Principal findings and possible explanations
The prevalence of polypharmacy and excessive polypharmacy increased year-by-year, in the entire Swedish population 2005-2008.
With the exception of the age group 0-9 years, the prevalence of polypharmacy and excessive polypharmacy increased in all age groups. The prevalence of excessive polypharmacy displayed a clear age trend, with the largest increase for the age groups 70 years and above. Generally, the increase in the prevalence of polypharmacy was approximately twice as high for men as for women, and the increase in prevalence of excessive polypharmacy was about 1.5 as high for men as for women.
The increase rate for both polypharmacy and excessive polypharmacy levelled out during the study period, but between separate years, we noted a variation in rate of increase. This variation refers to the different age groups and to both genders.
The increase in the prevalence of polypharmacy may have several different causes: changes in the recommended prescriptions for various drug treatments as well as the introduction of specific drugs for treatment of conditions/diseases regarding which they have previously not been applied. Furthermore, middle-aged individuals are increasingly informed, and become, consequently, more prone to request an increased amount of prescription drugs. Finally, more drugs are being prescribed for preventive use. All together, these factors may have resulted in a change in the physicians' prescription patterns.
The decrease in the prevalence of polypharmacy in the age group 0-9 years can be explained by the national interventions to reduce the prescribing of antibiotics to children, in order to prevent antimicrobial resistance. Nearly 80% of the children 0-9 years with polypharmacy received antibiotics in 2006, clearly indicating that antibiotics have the largest impact on the prevalence of polypharmacy in this particular age group .
Both the overall increase and the differences in the rate of increase between the years are puzzling. These increases suggest relatively rapid changes in prescription patterns among prescribers; changes that may have a variety of causes, e.g. the introduction of new clinical guidelines.
Prior to 2005, national clinical guidelines were available for only three different areas in Sweden. During the study period, 2005-2008, The National Board of Health and Welfare in Sweden introduced four new national clinical guidelines; Stroke, Chest-Colorectal cancer and Prostate cancer, Heart disease, and Addiction, and in 2009-2011 seven other new clinical guidelines are planned to be introduced (e.g. Depression, Dementias, Diabetics, Lung cancer).
Prior to being officially introduced, new clinical guidelines exist only in preliminary versions. Consequently, these guidelines might influence the prescription habits and the development of polypharmacy a number of years before the guidelines being officially introduced. The introduction of national clinical guidelines for heart diseases and prostate cancer might explain both the unequal increase between genders, and the variation in increase rate between the different years.
In a study from Sweden concerning general practitioners' (GPs') perceptions of multiple-medicine use , clinical guidelines were viewed as "medicine generators". GPs' expressed frustration concerning guideline recommendations for certain diagnoses, e.g. cardiovascular diagnoses that "immediately result in five medicines". Regardless of the patients' other diseases, many guidelines were perceived as too rigid, leading to a standard "kit" of medicines per indication, and thereby resulting in that individuals with multiple diseases received an increasing number of different drugs.
The introduction of new national guidelines might therefore also contribute to explaining the age trend in the development of excessive polypharmacy, as older patients are more often exposed to several diseases. The elderly may receive, as a result of the guidelines, more often than others, a number of different "kits" of drugs added .
Strengths and weaknesses of the study
Our study presents an overview of the development of polypharmacy in an entire national population. The applied 3-month period prevalence of dispensed drugs includes all drugs that are prescribed on a regular basis (e.g. drug used in diabetes), when needed (e.g. analgesics), and temporarily (e.g. antibiotics). The periodically used drugs have been shown to have a different impact on the prevalence of polypharmacy in different age groups .
As the study included all individuals in the population, we avoided certain known problems concerning sampling, recall, interview and confidence. On the other hand, when the register data regarding the dispensed drugs is used as an estimator of drug use and polypharmacy, over-as well as underestimations of actual drug use arises. The extracted data included dispensed prescription drugs only, corresponding to approximately 82% of all Defined Daily Doses (DDD) distributed in Sweden. Also, additional sources of drugs, such as OTC medications, in-hospital medications and non-institutional care medications, herbal and alternative remedies together with previously filled prescriptions (before the study period), gifts and elicit Internet sales, were not included in the study, and resulted in an underestimation of the total consumption of drugs.
In addition, generic duplication (intended and unintended duplication of dispensed drugs with the same substance) might also have caused an underestimation of polypharmacy in our data, as we calculated only the number of dispensed drugs comprised of different substances. In sample studies of drug use among individuals with polypharmacy, patients often have two or more drugs with the same substance [4, 19, 20]. In register studies, it is difficult to make distinction between generic duplication and generic substitution (an intended switch between two drugs with the same substance). If the generic duplicate had been taken into account, this would have resulted in an even larger prevalence of polypharmacy. Whether the generic duplicate could have any impact on the development of the prevalence of polypharmacy over the study period has not been addressed.
Conversely, dispensed drugs as an indicator of drug use might result in an overestimation, as it is well known that a certain proportion of all dispensed drugs will never be used .
Strengths and weaknesses in relation to other studies
The displayed increase of polypharmacy in the entire population in Sweden since 2005 is in line with studies focusing only on elderly individuals during the 1980's and 1990's [2, 22–25].
However, there are certain difficulties in comparing our results concerning the elderly population with some of the previous studies. Firstly, some studies have addressed the level of drug use for the same individuals over time, concluding that drug use and polypharmacy increase with increasing age, but without an increased prevalence over time [26–30].
Secondly, some studies have applied varying time periods, different definitions of drug use and polypharmacy or different samplings of individuals . Finally, certain studies are based on interviews, and their results might be influenced by the sampling, recall or interview bias impedes comparison with results from register-based studies [15, 31].
The displayed year-by-year increase in drug use, polypharmacy and the mean number of dispensed drugs in the present study is generally minor compared to the increase shown in previous studies of the development of drug use in the 1980 s and 1990 s, e.g. a displayed 3-fold increase in the prevalence of polypharmacy and mean number of drugs per person during a ten year period . This difference might be explained by the fact that our data included all individuals in the national population. Previous studies have often used samples of only the elderly admitted to hospitals or living in nursing homes. Relatively healthy individuals might, therefore, not have been included in these earlier studies. Another possible explanation is that the recent efforts to reduce the increases in drug use and polypharmacy actually have had an effect.
Implications for clinicians and policymakers
The substantial increase in the prevalence of polypharmacy and excessive polypharmacy occurs simultaneously with the introduction of new clinical guidelines aimed at increasing the benefits of the medical treatment. The increase also occurs when the potential risks with polypharmacy have been highlighted, and various efforts have been made to reduce the number of drugs prescribed to individuals with an excessive number of drugs, especially the elderly. In Sweden, efforts to reduce the prevalence of polypharmacy have been focused on, at in first hand, the reduction of unintended generic duplication.
The assessment of the increasing prevalence of polypharmacy is not interpreted in a unanimous manner. For certain clinicians and policymakers, the results of the present study may be interpreted as the regrettable further development of polypharmacy, and that, in particular, excessive polypharmacy is continuing in an undesirable direction. However, the results of our study may also be interpreted to imply that a larger proportion of patients are receiving recommended drug treatment in line with new clinical guidelines.
The prevalence of polypharmacy may hide the fact that the benefits and/or risks of polypharmacy can be evaluated at individual level only. For clinicians, recommendations are required as to the manner in which to combine and balance different clinical guidelines to achieve an appropriate drug therapy for patients with multiple diseases.
Unanswered questions and future research
Over the 4-year study period, the increase in the prevalence of polypharmacy and excessive polypharmacy was particularly notable for men, 12% and 20%, respectively, and was even more notable for elderly men. This increase in drug use remains to be analyzed, and can possibly be associated with the introduction of new national clinical guidelines during the period with special relevance for men (e.g. guidelines for Heart diseases and Prostate cancer).