key: cord-0872845-qt5hrun1
authors: Madsen, Kristoffer Panduro; Willaing, Ingrid; Rod, Naja Hulvej; Varga, Tibor V.; Joensen, Lene Eide
title: Psychosocial health in people with diabetes during the first three months of the COVID-19 pandemic in Denmark
date: 2021-01-20
journal: J Diabetes Complications
DOI: 10.1016/j.jdiacomp.2021.107858
sha: 20c5aca13be6eef9e30d53a28d577399f0c76ad3
doc_id: 872845
cord_uid: qt5hrun1

Aims To analyze trajectories of psychosocial health during the first three months of lockdowns and reopenings of the COVID-19 pandemic in Denmark. Methods An online longitudinal survey of 2,430 people with diabetes consisting of six questionnaire waves (Q1-Q6) was conducted between March 19 and June 25, 2020. Psychosocial outcomes assessed were COVID-19 worries, quality of life, feelings of social isolation, psychological distress, diabetes distress, anxiety, and general and diabetes-specific loneliness. Trajectories in psychosocial health were analyzed with linear multilevel mixed-effects models. Subgroup analyses were conducted. Results In total, 1,366 (56%) people with diabetes responded to the first questionnaire. COVID-19 worries, feelings of social isolation, psychological distress, anxiety and general loneliness had all improved at Q6 compared to Q1 (p < 0.001). In general, improvements in psychosocial health started after the first reopening phase (April 15); however, general loneliness increased up to the first reopening phase (p = < 0.001) before decreasing, and quality of life decreased up to the first reopening phase (p = 0.002), with no improvements to follow. Subgroup analyses revealed that women had larger decreases in feelings of social isolation (p < 0.001) and in psychological distress (p = 0.035) and increases in quality of life (p < 0.001), between Q1 and Q6, compared to men. Conclusions Psychosocial health in people with diabetes improved following reopening of society. However, increases in loneliness and decreases in quality of life during lockdown indicates a potential need to mitigate the acute effects of such policies.

The coronavirus disease 2019 (COVID- 19) was declared a global pandemic by the World Health Organization on March 11, 2020. 1 Since then, the pandemic has taken its toll on societies across the world, including economies, social relations, and physical and mental well-being. Prior to the pandemic"s emergence in the Western world, a study of 1,210 respondents documented the psychological impact of the COVID-19 pandemic in China. The study found that 54% of respondents rated the psychological impact of the COVID-19 outbreak as moderate or severe. 2 Furthermore, people with a chronic illness, such as diabetes, experienced higher levels of stress, anxiety and depression compared to people with no chronic illness. 2 Early studies of general populations in the Western world also documented psychological impacts of the pandemic. For example, a study in the general population of the UK found that the prevalence of clinically relevant levels of mental distress increased from 19% to 27% between 2019 and April, 2020 (one month into the lockdown). 3 Similarly, a study of the general population from the USA found that 36% of Americans reported that the COVID-19 pandemic seriously affected their mental health, while 48% were anxious about the risk of infection. 4 COVID-19 is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and presents, in most people, with symptoms like a common cold. 5 However, certain groups have an increased risk of experiencing more severe symptoms. As such, people with diabetes have been identified as a high-risk group with a two-fold higher mortality rate and a three-fold higher risk of a more severe COVID-19 infection when compared to people without diabetes. 6, 7 COVID-19-related mortality rates in people with diabetes are highest for those who are obese or underweight, and those who suffer from cardiovascular and/or renal complications of diabetes. 8- J o u r n a l P r e -p r o o f Journal Pre-proof 10 Whereas some studies suggest high HbA1c as an important risk factor for COVID-19 related mortality 8 , other studies dispute this. 11 To date, the role of COVID-19 in the interplay between diabetes and psychosocial health has mostly been studied in cross-sectional studies. [12] [13] [14] [15] [16] We found in a preceding study that having diabetes-specific worries about the COVID-19 pandemic was associated with poorer psychosocial health with regards to feelings of isolation, loneliness and diabetes distress. 12 Another study found that people with diabetes were more worried about being infected than people without diabetes. 15 Thus, it is likely that the pandemic exacerbated psychosocial challenges that are often experienced under "normal" circumstances by people with diabetes.

Research conducted among people with diabetes prior to the COVID-19 pandemic shows that women compared to men, younger compared to older people, people with type 1 compared to type 2 diabetes, and people with a history of mental illness compared to people without, experience higher levels of diabetes distress and lower levels of mental well-being. [17] [18] [19] [20] [21] [22] Social ties also impact psychosocial health in people with diabetes. For instance, people living alone have lower psychological well-being compared to people living with a partner. 23 The aim of this paper was to analyze trajectories of general and diabetes-specific psychosocial health during the first three months of lockdowns and reopenings of the COVID-19 pandemic in Denmark.

A longitudinal study was conducted using online questionnaires that were distributed via email to Danish adults (≥ 18 years) with diabetes who were members of user panels at either J o u r n a l P r e -p r o o f Journal Pre-proof Steno Diabetes Center Copenhagen, a specialist diabetes clinic, or the Danish Diabetes Association. Data were collected between March 19 and June 25, 2020. Thus, the information analyzed in this paper ranges from the first week to roughly three months after the first lockdown ensued in Denmark. Details of survey development, user panels and recruitment strategy have been reported previously. 12, 24 Briefly, the survey is a diabetes-specific version of the Copenhagen Corona-Related Mental Health questionnaire 24 , and consisted of six questionnaire waves (Q1-Q6). All respondents who answered the first questionnaire (Q1) received questionnaires at all subsequent waves (Q2-Q6) regardless of having missed one or more waves. The study was approved by the Danish Data Protection Agency (P-2020-271). Denmark, including details on key messages from the Danish Health Authority, information on lockdown and reopening phases, and questionnaire distribution dates.

Questions about psychosocial health were included in all six questionnaires; items on health and sociodemographic status were only included at Q1 and Q6. Wording of each item in the survey can be found in the supplementary material (S1).

Psychological distress was measured with a five-item questionnaire scale specifically developed to assess distress during the COVID-19 pandemic. 25 The scale consists of adapted items from the validated GAD-7, CES-D and Event Scale Revised questionnaire scales, and asks how often in the past week the respondent has experienced being 1) nervous, anxious, or on edge, 2) depressed, 3) lonely, 4) hopeful about the future, and 5) had physical reactions that can be linked to the Coronavirus. The scale has four response categories ranging from "rarely or none of the time (< 1 day)" to "most or all of the time (5-7 days)", weighing 1-4 points in the total score calculation (range 5-20), respectively.

Diabetes distress was measured with two items from the diabetes distress scale 26 : "feeling overwhelmed by the demands of living with diabetes" and "feeling that I am often failing with my diabetes routine" (DDS2). The response categories ranged from "not a problem" to "a very serious problem". The total distress score was calculated as the average of the two items, thus ranging from 1-6, where a score ≥2 indicates elevated diabetes distress. 27 Anxiety was measured with the Symptom Check List-revised anxiety subscale (SCL-ANX4). 28 The SCL-ANX4 contains four items about 1) feelings of fear, 2) nervousness, 3) terror/panic and 4) excessive worry, with response categories "not at all", "a little", "moderately", "quite a bit", and "extremely". To accommodate the rapidly changing situation during the pandemic, a one-week recollection period was chosen over the usual four-week timespan used for the scale. The final SCL-ANX4 score is the average of the responses to each item and ranges from 0-4. The likelihood of having an anxiety disorder at an average score of 0 corresponds to <10%, whereas it is 45% at an average score of 4.

General loneliness was measured with the three-item UCLA Loneliness Scale (UCLA). 29 The UCLA asks how much the respondent has felt starved for company, left out, and isolated from J o u r n a l P r e -p r o o f Journal Pre-proof others, and contains the response categories "never/rarely", "sometimes", and "often", providing a total score from 3-9. Diabetes-specific loneliness (UCLA-D) was assessed with two questions, similar to the format of the UCLA, about the degree of feeling alone with diabetes and missing someone to talk to about diabetes. 30 

The survey collected information on age, sex, education, cohabitation status, employment status, diabetes type, duration and complications, HbA1c, comorbidities and mental illness history.

Continuous variables are reported as means with standard deviations, and as medians with 25 th and 75 th percentiles. Categorical data are reported as frequencies and proportions. Marginal differences in participant characteristics between each questionnaire as well as between Q1 and Q6 were tested with Student"s t-tests and Wilcoxon tests for normally and non-normally distributed data, respectively, and with χ 2 -tests for categorical data.

Linear multilevel mixed-effects (LMM) regression models were used to analyze trajectories of psychosocial health over time while accounting for between and within-subject variation.

LMM models incorporate both fixed effects (population-level effects) and random effects (individual-level effects) by specification of a so-called "random intercept, random slope" model structure. The fixed effects (explanatory) part of each LMM was selected using the model- indicator variable for each respondent (random intercept) and a questionnaire wave indicator variable for each respondent (random slope). Missing data were handled using complete case analysis (listwise deletion).

To analyze differences over time in subgroups, interaction terms between the questionnaire wave indicator variable and explanatory variables of interest were added to the LMM models, using Q1 values as reference. Analyzed subgroups were defined by sex (women/men), cohabitation status (alone/with someone else), diabetes type (type 1/type 2), number of diabetes complications (0, 1 or 2+), and history of mental illness (yes/no). Age and HbA1c groups were also analyzed. The Danish Health Authority (DHA) categorizes age-related risk of severe COVID-19 infection into three age groups; < 65 years, 65-79 years, and ≥ 80 years. 31 The age variable was split into three categories accordingly. HbA1c groups were defined as high (> 70mmol/mol / 11.1%) or low (≤ 70mmol/mol / 11.1%) according to the DHA"s assessment of high-risk individuals with diabetes. Only statistically significant interaction terms at a 5% level (P ≤ 0.05) were explored further in within-group analyses to avoid erroneous inference. All models were estimated in Stata 15.1 using restricted maximum likelihood and assuming unstructured covariance matrices.

The first questionnaire was distributed to 2,430 individuals with diabetes of whom 1,366 (56%) responded (supplementary material S2). A response rate between 70-80% of the initial 1,366 respondents was maintained at Q2-Q4, dropping to 66% and 60% at Q5 and Q6, respectively. Table 1 ). There were roughly 25% who had moderate to severe diabetes distress, and 19.5% of the sample had a 20% risk or more for an anxiety disorder. More than half of the respondents often or sometimes felt either lack of company or isolated from others, while just under 1/3 often or sometimes felt left out. Responses to the UCLA-D revealed that one in four of the respondents often or sometimes missed someone to talk to about diabetes and that one in three did so with regards to feeling alone with their diabetes.

There were no statistically significant differences in socioeconomic or health status characteristics between questionnaires, nor any differences between Q1 and Q6. Table 2 show trajectories and regression coefficients, respectively, from Q1 through Q6 for each psychosocial outcome assessed by the VAS and the questionnaire scales. Psychological distress decreased at Q5 (Δ = -0.23, p = 0.020) and Q6 (Δ = -0.58, p < 0.001) compared to Q1 after having increased from Q2 to Q4 (0.14, p = 0.022). Diabetes distress decreased at Q2 (Δ = -0.06, p = 0.034) and Q4 (Δ = -0.08, p = 0.009) compared to Q1, but no statistically significant changes were found at Q5 and Q6 compared to Q1. Comparing Q1 to Q6, both anxiety (Δ = -0.11, p < 0.001) and general loneliness (Δ = -0.42, p < 0.001) decreased.

However, general loneliness increased from Q1 to Q4 (Δ = 0.21, p < 0.001) before decreasing.

Diabetes-specific loneliness remained stable when comparing Q1 to Q6, except for Q5 where a statistically significant increase was observed (0.08. p = 0.037).

Figures and underlying regression tables of subgroup analyses can be found in the supplementary material (S5-S11).

Age and sex. The oldest age group, compared to the youngest, had a statistically significant decrease in quality of life at Q6 compared to Q1 (between-group difference -1.59, p = 0.003).

Women had a larger decrease in feelings of social isolation (between-group difference -0.697, p < 0.001) and psychological distress (-0.44, p = 0.035) from Q1 to Q6 compared to men. There were statistically significant within-group improvements for both men and women with regards Cohabitation status. Respondents living with a partner and/or children had a larger increase in diabetes-specific loneliness at Q5 and Q6 relative to Q1 compared to respondents living alone (between-group differences: Q5: 0.24, p < 0.001 | Q6: 0.20, p = 0.031), and a larger increase in general loneliness at Q5 (0.331, p = 0.021).

Mental illness history. Compared to Q1, having a history of mental illness was associated with a greater decrease in COVID-19 worries at Q2 (-0.67, p = 0.010), Q3 (-0.59, p = 0.032) and Q4 (-0.55, p = 0.050) compared to having no history of mental illness; however, this betweengroup difference disappeared at Q5 and Q6. The same tendency was observed for anxiety, with a larger decrease in anxiety in respondents with a history of mental illness (Q2: -0.08, p = 0.065 | Q3: -0.18, p < 0.001 | Q4: -0.21, p < 0.001) compared to respondents without. Psychological distress decreased more in respondents with a history of mental illness at Q2 through Q4 compared to Q1 than in respondents without (between-group difference -0.73 --1.18, p < 0.001), but this difference was not statistically significant at Q5 and Q6. Within-group analyses revealed that psychological distress decreased only in the mental illness group from Q1 to Q6 (-0.61, p < 0.001). Diabetes distress initially decreased more in respondents with a history of mental illness compared to no mental illness; however, this between-group difference was statistically significant exclusively at Q4 (-0.32, p < 0.001).

Diabetes type, diabetes complications and HbA1c. Anxiety increased in respondents with type 2 diabetes compared to respondents with type 1 diabetes at Q2 (between-group difference 0.07, p = 0.024), Q3 (0.07, p = 0.025) and Q4 (0.10, p = 0.001) relative to Q1. Respondents with two or J o u r n a l P r e -p r o o f more diabetes complications had a larger increase at Q6 relative to Q1 in feelings of social isolation compared to respondents without complications (0.91, p = 0.014), a larger increase in diabetes distress at Q5 (0.47, p < 0.001) and a larger decrease in general loneliness at Q4 (-0.641 p < 0.001) Respondents with high HbA1c (>70 mmol/mol / 11.1%) had an increased diabetesspecific loneliness score compared to respondents with low HbA1c at Q5 (between-group difference 0.30, p < 0.004) and Q6 (0.28, p = 0.017) relative to Q1.

This study documents several changes in psychosocial health in a population of people with diabetes during the first three months of the COVID-19 pandemic in Denmark.

The initial level of psychological distress measured at the beginning of the pandemic in March was relatively low (7.7) compared to reported levels (8.7) in a study of 10,000 respondents from the US general population between March 19-24, 2020 (i.e., the same time as Q1 of this study), using the same distress measure. 25 The proportion of respondents initially feeling isolated did not differ when compared to levels before the pandemic, but a larger than usual proportion of respondents with diabetes felt left out and starved for company. 12 March 20 and April 16. 24 In the same study, feelings of social isolation increased slightly from March 20 to March 26, but then reversed. Comparing these findings to our results indicates that the psychosocial health of people with diabetes might be more affected and for a longer period of time compared to the general population. A study of the general German population, using the same measure of psychological distress as our study, found similar reductions in psychological distress between March 24 and April 21. 33 Other studies conducted in the general population of Germany and Austria between March and June showed that loneliness was low before the pandemic, highest during lock-down, and decreased during re-opening. 34, 35 Although this evidence suggests that loneliness during the pandemic was short-lived, an American survey of more than 1,000 individuals found that 43% experienced "high" loneliness in April 36 , which increased by June despite relaxations in societal restrictions. 37 Another study found markedly different levels in loneliness among general populations of the US, South Korea, France and Hong Kong over the course of the pandemic, adding further evidence of country-specific effects on psychosocial health. 38 Although psychosocial health of people with diabetes in our study was comparable to that of the Danish (and other) general populations, this is not necessarily the case in diabetes populations in other countries. 16 

Contrary to the many positive changes in general psychosocial health during the reopening phases, there were no improvements in diabetes-specific psychosocial health at Q6, on average, in the pooled sample. This may be due to the relatively low initial levels of diabetes distress and diabetes-specific loneliness. For example, the level of diabetes distress in the studied population corresponds with what is found in diabetes populations under normal circumstances. 17, 18 Whereas this may indicate that the first COVID-19 lockdown in Denmark had little if any impact on diabetes-specific psychosocial health, it may also reflect instead the result of the acute and relatively short lockdown period. That is, the three-month period may not have been enough to spike any real impact, with more general aspects (such as quality of life and loneliness) being more at the forefront of one"s psychosocial well-being than diabetes-specific ones. However, a study from the US found that diabetes-related stress increased along with negative impact on diabetes management during the pandemic. 16 These findings were based on participants" self-reported level of stress and worries related to their diabetes when compared to self-reported levels before the pandemic. The same study also reported that half of the participants felt that their diabetes was more difficult to manage during the pandemic than before and the authors link this to the increase in feelings of distress and isolation. 16 In our study, women had larger decreases in feelings of social isolation, anxiety and psychological distress, and larger increases in quality of life compared to men. Likewise, psychological distress was also higher in women in the aforementioned German study. 33 In our study, women experienced higher levels of anxiety and psychological distress at baseline, as also found in the American survey study 25 , suggesting an initially stronger psychological reaction to the pandemic. However, it cannot be ruled out that women had higher anxiety and psychological distress levels than men already before the pandemic. Adding to this, a Danish study found improve-J o u r n a l P r e -p r o o f ments in psychological well-being in both men and women between the beginning and end of April and attribute this improvement to reduction of symptoms of anxiety and depression. 39 We also observed larger improvements in psychosocial health for people with a history of mental illness during the study period. This finding can also potentially be attributed to higher baseline levels in this subgroup; however, as noted by Holingue et al., people with preexisting mental illness are more susceptible to psychological distress during pandemics. 40 Whereas we did not observe statistically significant changes in diabetes-specific psychosocial health in the pooled sample, we observed some interesting changes in diabetes-related subgroups. For instance, respondents with 2+ diabetes complications and those with high HbA1c had increased feelings of social isolation and diabetes-specific loneliness, respectively, compared to those without complications and those with low HbA1c. This suggests that the Danish Health Authority"s communication regarding risk factors in people with diabetes may have had an impact on psychosocial health (see Figure 1 ).

Strengths of this paper include the relatively large sample size and high retention rate across the study, and the unique and timely survey that was specifically designed to capture aspects of psychosocial health in individuals with diabetes throughout the COVID-19 pandemic. The most important limitation of this study is that data on psychosocial health was not collected in the same population before the pandemic, prohibiting direct before/after comparisons, and requiring that psychosocial health instead be compared to other pre-pandemic studies. Another limitation of this study is that the analyzed population is overrepresented by people of older age, retirees and people living alone, thus limiting the generalizability of the findings. The nature of the questionnaires used to assess diabetes distress and diabetes-specific loneliness may also be limiting.

For example, the DDS2 is mainly used in a clinical setting as a screening tool and may not be as J o u r n a l P r e -p r o o f suitable for research purposes as the full Diabetes Distress Scale. Nevertheless, we chose this and other abbreviated instruments to mediate potential psychological strain on respondents in an already straining situation. Finally, to our knowledge, no longitudinal studies using similar measures have been conducted in people with diabetes, preventing direct comparisons. However, it is also a strength of this study to be among the first to report on this topic, contributing with an extensive analysis that will provide useful in planning support for people with diabetes during the current and future pandemics.

To conclude, general measures of psychosocial health in people with diabetes improved, on average, during the first three months of the pandemic in Denmark; diabetes-specific measures did not. These findings both confirm and contrast those of other studies. Just as deteriorations in quality of life and increases in loneliness were observed immediately after lockdown, improvements in psychosocial health were observed in relation to reopening phases of society. However, given different trajectories of lockdown and re-openings in different countries, generalization of the presented findings may not be advisable. Nevertheless, our findings highlight the need for attention regarding the impact of the pandemic on psychosocial health, not least for people with chronic illness, and especially in relation to acute effects on loneliness and quality of life. Psychosocial health is highly and rapidly influenced by social restrictions that follow a pandemic, rendering it a crucial task to provide psychosocial support for people with diabetes during sudden and/or prolonged pandemics.

We wish to give thanks to the respondents for their participation in the study, and to Lone

Nielsen (Danish Diabetes Association) for their indispensable efforts in participant recruitment.

Finally, we would like to thank the research team at the University of Copenhagen behind "Standing together -at a distance"(www.coronaminds.ku.dk), whom the survey was developed in close collaboration with.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. J o u r n a l P r e -p r o o f 

WHO Declares COVID-19 a Pandemic

Immediate Psychological Responses and Associated Factors during the Initial Stage of the 2019 Coronavirus Disease (COVID-19) Epidemic among the General Population in China

Mental health before and during the COVID-19 pandemic: a longitudinal probability sample survey of the UK population

Especially for Loved Ones; Older Adults are Less Anxious

The epidemiology and pathogenesis of coronavirus disease (COVID-19) outbreak

Is diabetes mellitus associated with mortality and severity of COVID-19? A meta-analysis

Comorbid Chronic Diseases and Acute Organ Injuries Are Strongly Correlated with Disease Severity and Mortality among COVID-19 Patients: A Systemic Review and Meta-Analysis

Risk factors for COVID-19-related mortality in people with type 1 and type 2 diabetes in England: a population-based cohort study

COVID-19 in people with diabetes: understanding the reasons for worse outcomes

Obesity as a risk factor for COVID-19: an overview

Preadmission Diabetes-Specific Risk Factors for Mortality in Hospitalized Patients With Diabetes and Coronavirus Disease

Diabetes and COVID-19: psychosocial consequences of the COVID-19 pandemic in people with diabetes in Denmark-what characterizes people with high levels of COVID-19-related worries

Coping with diabetes during the COVID -19 lockdown in India: Results of an online pilot survey

Effects of COVID-19 lockdown on type 2 diabetes, lifestyle and psychosocial health: A hospital-based cross-sectional survey from South India

Perceived risk, behavior changes and Health-related outcomes during COVID-19 pandemic: Findingsamong adults with and without diabetesin China

The early impact of the COVID-19 pandemic on adults with type 1 or type 2 diabetes: A national cohort study

The prevalence of diabetes-specific emotional distress in people with Type 2 diabetes: a systematic review and meta-analysis

The detection and management of diabetes distress in people with type 1 diabetes

Understanding the sources of diabetes distress in adults with type 1 diabetes

Diabetes distress in Type 1 diabetes--a new measurement fit for purpose

Predictors of psychological well-being in types 1 and 2 diabetes

Functional and self-rated health mediate the association between physical indicators of diabetes and depressive symptoms

Household composition and psychological health: Results of the Second Diabetes Attitudes, Wishes and Needs (DAWN2) study

Standing together -at a distance': Documenting changes in mental-health indicators in Denmark during the COVID-19 pandemic

Mental distress during the COVID-19 pandemic among US adults without a pre-existing mental health condition: Findings from American trend panel survey

Development of a brief diabetes distress screening instrument

Addressing diabetes distress in clinical care: a practical guide

A brief case-finding questionnaire for common mental disorders: the CMDQ

Reliability and validity of the Danish version of the UCLA Loneliness Scale

The effect of peer support in adults with insulin pump-treated type 1 diabetes: a pilot study of a flexible and participatory intervention

Danish Health Authority. Personer i øget risiko

En undersøgelse om ensomhed -ensomhed blandt mennesker med diabetes og pårørende til mennesker med diabetes [An investigation of loneliness -loneliness among people with diabetes and relatives to people with diabetes

Psychische Belastungen, Resilienz, Risiko-und protektive Faktoren während der SARS-CoV-2-Pandemie

The impact of COVID-19 restriction measures on loneliness among older adults in Austria

Health behaviors and mental health before and during the COVID-19 pandemic: A longitudinal population-based survey

Loneliness: A signature mental health concern in the era of COVID-19

Three months of loneliness during the COVID-19 lockdown

Cross-cultural comparisons of psychosocial distress in the USA, South Korea, France, and Hong Kong during the initial phase of COVID-19

Increased psychological well-being after the apex of the COVID-19 pandemic

Mental Distress in the United States at the Beginning of the COVID-19 Pandemic

The authors have no conflicts of interest to declare.J o u r n a l P r e -p r o o f