Study design and population

The COCO study is a cross-sectional multicenter questionnaire survey designed and conducted by the Working Group on Self-Care for Common Colds of the European General Practice Research Network (EGPRN). The Ethics Committee of the Faculty of Medicine at the University of Duisburg-Essen, Germany, provided the first ethical approval (13-5495-BO) which was subsequently used by all study site coordinators to obtain additional ethical approval according to regional requirements. Participants were informed about their anonymous and voluntary participation. Completion of the questionnaire was regarded as an indication of participants’ consent to the study. Details of the rationale of the COCO study design have been published elsewhere [11]. Briefly, in order to ensure a random sample, 120 consecutive patients aged ≥18 years visiting their general practitioner for any reason in fall/winter 2013/14 were asked to participate. Patients were asked to answer the following question: ‘When you had a common cold during the last year: what did you do?’ A total of 3,074 questionnaires were returned from 27 sites in 14 European countries. After excluding data from sites with different sampling strategies (physician interviews, older questionnaire version, incomplete questionnaire distribution) the study population available for analysis comprised 2,724 patients.

Study instrument

The patient questionnaire was developed based on a survey among 10 primary care physicians from seven European countries (Austria, Bosnia and Herzegovina, Germany, Israel, Italy, Macedonia, and Poland) and one EU-associated country (Turkey). Physicians were asked to list typical self-care items used by their patients for common colds [11]. The final patient questionnaire consisted of 27 questions which included 94 self-care practices plus free text options in 11 categories: over-the-counter medication (11 items), specific foods or drinks (11 items), herbal teas (18 items), alcoholic drinks (3 items), self-prepared special recipes (7 items), pastilles or drops (10 items), something for the nose (4 items), inhalations (8 items), gargle or sprays for the throat (4 items), items applied externally (5 items), and extras at home (13 items). Additionally, various patient characteristics were requested, e.g. age, gender, health insurance status, number of school years, number of pills taken daily, and frequent chronic conditions (depression, chronic kidney disease, chronic pain/arthritis/osteoarthritis, asthma/chronic bronchitis, diabetes, heart disease, high blood pressure). Subjective discomfort was measured by asking whether the patient ‘felt very poorly during the last common cold’ (answer options: yes, no, don’t know). Knowledge on the self-limited nature of the disease was assessed by asking whether ‘a common cold goes away by itself’ (answer options: yes, no, don’t know).

Classification of items

Two researchers classified all single self-care items by their mode of application: ‘intestinal absorption’, ‘intranasal application’, ‘local oral effects’, ‘inhalation’, ‘topical use in throat’, ‘external use’, ‘foodstuffs’, and ‘extras at home’. All pharmacological items were further sub-classified using the international ATC (WHO Anatomical Therapeutic Chemical) classification system for pharmacological substances, which divides active substances into groups according to the organ or system on which they act as well as their therapeutic, pharmacological and chemical properties [12]. Items not listed were assorted in plausible groups and sub-categories based on content (e.g., ‘teas’, ‘fruit’, ‘vegetables’, ‘lozenges’ or ‘alcohol’): in this sub-classification, items were listed individually if they had a utilization rate of at least 1% (n ≥ 27) of the sample; otherwise they were summarized as ‘other’ within each mode. All items with an assigned ATC code and the unspecific answer options ‘antibiotics’ and ‘pain medication’ were called ‘pharmaceutical products’, all other items were called ‘non-pharmaceutical products’. Disagreement between researchers was solved by discussion until a consensus was reached.

Statistical analysis

Descriptive analyses were performed for the total study population and stratified by country. The response rate was calculated per site and country. For this analysis, percentages reported for ‘knowledge on the self-limited course’ (yes/no) and ‘subjective discomfort’ (yes/no) were stratified by country and site.

Four groups were stratified based on ‘knowledge’ yes/no and ‘discomfort’ yes/no (note: ‘yes’ indicated as ‘+’ and ‘no’ as ‘-’): group 1: +knowledge/-discomfort; group 2: -knowledge/-discomfort; group 3: +knowledge/+discomfort; group 4: -knowledge/+discomfort. We hypothesized that groups 1 and 4 would show the largest contrast, with group 1 having the most beneficial characteristics in terms of knowledge and discomfort. Groups 2 and 3 were considered intermediate groups. Analyses for the four groups included: a) the mean total item use and standard deviation; b) the mean item use for the categories non-pharmaceutical/pharmaceutical products and self-care with/without proof of evidence; c) descriptions of the modes of application calculated as the percentage of patients using at least one item per mode, and the mean item use for each mode; d) depending on the properties of the scales, either χ²-tests for categorical data or ANOVA and non-parametric tests (Kruskall-Wallis) for continuous data were used to compare groups.

In order to identify factors associated with using more self-care items than 50% of the respective country sample, multivariate logistic regressions were performed. Independent determinants were discomfort (yes), knowledge (no), gender (female), age (48), years of education (above country-specific median), chronic condition(s) (yes), daily pill intake (yes), and smoking status (yes). It was tested for collinearity (Pearson’s r), and determinants significant in bivariate analyses were included in the multivariate regression model. To prevent the multiple inclusion of the chronic disease state in the multivariate regression analysis, only chronic pain/arthritis/osteoarthritis was included. This condition was chosen because these patients typically have analgesics available for regular or occasional use. Their use of self-care including analgesics is likely to be higher than that of participants without access to this medication. The nominal statistical significance level was p<0.05. To correct for multiple testing, the Benjamini-Hochberg procedure was applied [13]. All statistical analyses were performed using IBM® SPSS® Statistics, version 20.

Characteristics of the participating countries, sites and patients

A total of 2,724 patients were available for data analysis. The response rate differed between countries, ranging from 53.3% to 89.7%. These numbers are not precise estimates as some sites diverted from the sampling design (the response rate for Macedonia could not be calculated). See S1 Table for the number of participants included per site.

222 patients with missing data for knowledge and/or discomfort and 298 patients who answered ‘don’t know’ for knowledge and/or discomfort were excluded. The final analysis included 2,204 patients (mean age of 46.5±16.41 years, 62.5% female, 12.8±4.46 years of education). Excluded patients did not differ significantly from the final sample with regard to age, gender, years of school, daily pill intake, and chronic conditions (depression, chronic kidney disease, chronic pain/arthritis/osteoarthritis, asthma/chronic bronchitis, diabetes, heart disease, high blood pressure). See Table 1 for baseline characteristics; site characteristics have been published elsewhere (see [10]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the COCO study sites and participants (n = 2,204).

https://doi.org/10.1371/journal.pone.0195564.t001

Knowledge of the self-limited course and subjective discomfort during a common cold

One third (36.3%) of the participants did not know about the disease’s self-limited course, with a variation of 6.5% to 60.5% between countries and sites from the same countries. More than two thirds of the participants (70.6%) reported high discomfort during their last common cold, with a range of 40.3% to 90.3% between countries and sites. See additional file 1 in the Supplementary Material for characteristics per data sampling site/primary care practice.

3.3. Self-Care Stratified by Knowledge and Discomfort (n = 2,204)

Stratified by knowledge/discomfort, the size of the four groups was as follows: group 1 (no discomfort/has knowledge) consisted of 23.1% of the participants, group 2 (no knowledge, no discomfort) of 6.3%, group 3 (has discomfort/has knowledge) of 40.6%, and group 4 (has discomfort/no knowledge) consisted of 30.0%.

The mean item use increased with increasing group number, ranging from 8.8 to 13.3. In general, patients with discomfort (groups 3 and 4) used more self-care measures than those in groups 1 and 2. The largest differences for self-care were identified when comparing groups 1 (no discomfort/has knowledge) and 4 (has discomfort/no knowledge). For five of eight application modes, patients in group 4 (has discomfort/no knowledge) reported a higher use than those in group 3 (has discomfort/has knowledge). Across the four groups, patients used about 9.1 more non-pharmaceutical than pharmaceutical products. Interestingly, group 3 used the highest number of pharmaceuticals. Patients in group 4 (has discomfort/no knowledge) used 53% more non-pharmaceutical products than those in group 1 (no discomfort/has knowledge). See Table 2 for details.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Self-care behavior stratified by four patient groups based on the presence and/or absence of knowledge about the self-limited disease course and subjective discomfort during the last common cold (n = 2,204).

https://doi.org/10.1371/journal.pone.0195564.t002

Factors associated with higher use of self-care practices

Bivariate analyses showed that a higher number of self-care items used was associated with lack of knowledge (33.2% vs. 40.1%, p = 0.001), having discomfort (64.4% vs. 78.3%, p<0.001), female gender (56.7% vs. 69.8%, p<0.001), younger age (51.4% vs. 58.4%, p = 0.001), more years of education (39.6% vs. 45.9%, p = 0.003), chronic pain/arthritis/osteoarthritis (8.4% vs. 11.4%, p = 0.017), high blood pressure (23.8% vs. 20.1%, p = 0.041; n.s. after correcting for multiple testing), diabetes (8.7% vs. 6.4%; p = 0.044 n.s. after correcting for multiple testing), and depression (4.2% vs. 6.6%, p = 0.013). Only chronic pain/arthritis/osteoarthritis was included. The multivariate regression analysis (Fig 1) showed that high self-care use was most strongly associated with having discomfort (OR 1.8, CI 1.5–2.2), followed by female gender (OR 1.7, CI 1.4–2.0), chronic pain/arthritis/osteoarthritis (OR 1.6, CI 1.2–2.1), more years of education (OR 1.3, CI 1.1–1.6), younger age (below 48 years) (OR 1.3, CI 1.0–1.5), and lack of knowledge (OR 1.2, CI 1.0–1.5) (Nagelkerke R2 = 0.66). A sub-analysis excluding patients with chronic conditions yielded comparable estimates. Regression analyses showed similar estimates for discomfort (OR 2.0, CI 1.7–2.4) and knowledge (OR 1.4, CI 1.1–1.6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Logistic regression analysis for high self-care use (n = 2,084).

https://doi.org/10.1371/journal.pone.0195564.g001

Comparison with other studies

The literature suggests that self-care behavior is illness-specific, e.g. differing between self-care for upset stomach and bowel irregularity [14]. Therefore, caution is needed when comparing our results to self-care for other illnesses. Surprisingly few studies are available for comparison, although common colds are the most frequent disease and are associated with tremendous socio-economic costs. Although several studies report that symptoms of common colds are the most frequent reason for self-care [14–17], the illness-specific literature on self-care targets predominantly chronic rather than acute, self-limited conditions.

Our approach of distinguishing four groups based on the characteristics discomfort and knowledge is a novelty in the field of self-care for acute diseases. However, illness perception, namely patients’ subjective beliefs and emotional responses to their illnesses, are known to influence patients’ coping, self-management, and thereby outcomes for chronic diseases, e.g. in asthma [18]. Our finding that subjective discomfort influences self-care corresponds to a nationwide survey on self-medication in the German population: the key motive for self-medication (88.3% of the respondents) is patients’ perception of the severity and the assessment that minor health conditions do not require immediate physician consultation [16]. Similarly, a recent Japanese study showed that the health-related quality of life was significantly higher among patients with colds relying on self-medication first compared to patients who visited a physician immediately [19], although further research is needed to better understand the cause-effect relationship.

We identified no study comparable to COCO that assessed knowledge of the self-limited course. However, other studies addressing knowledge and/or misconceptions about the effectiveness of various measures used for colds are well described in the British general population, and showed that up to 35% of participants had misconceptions about the effectiveness of garlic, chicken soup and antibiotics, and 53.2% did not know that viruses caused colds [20]. Addressing reasons for physician consultations, the study showed that neither knowledge about the effectiveness of various measures nor age or education were significant, while there was a positive association for a physician consultation-prone attitude (OR 3.6, p<0.01) and a poorer self-rated health status (OR 1.5, p<0.05).The level of discomfort, however, was not addressed.

Our finding that the current health status (here: the presence of chronic pain and/or arthritis and/or osteoarthritis) is associated with higher self-care use is in agreement with a Spanish National Health Survey with 20,311 participants of the general population, which showed that the prevalence of self-medication was greater among those with a chronic illness (PR 1.2, CI 1.1–1.3) [21]. A Canadian study on self-medication specifically for upset stomach also reported a higher use in participants with a generally poor health status [14]. However, the association between health status and self-care was not consistent across studies. For example, the same Canadian study failed to show a significant association between health status and self-medication of bowel irregularity [14], as did a longitudinal study in Australians (age 65+) separately for the use of self-medication and complementary alternative medicine in general [22]. Neither of the studies addressed discomfort and knowledge, which might explain these heterogeneous results.

Similarly, our result that females, patients with more years of education, and those of younger age exhibit a higher use of self-care for various conditions is not consistently supported across studies. Prior studies addressing self-care in general reported that females use more self-medication than males in Spain (PR 1.21, 95% CI 1.10–1.33) [21], use more home remedies in Canada [14], use a larger variety of home remedies in Germany [15], and are more regular users of over-the-counter (non-prescription) analgesics in Northern Ireland (48% vs. 40.5%, p<0.01) [23]. However, no association between gender and over-the-counter use was found in a random sample of the German general population [16] and in elderly Australians [22]. Our association between younger age and higher self-care use is in line with an Australian and a Japanese survey [19, 24]. The latter considers a better physical condition and less fear of a more serious illness as explanations of why younger patients are more likely to rely on self-care before consulting a physician [19]. In accordance with our finding, the Spanish national survey [21] identified an association of university-level education with the use of self-care (PR 1.25, 95% CI 1.03–1.53). In contrast, Canadians with less formal education were more likely to use self-medication for an upset stomach and bowel irregularity [14]. To our surprise, years of education and knowledge of the self-limited course of common colds were not correlated in our study. And, even more interesting, those with better knowledge and discomfort were more likely to use pharmaceuticals than those who did not know about the disease’s natural course, although non-pharmaceuticals were used about six to ten times more frequently than pharmaceuticals in all four groups.

Strengths and limitations of the study

This is the first study to analyze self-care for common colds on a European scale. The strength of our study is that it differentiates the influence of discomfort and knowledge on self-care for common colds using four patient groups. However, knowledge was assessed with only one question and necessitates further research. As this study was not designed to provide country-representative results, we abstained from country-specific comparisons. In addition, participants were not asked for their intention to use certain practices and their appraisal of the evidence for various self-care practices. Also, a recall bias may apply as participants were asked about their self-care behavior during their last common cold irrespective of the reason for their practice visit.