Abstract
Background:
Drug-related problems (DRP) are common in the elderly population, especially in people living with dementia (PwD). DRP are associated with adverse outcomes that could result in increased costs.
Objective:
The objective of the study was to analyze the association between DRP and healthcare costs in PwD.
Methods:
The analysis was based on the cross-sectional data of 424 PwD. Compliance, adverse effects, and drug administration of prescribed and over-the-counter drugs taken were assessed. DRP were identified and classified by pharmacists using an adapted German version of “PIE-Doc®”. Healthcare utilization was assessed retrospectively used to calculated costs from a public payer perspective using standardized unit costs. The associations between DRP and healthcare costs were analyzed using multiple linear regression models.
Results:
394 PwD (93%) had at least one DRP. An inappropriate drug choice was significantly associated with increased total costs (b = 2,718€; CI95% 1,448–3,988) due to significantly higher costs for hospitalization (b = 1,936€; 670–3,202) and for medications (b = 417€; 68–765). Problems with medication dosage and drug interactions were significantly associated with higher medication costs (b = 679€; 31–1,328; and b = 630€; 259–1,001, respectively).
Conclusions:
DRP could significantly lead to adverse outcomes for PwD and healthcare payers, reflected by a higher hospitalization and costs, respectively. Further research is needed to clarify on interventions and approaches efficiently avoiding DRP and on the effect on patient-reported and economic outcomes.
INTRODUCTION
At present, dementia is an incurable syndrome that goes along with deterioration in cognitive function [1]. There were 47 million people with dementia (PwD) worldwide in 2016 [2]. It is estimated that the prevalence of dementia will double every twenty years, reaching 74.7 million PwD in 2030 and finally 131.5 million in 2050 [3]. Considering this increasing prevalence and the fact that dementia goes along with reductions of functioning in everyday living [4], dementia disease represents a giant economic burden as well—for the public healthcare payers and for the society. Total worldwide costs valued US$1 trillion in the year 2018 [5]. Healthcare cost will increase continuously as well—US$2 trillion are estimated for 2030 [3]. In Germany, €15.1 billion were spent on dementia in 2015, representing 5% of total healthcare costs in Germany [6].
Dementia is common in people over the age of 65 [7] and thus, a special concern of the elderly [8]. Patients could have several co-existing illnesses next to dementia diseases (multimorbidity) [9] and usually take several medications (polypharmacy) [10]. Polypharmacy, in turn, could be associated with drug-related problems (DRP) [11, 12], which are defined as events that actually or potentially interfere with desired health outcomes [13]. DRP can be related to drug selection, dose selection, or can be patient-related like insufficient storage of drugs or application [13, 14]. Based on data from a randomized controlled clinical trial exploring the effects of pharmacist intervention in patients 65 years and older with dementia or cognitive impairment, Pfister et al. estimated that two-thirds of PwD have at least one DRP [15]. Other studies underlined that DRP are highly prevalent in PwD [16, 17]. Among DRP, potentially inappropriate medication (PIM) [15, 18–21] and inappropriate drug storage [22] are the most common DRP in the elderly.
DRP could cause adverse outcomes, like the reduced quality of life [23] and increased hospitalization [24, 25]. Therefore, DRP could be associated with higher healthcare cost as well. Only a few studies evaluated the impact of DRP on healthcare costs. Bootman et al. estimated that a reduction of DRP is associated with a lower institutionalization and thus, cost for nursing homes [26]. However, little is known about the association of DRP and its impact on total healthcare costs, especially in community-dwelling PwD. Therefore, the present study aimed to analyze the association of DRP and healthcare costs in community-dwelling PwD from a public payers’ perspective.
MATERIALS AND METHODS
Study design, setting, and sample
The analysis was based on cross-sectional data of the general practitioner (GP) based cluster-randomized, interventional trial DelpHi-MV (Life- and person-centered help in Mecklenburg-Western Pomerania, Germany) [27]. In this trial, 6,838 patients were screened for dementia in 125 participating GP practices in Mecklenburg-Western Pomerania, Germany, using the DemTect procedure [28, 29]. Patients who were eligible (being at the age of 70 and above, living community-dwelling, and being positively screened for dementia) were asked to participate in the trial and to provide written informed consent. In the case that the patient was unable to give an informed consent, his or her legal representative was asked to sign the consent on his or her behalf (as approved by the Ethical Committee of the Chamber of Physicians of Mecklenburg-Western Pomerania, registry number BB 20/11). The study design has been published in more detail elsewhere [27].
After screening, 1,166 patients (17.1%) of 94 general practitioners were eligible and 634 (54.4%) of them agreed to participate. 118 participants dropped out due to withdrawal of informed consent (n = 85), death (n = 19), relocation (n = 5), or other reasons (n = 9) before starting the baseline assessment. A total of n = 516 participants started the baseline assessment. Of these, n = 92 participants were excluded from the present analysis due to missing data concerning activities of daily living, cognitive impairment, depressive symptoms, resource utilization, and DRP. Thus, this analysis was based on a sample of n = 424 PwD. A detailed description of the study sample is published elsewhere [30].
Socio-demographic and clinical variables
To assess the association between DRP and healthcare cost, we included the following socio-demographic and clinical variables: Activities of daily living (ADL) were measured using the Bayer-Activity of Daily Living Scale (B-ADL), developed to assess deficits in daily living activities [27, 28]. To assess the participant’s cognitive impairment, the Mini-Mental State Examination (MMSE) was used [31, 32]. According to the German S3-guidelines for dementia, the cognitive impairment was divided into the following four groups: without (MMSE values 27–30), mild (MMSE values 20–26), moderate (MMSE values 10–19), and severe (MMSE values 0–9) cognitive impairment [33]. Depressive symptoms were measured using the Geriatric Depression Scale (GDS), a 30-item questionnaire to assess depression. A score of five and higher suggest a hint for a depression. [34]. Patients' comorbidity was assessed using patients’ ICD-10 codes listed in the GP’s medical records as well as the Charlson Comorbidity Index [35].
Healthcare utilization and costs
Patients’ resource utilization of medical treatment and care services were assessed using a standardized computer-assisted interview [27]. The utilization of out-patient physician treatments, in-patient treatments, rehabilitation, medications, medical aids, therapies (e.g., occupational, physical and speech therapy), and ambulatory care services were assessed retrospectively for a period of 12 months.
Healthcare costs were calculated using a bottom-up and prevalence-based cost-of-illness design as well as published unit costs [36], representing the average cost per patient from a public payers' perspective for a retrospective period of 12 months. Informal healthcare costs as well as productivity losses of patients' caregivers have not been considered for the present analysis. Detailed information concerning the calculation of healthcare costs is demonstrated in the Table 1.
Methods and used unit costs for monetary valuation of medical and formal healthcare resources and services
GP, general practitioner; AOK, allgemeine Ortskrankenkasse; ADL, activities of daily living; IADL, instrumental activities of daily living; *care level one: mild functional impairment, care level two: moderate functional impairment, care level three: severe functional impairment; ‡ when drugs, aids or services were unknown or market prices were not available; † inflation included.
Medication review and DRP classification
Dementia-specific qualified study nurses interviewed patients at their own homes and assessed all drugs that were taken regularly (including prescribed and over-the-counter drugs) [37]. Furthermore, the compliance, any adverse effects, and the drug administration were assessed [38]. Afterward, the medication reviews were conducted by study pharmacists. Active substances were coded according to the Anatomical Therapeutic Chemical (ATC) classification system. Topical agents and homeopathic medicines were not considered in this analysis. Subsequently, the DRPs were categorized into five main groups according to the PIE-Doc®” system (Problem-Intervention-Result-Documentation), which was originally conceptualized in Germany by Schaefer et al. [14]: 1) inappropriate drug choice, 2) problems with administration and compliance, 3) problems with the dosage, 4) adverse drug events, and 5) problems with drug interactions. Drug interactions, drug-food interactions, and double prescriptions were identified by the Risk-Check tool CAVE of the ABDA-Database. A “traffic light system” was employed to classify the interactions into six categories of severity: “serious consequences probable, contraindicated”, “contraindicated as a precaution”, “monitoring or adjustment is needed”, “monitoring and adjustment is necessary in some cases”, “supervise as a precaution”, and “no action is normally required”. The first three categories of severity were considered during the medication reviews. The clinical relevance of drug-drug interactions was assessed by the pharmacists during the medication review implementation. In the analysis, the interactions of category “monitoring or adjustment is needed” were described as “potential drug interactions of moderate severity”. The interactions of categories “serious consequences probable, contraindicated” and “contraindicated as a precaution” were summed as “potential drug interactions, clinically relevant”. PIMs were determined using a list of PIM in the elderly (Priscus list). Due to the fact that PIM is very common in the elderly [39] and a huge part of the DRP, PIM according to the German Priscus list [40] was taken into account as a subcategory of inappropriate drug choice. The German Priscus list was established in line with the international PIM lists and published in 2010, aiming to reduce the rate of adverse drug events and to provide higher medication safety.
More detailed information concerning DRP in people who participated in this study were published by Wucherer et al. in 2017 [22]. A detailed description and definition of each category are demonstrated in Table 2.
Classification of drug related problems and its subcategories
Statistical analysis
To analyze the association between DRPs and healthcare costs, we used different multiple linear regression models. In these models, the association with costs was assessed for total DRPs (dichotomous: no DRP versus at least one DRP) and for each category of DRP, separately (no DRP in a specific category versus at least DRP in that category). Total healthcare cost, formal care cost, as well as medical care cost and its subcategories, cost for medication, cost for hospitalization, and physician costs were used as dependent variables. Each subcategory of DRP was used as independent variable. We furthermore adjusted each model for the different socio-demographic and clinical variables mentioned above. Standard errors were assessed using non-parametric bootstrapping with 2000 replications. We additionally included random effects for each GP to adjust for the effect of each cluster (GP practice). STATA/IC version 13.0 was used for statistical analysis.
RESULTS
Socio-demographic and clinical variables
A detailed description of patients’ socio-demographics and clinical variables is demonstrated in Table 3.
Study sample’s socio-demographic and clinical characteristic
ICD-10, International Classification of Diseases and Related Health Problems; B-ADL, Bayer Activities of Daily Living Scale, score 1–10, higher score means more difficulties; MMSE, Mini-Mental State Examination, score 0–30, higher score indicates a higher severity of cognitive impairment; GDS, Geriatric Depression Scale, hint for depression if score > 5; SD, standard deviation; GP, general practitioner.
To sum up, 56% (n = 239) of the PwD were females, 80 (SD 5.4) years old, and mildly functionally (B-ADL mean score 4; SD 2.4) and cognitively (MMSE mean score 23; SD 4.9) impaired. 49% (n = 206) of the PwD were living alone.
Number of drugs taken and drug-related problems
On average, 6.5 (SD 3.17; range 0 – 21) drugs were taken. Most PwD were taking between 6 and 10 drugs (n = 234, 55.2%). 92.7% of the PwD had at least one DRP. On average, PwD had 2.5 (1.49) DRP. Inappropriate application by the patient/problems with administration and compliance was the most common DRP, appearing in 80.2% (n = 341) of cases. Problems with drug interaction was observed in 38.4% (n = 163), inappropriate drug choice in 31.6% (n = 134), inappropriate dosage/problems with dosage in 10.4% (n = 44) and adverse drug events in 6.4% (n = 27) of PwD. In addition, 95 PwD (22.4%) took at least one PIM according to the “Priscus list” [40]. For example, two percent of the study participants (n = 9) took acetylcholinesterase inhibitors and anticholinergic drugs (quetiapine: n = 5 cases; amitriptyline: n = 3; doxepin, n = 1; tolterodine, n = 1) at the same time. A detailed description of the prevalence of DRP is demonstrated in Table 4 and Supplementary Table 1.
Occurrence of drug related problems
*Number of each DRP per person with dementia; **Percentage of person with dementia having at least one DRP in each category. DRP, drug related problem.
Healthcare costs and association between drug-related problems and costs
Total costs from the payers’ perspective valued 7,713€ (SD 399), summarizing medical care cost of 5,452 € (SD 325) and formal care cost of 2,260.57€ (SD 191).
In the multivariate models, inappropriate drug choice was significantly associated with higher total costs of 2,718 € from the payers’ perspective (CI95% 1,448€ to 3,988€). This was caused by a significantly higher cost for hospitalizations (b = 1,936€; CI95% 670€ to 3,202€) and medication costs (b = 417€; CI95% 68€ to 765€). In addition, taking a drug of the Priscus list was as well significantly associated with 3,002€ (CI95%1,227€ to 4,775€) higher total costs from a payers’ perspective, especially due to significantly higher cost for hospitalization and medication cost as well. These higher costs were not caused by some outlying high-cost patients. Moreover, the distribution of comorbidities in patients having at least one inappropriate drug choice and those who have not were furthermore analyzed considering patients’ comorbidity using the Charlson Comorbidity Index [35]. Apparently, the percentage of participants having diabetes without complications or renal diseases was significantly higher in participants who had at least one inappropriate drug choice. However, the association between an inappropriate drug choice and higher cost remains significantly after adjusting for several comorbidities.
Furthermore, medication costs significantly increased in PwD with problems with dosage (b = 679€; CI95% 31 to 1,328) and problems with drug interactions (b = 630€; CI95% 259 to 1,001). According to the utilization of formal care services, costs significantly decreased in PwD with adverse drug events (b = –1022€ 95% CI –1893 to –151). The results of the multivariate analyses are demonstrated in Table 5.
Association between drug related problems and different cost categories and summarized total costs Results for healthcare costs from the payers’ perspective depending on drug-related problems, costs in Euros
*p value < 0.05; **p value < 0.01; ***p value < 0.001; 1includes cost for physician visits, medication, in-hospital treatments, medical aids and out-patient therapies; 2includes medical treatment cost and formal care cost. Multiple linear regression models adjusted for socio-demographic and clinical variables. Bootstrapping with 2000 replications and random effects for general practitioners were used. Please note that Medical Treatment Costs as well as Formal Care Costs are partial amounts of total costs. Moreover, Costs due to Hospitalization as well as Medication Costs are partial amounts of Medical Treatment Costs.; b, observed coefficient; SE, standard error; CI, confidence interval.
DISCUSSION
The results of this analysis demonstrated that DRP are very common in PwD and that DRP could significantly increase total healthcare costs. Especially cost for hospitalization and medication increased due to an inappropriate drug choice or PIM taken as well as problems with the dosage and drug interactions, respectively. Contrary to this, the occurrence of adverse drug events was associated with reduced costs for formal care services, which could be an inverse causality.
Wucherer et al. revealed for the same sample used for this analysis that a higher number of DRP is significantly associated with a higher number of drugs taken, obviously representing higher costs for medication [22]. This could be the reason for the revealed link between DRP and the higher cost for medication. Therefore, it is very likely that the higher number of drugs taken and thus, the higher total cost for the medication of a PwD represent a higher risk of DRP and not the other way around. However, such DRP are avoidable by using medication management as well as by improving the interdisciplinary cooperation between practitioners and pharmacists [41, 42].
However, according to our results, inappropriate drug choice and the intake of PIM seem to be associated with increased cost for in-hospital treatments up to 1,936 € from the payers’ perspective. Harrison et al. highlighted in 2018 as well that PIM often appeared in the elderly and caused increasing costs within the Australian INSPIRED study [39]. Through the fact that total costs in dementia from the payers’ perspective valued around 7,016€ [37] to 9,207€ [43], the increase of costs for the hospitalization due to DRP represents a major part of total healthcare costs in dementia that have to be paid by healthcare insurances. Therefore, DRP and PIM are major cost drivers in PwD and should be a priority of health care payers and insurances.
Since PIM are very common in the elderly population [22, 44], which will rapidly grow within the next decades [45, 46], the findings of this study might have significant relevance not only for the public payers but also for PwD therapy. Pohl-Dernick et al. estimated that prescribing the recommended substitution instead of the PIM would cause higher reimbursement cost [47]. However, this analysis did not include patient-related and economically relevant long-term effects like a subsequent hospitalization or further adverse outcomes. Even though a substitution could cause higher cost for the patient’s medication, the higher cost would not exceed the cost for hospitalization caused by PIM, as demonstrated in this analysis. For example, Pohl-Dernick et al. reported that substitution of Piracetam (annual cost of 164€) by acetylcholinesterase inhibitors (annual cost of 1,500€), like donepezil, galantamine, and rivastigmine, would not exceed the reported higher cost for hospitalization of 1,500€ in this paper [47]. Thus, there are sound economic, medical, and pharmaceutical reasons to avoid PIM.
The association between adverse drug events and lower cost for formal care is unclear, unexpected, and never confirmed by previously published studies. It has to be noticed that this finding was based on a very small subsample of PwD (n = 27) having at least one adverse drug event. According to the Charlson Comorbidities Index, there were no significant differences between people with or without adverse drug events. However, PwD that had at least one adverse drug event were more functionally impaired according to the B-ADL (4.0 versus 3.5, p = 0.182), which could explain the higher utilization of formal care services. However, this association could be an inverse causality: Community-dwelling PwD without any assistance by ambulatory care services that could better monitor and supervise the medication of the patient could, therefore, have more often adverse drug events. Further research is needed to clarify if home support could significantly decrease DRP, especially adverse drug events.
In light of the result of this analysis, physicians and pharmacists should be encouraged to better survey the prescription of PIM in order to avoid adverse outcomes for the PwD, like unplanned in-hospital treatment that could have further aftereffects, as well as to avoid the waste of already scarce resources. Different approaches are currently available and needed to improve medication management in PwD. Furthermore, physicians should be sensitized about PIM and another inappropriate drug choice and their impact on patient-related treatment outcomes as well as on long-term cost. Implementing interdisciplinary networks and cooperation between practitioners and community pharmacies could improve the current situation [48, 49]. Further qualification and sensitization of caregivers could enhance such effects [50].
However, until now, there is a paucity of knowledge about the effectiveness of interventions that aimed to improve the medication of PwD and to reduce DRP. Only a few studies revealed benefits, for example, of a comprehensive geriatric assessment including an assessment of patients’ medication [41] and in-hospital medication management [42]. Therefore, further research is needed to assess the effectiveness as well as the cost-effectiveness of such approaches in the long run [51, 52].
Limitations
The DelpHi-trial was conducted in mildly cognitively impaired patients in a mostly rural setting in Germany, which limits the generalizability of the presented results as well as the reproducibility of findings in older and more severely cognitively impaired patients. Solely, 634 (54%) of 1,166 eligible patients participated in this study, which could indicate a selection bias. However, we found no significant differences in age, sex, and screening score (DemTect) between participants and non-participants. Therefore, it seems to be that the sample under analysis is representative for community-dwelling primary care patients living with dementia. Furthermore, the validity of the assessed health resource data might be limited with regard to completeness and accuracy, due to recall bias among PwD and caregivers. In addition, DRP like adverse drug events are in part subjective inceptions and may be estimated incorrectly.
Conclusions
The results of this study add important evidence to the current paucity of knowledge about the impact of DRP, demonstrating that DRP could significantly lead to adverse outcomes for PwD and healthcare payers. Most of those problems could be prevented using innovative approaches, like medication management or medication review and better cooperation between pharmacists and practitioners. Therefore, further research is needed to clarify if interventions and approaches could efficiently avoid DRP and how this would affect patient-reported and economic outcomes.
CONFLICT OF INTEREST
Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/19-0819r1).
