Article Text

Original research
Depression and glycaemic control among adult patients with type 2 diabetes: a cross-sectional study in a comprehensive specialised hospital, Jigjiga, Ethiopia
  1. Seid Yimam Ali1,
  2. Ahmed Muhye Seid2,
  3. Kalkidan Hassen3,
  4. Samuel Taddese Abebe4,
  5. Zelalem Banjaw4,
  6. Mohammed Ibrahim4
  1. 1Medicine, Jigjiga University, Jigjiga, Somali, Ethiopia
  2. 2Public Health, Dire Dawa University, Dire Dawa, Ethiopia
  3. 3Nutrition and Dietetics, Jimma University College of Public Health and Medical Sciences, Jimma, Oromia, Ethiopia
  4. 4Biomedical Science, Jimma University College of Public Health and Medical Sciences, Jimma, Oromia, Ethiopia
  1. Correspondence to Mr Seid Yimam Ali; seidyimam4500{at}gmail.com

Abstract

Objectives The primary objective of this study was to evaluate the prevalence of depression and its sociodemographic predictors, clinical predictors and glycaemic control among adult patients with type 2 diabetes at Sheik Hassan Yabare Comprehensive Specialized Hospital (SHYCSH), Jigjiga, Ethiopia.

Design A hospital-based cross-sectional study was conducted.

Setting Patients with type 2 diabetes mellitus (T2DM) at Sheik Hassan Yabare Comprehensive Specialized Hospital, for chronic follow-up from 3 October 2022 to 13 November 2022.

Participants Randomly selected 278 patients with T2DM age 18 years and older, with a duration of 1 year or more since diagnosis, who had a diabetic follow-up at SHYCSH.

Main outcome measures Depression was assessed using the Patient Health Questionnaire.

Results A total of 263 participants were included, with a response rate of 94.6%. Of the respondents, 134 (51%) were male, making up more than half of the total. The overall prevalence of depression was 47.1% (95% CI 41.1 to 53.2). Depression was further classified, as follows, based on its severity: the majority (66 or 25.1%) had mild depression, followed by 44 (16.7%) with moderate depression, 9 (3.4%) with moderately severe depression and 5 (1.9%) with severe depression. A multivariable logistic regression analysis indicated that poor glycaemic control (adjusted OR (AOR)=1.93; 95% CI 1.05 to 3.53), DM complications (AOR=2.02; 95% CI 1.09 to 3.74) and DM duration of 6–10 years since diagnosis (AOR=2.29; 95% CI 1.21 to 4.34) were independently associated with depression.

Conclusions Our study revealed a significant burden of depression among patients with T2DM receiving follow-up care at the hospital. Glycaemic control, the presence of complications and a longer duration of diabetes were identified as predictors of depression. Therefore, concerned stakeholders should work to improve blood sugar control and promote healthy behaviour, particularly among those with complications or who have been sick for an extended period of time.

  • Depression & mood disorders
  • DIABETES & ENDOCRINOLOGY
  • NEUROLOGY

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • The level of glycaemic control was determined using the glycated haemoglobin A1c test, which is a more reliable predictor of long-term glycaemic control than fasting blood sugar.

  • The study used a cross-sectional design, which can only establish associations between variables and cannot determine causality.

  • The study was conducted in a single comprehensive specialised hospital in Jigjiga, Ethiopia, which may limit the generalisability of the findings to other settings.

Introduction

According to the International Diabetic Federation, in 2021, 537 million people worldwide had diabetes, with type 2 diabetes mellitus (T2DM) accounting for 90% of all cases. Depression is a common and serious medical illness that causes feelings of sadness and/or a loss of interest in activities, and around 280 million people worldwide suffer from depression. In terms of disease burden, depression is currently ranked third.1 However, it is expected to overtake all other diseases by 2030.2 Despite their mutual relationship, 15%–20% of those who have T2DM struggle with a moderate to severe form of depression worldwide.3 Patients with T2DM experience depression due to complications and poor glycaemic control.4 Environmental stressors and heritable genetic factors cause structural and functional changes in many brain regions, resulting in dysfunctional neurogenesis and neurotransmission. Affected areas include the prefrontal cortex, cingulated gyrus, amygdala, hippocampus, thalamus and hypothalamus, which are involved in the regulation of motivation, eating, sleeping, energy level, circadian rhythm and responses to rewarding and aversive stimuli.5

Hyperglycaemia is a common cause of depression, as seen by glycated haemoglobin A1c (HbA1c) levels.6 This is due to the brain’s susceptibility to high plasma glucose levels due to the lack of an active glucose transporter.6 Hyperglycaemia also has the potential to activate the polyol pathway, resulting in oxidative stress and increased production of advanced glycation end products, which can lead to neuronal damage and eventually depression.7

Likewise, excessive stress causes negative feedback loops of the hypothalamic-pituitary-adrenal (HPA) axis to become dysfunctional, leading to elevated cortisol levels in the adrenal cortex and reduced glucocorticoid receptor (GR) expression. Antidepressants have been shown to restore the negative feedback control of the HPA axis and increase GR expression.8 Hyperglycaemia also reduces brain-derived neurotrophic factor levels and decreases hippocampal volume, leading to increased cortisol release, which is involved in depression.9

Depression and T2DM are causally related and deserve attention from clinicians to ensure better management. It is a comorbid condition with DM, and the most substantial increase in its prevalence is observed among individuals with T2DM in developing countries, including Ethiopia.10–12 T2DM prevalence is increasing globally, with the highest rate of increment in developing countries. It has multidimensional burdens ranging from physical and mental to socioeconomic consequences. While most of these burdens had due consideration, the mental health impact of the disease, such as depression, is often unnoticed, undiagnosed or untreated.13 More importantly, depression is known to have a higher likelihood of occurrence in patients with T2DM, but its recognition and treatment are less than optimal.14 Targeted efforts are needed to improve depression detection among adults, mainly those with diabetes.

Despite the controversial association between depression and glycaemic control, several cross-sectional studies across the world, including Ethiopia, showed a significant association between poor glycaemic control and depression.15–17 It is highest in the Gulf countries and Africa, particularly in Ethiopia, where more than two-thirds of patients with T2DM have poor glycaemic control.18–20 In sub-Saharan countries, the costs of poor control of diabetes are 3–4 times higher than those with good glycaemic control.21

Chronic disease management often focuses on single conditions, such as diabetes and depression. Although some efforts have been made to estimate the prevalence of depression and its associated predictors among patients with T2DM in Ethiopia, limited studies have addressed the association of depression with glycaemic control. These studies did not measure glycaemic control using the standard measurement, HbA1c, which better indicates the chronic level of hyperglycaemia.17 22 Hence, measuring the prevalence of depression and its sociodemographic predictors, clinical predictors and glycaemic control remains a critical undertaking. Generally, the literature now shows a significant association with various predictors of depression among patients with DM, but it often goes unrecognised and untreated in developing countries, particularly in the study area.23 As a result, this study aims to determine the prevalence of depression and its sociodemographic predictors, clinical predictors and glycaemic control among adult patients with T2DM.

This cross-sectional study will help health development planners consider depression in patients with T2DM when designing diagnosis and management strategies. It will also provide additional knowledge and baseline information for other researchers and provide information to the diabetic population. This cross-sectional study will help health development planners consider depression in patients with T2DM when designing, diagnosing and managing strategies. It will also provide additional knowledge and baseline information for other researchers and provide information to the diabetic population.

Methods and materials

Study design and setting

From 3 October 2022 to 13 November 2022, a hospital-based cross-sectional study was conducted at Sheik Hassan Yabare Comprehensive Specialized Hospital (SHYCSH), located in Jigjiga, the Somali Region of eastern Ethiopia. Jigjiga is situated 626 km east of Addis Ababa, the capital of Ethiopia. SHYCSH started providing services in 2017 and serves Somalia and neighbouring regions. The hospital has 282 beds and offers a broad range of services at general and specialty levels, including ENT, neurology, plastic surgery, psychiatry, dialysis centres and chronic follow-up care. Specific services provided by the hospital include endocrinology, cardiology and gastroenterology. Annually, nearly 712 patients with T2DM receive follow-up care at SHYCSH. At the time of data collection, 474 patients with T2DM had regular follow-up appointments.

Study participants and sample size

Study subjects

Randomly selected patients diagnosed with T2DM and who had follow-up at the chronic outpatient department unit.

Inclusion and exclusion criteria

All patients with T2DM aged 18 years and older, with a duration of 1 year or more since diagnosis, who have had diabetic follow-up at SHYCSH were included. However, individuals with a history of depression prior to T2DM, those diagnosed with any mental disorder, individuals with a current family member or very close person’s recent death or loss, patients with hypothyroidism, mothers 1 month before and after delivery and those taking corticosteroids were all excluded.

Sampling and data collection

The sample size was calculated using a single population proportion formula by considering the p value from the prevalence of depression in patients with T2DM in Harar, Eastern Ethiopia: 48.9%.24 The 95% CI, 5% margin of error and estimated total T2DM population during the study period were n=712.

Embedded Image

Finally, 278 patients with T2DM, aged at least 18 years, were enrolled in the study using a systematic random sampling technique. The sampling interval (k) was determined by dividing the number of patients on follow-up during the data collection period by the sample size of the study. The first patient was selected by lottery based on their visit order, and subsequently, every other patient was recruited for the study.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Variables and measurements

The data collection was done using an interviewer-administered structured questionnaire. The Patient Health Questionnaire-9 (PHQ-9) was implemented to diagnose depression and grade its severity. The PHQ-9 is a self-administered and/or interviewer-administered measure of depression symptoms based on the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, criteria for the diagnosis of depressive episodes and has been found to have a sensitivity and specificity of 88% for a diagnosis of major depression based on structured interviews.25 26 It has been validated in sub-Saharan Africans and in Ethiopia in patients with diabetes for the detection of general depressive symptoms.27 28 The PHQ-9 assesses how often the respondent has experienced specific symptoms more than half the days over the past 2 weeks, assigning values of 0–3 points (0=not at all, 1=several days, 2=more than half of the days and 3=nearly every day).26 Depression is diagnosed if there is a score of at least 5 reported for the past 2 weeks. Depression was further classified according to severity: 0 indicates no depression, 1–4 indicates minimal depression, 5–9 indicates mild depression, 10–14 indicates moderate depression, 15–19 indicates moderately severe depression and 20–27 indicates severe depression.26 Questionnaires were translated into local languages to avoid biased responses. Fourteen patients with T2DM were given a pretest, and modifications were introduced depending on feedback. Blood was taken after overnight fasting for biochemical analysis. Three millilitres of venous blood were collected in a vacutainer with an EDTA anticoagulant tube and refrigerated at 4°C. The blood was subjected to HbA1c determination in the same laboratory by the same technician applying the same technique using a fully automated Cobas C 311 chemistry analyser. HbA1c levels of 7% were considered poor glycaemic control, and HbA1C levels between 5.7% and 6.5% indicated that blood sugar levels were in the prediabetes range.29

Data processing and analysis

Data were edited, coded and entered into EpiData, exported to and analysed by SPSS V.26.64. Bivariate analysis was performed separately using binary logistic regression to rank the relative importance of exposure variables with the outcome variable. Multiple logistic regressions with the backward elimination method were performed to control the effect of confounding variables. Exposure variables with a p value <0.05 and a 95% CI were declared to be significantly associated factors for depression. The Hosmer and Lemshow test was non-significant and had a small χ2 value.

Results

Socioeconomic characteristics of the study participants

Of the 278 patients with T2DM selected for the study, 263 consented, yielding a response rate of 94.6%. More than half (134 or 51%) of the respondents were male. The mean age of the respondents was 50.21±14.81 years, with a minimum age of 19 and a maximum of 91 years. The majority of respondents (127 or 48.3%) were in their late adulthood (41–60 years), and 182 (69.2%) were married. Three-fourths (197 or 74.9%) of the respondents lived in cities, and half (132 or 50.2%) had an estimated monthly income of 4000–8000 Ethiopian Birr (table 1).

Table 1

Socioeconomic characteristics of adult patients with T2DM on chronic follow-up at Sheik Hassan Yabare Comprehensive Specialized Hospital, Ethiopia, 2022 (ո=263)

Measurement of clinical predictors

More than half of the participants (139 or 52.9%) had comorbidity, with the majority (100 or 71.9%) having only one type of comorbidity. And the major type of comorbidity was hypertension in 95 respondents (68.3%). In addition, 99 (37.6%) respondents had complications. Among those complications, more than half (56 or 56.6%) had only one type of complication. The majority were microvascular complications (68.7%). About 121 (46.0%) respondents had a diabetic duration of 1–5 years since diagnosis, and 122 (46.4%) used both insulin and oral hypoglycaemic drugs as diabetic medications (table 2).

Table 2

Clinical measurements of adult patients with type 2 diabetes mellitus on chronic follow-up at Sheik Hassan Yabare Comprehensive Specialized Hospital, Ethiopia, 2022 (ո=263)

Glycaemic control

Among 263 enrolled patients with T2DM, two-thirds of 175 (66.5%) participants had poor glycaemic control, while one-third of 88 (33.5%) participants had good glycaemic control. The mean HbA1c level was 8.8%±2.7%, with a low of 4% and a high of >14% (figure 1).

Figure 1

Glycaemic status of adult patients with type 2 diabetes mellitus on chronic follow-up at Sheik Hassan Yabare Comprehensive Specialized Hospital, Ethiopia, 2022 (ո=263).

Prevalence and severity of depression among patients with diabetes

The overall prevalence of depression was 124; 47.1% (95% CI 41.1 to 53.2). Then, 139 (52.9%) patients with T2DM had a PHQ-9 score of <5 means clinically no depression (figure 2). Furthermore, the average PHQ-9 score was 6.8±4.5 out of 27.

Figure 2

Depression status among adult patients with type 2 diabetes mellitus on chronic follow-up at Sheik Hassan Yabare Comprehensive Specialized Hospital, Ethiopia, 2022 (ո=263).

Among 124 (47.1%) depressed respondents who scored PHQ-9 ≥5, the majority (66 or 25.1%) had mild depression, followed by 44 (16.7%) with moderate depression, 9 (3.4%) with moderately severe depression and 5 (1.9%) with severe depression (online supplemental figure).

Factors associated with depression among patients with T2DM

For the bivariate analysis, variables from sociodemographic predictors, clinical predictors and glycaemic control were included. Among those variables associated with depression among patients with T2DM, five variables (age category, history of depression, complication, DM duration since diagnosis and HbA1c) had a p value <0.25 on binary logistic regression analysis. Hereafter, it is included in multiple logistic regression analyses for the backward elimination method.

From the total variables included in the multiple logistic regression model, three variables were found to be statistically significant at the level of p<0.05. Hence, the presence of DM-related complications, DM duration since diagnosis and HbA1c had an independent and significant association with depression among patients with T2DM.

As a result, patients with DM-related complications were twice as likely to suffer from depression (adjusted OR (AOR)=2.02; 95% CI 1.09 to 3.74; p=0.026) as patients without complications. Furthermore, patients with DM duration of 6–10 years since diagnosis were 2.3 times (AOR=2.29; 95% CI 1.21 to 4.34; p=0.011) more likely to experience depression compared with those with 1–5 years of DM duration. Moreover, those patients with poor glycaemic control were about twice (AOR=1.93; 95% CI 1.05 to 3.53; p=0.034) more likely to suffer from depression than patients with good glycaemic control. Some variables that established a statistically significant association with depression in bivariate analysis, such as previous history of depression and diabetes duration of more than 10 years since diagnosis, did not confirm a statistically significant association with depression in the final model. Therefore, they were not considered potential independent factors of depression among adult patients with T2DM (table 3).

Table 3

Bivariate and multivariable analysis to identify independent predictors of depression among adult patients with type 2 diabetes mellitus on chronic follow-up at Sheik Hassan Yabare Comprehensive Specialized Hospital, Ethiopia, 2022 (ո=263)

Discussion

Magnitude of depression

This study attempted to measure the prevalence of depression and its sociodemographic predictors, clinical predictors and glycaemic control among patients with T2DM on chronic follow-up. The study revealed that 47.1% (95% CI 41.1% to 53.2%) of patients with T2DM had been diagnosed with depression. This finding was in line with previous studies conducted in Ethiopia: Harar (48.9%),24 Ambo (47%),30 Eastern Province (Saudi Arabia) (49.6%)31 and Lorestan Province (Iran) (46.3%)32 Likewise, meta-analyses in Africa and Ethiopia showed a pooled depression prevalence of 40% and 39.7%, respectively.11 12

However, the current prevalence finding was relatively higher than recent large hospital-based study results from 21 developing countries, which reported 33.1%,15 and a cohort study in China, which reported 29% depression prevalence among patients with T2DM.33 The current prevalence is also higher than the results of studies done in the USA (26.8%),34 Vietnam (23.2%),35 Ghana (31.3%)36 and North Sudan (35.6%),37 and Ethiopian studies in Mizan (37.0%)38 and Halaba (29.3%).17 Different age groups in a study of Ghana and using a higher PHQ-9 cut-off point in studies conducted in Mizan and Halaba, respectively, might explain the lower depression prevalence in these settings. Conversely, the prevalence of depression in the current study is lower than in a meta-analysis study conducted in Iran (61%),39 and in several other studies in Riyadh (Saudi Arabia) (55%),40 Indonesia (57%),41 Tanzania (87%),42 Egypt (69.0%)16 and Cameroon (60%).43 The difference could be attributed to the study design in China, the enrollment of a small sample using a consecutive sampling technique in Indonesia and the inclusion of minimal depression in a Tanzanian study, which could amplify its prevalence. In addition, variations in sample characteristics and overstating of the outcome due to the self-administered nature of PHQ-9 might create this gap among settings.

The indicated pathophysiological explanation in several studies for depression among patients with T2DM was activation of the HPA, which increases cortisol secretion and the inflammation/cytokine hypothesis of depression.33 44 45 Factors identified in the current study also have a pathophysiological linkage to portions of this hypothesis, as described below.

Factors significantly associated with depression

In this study, poor glycaemic control, the presence of complications and DM duration of 6–10 years since diagnosis were significant and independent predictors of depression among patients with T2DM in multivariable logistic regression analysis.

Although the association between depression and glycaemic control remained controversial, we found an independent and statistically significant association among patients with T2DM. Poor glycaemic control in T2DM was about twice (AOR=1.93; 95% CI 1.05 to 3.53) as risky for depression compared with those who had good glycaemic control. This finding is consistent with a large-scale study in 21 developing countries15 and independent studies in Peru,46 Indonesia,41 Egypt,16 North Sudan37 and Halaba.17 In contradiction to this finding, a study conducted in the USA34 and Ireland47 and a meta-analysis in China33 and Ghana36 did not find any significant associations in the multivariable model. The variation in the results could be due to a missed HbA1c result in the USA, HbA1c not being measured in China and only patients over 50 years old being included to see the association in Ireland. A main candidate pathway could be the activation and disturbance of the stress system. A main candidate pathway could be the activation and disturbance of the stress system. Chronic stress activates the HPA axis, increasing the production of cortisol as much as 20-fold, which lowers the brain’s 5-hydroxytryptamine function, which in turn leads to the manifestation of depressive symptoms.48 Long-term sympathetic nervous system activation promotes insulin resistance, visceral obesity and metabolic syndrome, which leads to brain damage, cognitive dysfunction and behavioural changes.49 Excess cortisol disturbs neurogenesis in the hippocampus, a region involved in depression.50 In addition, the brain is highly susceptible to high plasma glucose levels.6 Hyperglycaemia also activates the polyol pathway, which causes neuronal damage that may eventually result in neurotoxicity.7 Alternatively, chronic stress has behavioural consequences through cortisol and other hormones, which activate the fear system, determining anxiety, anorexia or hyperphagia. The same mediators cause a rapid collapse of the reward system, producing depression and cravings for food.49

In addition, the duration of DM was another significant predictor of depression. Patients with diabetes for 6–10 years were about 2.3 times (AOR=2.29; 95% CI 1.21 to 4.34) more likely to develop depression than those with diabetes for 1–5 years. This finding is consistent with studies in Saudi Arabia,51 Bangladesh,52 Mizan Tape (Ethiopia)38 and Ambo (Ethiopia)30 which pointed out that the duration of DM for more than 6 years increases the likelihood of depression occurrence 2–6 times. A longer time since illness diagnosis will almost certainly result in more complications and stressful DM management, which will contribute to depression.

Furthermore, patients with at least one DM complication were twice as likely to suffer from depression (AOR=2.02; 95% CI 1.09 to 3.74) as those without a complication. Numerous cross-sectional studies were in agreement with this finding. The presence of either microvascular or macrovascular complications among patients with T2DM was significantly associated with depression in studies conducted in Saudi Arabia,51 Heidari (Iran),53 the Philippines,54 Bangladesh52 and Ambo (Ethiopia).30 Perhaps it is due to frequent ulcers, a long healing time, hospitalisation, the possibility of amputation, blindness and chronic pain.55 These daily stressors can be overwhelming and, in turn, precipitate or worsen depression. Additionally, neuroendocrine and inflammatory responses that accompany depression may also play similar roles in the progression of microvascular and macrovascular complications among patients with T2DM. Although several studies have shown a significant association between specific types of complications and depression,32 38 51 we are unable to notice any significant association in this aspect.

Finally, it is crucial to acknowledge that time constraints imposed our use of a cross-sectional design, limiting our ability to establish causality, with longitudinal studies offering a stronger approach. Additionally, resource constraints confined our study to a single specialised hospital in Ethiopia. While we were able to collect valuable data and insights, differences in healthcare practices, patient demographics and resource accessibility across various settings should prompt caution when extending our results to broader populations. Despite these limitations, our study uses HbA1c as a measure of glycaemic control, as HbA1c is recognised as a more reliable indicator of long-term glycaemic control than fasting blood sugar due to its stability over time. This provides important insights into the relationship between depression with glycaemic control and various factors in a specific clinical setting, offering a valuable foundation for further investigations and potential interventions.

Conclusions

The prevalence of depression was found to be high among patients with T2DM in SHYCSH.

Risk factors for depression among patients with T2DM were identified as poor glycaemic control, the presence of complications and the long duration of DM.

Therefore, it is recommended that SHYCSH and the Somali Regional Health Bureau integrate depression screening strategies for patients with T2DM during follow-ups at health facilities.

A holistic approach that prioritises blood sugar control and treatment of DM complications among patients with T2DM is also recommended.

Additionally, we suggest that further research should examine the relationship between these predictors and depressed patients with T2DM, non-depressed patients with T2DM, and healthy controls in a multicentre or longitudinal study to ensure generalisability and determine causality.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Ethics approval

Ethical approval for this study was obtained from the Institutional Ethical Review Board of Jimma University, Institute of Health (IRB No. 79/22). Then, an official letter of cooperation was taken from the Jimma University Institute of Health Ethical Review Board and submitted to the SHYCSH directorate office. Details about the study and its benefits and potential risks were explained. Patients were recruited into the study after a written informed consent was obtained. Data obtained during the study was kept confidential. Patients found to have severe depression during data collection were linked to the psychiatric department for further assessment and treatment.

Acknowledgments

We would like to express our heartfelt gratitude to Jimma University for providing the opportunity and technical support to carry out the research. We would also like to thank the SHYCSH team for their assistance. Finally, we would like to express our deepest appreciation to data collectors and research participants for their cooperation and consent to share personal information and medical data for this study.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors SYA and MI substantially contributed to the conception, design, acquisition, and interpretation of data, as well as the write-up of the study. AMS and STA were actively involved in the analysis and interpretation of data. KH and ZB participated significantly in the interpretation of data and provided valuable comments. SYA prepared the manuscript for publication and, acting as the guarantor, took responsibility for the overall content. All authors read and approved the final manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.