Do Everyday Problems of People With Chronic Illness Interfere With Their Disease Management?

Lieke van Houtum; Mieke Rijken; Peter Groenewegen


BMC Public Health. 2015;15(1000) 

In This Article


Study Sample

The sample of the present study consisted of members of the National Panel of people with Chronic illness or Disability (NPCD), a nationwide prospective panel-study on the consequences of chronic illness in the Netherlands.[16] For this study, we only included the participants with chronic illness. Participants with chronic illness were recruited from more than a hundred general practices (random samples of general practices drawn from the Dutch registration of General Practices[17]). These panel members were selected according to the following criteria: diagnosed with a somatic chronic disease by a certified medical practitioner, aged ≥15, not permanently institutionalised, aware of the diagnosis, not terminally ill (life expectancy >6 months according to the general practitioner), mentally capable to participate, and sufficiently proficient in Dutch. Potential panel members received an information letter about the panel and were asked to fill in a reply form whether or not they want to join the panel. If they were interested, they received a questionnaire on their demographic characteristics. When that questionnaire was returned, they were considered members of the panel. Annually, 500 new panel members were selected via the standardised procedure to replace panel members who withdrew or who had participated for the maximum term of 4 years. NPCD is registered with the Dutch Data Protection Authority; all data were collected and handled in accordance with the privacy protection guidelines of the Authority.

At inclusion, NPCD participants received a questionnaire on their socio-demographic characteristics. In addition, general practitioners (GP) provided (with patients' permission) medical information about the panel members (chronic diseases diagnosed, dates of diagnoses, health status etc.). In April 2013, a questionnaire about self-management, everyday life problems and perceived general health was sent to the panel members. A total of 1731 patients diagnosed with a chronic disease completed this questionnaire (response = 80 %).


Self-management. We used the Dutch version of the Partners in Health Scale (PIH-Dutch) to measure patients' self-management knowledge and behaviour. This scale was originally developed as part of the 'Flinders Program of Chronic Care Self-Management.[18,19] The PIH-Dutch scale consists of 12 items, which were answered on a scale ranging from 0 'low self-management' to 8 'high self-management'. Examples of items are: 'I have the ability to take action when my symptoms get worse', 'I have the ability to arrange appointments as recommended by my healthcare provider' and 'I have the ability to manage the impact of the illness on my social life'. As the first answering options of the original scale were all at a very close distance from each other resulting in a distribution very skewed to the right, we recoded the lower scores (0–3 = 0, 4–5 = 1, 6 = 2, 7 = 3 and 8 = 4). Four components of self-management were distinguished, namely knowledge (two items; knowledge about illness and treatment), recognition and management of symptoms (two items; monitor symptoms and act when symptoms worsen), active involvement in treatment (four items; taking medications as prescribed, attend appointments, shared decision-making) and coping with consequences (four items; dealing with effects on physical, emotional and social wellbeing and progress towards healthy life). As we focused in this study solely on self-management behaviour, we did not include the knowledge scale. Scale scores were computed by dividing the sum of respondents' item scores by the number of items filled in, and range from 0 to 4, with higher scores indicating better self-management.

Everyday Problems. To assess everyday problems we used the biographical list of problems (BIOPRO), developed by Hosman.[20] In this questionnaire, respondents were asked to indicate whether or not they have recently (no specific time frame given) encountered any of the following problems: financial, housing, employment, with partner, with children, with other people, sexual, with leisure time. Based on an exploratory factor analysis using principal component extraction with varimax rotation, we distinguished two types of problems, namely problems related to basic needs (financial, housing, employment) and problems related to social needs (partner, children, other people, sexual, leisure time). Based on this distinction, we constructed two dichotomous variables: basic problems and social problems (both scored into 0 'having no problems' and 1 'having problems').

Socio-demographic, Illness and Health Status Characteristics. In our study, we included the following socio-demographic characteristics of the participants: age, gender and highest level of education, classified as low (primary education, lower secondary and lower vocational education), intermediate (intermediate secondary and intermediate vocational education) and high (higher vocational education and university). We included these socio-demographic characteristics as we expected that these characteristics would have an effect on having everyday problems as well as self-management behaviour.

In addition, we used data provided by their GPs: type of chronic disease(s) diagnosed (coded by means of the International Classification of Primary Care[21]) and presence of more than one chronic disease (multimorbidity). Patients' self-rated general health was measured by the general health scale of the RAND-36 Short Health Status Survey, ranging from 1 'poor health' to 100 'excellent health'.[22] Finally, the severity of physical limitations was assessed by the SCP physical disability indicator,[23] a self-report questionnaire distinguishing four levels: none, mild, moderate and severe. This indicator assessed people's ability to perform different tasks and activities, such as the ability to walk for short period of time, walk for a longer period of time, do odd jobs around the house, read the newspaper, hear what is being said during conversations, etc.


Descriptive analyses were performed to provide information about the characteristics of the study sample and to describe the everyday problems people with chronic illness encounter. To assess whether having basic or social everyday problems was related to the socio-demographic, illness and health status characteristics of people with chronic illness, we performed two multivariate logistic regression analyses (one with basic problems as dependent variable and the other with social problems as the dependent variable).

Next, we assessed the relationship between having basic or social problems and level of self-management. The problem in assessing this relationship is that having basic or social problems is not exogenous. Personal and health characteristics are related to self-management and to having problems. This makes it difficult to estimate the relationship of having problems and self-management. In other words, there are confounding variables that might influence both the outcomes (in our study, level of self-management of people with a chronic illness) and comparison groups (people having everyday problems versus those not having everyday problems). We used propensity score matching (PSM) to solve this problem as much as possible. The propensity score is a balancing score: conditional on the propensity score, the distribution of observed covariates will be similar between chronically ill people with and without everyday problems.[24,25] Models were adjusted for age, sex, education, comorbidity, perceived general health and physical limitations and we inspected the diagnostics for propensity score analysis (checking for balance in the covariates). PSM is one way of approaching the problem and has its own assumptions that are not perfectly met in our study. Our assumption was that given similar background characteristics having everyday problems (the 'treatment') or not (the 'controls') could be considered as randomly assigned. Therefore, we also conducted multivariate linear regression analyses as a sensitivity analysis of our findings in the propensity score matching. In addition, as we could not include an interaction effect in the PSM analysis between having basic and social problems on the level of self-management, we also conducted a multivariate linear regression analyses in which we included this interaction effect as well.

The panel members were originally selected from general practices, resulting in a hierarchical data structure. Since intra-class correlations showed hardly any clustering of self-management behaviour within general practices (mean 0.01), and the likelihood ratio test did not show that multilevel analyses had an advantage over ordinary regression analyses, single-level regression analyses were conducted. All analyses were performed using Stata 13.0.