Abstract
Background Many older patients suffer from chronic diseases for which medicines should be used. Because of the higher number of medicines used and decline in hepatic and renal function, older patients are more prone to problems caused by these medicines. Therefore, it is important to review pharmacotherapy concerning older patients in primary care in a reliable way.
Aim To determine the nature, volume and clinical relevance of prescription-related points of attention in the elderly.
Design of study Analysis of pharmacotherapy by a multidisciplinary expert panel consisting of GPs, geriatric specialists, clinical pharmacists and community pharmacists.
Setting Pharmacotherapy of 102 home-dwelling older patients on polypharmacy (≥75 years, using ≥4 medicines continually) living in the Netherlands.
Method The analysis of medication-profiles was based on a two-round consensus method.
Results When performing medication reviews for older people it seemed that for almost all (98%) improvement in pharmacotherapy could be made. For 94% of all patients points of attention could be identified in prescribed medicines, of which 30% was considered to be of direct clinical relevance. In 61% of all patients a medicine could be added to improve pharmacotherapy, 25% of these prescribing omissions were considered to be of direct clinical relevance.
Conclusion The regular performance of medication reviews should be part of routine in primary care as it yields significant numbers of prescription-related points of attention. Although they were not all considered to be of direct clinical relevance, all points of attention do ask for a signal to the prescribing physician. This paper is not implying poor practice or poor reviewing practice but documenting the need for performing regular medication reviews.
INTRODUCTION
In the Netherlands 14% of the population consists of older people (≥65 years old). This proportion of the population is responsible for as much as 39% of all expenses on medicines as delivered by community pharmacies. People aged 65 years or over use three times as many medicines as compared to the whole population in the Netherlands (three medicines daily on average). People of 75 years or over use on average as many as four medicines daily.1 Older people use many medicines because they suffer from more chronic conditions that need treatment by means of pharmacotherapy. However, older people are more prone to adverse drug reactions, resulting from age-related factors such as changes in drug distribution, metabolism and excretion, and in receptor sensitivity as well as from drug–drug interactions and drug–disease interactions caused by prescribing of multiple drugs.2–5 In other words, prescribing in older patients involves balancing conflicting demands, and the benefit:risk ratio should be considered when deciding whether to initiate pharmacotherapy.
Although it is not possible to prevent all prescription-related problems in older people, several studies have shown that it is possible to reduce the occurrence of prescription-related problems by means of a medication review.6–9 In such a medication review, complete pharmacotherapy of an individual patient is assessed by a trained professional (GP and/or pharmacist). In the UK regular medication reviews for older people on long-term medication were recommended by the Department of Health to maximise therapeutic benefit and minimise potential harm,10 and this practice has been included in the Community Pharmacy Contractual Framework for all patients on long-term medication in the UK.11
In this article we describe the occurrence and clinical relevance of prescription-related points of attention found in older patients when use is made of an in depth and comprehensive approach with medication reviews performed by both prescribers and pharmacists. The occurrence of user-related pharmaceutical care problems in the same group of older patients had been determined in a previous study,12 creating insight in to whether it appears more effective to focus quality improvement interventions on prescribers (in particular GPs), or on the users of medicines.
This study is the first in-depth analysis by a large expert panel and focuses on a wider and more comprehensive set of prescription-related points of attention than previous studies have done.13–21 It therefore provides a more complete and accurate picture of the size and types of prescription-related points of attention faced by older patients as well as the clinical relevance of them. Whether or not precautions were taken by the prescribing physician (such as regularly checking potassium levels) to prevent these potential problems is not included in this study. However, the results of this study should give some insight in to the process of medication review that can be used for setting up better and more reliable medication reviews in the future.
METHOD
Study design and population
An analysis was performed of pharmacotherapy of 107 older people living in the community in the southeast of the Netherlands. Pharmacy dispensing data were collected from November 2001 to December 2002. The assessment of pharmacotherapy by the expert panel was based on a consensus method.
Patients were selected from the participants of a study on user-related problems12 with 298 home-dwelling participants of ≥75 years old who were being prescribed four or more medicines chronically, and were living in the south of the Netherlands. In the previous study, nine pharmacies were included (convenience sample). These pharmacies each contacted one to three GPs. The pharmacists and GPs invited eligible patients to participate in the study: patients were included if they returned the application form, including their informed consent.
How this fits in
Older people are more prone to adverse drug reactions, but they also use a higher number of medicines. Improvement in pharmacotherapy for almost all older patients can be made.
For each GP participating in this study (n = 18), six patients were picked at random, resulting in a total of 107 patients (for one GP only five eligible patients could be pointed obtained).
Variables and instruments
Types of prescription-related points of attention
Inappropriate prescribing was assessed based on the aspects described in Table 1.
Table 1 Aspects of inappropriate prescribing including examples for each aspect.
Clinical relevance of prescription-related points of attention
Panel members rated the clinical relevance of points of attention and prescribing omissions by means of a score from zero to three. Points of attention were considered as having clinical relevance if they could lead to a deterioration in general health status of the patient (see Table 2).
Table 2 Levels of clinical relevance for prescription related pharmaceutical care problems, including examples for each score of clinical relevance.
Procedures
Expert panel
The expert panel consisted of two GPs, two community pharmacists, two older-patient specialised internal medical specialists and two clinical pharmacists. Panel members were selected on the basis of their nationally recognised expertise in pharmacology and/or clinical older patient pharmacology.
Individual scoring
For each of the 107 participating older patients the panel members received a pharmacy record, a graphic medication record, the reasons for prescribing the medicines (provided by the GP), and a scoring form, containing all medicines regularly taken as determined by pharmacy records and the previously named aspects (see Table 1).10 The scoring forms were completed and sent back to the researcher by individual panel members. Before the consensus meetings, panel members received overviews in which their own scores were reflected in the light of the scores of the other panel members.
Consensus meeting
During the consensus meetings aspects of medicines were discussed that indicated a lack of consensus or were of clinical relevance. The researcher (a pharmacist) selected the points of attention that needed further discussion, including all items that had a score of at least six (when taking scores of all experts together) and all items that had scored at least a single three (clinically relevant item). An independent chairperson led the meeting. Panel members were invited to raise any additional topic that they considered of concern.
In case panel members were not able to join the meeting the researcher held an individual interview with the panel members to discuss his/her scores, and brought it into the discussion during the group meetings.
After the panel meeting, reports of the meeting, made by the researcher, were sent by email to all panel members, so that they could give their comments. Issues that remained unclear and comments of panel members were discussed again during the next consensus meeting, until consensus was reached.
Data analysis
After the panel discussions the scored points of attention (consensus) were analysed with SPSS 11 (SPSS Inc. Chicago, Illinois, US), and an inventory of all prescription-related points of attention was made. During the panel discussions it seemed that a score of 1 was not always used consequently; when an aspect was not relevant it was not scored at all. Therefore, in the results, only points of attention with a score of 2 (potential clinical relevant) or 3 (clinically relevant) are included.
RESULTS
Consensus meetings
In total, five panel discussions (four on telephone and one in person) took place during which the medications of 107 patients were discussed. On average, there were more than six panel members present during the panel discussions (one time, all experts were present, one time only five experts were able to participate, for two discussions six panel members participated and in one instance seven panel members were present).
On average, the total panel consensus contained more (and other) points of attention than the individual scoring lists. It appeared that each panel member had his/her own area of expertise. The individual written score κ value showed a variation for each item and each panel member (range 0.01–0.88). The average κ-value after the round in writing for all items and all panel members was 0.34 (slight agreement). The discussion sometimes yielded additional points of attention because of the interaction between panel members of different professions. During the consensus meetings, however, consensus was reached for all items.
Patients
In the panel discussion the medications of 107 elderly patients were discussed. After an evaluation of medicine use, five older patients were excluded because they used fewer than four medicines. The included patients were on average 81 years of age, were almost two-thirds female (62%), and used on average 6.8 medicines chronically. Forty-one per cent of the included patients got their prescriptions only from one physician.
In total, 102 older patients used 755 medicines. Medicines for cardiovascular diseases were prescribed most frequently (36% of the total number of medicines used), followed by medicines for the central nervous system (13%), the alimentary tract and metabolism (12%), and blood and blood-forming organs (10%).
In the medication records of 98% of all patients, points of attention were identified. In 4% of these medication profiles the expert panel had no comments on the medicines currently used, but one or more medicines could possibly or should be added to improve pharmacotherapy.
Number, type and clinical relevance of prescription-related points of attention
Panel members rated 457 points of attention considering prescribed medicines used by 96 older patients. Thirty per cent of these recommendations were considered to be of direct clinical relevance, the remaining 70% was considered to be of potential clinical relevance. The latter category of problems can possibly partly be solved by reviewing the medical records (such as measures of potassium or blood pressure), but whether or not these measures were regularly performed by the GP was not registered in our study.
Table 3 shows the distribution of the points of attention by various problem categories.
Table 3 Number and types of prescription-related points of attention (including points of attention with potential clinical relevance and points of attention with direct clinical relevance).
Medicines considered as being not useful are reported most frequently, seen in anatomical therapeutic classification (ATC) group N (medicines for the nervous system, 21%) and C (medicines for the cardiovascular system, 20%). The problem category seen second most frequently is prescribing medicines for an incorrect period of time, almost exclusively (88%) seen in ATC group N, and prescribing medicines in a dose not appropriate for older people, seen in group C (56%) and N (40%). Drug–drug interactions are also reported frequently, drug–drug interactions are mainly (57%) caused by medicines from ATC group C, medicines for the cardiovascular system.
Table 4 shows the percentages of medicines out of main ATC groups having at least one prescription-related point of attention of potential clinical relevance. The main ATC group R (medicines for the respiratory system) is the group with the highest number (relatively); this is mainly caused by concerns of panel members about the suitability of the inhalation devices for elderly patients, but also about the use of mucolytics. There is some doubt whether these preparations are effective. At the time of our study, the leading Dutch Drug Compendium (Farmacotherapeutisch Kompas) discouraged the use of oral mucolytics, this discouragement is still present in the 2006 edition.22 The panel felt that their use should, at the very least, be carefully considered.
Table 4 Number of recipes within a main anatomical therapeutic classification (ATC) group.a
Main ATC group M (medicines for the musculoskeletal system) is the group with the second highest number of points of attention, mainly caused by drug–drug interactions caused by NSAIDs (26%), use of hydroquinine or NSAIDs being less appropriate for the elderly (26%), and use of NSAIDs when other analgesics are indicated (18%).
In the main ATC group G (medicines for the genitourinary system and sex hormones) recommendations were related to medicines for incontinence with a marginally proven effectiveness (while leading to side effects) for which alternatives exist causing fewer side effects (42%) and inappropriateness for older people because of anti-cholinergic side-effects (33%). In group N (medicines for the nervous system) points of attention were mainly related to prolonged prescribing of benzodiazepines (50%). Points of attention in this group were aimed at prescribing long-acting benzodiazepines that are less suitable for use in the elderly (27%) and prescribing drugs—mainly benzodiazepines—in dosages exceeding the geriatric daily dose (23%).
In some ATC groups, high percentages of prescriptions have at least one recommendation.
These recommendations can be categorised into specific groups of points of attention, more than half of all points of attention can be identified by looking at these specific medicines or groups of medicines.
Prescribing omissions
By reviewing the complete medication profiles, it appeared that 101 medicines might have been needed to improve the quality of medication therapy in 62 patients (61% of all older patients). Score 2 (a medicine might be added to improve pharmacotherapy depending on the general condition of the patient) was scored in 76% of all cases and seen in 52% of all older patients. Twenty-five per cent of the omitted medicines had a score of 3, meaning that according to prescription guidelines a medicine should be added to improve pharmacotherapy. These prescribing omissions were seen in 23% of all elderly patients.
More than half of all prescribing omissions (60%) were found in main ATC group C (medicines for the cardiovascular system), for example, the need of adding an ACE-inhibitor to pharmacotherapy of an elderly patient with heart failure. Twenty-two per cent of the prescribing omissions could be categorised in main ATC group B (blood and blood forming organs), such as adding a thrombocyte-aggregation-inhibitor to the pharmacotherapy of an older patient with angina pectoris. Ten per cent of all omitted medicines belonged to main ATC group R (respiratory system); a medicine should probably be added to optimise ATSMA/COPD treatment, such as rescue-medication (short-acting β2-sympathicomemetica) for the treatment of a patient only using long-acting β2-sympathicomimetica.
DISCUSSION
Summary of main findings
In this study, prescription-related points of attention of potential clinical relevance were found in pharmacotherapy of almost all included patients. One-third of the points of attention found in prescribed medicines were considered to be of direct clinical relevance, implying that these prescriptions should be changed unconditionally. The remaining two-thirds were potentially relevant, meaning that adjustment would depend on clinical measurements or specific clinical parameters of the patient, whether or not these precautions were taken by the physician was not registered in our study. In addition, the panel determined that a relevant medication was missing or potentially missing in almost two-thirds of the patients.
Strengths and the limitations of the study
This study is the first in-depth analysis by a large expert panel and focuses on a wide and comprehensive set of prescription-related points of attention. It provides a complete and accurate picture of the number and types of prescription-related points of attention faced by older patients as well as the clinical relevance of these problems.
Our study is not without limitations. First, the patients in our study consisted of a limited sample. Although their number was quite high for such a comprehensive method of evaluation, some types of prescribing problems—in particular those that occur rarely—may be underrepresented. Second, consensus approaches always entail a risk that some panel members are more influential than others. Third, our expert panel has identified points of attention on the basis of a medication record and the indications for the medicines as given by the physician. Our panel had no medical records at their disposal. In most instances, regular checks and measurements will be performed by the physician and in some instances a second choice medicine will be optimal treatment because other medicines will not be tolerated by the particular patient. Our study does indicate a high number of points of attention in daily practice. However, a part of these points of attention will be dealt with already by means of regular checks. This paper is not implying poor practice or poor reviewing practice but documenting the need for regular medication reviews.
Comparison with existing literature
Recommendations were mainly seen in the medicines for the respiratory system, the cardiovascular system and the nervous system. Points of attention regarding medicines for the cardiovascular system were mainly caused by drug–drug interactions, which were in most instances not of direct clinical relevance. In daily practice, high numbers of drug–drug interactions are seen within this group, and many problems caused by these interactions will be prevented by regularly measurements (such as potassium levels or blood pressure).23–24
Recommendations regarding medicines for the respiratory tract were mainly aimed at the suitability of inhalation devices used for older patients. This is consistent with other studies that also found that older patients frequently have problems taking inhaled medication,25,12 therefore such a signal to the physician may be relevant.
Most points of attention of direct clinical relevance were seen in the group of medicines for the central nervous system, which were in particular related to benzodiazepine use. Problems included the use for an incorrect period, in dosages exceeding the geriatric daily dosage and use of substances with a long half-life time that are not suitable for use in older patients. Prolonged use of hypnotics, particularly in the elderly, is a widespread problem, as numerous studies concerning inappropriate prescribing for the elderly have shown.13,14,20,26
In almost two thirds of the patients, prescribing omissions were identified, of which one out of four were of direct clinical relevance. Prescribing omissions are only scarcely described in studies concerning inappropriate prescribing for the elderly,27 in spite of studies that prove that a substantial number of older patients is not receiving omitted but necessary pharmacotherapy for established diagnosis.28–31 Prescribing omissions may place older patients at higher risk for preventable adverse consequences. Hence, medication reviews should point at the quality of complete medication profiles and not only at the quantity of drugs prescribed.
Implications for future research or clinical practice
Over half of detected points of attention recurred in only a handful of drug classes, suggesting that medication reviews of older outpatients on polypharmacy may benefit from a computerised screening tool. Although such a computerised screening tool could detect a large proportion of potential problems, the detection of various other problems in our analysis shows that such a tool should be supplemented with a more implicit method of assessment. The professional judgement of a complete medication profile by an experienced healthcare provider can detect problems that would go unnoticed if one would rely solely on computerised screening. The overall κ-value indicated slight agreement after the round in writing. All panel members seemed to have their own speciality. During the consensus meetings, however, consensus about all aspects was reached. In some instances panel members had to make out their case, in other instances consensus was reached quickly because other panel members realised they had overlooked a particular problem. Another interesting observation (data not shown) was that about 15% of the points of attention could only be detected because the panel was not only supplied with the medications prescribed but also with the reasons for prescribing them. Together these findings raise the possibility that medication reviews ideally should be performed by more than one healthcare professional, ideally of different professions, with the medical record at their disposal. Further research is needed to confirm these assumptions.
All in all, we conclude that it appears advisable to perform medication reviews for home-dwelling older patients by GPs, community pharmacists and other specialists. It yields significant numbers of relevant prescription-related points of attention and a potential for quality improvement of prescriptions for older patients living in the community.
Acknowledgments
The authors thank YA Hekster, PharmD, PhD; P van den Hombergh, MD, PhD; PAF Jansen, MD, PhD; JR van der Laan, MD: CPJM Lemmens, PharmD; CK Mannesse, MD, PhD and EC Weening PharmD for their contribution to the expert panel. Furthermore, we want to thank the participating pharmacists and GPs for their contribution to our study.
