.18 (1.06–1.30) 0.002 1.15 (1.01–1.31) 0.037
>30,000 1.25 (1.09–1.43) 0.001 1.10 (0.91–1.32) 0.326 Living alone2 0.95 (0.89–1.01) 0.075 0.97 (0.91–1.04) 0.397 Medication use3
Anti-diabetics 1.05 (0.95–1.15) 0.359 1.11 (0.99–1.24) 0.066
Psychotropics 1.38 (1.29–1.47) <0.001 1.21 (1.13–1.29) <0.001 Cardiovascular medications 1.04 (0.98–1.09) 0.194 0.97 (0.90–1.04) 0.377
Anti-dementia medications 0.68 (0.42–1.09) 0.107 0.95 (0.71–1.27) 0.734 Excessive polypharmacy3 1.66 (1.52–1.80) <0.001 1.55 (1.43–1.69) <0.001
Hospital district1,4 0.009 0.007
Number of subjects 15,080 13,064
Number of failures (PIM initiations) 6,380 4,236 PIM indicates potentially inappropriate medication; HR, hazard ratio; CI, confidence interval
1At the start of the follow-up (1 January 2002)
2Year 2000
3At the washout period (years 2000–2001)
4Including 20 hospital districts
Accepted Article
Table 3. Multilevel models for physician effects on PIM initiation in years 2002 and 2013.
Year 2002 Year 2013
Constant-only model
Persons aged 65–74
years
Persons aged ≥75
years
Persons aged >75 years1
Observations 173,052 184,362 309,670
Number of physicians 9,014 8,530 11,869
OR (95% CI) p-value OR (95% CI) p-value OR (95% CI) p-value cons 0.01 (0.01–0.01) <0.001 0.01 (0.01–0.01) <0.001 0.01 (0.00–0.01) <0.001
Variance (95% CI) Std.error Variance (95% CI) Std.error Variance (95% CI) Std.error Physician 0.61 (0.45–0.84) 0.10 0.41 (0.27–0.62) 0.09 0.31 (0.20–0.48) 0.07 LR test vs. logistic model p<0.001 p<0.001 p<0.001
ICC 95% CI ICC 95% CI ICC 95% CI
Intraclass correlation 0.16 0.12–0.20 0.11 0.07–0.16 0.09 0.06–0.13 Random-intercept
model
Observations 173,052 184,362 309,670
Number of physicians 9,014 8,530 11,869
OR (95% CI) p-value OR (95% CI) p-value OR (95% CI) p-value Gender
Male 1.00 (reference) 1.00 (reference) 1.00 (reference) Female 1.24 (1.11–1.39) <0.001 1.06 (0.94–1.19) 0.365 1.15 (1.03–1.27) 0.011 Socioeconomic position
(income)
<9,999 1.00 (reference) 1.00 (reference) 1.00 (reference)
10,000–19,999 1.02 (0.89–1.17) 0.788 1.17 (1.03–1.33) 0.013 1.06 (0.95–1.20) 0.307 20,000–29,999 1.10 (0.88–1.36) 0.408 1.46 (1.17–1.82) 0.001 1.14 (0.94–1.38) 0.178
>30,000 1.08 (0.81–1.46) 0.594 1.04 (0.73–1.48) 0.819 1.45 (1.15–1.84) 0.002 Living alone 0.85 (0.75–0.97) 0.017 0.97 (0.86–1.10) 0.682 1.00 (0.90–1.12) 0.990 Medication use
Accepted Article
Antidiabetics 0.64 (0.52–0.78) <0.001 0.81 (0.67–0.98) 0.030 0.79 (0.64–0.97) 0.027 Psychotropics 1.16 (1.02–1.33) 0.021 0.95 (0.84–1.06) 0.355 1.16 (1.03–1.31) 0.014 Cardiovascular
medications 0.46 (0.41–0.52) <0.001 0.53 (0.46–0.61) <0.001 0.86 (0.78–0.96) 0.005 Anti-dementia
medications 0.36 (0.13–0.98) 0.046 0.78 (0.50–1.21) 0.262 1.42 (0.44–4.57) 0.557 Excessive polypharmacy 0.87 (0.75–1.02) 0.082 0.90 (0.79–1.04) 0.152 0.97 (0.83–1.14) 0.726 cons 0.01 (0.01–0.01) <0.001 0.01 (0.01–0.01) <0.001 0.00 (0.00–0.01) <0.001
Variance (95% CI) Std.error Variance (95% CI) Std.error Variance (95% CI) Std.error Physician 0.53 (0.38–0.75) 0.09 0.40 (0.26–0.61) 0.09 0.31 (0.20–0.48) 0.07 LR test vs. logistic model p<0.001 p<0.001 p<0.001
ICC 95% CI ICC 95% CI ICC 95% CI
Intraclass correlation 0.14 0.10–0.19 0.11 0.07–0.16 0.09 0.06–0.13 OR indicates odds ratio; CI, confidence interval; ICC, Intraclass correlation
1All persons were aged >75 years in the year 2013
