Design and study sample
Baseline data from the European Project on OSteoArthritis (EPOSA) were used. The EPOSA study focuses on the personal and societal burden of OA and its determinants in older persons. A detailed description of the study design and data collection of the EPOSA study is described elsewhere [13]. In summary, random samples were taken from existing population-based cohorts in five European countries (Germany, the Netherlands, Spain, Sweden and the United Kingdom (UK)). In Italy, a new sample was drawn. A total of 2942 respondents (response rate, ranging from 64.6% to 82.2%, averaging 72.8%) were included. The age-range was between 65-85 years in most countries except for the UK, which had an age-range of 71-80 years. All participants were interviewed by a trained researcher at home or in a clinical center, using a standardized questionnaire and a clinical exam. The interview lasted about one and a half hours. All participants completed an informed consent. For all six countries, the study design and procedures were approved by the Medical Ethics committee of the respective centers (Germany: Ethical Committee of Ulm University; the Netherlands: Medical Ethical Committee of the VU University Medical Center; Spain: Ethic Committee for Clinical Research of University Hospital La Paz of Madrid; Sweden: Ethics Board of Karolinska Institutet; UK: The Hertfordshire Research Ethics Committee; Italy: Comitatio etico ULSS7).
In the EPOSA study, clinical classification criteria, developed by the American College of Rheumatology (ACR) [14], were used to determine clinical OA. The ACR criteria for any clinical knee, hip or hand OA was satisfied in 889 participants (31.7%). Of these participants, 727 persons completed all 14 days of the pain calendar. Data on self-perceived weather sensitivity was available for 712 subjects. These participants were included in the final study sample of the current study. The excluded participants with clinical OA (n = 177) were older, lower educated and more depressed than the included subjects. In addition, they had a lower sense of mastery and used less (additional) pain medication than the included participants. The two groups did not differ in sex, partner status, anxiety, body mass index (BMI), number of chronic diseases and outdoor physical activity.
Measures
Dependent variable
Self-reported joint pain
Joint pain was assessed prospectively with a two-week pain calendar. After the baseline-interview, participants were asked to complete this pain calendar. Per day respondents indicated how much joint pain they experienced on a 11-point rating scale from 0 to 10 with 0 representing no pain and 10 representing the greatest pain intensity. For each respondent, the average self-reported joint pain in the pain calendar period was calculated as the sum of all noted pain intensity levels divided by 14.
Independent variable
Self-perceived weather sensitivity
To assess self-perceived weather sensitivity, participants were asked which specific weather condition(s) affects their joint pain. There were four response categories: my joint pain is affected by (1) damp/rainy weather, (2) cold weather, (3) hot weather, and (4) my joint pain is not affected by one of these weather conditions. Participants were allowed to indicate more than one answer. Participants were considered as weather-sensitive persons when, in their opinion, damp/rainy, cold and/or hot weather affected their joint pain. Subjects who noted that their joint pain is not affected by one of these weather conditions were considered as non-weather-sensitive persons.
Potential confounders
Socio-demographic variables
Prior studies revealed that socio-demographic factors are associated with pain intensity in people with OA [12]. Socio-demographic information was obtained on participants’ age, sex, partner status and education level. Partner status referred to whether participants have a partner at the moment (yes/no). Education was measured by the highest level of education completed (elementary school not completed, elementary school completed, vocational education/general secondary education, and college or university education) and dichotomised into “better educated than secondary education” (yes/no).
Pain medication use
Pain medication use (yes/no) referred to the use of analgesics (ATC N02 subgroup) and/or anti-inflammatory products (ATC M01 subgroup). In addition, participants were asked whether they used additional pain medication on the day of pain report because of joint pain. For each participant, the total number of days on which they used additional pain medication was calculated.
Emotional distress: anxiety and depression
Emotional distress, such as anxiety and depression, is associated with more pain in people with OA [11, 12]. Anxiety and depressive symptoms were examined by the Hospital Anxiety Depression Scales (HADS) [15]. HADS is a self-report questionnaire comprising 14 four-point Likert scaled items, 7 for anxiety (HADS-A) and 7 for depression (HADS-D). Both scales have a range from 0 to 21. A higher score on the HADS-A and HADS-D indicates greater anxiety and depression respectively.
Mastery
Mastery is the extent to which individuals consider themselves to be in control of events and ongoing situations [16]. Mastery is considered as a psychological resource when coping with stressful life events. A high sense of mastery reduces psychological distress and therefore it may affect pain perception in people with OA.
Mastery was measured by means of an abbreviated 6-item version of the Pearlin Mastery Scale [16]. The questionnaire consists of six statements such as “I can do almost everything, if I want to”. Response categories range from 1 = strongly disagree to 5 = strongly agree. The summed items range from 6 to 30, but for ease of interpretation 6 is subtracted, so the final scale ranges from 0 to 24, with higher scores indicating more mastery.
Outdoor physical activity
It has been shown that physical activity is beneficial for reducing pain in people with OA [17]. Physical activity was measured using the LASA Physical Activity Questionnaire (LAPAQ), an instrument validated against diaries and pedometer measurements in older persons [18]. Frequency and duration of activities over the past two weeks were asked for walking, cycling, gardening, light and heavy household work and a maximum of two sports. In order to calculate the daily outdoor physical activity, the frequency and duration of walking, cycling and gardening were multiplied and divided by 14 days. A total outdoor activity score was calculated in minutes per day.
Body mass index
Body mass index (BMI) affects pain in OA-patients. Pain increases with patients’ weight [19]. BMI was calculated as weight in kilograms divided by height in squared meters. Weight was measured to the nearest 0.1 kg using a calibrated scale. Height was measured to the nearest 0.001 m using a stadiometer.
Number of chronic diseases
It has been shown that number of comorbid conditions, including chronic diseases, influences pain in OA-patients [12]. Number of chronic conditions was measured through self-reported presence of the following chronic diseases or symptoms that lasted for at least three months or diseases for which the participant had been treated or followed by a physician: chronic non-specific lung disease, cardiovascular diseases, peripheral artery diseases, stroke, diabetes, cancer, and osteoporosis. If participants answered “yes” then they were asked to specify which diseases or type. Chronic conditions were evaluated as the number of diseases and multimorbidity was defined as the occurrence of 2 or more coexisting conditions.
Local climate
Local climate of the residences of the participants in the six population-based cohort studies were classified by the Köppen-Geiger climate classification system. The Köppen-Geiger climate classification system is applied in various disciplines and is the most frequently used climate classification system in the world [20]. Based on criteria about vegetation, annual and monthly precipitation and temperature, this classification system distinguishes thirty possible climate types [21]. In the current study, three different climate types were classified. The residence locations of the participants in Germany, Italy, the Netherlands and the UK are characterized by a temperate warm climate without dry seasons and a warm summer (relatively warm and wet climate). The residence location in Spain is characterized by a temperate warm climate with a dry and hot summer (relatively warm and dry climate). The Swedish residence locations represent a cold climate without dry seasons and a warm summer (relatively cold and wet climate).
Seasonal weather patterns
Seasonal weather patterns affect pain perception in weather-sensitive people. Additionally, weather patterns may influence mood in certain individuals and thereby indirectly affect pain perception [9, 10]. The season (spring, summer, autumn or winter) in which the pain calendar is completed by the participant may have an effect on pain perception in older people with clinical OA. Information was obtained concerning the astronomical season in which participants completed their pain calendar.
Statistical analyses
Differences in characteristics between weather-sensitive and non-weather-sensitive participants were examined with independent sample t-tests for continuous data and chi-square tests for categorical data. Differences between weather-sensitive and non-weather-sensitive persons were tested with a Mann-Whitney U test for skewed continuous variables. Descriptive analyses were used to examine the percentages of weather-sensitive persons who reported to be sensitive to a particular weather condition or a combination of specific weather conditions.
To examine differences in self-reported joint pain between weather-sensitive and non-weather-sensitive people with clinical OA, an independent sample t-test was performed. Self-perceived weather sensitivity and self-reported joint pain were used as independent and dependent variable respectively. Linear regression analyses were performed to correct for socio-demographic characteristics (sex, age, partner status, education and country) and other potential confounders (anxiety, depression, mastery, outdoor physical activity, medication use, BMI, number of chronic diseases, seasonal weather patterns and local climate).
Logistic regression analyses were performed to determine those variables that best predicted self-perceived weather sensitivity. First, each variable was examined for significantly predicting self-perceived weather sensitivity. Subsequently, all variables with a p-value below 0.20 were included in a multivariable model. Level of significance was α = 5.0%. Statistical analyses were performed in IBM SPSS Statistics (version 20.0).