Depression affects 10% to 30% of the people with diabetes worldwide [1]. A recent survey indicated that, globally, individuals with diabetes had greater odds of experiencing an episode of depressive symptoms than those without diabetes [2]. Diabetes patients with depression are at increased risk for poor glycemic control and development of cardiovascular complications and have higher mortality rates [1,3,4]. One of the behavioral mechanisms that may link depression with poor outcomes in diabetes is suboptimal diabetes self-care behaviors. In developed countries, patients with comorbid depression and diabetes have been found to have less frequent fruit and vegetable intake, more frequent fat intake, more sedentary behaviors, and fewer exercise sessions per week [5,6,7]. Moreover, smoking was twice as prevalent among depressed than among non-depressed diabetes patients [4]. However, studies on the association between depression and self-care in diabetes patients are scarce in the Sub-Saharan African context. Nevertheless, studies that have looked at the association between depression and other chronic illnesses, especially human immunodeficiency virus (HIV) and acquired immune deficiency syndrome (AIDS), have consistently found that depression is associated with suboptimal self-care in patients [8,9,10].

Diabetes care mostly consists of self-care behaviors [11]. After consultations with their health care providers, patients are responsible for their diabetes care 24 hours per day. Optimal outcomes in diabetes require diligent daily self-care [4]. Diabetes self-care comprises the proper use of medication (insulin or oral medication, correct timing, and correct doses), counting carbohydrates, maintaining a healthy diet (e.g., low in saturated fat, low in salt, high in vegetables and fruits, and moderate alcohol intake), being physically active, monitoring blood glucose levels, maintaining proper foot care, and refraining from smoking, among others [12].

Major depressive disorder is a medical illness that affects feelings, thoughts, and behaviors. Depression results from persistent feelings of sadness and loss of interest in previously enjoyed activities [13,14].

Cross sectional as well as systematic review studies showed that baseline levels of depression symptoms significantly predicted medication non-adherence at follow-up [15,16]. For an increase of one point on a baseline depression score, there was a 1.08-fold increase in the odds of non-adherence to medication at follow-up [15]. Similarly, Lin and colleagues [4] observed that on average, depressed patients did not adhere to oral hypoglycemic medication 80 days in the previous year, compared to 62 days for non-depressed patients. The average percentage of non-adherent days was 25% in depressed patients compared to 19% in non-depressed patients [4].

A recent study conducted among Dutch patients showed that, in particular, depressed mood, sleeping difficulties, appetite problems, and suicidal ideation were significantly related to higher glycosylated hemoglobin (HbA1c) in patients with type 1 or type 2 diabetes mellitus (T1DM/T2DM) [14]. Therefore, the aim of the present study is to explore the associations between depression symptoms and diabetes self-care in Zambian individuals with diabetes.

The aim of the present study was to explore the association between depression as measured by the MDI and diabetes self-care as measured by the SCI. Our first model showed that there was no association between depression total score and diabetes self-care total score. Diabetes self-care involves a number of activities, many of which are measured by the SCI. All of these behaviors or activities (proper use of medication such as insulin or oral medication; counting of carbohydrates; eating a healthy diet, e.g., low in saturated fat, low in salt, high in vegetables/fruits, and moderate alcohol intake; being physically active; and monitoring of blood glucose levels) are designed to improve glycemic control and reduce risk of developing complications caused by poorly managed diabetes [12]. Therefore, we decided to run a linear regression model assessing the association between depression and the 13 self-care activities measured by the SCI. Results showed that depression was associated with decreased glucose testing and not eating meals on time. Other activities were not significantly associated with depression. The most important objective of monitoring blood glucose level is the assessment of overall glycemic control and initiation of appropriate steps in a timely manner to achieve optimal control. In our sample, glucose testing was an important self-care activity that helps in glycemic control and was associated with depression. Moreover glucose testing was one of the highly self-reported problems (61%) among our sample. Self-monitoring of glycemic control is a cornerstone of diabetes care that can ensure patient participation in achieving and maintaining specific glycemic targets [12,22]. Self-monitoring provides information about current glycemic status, allowing for assessment of therapy and guiding adjustments in diet, exercise, and medication to achieve optimal glycemic control. It was also not surprising that depression was associated with not eating meals on time. Eating meals on time helps reduce hypoglycemia. Hypoglycemia occurs when a person's blood sugar (glucose) level falls too low. These two self-care activities (glucose testing and eating a healthy diet on time) are necessary to help a physician and patient make adjustments of other self-care activities such as exercising, going for appointments, carrying quick-acting sugar to treat reactions, administering correct insulin, and adjusting insulin intake.

Although only two self-care activities were associated with depression, the majority of individuals in our study reported below average self-care activities on ketone testing (80%), glucose recording (62%), and glucose testing (60%), eating regular snacks (57%), and carrying quick-acting sugar to treat reaction (56%). This was not surprising because patients in Zambia can only test for ketone and sometimes glucose at the hospital because they cannot afford to have a glucometer of their own or to buy ketone strips. Most of the patients test their glucose at the hospital, so glucose recording is not possible for most of them at home. Their glucose monitoring history is only avail-able in their hospital record files. This is a challenge, given that most of the patients have a hospital appointment every 3 months (pediatric section) unless there is an emergency or they experience recurrent hypoglycemia or are requested to appear for an early appointment by a physician. A qualitative study among adolescents with T1DM in Zambia showed that most patients could not afford to have meals or even carry a snack or quick-acting sugar to treat reactions, “So we all have to come here for reviews and some come from very far and transport is very expensive. But also money for food is a problem. Especially if I am here and I am hungry I need money to buy food, I was told to be eating but I cannot get money here to buy food” [23].

Similar to other studies [24], we found that age, BMI, and to some extent, SES were associated with self-care. Being young was negatively associated with self-care. We speculate that being young was negatively associated with self-care because young people often experience conflicts of priorities that negatively affect their self-care behaviors. Young people often forget to adhere to their regimen because they engage in many competing activities, feel embarrassed in the presence of their peers, and are less inclined to engage in self-care behaviors compared to adults. These results must be interpreted with caution because other studies suggest that older people have elevated depression, which may interfere with their diabetes self-care behaviors compared to young people [25]. Increase in BMI was associated with poor self-care. Similar results were found in Australia among 1,796 respondents with T2DM when a group with BMI ≥35 kg/m² was found to be less likely to achieve healthy diet and exercise targets and placed less importance on diet and exercise recommendations [26]. An increase in BMI is associated with low self-esteem, which may also affect self-care activities. BMI is particularly documented to predict poor glycemic control [27]. Although females are more likely to be depressed than men [2,28], we did not find any association between sex of the participants and self-care activities. We also found that high SES was positively associated with self-care. Although SES is a proxy for a wide variety of factors, these results could suggest that individuals with high SES have readily available money to buy healthy food and required medication and engage in social activities that do not side-track them from self-care activities compared to patients with low SES.

The strength of the present study is that it used the MDI, which is based on the DSM-IV, ICD-10 symptoms of major depression, and the SCI, which covers all recommended critical areas of diabetes self-care [18,26]. Limitations of the study are that we do not know the refusal rate, so there might have been a selection bias, although it is difficult to estimate how big this selection bias was. The sample size was small due to recruitment difficulties also; as a consequence, probability sampling could not be used. In addition, we were unable to screen for other illnesses, such as HIV, and assess the effects of different treatment modalities and biochemical markers such as HbA1c, which could have a confounding effect on our results. Finally, we cannot rule out a possible overestimation of self-care activities and underestimation of depressive episodes by our respondents due to social desirability often found in self-report measures. Moreover, the SCI and the MDI have no norms; also, the former was not validated for the Zambian sample. However, in its form, the SCI had items that tapped into common self-care activities except for one that was dropped, “wearing a medic alert ID,” because no subjects who participated in the study had one. This measure can be made more robust by adding items such as “receiving reminders from family/friends to take medication” and “accessing medical essentials, e.g., syringes.”

Depression is associated with less frequent glucose testing and lack of adherence to meal plans by patients with diabetes. Some variance in poor self-care is explained by demographic characteristics, specifically, age, BMI, and to some extent SES. Therefore, clinicians need to screen and treat individual depressive symptoms in patients with diabetes, especially among young people, patients with low SES, and those with high BMI. Given the lack of manpower in cognitive behavioral therapy in Zambian hospitals, we, in the meantime, recommend the use of anti-depressants in treating depression; these have not only been found to be effective in treating depression [29,30,31], but are free and readily available in Zambian major hospitals. Specifically, if drugs are to be prescribed, we recommend the use of selective serotonin reuptake inhibitors, which has been found to be associated with improved glycemic control [32]. Not only is depression management meaningful from the perspective of a patient's quality of life, it may also have beneficial implications for patients' ability to achieve and maintain diabetes self-care goals [29]. In addition to depression treatment, patients with diabetes ought to be helped to optimize self-care by enhancing self-efficacy, healthy eating, physical activities, monitoring of blood sugar, good problem-solving skills, healthy coping skills, risk-reduction behaviors, and compliance with medication instructions [12]. Therefore, interventions should simultaneously address depression symptoms and self-care skills to achieve optimal diabetes outcomes.