The authors have declared that no competing interests exist.
Conceived and designed the experiments: JG YN WR. Performed the experiments: HG EB. Analyzed the data: JG YN. Contributed reagents/materials/analysis tools: HG WR EB. Wrote the paper: JG YN WR HG EB. Conceived and designed the experiments: JG YN WR. Performed the experiments: HG EB. Analyzed the data: JG YN . Contributed reagents/materials/analysis tools: HG WR EB. Wrote the paper: JG YN WR HG EB.
Adherence to medication is low in specific populations who need chronic medication. However, adherence to medication is also of interest in a more general fashion, independent of specific populations or side effects of particular drugs. If clinicians and researchers expect patients to show close to full adherence, it is relevant to know how likely the achievement of this goal is. Population based rates can provide an estimate of efforts needed to achieve near complete adherence in patient populations. The objective of the study was to collect normative data for medication nonadherence in the general population.
We assessed 2,512 persons (a representative sample of German population). Adherence was measured by Rief Adherence Index. We also assessed current medication intake and side effects. We found that at least 33% of Germans repeatedly fail to follow their doctor's recommendations regarding pharmacological treatments and only 25% of Germans describe themselves as fully adherent. Nonadherence to medication occurs more often in younger patients with higher socioeconomic status taking short-term medications than in older patients with chronic conditions. Experience with medication side effects was the most prominent predictor of nonadherence.
The major strengths of our study are a representative sample and a novel approach to assess adherence. Nonadherece seems to be commonplace in the general population. Therefore adherence cannot be expected per se but needs special efforts on behalf of prescribers and public health initiatives. Nonadherence to medication should not only be considered as a drug-specific behaviour problem, but as a behaviour pattern that is independent of the prescribed medication.
The World Health Organisation identified medication non-adherence as one of the major causes of morbidity, mortality and health care costs
Numerous researchers put effort into examining the rates and predictors of adherence to contribute to the development of adherence boosting interventions
However, adherence to medication is also of interest in a more general fashion, independent of specific populations or side effects of particular drugs. If clinicians and researchers expect patients to show close to full adherence, it is relevant to know how likely the achievement of this goal is. Population based rates can provide an estimate of efforts needed to achieve near complete adherence in patient populations.
Therefore the issue of adherence in general raises a couple of important questions, which can only be answered with appropriate investigations in the general population:
How many people, regardless of suffering from a chronic condition or just having a common cold from time to time, generally follow their doctor's advice?
What are the base rates for adherence that can be expected in the general population?
How certain can a family doctor be that an average patient will show adherent behaviours after leaving their office?
What are the factors generally associated with adherence?
Is nonadherence a drug specific behaviour or rather a behavioural pattern?
And, last but not least: what is “normal adherence” with regards to the general population?
Only a fistful of studies have assessed adherence to medication independently from a specific drug or a particular population. Bardel et al., for example, assessed adherence to “prescribed drugs” in women in central Sweden
The aim of the present study is to describe and categorize general adherence to medication in a representative German sample. The study is of novelty in its field because it focuses on neglected questions in adherence research:
First, to our knowledge, this is the first study representatively assessing adherence to medication in the general population
Secondly, adherence was assessed independently of medication groups as a behavioural disposition
Thirdly, we will identify predictors of nonadherence to medication in a representative sample
The results from a population based study on adherence may open new field of research that focuses on general behavioural patterns associated with non adherence. Based on this research behavioural trainings for patients and their doctors could be developed in order to improve adherence to medication independently from specific drugs.
A representative sample of the German general population was selected with the assistance of a demographic consulting company (USUMA, Berlin, Germany). The population based survey met the ethical guidelines of the international Code of Marketing and Social Research Practice by the International Chamber of Commerce and the European Society for Opinion and Marketing Research. The market research company conducting data sampling is member of a group with general ethical approval from the government to conduct these types of surveys. According to the Federal Data Protection Act (Bundesdateschutzgesetz BDSG, § 30a), the need for consent from a specific ethics comitee is waived for USUMA surveys. A total of 2,512 people agreed to participate and signed written informed consent forms.
The area of Germany was separated into 258 sample areas representing the whole country. After selecting a sample area, households of the respective area and members of these households fulfilling the inclusion criteria were selected using the Kish-selection-grid technique. The Kish-selection-grid technique is aimed at sampling individuals on the doorstep among household residents. The system is devised so that all individuals in a household have an equal chance of selection. The sample was aimed to be representative in terms of age, gender, and education based on data from German Federal Bureau of Statistics on German population from 2009
Adherence to medication was assessed using a four-item self-report scale, the Rief Adherence Index (RAI). The participants were questioned on their general past and present behaviours concerning medication intake, independently of current medication intake. The participants were instructed to consider “all past behaviours concerning any prescribed medication” in order to assess a behaviour pattern that is independent of the prescribed medication.
The RAI consists of 4 items:
I stored or threw away prescribed medication without unwrapping it
I changed the doses of my medication without doctor's authorisation depending on my well-being
I discontinued my medication earlier then the doctor recommended
I discontinued my medication because of mild side-effects
Respondents were instructed to indicate their agreement with each statement on a five-point Likert scale. Item responses were:
1 = (almost) never happened (in 0–20% of cases)
2 = rarely happened (in 20–40% of cases)
3 = often happened (in 40–60% of cases)
4 = happened most of the time (in 60–80%)
5 = (almost) always happened (in 80–100% cases).
The RAI's maximum score is 20, reflecting reports of generally very high non-adherence, the minimum score is 4, reflecting reports of high adherence to prescribed medication with non-adherent behaviours in under 20% of cases.
The use of percentage scores is a novelty and allows an estimation of the health-economic relevance of nonadherece.
Participants were asked whether they were currently taking any medications and if so, the type of medication taken. Items included the ten most common drug classes (anti-diabetic drugs, pain killers, lipid-lowering drugs, antidepressants, antihypertensives, asthma medication, contraceptives, antibiotics, tranquilizers, and sleep medication
The GASE
Age, gender and monthly income were demographic variables of interest for our analyses.
Factor analysis and internal consistency analysis were performed to validate the factor structure and consistency of RAI. Descriptive statistics were carried out to describe the distribution of adherence to medication in the general population. A linear regression analysis was chosen to assess the predictors of adherence to medication. All analyses were carried out with SPSS 17 for Windows ™.
The population based survey met the ethical guidelines of the international Code of Marketing and Social Research Practice by the International Chamber of Commerce and the European Society for Opinion and Marketing Research.
The mean age of the final sample was 49.4 years (SD 18.2), and 55.8% of the total sample was female. 53% had more than 9 years eduacation.
Males (in %) | Females (in %) | |||
Age | Study | German pop. | study | German pop. |
18–30 | 17.2 | 18.1 | 15.7 | 16.2 |
31–40 | 16.0 | 15.8 | 18.3 | 14.4 |
41–50 | 21.2 | 21.5 | 20.0 | 19.4 |
51–60 | 17.8 | 17.0 | 16.9 | 16.1 |
61–70 | 17.3 | 13.7 | 16.6 | 13.7 |
>70 y. | 10.4 | 13.9 | 12.5 | 19.8 |
Cronbach's Alpha for RAI was
RAI was validated by Pearson's r correlations with a German version of the Beliefs about Medicines Questionnaire (BMQ)
There was also positive correlations of RAI with the BMQ General Overuse Scale (
The average severity of current side effects measured by GASE on a four-point Likert scale from 0 (“not present”) to 3 (“severe”) was .43 (SD = .68).
2,452 of 2,512 participants responded to all 4 items of RAI with a mean of 7,5 (SD = 3,35). The distribution of the sum scores of RAI is presented in
According to the distribution of RAI sum scores as displayed in
With respect to the defined cut-off, 66.8% (1638 of 2452 participants) reported to be generally rather adherent to medication, while 33.2% (814 of 2452 participants) reported to be generally rather non-adherent, acting contrary to their doctor's recommendation in 40–100% of cases.
1,283 of 2,452 participants (51.1%) reported to be “currently taking medication”. Age, gender and socioeconomic status (represented by monthly household income), current intake of different medication classes (anti-diabetic drugs, pain killers, lipid-lowering drugs, antidepressants, antihypertensives, asthma medication, contraceptives, antibiotics, tranquilizers, and sleep medication) and general intensity of side-effects were considered as potential predictors of adherence to medication.
A linear regression analysis (
All other variables were excluded through the forward procedure during the regression analysis.
This study is the first to address the neglected subject of adherence to medication in the general population. We found that at least 33% of Germans repeatedly fail to follow their doctor's recommendations and only 25% of Germans describe themselves as fully adherent.
At first sight this result appears familiar – nonadherence rates of 30–40%, depending on the definition and assessment of adherence as well as type of drug and disease, are commonly described
Our results show that these nonadherence rates are not drug specific but resemble a more general behavioural pattern.
The data presented here clearly show that full medication adherence is uncommon. Only 25% of the German population report following or having followed their doctor's advice in over 80% of cases and thus meet the definition of “full adherence”
Thus, our results indicate that high nonadherence to medication is not only a problem of specific population, medication class or costs: an average German person has the same probability of nonadherent behaviours concerning prescribed medication as, for example, an HIV positive drug addict
In addition to rates of medication adherence in the general population, we also investigated predictors of adherence in a subgroup of participants who reported to take medication at the time point of the survey.
We found that experiences with side effects limited medication adherence. Our results indicate that younger males with higher socioeconomical status fail to follow their doctor's recommendations more often than older, less well situated females. Additionally, we showed that patients with long-term medication regiments like antihypertensive treatment were more likely to take their medication as prescribed. Patients with short-term prescriptions such as antibiotics were more likely to ignore doctor's advice or patient information sheets and thereby contributing to the international public health challenge of antimicrobial resistance
These results are in line with the conclusions from another population-wide study: Bardel et al.
Our study has several limitations. A potentially controversial point is the assessment of adherence. Self reports can be susceptible to errors, generally overestimating patient's adherence
A potential other limitation is that we did not differentiate between adherence and persistence as requested by some authors
A further limitation of our study is that we only assessed a selection of possible predictors of adherence. The explained variance in adherence was 12%, similar to what was found in numerous other predictor analyses
The adherence cut-off chosen for this study might be regarded as rather liberal. Some studies define optimal adherence as 100% uptake of prescribed doses and the most liberal cut-offs used in various studies define uptake greater than 80% as “adherence”
Although the sample is representative in comparison to the data from German Federal Bureau of Statistics, one possible limitation is a risk of bias since a part of approached households did not agree to participate.
The major strengths of our study are a representative sample and a novel approach to assess adherence. RAI's items additionally allow quantifying nonadherence percentagewise and thus allowing for an estimation of the economic costs associated with discontinuation of prescribed medications.
These results show that nonadherence is neither a problem of specific patients nor of drug classes. Despite several papers on predictors of adherence there is no study that could explain a sufficient amount of variance