Over-indebtedness and its association with sleep and sleep medication use

Over-indebtedness has been increasing steadily across Europe [1], and beyond [2]. In Germany alone, currently 6.9 million individuals are estimated to be over-indebted, defined as the situation when household income and assets are insufficient to both meet all payment obligations and cover living expenses over a longer period of time [3]. Those affected are facing severe financial stress and are prone to experience social exclusion and stigmatization [4]. While negative consequences of common measures of socioeconomic status (SES), i.e. income, educational attainment and occupation, on health have been well documented [5, 6], over-indebtedness has been largely neglected in health research [7]. Contrary to common belief, however, over-indebtedness is not limited to those in lower socioeconomic positions but may affect individuals at all income and education levels or occupational status [1, 4]. Recent research has revealed that over-indebtedness is a relevant determinant of health independent of conventional socioeconomic measures. Available studies indicate an association between over-indebtedness and mental and physical morbidity [8], such as depression [9], diabetes [10], obesity [11, 12], and back pain [13]. More specifically, a first longitudinal register-based study of 48778 Finnish adults during 1995–2010 has recently shown an association between over-indebtedness and an increased incidence of various chronic diseases [10]. Thus, over-indebtedness has been shown to increase the risk of serious adverse health effects, but its association with sleep problems and associated sleep medication use has not yet been examined.

Recent population studies suggest a steady increase in sleep problems as well as the use of sleep medication in industrialized countries [14‐19]. Common sleep disturbances and sleep disorders such as insomnia encompass problems with initiating sleep, remaining asleep, poor sleep quality or insufficient sleep duration [20, 21]. The prevalence rate of these symptoms amounts to approximately 30% whereas specific sleep disorders affect 5 to 10% of adult populations [22‐26]. In clinical practice, sedative medications such as hypnotics have a long-standing history in the treatment of sleep problems [27‐29]. Between 4 and 10% of adult populations report chronic and current use of sleep medications [29, 30].

Although a growing body of evidence supports an association between physical, psychological as well as socioeconomic parameters and sleep problems and sleep medication use, none of the previous studies has yet considered the potential influence of over-indebtedness.

Sleep is an essential requisite for physical and mental well-being and functioning [21, 56]. Therefore, it is a highly relevant public health issue to address the complexity of socioeconomic disparities in sleep that are assumed to go beyond common measures of socioeconomic status or temporary economic difficulties: Substantial societal cost have been suggested to arise from sleep problems in terms of accidents [57] and occupational injuries [58], absenteeism, productivity [59‐66] as well as health care utilization [67, 68]. Moreover, studies have identified associations between sleep problems and a wide range of adverse health effects: Poor sleepers have been shown to have an increased risk of weight gain and obesity [69, 70], hypertension [71‐73]; hyperlipidaemia [74, 75]; inflammation [76]; diabetes [71, 74, 77, 78], stroke [75, 79], heart attack [75, 80] mortality [54, 81‐83], and reduced quality of life [84].

The aim of this study was to examine the association between over-indebtedness and sleep problems as well as sleep medication use, to contribute to broadening the understanding of underlying mechanisms of sleep disparities. In order to understand the role of over-indebtedness with regards to difficulties with sleep, the prevalence and factors associated with sleep complaints and sleep medication use among over-indebted individuals are assessed, and compared to nationally representative data from the German population.

The present study is based on a cross-sectional survey among over-indebted individuals (OID survey) [85] that was combined with to the first wave of the German Health Interview and Examination Survey for Adults (DEGS1) [86] (see detailed comparison in Fig. 1).

An anonymous health survey using a self-administered written questionnaire was conducted among clients visiting approved debt advisory centres in North Rhine-Westphalia (NRW), Germany, between July and October 2017. Of 145 non-profit debt advisory centres that were invited to act as recruiters by their umbrella organisation, 70 centres agreed to participate. All debt advisory centres were associated with the local German Consumer Organisation or one of the member organisations of the ‘Expert Committee Debt Counselling of Non-statutory Welfare NRW’ (German: Fachausschuss Schuldnerberatung der Freien Wohlfahrtspflege NRW). Counselling services offered by debt advisory centres across Germany are similar. North Rhine-Westphalia is the most populous of the 16 federal states in Germany (17.5 million inhabitants, 2011). Its demographic structure (gender, age distribution, foreigners) is similar to the national average [87]. We chose to invite advisory centres in North Rhine-Westphalia to participate in the study due to the location of our study centre in that federal state which facilitated contact to both the local umbrella organisations and advisory centres.

Eligibility criteria comprised a minimum age of 16 years due to limited contractual capability. Eligible persons were invited to participate in the survey by debt advisors as of the second consultation because the initial counselling interview reflects a sensitive moment required to build trust. All nationalities were considered but only one respondent within each household. Language, reading and writing skills were required due to the data collection method. Debt advisors received both the study material and a comprehensive information letter which illustrated the study objectives, procedures and eligibility criteria in order to standardise the recruitment and survey. The anonymous self-administered questionnaire that was specifically developed for the target group and stamped addressed envelopes were handed to eligible clients by the debt advisors. The health survey focused on the assessment of medication use and self-medication use among over-indebted individuals. Sociodemographic parameters, including age, sex, education, and measures of over-indebtedness, as well as health status, illnesses and utilization of health care services were assessed. Only respondents who reported sex and age were included for analyses. In the OID survey, a total of 1393 individuals were offered the study material by debt advisors. 699 of these returned the questionnaire with complete data on sex and age which reflects a response rate of 50.2%. This study was approved by the ethical committee of the University Medical Faculty in Bonn (No. 167/17).

The data collected among over-indebted individuals was combined with nationally representative health data. As part of the German health-monitoring programme by the national public health agency, the Robert Koch Institute (RKI), data on adults aged 18 to 79 years was collected in DEGS1 between 2008 and 2011. Methodological considerations of DEGS1 have been published depicting recruitment of participants, data collection and data management in great detail [88, 89]. The random DEGS1 sample selected from local population registries comprised 7987 individuals that were available for public use. While the response rate among newly recruited study participants was 42%, the response rate among adults that had already participated in a previous national health survey in 1997–1999 (German National Health Interview and Examination Survey 1998, GNHIES98) was 62% [88]. The DEGS1 study protocol was approved by the Charité-Universitätsmedizin Berlin ethics committee in September 2008 (No. EA2/047/08).

A broadly accepted definition of over-indebtedness is not yet available. However, over-indebtedness commonly refers to a household’s persistent and ongoing difficulties meeting financial commitments that can be measured by using data on arrears, debt settlement, financial burden or consulting debt counselling services [90]. In our study, we classified all clients of debt advisory centres that were eligible to participate in the survey as over-indebted whereas all participants of the DEGS1 survey were classified as non-over-indebted. Due to the lack of information on debts in the DEGS1 survey, just as in other available population-based surveys, this procedure may introduce attenuation in terms of a bias toward the null. We limited analyses to participants with complete data on all sleep variables (OID: n = 538; DEGS1: n = 7447). Thus, all participants of the OID survey and DEGS1 with missing data on sleep-related outcome variables, i.e. problems concerning sleep onset, sleep maintenance and/or sleep medication use (OID survey: n = 161; DEGS1: n = 540) were excluded from analyses. Following the merging of data from the OID survey (n = 538) and the DEGS1 survey (n = 7447), the combined dataset used for analysis comprised 7985 individuals in total.

Both surveys captured the key outcomes frequency of sleep problems and sleep medication use by the same items: Sleep problems were assessed in terms of difficulties with sleep onset and sleep maintenance in the previous four weeks. The frequency of these sleep problems was rated on a 4-point scale (not at all, less than once a week, once or twice a week, three or more times per week). Likewise, self-reported sleep medication use in the past four weeks was assessed. For logistic regression analysis, the outcome variables were dichotomised referring to the experience of problems with sleep onset, sleep maintenance and sleep medication use (not at all, yes) to assess any complaints related to sleep problems and sleep medication use rather than to identify insomnia disorder.

Based on previous studies, sociodemographic variables were included as covariates in logistic regression analyses to control for potential confounding and systematic differences between the two study populations: Besides sex, age that was classified into four age groups (18–29 years, 30–49, 50–64 and 65–79 years) to differentiate phases of life was considered. Educational level according to the International Standard Classification of Education (ISCED) [91] was classified into three categories (low, medium, high). We dichotomized current employment status as “employed” or “unemployed” to control for occupational factors such as work stress, workload and unemployment that might influence sleep. The data on the current employment status of OID and DEGS1 respondents were derived from multiple answers questions. When participants reported any kind of full or part-time employment, we considered these as currently employed. Marital status was classified into three groups a priori: we compared the married that were cohabiting (Reference) with individuals that were divorced, widowed or living separately and singles to account for potential differences in the frequency of regularly sleeping in a shared bed and the effect of socioeconomic advantages of marriage. Moreover, the subjective health status was included as potential confounder and dichotomised a priori into two groups, i.e. “good” to “very good” versus “fair” to “(very) poor”. The presence of a psychological disorder (absent versus present) was also considered in order to control for the potential impact of such conditions on sleep. In DEGS1, available data on psychological disorders comprised the self-reported 12-month prevalence of diagnosed depression and anxiety disorder. In the OID survey data on both self-reported chronic illnesses and medical indication for medication use in the last seven days was used to capture the presence of a psychological disorder according to the assessment in DEGS1. These data were first classified according to ICD-10-GM (German adaptation of the International Statistical Classification of Diseases and Related Health Problems) by medical professionals. Subsequently, illnesses were further classified to indicate presence or absence of a depression (F32, 33) or anxiety disorder (F40, 41).

Descriptive statistics were used to illustrate population characteristics and to examine differences in the distribution of sociodemographic and health characteristics between the OID and DEGS1 samples using chi-squared test. The prevalence of sleep problems and sleep medication use associated with over-indebtedness was calculated. Multiple logistic regression analysis was used to identify factors that predict sleep problems and use of sleep medication (see Additional file 2: Tables S2 for further analysis). Within covariates missing values that were all below a threshold of 5% were assigned to the most frequent category in the combined dataset of the OID (n = 538) and DEGS1 (n = 7447) sample. As a sensitivity analysis, we conducted complete case analysis to validate the approach to handle missing data (Additional file 1: Table S1). All independent variables were entered into the model simultaneously. The reference group was defined as the most frequent category, except for the reference category of sex (male) and age (youngest age group) to simplify interpretation. The level of statistical significance was set at 0.05 for all analyses. Analyses were carried out using IBM SPSS statistics (version 25).

Altogether 7985 persons, aged 18 to 79 years, were included in the combined dataset of the OID (n = 538) and DEGS1 (n = 7447) samples. Female and male participants were represented in both samples in nearly equal shares (Table 1).

The sample of over-indebted individuals and the DEGS1 sample differed with regards to a number of sociodemographic variables. The over-indebted individuals were significantly younger (70.8% 18–49 years) than the nationally representative sample (46.9% 18–49 years). Educational attainment in the over-indebted was significantly lower whereas the majority of participants in both samples was employed. In contrast to the DEGS1 sample, the majority of over-indebted individuals was not married but single, divorced or widowed (OID survey: 71.4%; DEGS1: 33.9%). Over-indebted individuals had a significantly poorer subjective health status than the general population. Accordingly, the over-indebted population had a significantly higher prevalence of psychological disorders in the form of depression and/or anxiety (16.2%) compared to the general population (6.3%).

In the present study, an increased risk of problems with sleep onset and sleep maintenance as well as sleep medication use was observed for over-indebted individuals compared to the general population.

In view of the increasing trend of over-indebtedness of individuals across high-income countries and adverse health effects of inadequate sleep, the study results highlight over-indebtedness as a public health concern. Until today, this is the first study that considers over-indebtedness as a determinant of sleep problems and sleep medication use in health research at the global level. The results suggest that conventional measures of socioeconomic status are insufficient to describe the complexity of financial hardship with regards to its association with health, and sleep specifically.

The present study reveals a strong association between over-indebtedness and sleep problems in terms of sleep onset and maintenance as well as sleep medication use. These associations were independent of standard socioeconomic measures and were not accounted for by other sociodemographic or health factors. The results contribute to the increasing evidence on the vital role of socioeconomic factors with regards to sleep problems and sleep medication use. As sleep is essential to good health and functioning, the understanding of risk factors of problems related to sleep is crucial to improve health and reduce health disparities. Considering over-indebtedness as a risk factor for sleep problems helps to provide a basis for early intervention in clinical practice that needs to address both unmanageable debt burden and sleep problems. Specifically, when health-related causes are not detectable, referring over-indebted primary care patients that seek medical consultation for sleep problems to social support and debt counselling services can contribute to alleviating sleep disturbances. Thus, awareness of the association between over-indebtedness and sleep problems and sleep medication use in epidemiologic research and health care can help to avert detrimental health effects and societal cost of sleep problems.