Anxiety, Distress and Stress among Patients with Diabetes during COVID-19 Pandemic: A Systematic Review and Meta-Analysis

The prevalence of mental health disorders has increased during the COVID-19 pandemic. Patients with chronic diseases, such as diabetes, are a particularly vulnerable risk group. This study aims to assess the levels and prevalence of anxiety, distress, and stress in patients with diabetes during the COVID-19 pandemic. A systematic review was conducted in CINAHL, Cochrane, LILACS, Medline, SciELO, and Scopus in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Thirty-seven articles with a total of 13,932 diabetic patients were included. Five meta-analyses were performed. The prevalence of anxiety was 23% (95% CI = 19–28) in T1DM and 20% (95% CI = 6–40) in T2DM patients. For diabetes distress it was 41% (95% CI = 24–60) for T1DM and 36% in T2DM patients (95% CI = 2–84). For stress, the prevalence was 79% (95% CI = 49–98) in T1DM patients. People with diabetes have significant psychiatric comorbidity as well as psychological factors that negatively affect disease management, increasing their vulnerability in an emergency situation. To establish comprehensive care in diabetic patients addressing mental health is essential, as well as including specific policy interventions to reduce the potential psychological harm of the COVID-19 pandemic.


Introduction
The coronavirus infection  has become a global health problem since the beginning of 2020 [1]. The lockdown as well as the restrictions in the different waves of contagion have caused a negative impact on the health of the general population and especially on people who suffer from chronic diseases such as people with diabetes [2]. People with diabetes mellitus (DM) are a risk group, with high hospitalization and mortality rate, and this risk increases when there is COVID-19 infection [3].
The prevalence of mental health disturbances has increased at an alarming rate during the COVID-19 pandemic [4]. Patients with DM present multiple psychosocial factors, which together with the psychological stressors of a pandemic, such as quarantine, social J. Pers. Med. 2022, 12, 1412 2 of 20 distance, and fear of contagion, make this group even more vulnerable [5]. Mental disorders in DM patients reach figures of up to 50%, which predisposes to an increase in mental health disorders in the face of a pandemic situation that leads to difficulties in adapting psychologically [6]. Some reports show that up to 87% of DM type 2 patients indicate being "psychologically affected" [7].
Among the possible issues in psychological health, we can find a greater susceptibility to severe symptoms of depression and a feeling of loneliness, anxiety, stress, or diabetes stress, referring to negative emotions related to the disease such as feeling frustrated, desperate, or angry [8][9][10]. These comorbidities in DM patients can reduce self-care, adherence to treatment and engagement with health professionals, with a negative impact on disease management [11,12]. Several studies indicate that up to 50% of DM patients were afraid of possible contagion [7]. This situation, together with medical distrust, and frustration due to the difficulties in DM management, is related to a reduction in control visits and even more in the demand for assistance in non-emergencies problems, especially those related to mental health [13,14].
The lockdown and successive waves of restrictions have disrupted healthy lifestyle patterns and the ability to self-care [14]. Some studies report that up to 54% of chronic patients claim to have problems related to their usual treatment [15], and data from a survey conducted in 155 countries by the World Health Organization showed that diabetes treatment was partially or completely interrupted in 49% of the countries surveyed [16]. Unhealthy behaviours in DM patients with higher consumption of sugary drinks as well as a reduction in physical activity have also been reported [7]. Other studies report a reduction in self-monitoring of blood glucose; only 28% of patients regularly monitored glucose levels during the COVID-19 lockdown [17]. Given these data, some authors show a clear relationship between self-care deficit and an increase in the number of mental disorders [18].
Although there are several studies that analyse mental health in the general population, data about chronic disease patients and more specifically in patients with DM are still limited. There are studies focused on the treatment of diabetes and associated complications during the COVID-19 pandemic [5,19,20]; however, no systematic review and meta-analysis address psychological disturbances.
An analysis of levels of these variables, looking at the definition by the Medical Subject Headings, anxiety ("feelings or emotions of dread, apprehension and impending disaster"), distress ("negative emotional state with emotional and/or physical discomfort"), and stress (with emotional factors predominating) in the population with DM is necessary, since the number of DM patients affected by these problems before the COVID-19 pandemic was important [21] and these levels may have increased. This review analyses the data currently available in the pandemic scenario, in order to establish intervention strategies and address a psychosocial approach in people with DM during COVID-19. Therefore, the objective of this systematic review and meta-analysis was to analyse the levels and prevalence of anxiety, distress, and stress during the COVID-19 pandemic in diabetic patients.

Design
The review and meta-analysis were reported according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines [22] (see Supplementary  Materials Table S1 for further information). The protocol was registered in PROSPERO (International Prospective Register of Systematic Reviews) with the registration number CRD42022325197.

Search Strategy
A search was performed in the following databases: the Cumulative Index to Nursing and Allied Health Literature (CINAHL) (EBSCO), the Cochrane Central Register of Controlled Trials (CENTRAL), LILACS (BIREME), Medline (Ovid), SciELO (BIREME), and Scopus (Elsevier). The search was done in July 2022 without restriction by language or publication date. The search terms used were: "(anxiety OR psychological distress OR stress) AND (diabetes OR chronic illness OR chronically ill OR non-communicable diseases) AND (SARS-CoV-2 OR coronavirus OR COVID-19)".

Eligibility Criteria
Studies conducted during the COVID-19 pandemic were included with the following inclusion criteria: (1) original studies, (2) type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM), (3) assessing anxiety, distress, or stress symptoms (percentages, means, or median levels), (4) use of anxiety, distress, and stress validated measurement tool. There was no restriction by language or publication date.
Studies were excluded if they were: (1) letters to editors, conference paper review articles, and case reports, (2) articles with other types of diabetes (gestational, MODY, LADA), (3) articles including different chronic pathologies without indicating a number of participants with diabetes, (4) sample of patients with serious cognitive/neurological impairment or mental/physical disability.

Study Selection and Data Collection
First, two independent reviewers analysed titles and abstracts and then the full texts according to the inclusion criteria ( Figure 1). A third author was consulted in case of disagreement.

Search Strategy
A search was performed in the following databases: the Cumulative Index to Nursing and Allied Health Literature (CINAHL) (EBSCO), the Cochrane Central Register of Controlled Trials (CENTRAL), LILACS (BIREME), Medline (Ovid), SciELO (BIREME), and Scopus (Elsevier). The search was done in July 2022 without restriction by language or publication date. The search terms used were: "(anxiety OR psychological distress OR stress) AND (diabetes OR chronic illness OR chronically ill OR non-communicable diseases) AND (SARS-CoV-2 OR coronavirus OR COVID-19)".

Eligibility Criteria
Studies conducted during the COVID-19 pandemic were included with the following inclusion criteria: (1) original studies, (2) type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM), (3) assessing anxiety, distress, or stress symptoms (percentages, means, or median levels), (4) use of anxiety, distress, and stress validated measurement tool. There was no restriction by language or publication date.
Studies were excluded if they were: (1) letters to editors, conference paper review articles, and case reports, (2) articles with other types of diabetes (gestational, MODY, LADA), (3) articles including different chronic pathologies without indicating a number of participants with diabetes, (4) sample of patients with serious cognitive/ neurological impairment or mental/physical disability.

Study Selection and Data Collection
First, two independent reviewers analysed titles and abstracts and then the full texts according to the inclusion criteria ( Figure 1). A third author was consulted in case of disagreement. Two authors extracted data from selected studies into an Excel spreadsheet, consulting with a third author in case of discrepancies. The following information was extracted from each study: (1) author, year of publication, country, (2) study design and period, (3) sample, (4) setting, (5) measuring instrument, (6) type of diabetes, (7) levels of anxiety, distress, or stress (percentage, mean, median) ( Table 1). Two authors extracted data from selected studies into an Excel spreadsheet, consulting with a third author in case of discrepancies. The following information was extracted from each study: (1) author, year of publication, country, (2) study design and period, (3) sample, (4) setting, (5) measuring instrument, (6) type of diabetes, (7) levels of anxiety, distress, or stress (percentage, mean, median) ( Table 1).

Quality Assessment, Evidence Level and Grade of Recommendation
A quality assessment and bias analysis were carried out by two reviewers independently with a third reviewer consulted in case of disagreement.
The National Heart, Lung and Blood Institute quality assessment scale was used for bias assessment of observational studies [23] (Appendix A). The recommendations of the OCEBM were also used (Oxford Centre for Evidence-Based Medicine) to analyse the levels of evidence and grades of recommendation [24] (Table 1).

Data Analyses
A descriptive analysis was performed for the systematic review, extracting the variables in a data table.
For the meta-analysis, all the studies that presented data on the percentage of anxiety, diabetes distress, or stress measured through the same tool were used. Heterogeneity was assessed using the I 2 index. Random effects meta-analysis were performed [25]. Sensitivity analysis and Egger's regression test were used to assess bias in the studies.
Five meta-analyses were performed to estimate the prevalence of anxiety, diabetes distress or stress, and the corresponding confidence interval. StatsDirect software (StatsDirect Ltd., Cambridge, UK) was used for all statistical calculations.

Characteristics of the Studies Included
The initial search found 3157 results. After deleting duplicates and reading the title and abstract, a total of 614 articles were selected. Finally, after reading the full text and analysing the inclusion criteria, 37 articles were included. The study search and selection process are shown in Figure 1.
All the studies found were observational (cross-sectional, retrospective, or prospective) and one was a case-control study. The total sample population consisted of 13,932 type 1 and type 2 diabetic patients. Most studies were conducted in Italy (n = 5), US (n = 5), followed by Saudi Arabia (n = 3), and Turkey (n = 3) ( Table 1).
To measure anxiety, the most used questionnaires were the General Anxiety Disorder-7 (GAD-7) (n =7) and the Visual Analog Scale (VAS) for anxiety (n = 3). The remaining questionnaires used for anxiety were the Hospital Anxiety and Depression Scale (HADS), the Test of Depression and Anxiety Scale (TAD), Spence Children Anxiety Scale (SCAS), the Symptom Check List-revised anxiety subscale (SCL-ANX4), the General Health Questionnaire-12 items (GHQ-12), and the State-Trait Anxiety Inventory (STAI) (see Table 1).
Finally, the stress measurement tools used were the Perceived Stress Scale (PSS) (n = 11), the Visual Analog Scale (VAS) for stress (n = 3), and the Impact of Event Scale Revised (IES-R) ( Table 1).
The data were collected in different settings that included the collection of information through telephone surveys, online forms or through face-to-face at outpatient clinics, hospitals, or primary care centres. Most of the studies (n = 21) collected data during the first phase of the pandemic (January-June 2020).
The studies included had an adequate level of quality; according to the measurement tools applied there were no exclusions. The assessment and characteristics of the studies are represented in Table 1.

Meta-Analysis
Five random effects meta-analyses were performed with a total of 1024 T1DM patients and 4238 T2DM patients.
Finally, stress levels measured with the PSS questionnaire showed a prevalence in T1DM patients for moderate and high levels (PSS ≥ 14) of 79% (95% CI = 49-98) with high heterogeneity (I 2 = 97%). Egger's test showed no publication bias, and no study was removed after sensitivity analysis. Figures 2-4 summarize the findings in relation to anxiety, distress and stress prevalence.

Meta-Analysis
Five random effects meta-analyses were performed with a total of 1024 T1DM patients and 4238 T2DM patients.
Finally, stress levels measured with the PSS questionnaire showed a prevalence in T1DM patients for moderate and high levels (PSS ≥ 14) of 79% (95% CI = 49-98) with high heterogeneity (I 2 = 97%). Egger's test showed no publication bias, and no study was removed after sensitivity analysis. Figures 2-4 summarize the findings in relation to anxiety, distress and stress prevalence.

Discussion
This study suggests relevant data about psychological disorders in the diabetic population during the pandemic, with a meta-analytical prevalence estimation of anxiety of

Discussion
This study suggests relevant data about psychological disorders in the diabetic population during the pandemic, with a meta-analytical prevalence estimation of anxiety of

Discussion
This study suggests relevant data about psychological disorders in the diabetic population during the pandemic, with a meta-analytical prevalence estimation of anxiety of 23% in T1DM patients and 20% in T2DM patients, diabetes distress of 41% in T1DM and 36% in T2DM, and stress of 79% in T1DM.
Studies before the pandemic reported a prevalence of anxiety symptoms of 17.7% for T1DM patients [63], and of 18% for T2DM [64] being for diabetes distress of 42.1% in T1DM [65] and 29.4% in T2DM [66], and for the stress of 50% in T1DM [67]. These data suggest a significant increase in symptoms.
In addition, the prevalence of anxiety found in DM patients was higher than that of studies performed in other groups during the pandemic. In the elderly population, the prevalence of anxiety symptoms found ranged from 10.10% [68] to 21.6% of moderate/severe anxiety in general population [69,70]. Other studies in the general population stated DM as one of the main psychosocial problems with a prevalence of up to 40% [71]. Even a recent meta-analysis in the general population showed that the mean prevalence of anxiety and psychological stress was 38.1% and 37.5%, respectively [72].
More than half of the population with chronic pathology wished to have received additional information about the risks associated with their medical condition during the pandemic [15]. Several authors indicate that the provision of diabetes care was significantly disrupted during the pandemic [73], as corroborated by studies conducted in chronic patients where 52% of adults and 38% of children worsened their health condition during confinement [74].
During the pandemic, the psychological disorders of diabetic patients are often not recognized or underestimated, which can impair the quality of life and self-management of the disease [75]. Greater support for self-care is related to higher adherence to the expected regimen and life changes [12]; however, psychological stressors can have an adverse effect, for example in the loss of good glycaemic control [76].
This study suggests a higher prevalence of anxiety and stress diabetes in T1DM patients, as corroborated by other studies that found several factors related to worse mental health such as T1DM or the female gender [70,77]. Other factors such as age remain controversial; some studies reported worse data in younger patients [70,[77][78][79][80], while for others the levels were higher in older age groups [81].
Regarding the negative results of the pandemic involving mental health, other related factors were the fear of contagion by COVID-19 [82,83] and COVID-19 anxiety syndrome [84]. Studies reported that up to 27.3% of people with DM experienced stress due to the spread of the COVID-19 pandemic and 20% experienced stress due to fear of drug shortages [85]. Even in hospitalized patients, stress levels reached up to 39.3% [75], being lower than those found in our meta-analysis.
Several studies highlight these facts as a reason for greater concern and related them to a reduced capacity in the provision of psychological support to this group [73]. Therefore, finding strategies to identify and reduce anxiety, distress, and stress, as well as multiple other possible disorders such as depression or loneliness should be a priority for diabetes services [86]. In this sense, several studies support the routine implementation of telemedicine [87], as well as increasing the capacity of primary care to provide telehealth services for diseases related to COVID-19 and for several other chronic medical conditions [88]. Studies that have used the telemedicine care model have found positive benefits, for example in a higher mean reduction in the HbA1c level compared with traditional care model [89], so it could also have positive results in the treatment of mental health disorders.
Although a large number of protocols have been developed to identify and recover people with DM infected by COVID-19, there is still a large gap in mental health care. Managing DM in the midst of the COVID-19 pandemic has proven to be a real challenge. To date little is known about how pandemics globally affect the psychosocial health of people with DM. This study is the first meta-analysis to provide an assessment of current levels of anxiety, distress, and stress since the onset of COVID-19 exclusively in patients with DM. It is necessary to clarify the current situation of mental health disorders in these patients in order to establish intervention strategies.

Limitations
This study has several limitations. First, the inclusion of T21DM and T2DM patients from different countries could increase the heterogeneity given the differences in the conditions of the health system, the management and follow-up of the disease, and also clinical variability in the percentage of female, type of diabetes, or measurement instrument. The heterogeneity in the meta-analyses were also high. However, the results of this study may allow understanding the impact of the pandemic on these patients as a start for future research. Another limitation is the inclusion of all the data since the start of the pandemic (different restrictions and waves of contagion), which could increase the heterogeneity. Finally, the different methods of data collection (by telephone, online, or face-to-face interviews) could lead to bias. This review has shown that there are important levels of anxiety, distress, and stress in people with diabetes during the COVID-19 pandemic. Future research should analyse which factors are related with these problems and how those levels can be reduced.

Implication for Practice and Research
The COVID-19 infection and confinement have a diverse impact on access to health services, psychosocial well-being, and self-management of people with diabetes, which must be contextualized to the responses and preparation of each country. Diabetes significantly increases the risk of emotional and behavioural disorders, especially in times of social crisis such as the one experienced with the COVID-19 pandemic [90]. Improving effective self-care behaviours that include healthy coping (healthy eating, being active, blood glucose control) are essential components to establishing optimal behaviour goals, which in turn will improve mental health outcomes [5]. Future research is needed to analyse the monitoring of levels as the pandemic progresses, as well as large multicentre longitudinal studies to avoid the above-mentioned limitations.

Conclusions
The prevalence found during the COVID-19 pandemic for anxiety ranged between 23% and 20%, for diabetes distress between 41% and 36%, and for stress it was 79%. People with diabetes have significant psychiatric comorbidity as well as psychological factors that negatively affect disease management, increasing their vulnerability in an emergency situation. To establish comprehensive care in diabetic patients addressing mental health is essential, as well as including specific policy interventions to reduce the potential psychological harm of the COVID-19 pandemic. Moreover, assessing the variables that can prevent or reduce the development of anxiety, distress, and stress in this population would be important.   Acknowledgments: This study forms part of the Doctoral Thesis of the first-named author within the Health Sciences Doctoral Program from the University of Murcia (Spain).

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. Observational studies quality assessment with National Heart, Lung, and Blood Institute.