Association between self-care practices and glycemic control among adults with type 2 diabetes mellitus in Douala, Cameroon: a mixed-methods study

¹Department of Community Health, AMREF International University, Nairobi, Kenya, ²Pan African Organization for Research and Protection of Violence on Women and Children, Tamale, Ghana, ³Department of Non-Communicable Diseases, Research Health and Development (RSD) Institute, Yaoundé, Cameroon, ⁴Department of Clinical and Biomedical Sciences, University of Exeter Medical School, Exeter, United Kingdom

^&Corresponding author
Ngenche Comfort Tangang, Department of Community Health, AMREF International University, Nairobi, Kenya

Introduction    

Diabetes is a chronic condition characterized by high levels of glucose in a patient´s bloodstream due to the body´s inability to produce enough insulin to effectively use the insulin produced [1]. Diabetes mellitus is a disease in which the pancreas fails to produce adequate insulin (Type 1), or in which the body does not appropriately use the insulin (Type 2) [2]. T2DM contributes to 90%-95% of diabetes cases and is most prevalent in low- and middle-income countries [3]. Type 2 diabetes mellitus is the 8th leading cause of disease burden worldwide [4] and is expected to be the second leading cause by 2050 [5].

According to the United Nations high-level meeting (2018), diabetes mellitus is one of the five leading non-communicable diseases that have been identified as global health issues, which, individually or in combination with cardiovascular diseases, chronic respiratory diseases, cancers, and mental health disorders, account for around two-thirds of all deaths globally [6]. Although the African region has the lowest prevalence estimate of 5.0% among International Diabetes Federation (IDF) regions, Africa is estimated to rank highest in the increase by 2050, with an estimated percentage rise of 142%, reaching 60 million cases [1]. Africa also has the highest rate, 72.6%, of undiagnosed diabetes of all International Diabetes Federation regions [1]. Approximately 6-8 % of urban Cameroonians are diabetic, although 80 % of these individuals remain undiagnosed, and only 25 % had achieved blood glucose control in 2002 [7]. A tenfold (1994 to 2004) increase in diabetic cases was documented among Cameroonian adults through three cross-sectional surveys [7], with a prevalence of 5.5% reported as of 2021 [8].

The urban population of Cameroon struggles with T2DM management, with 61.8% of patients having poor glycemic control [9]. In spite of the need for diet, medication, and exercise [10], research on self-care behaviors and their influence on glycemic outcomes is limited. In Cameroon, most studies conducted have focused on complications of diabetes [11] or public knowledge about diabetes [12]; however, none have assessed the effect of self-care management practices on glycemic outcomes. Although organizations like the World Health Organization and the International Diabetes Federation have partnered with Cameroon to enhance diabetes management, recent studies have shown insufficient self-care practices (SCP) among adult T2DM patients despite national healthcare programs [13]. Additionally, studies in Limbe and Buena-Cameroon revealed a substantial discrepancy between the diabetes self-care education provided and actual patient practices [14]. Thus, this study sought to fill the existing research gap by firstly assessing the levels of glycemic control among T2DM patients, and secondly determining the association between self-care management practices (medication compliance, dietary practices, physical activity, self-monitoring of blood glucose, foot care, smoking, and overall self-care) and glycemic control.

Methods     

Study design: this study design was a cross-sectional descriptive survey that utilized mixed-methods with quantitative and qualitative approaches. Data was collected from January to March 2025 in Douala, the capital of the Littoral region in Cameroon. The revised version of the Summary of Diabetes Self-Care Activities (SDSCA) questionnaire and glycated hemoglobin (HbA1c) test were used to gather quantitative data. Two focus group discussions (FGD) gathered qualitative data, analyzing the experiences of the participants, their motivation, and barriers to the self-care practices. This research included T2DM outpatients at Laquintinie Hospital and Presbyterian Health Complex (PHC) Bepanda, both having substantial populations and diversity.

Study population and eligibility criteria: the study population comprised individuals living in Douala, diagnosed with T2DM who were receiving regular follow-up at Laquintinie Hospital and the Presbyterian Health Complex Bepanda during the study period. Outpatients who were 21 years and older, diagnosed with T2DM for at least six months, and had been living with the disease before the time of recruitment, were included in the study. Outpatients with cognitive impairment, severe illness at the time of data collection, those who needed urgent medical attention, and those who were pregnant were excluded from the study.

Sample size determination: The sample size was computed using Cochran´s formula, because the actual prevalence of T2DM in Douala was not known, and the proportion (p) was assumed to be 0.5, giving maximum variability.

Where: e is the desired level of precision (i.e. margin of error); n₀ is the initial sample size for large populations; p is the estimated proportion of the population with T2DM; Z is the z-value (e.g., 1.96 for 95 % confidence interval); q is 1 - p. Substituting, n₀ =(1.96² × 0.5 × 0.5)/0.05² = 385. Since the population for the study was not large, the Finite Population Correction was applied, adjusting the sample size: N =total population size (estimated at 800 T2DM patients on follow-up at Laquintinie and PHC Bepanda). n =385/(1 + (384/800)) = 260.135.Therefore, the final number of participants was 261. In order to handle dropouts and uncertainties, 267 participants were invited to participate, out of which 230 successfully provided complete responses for the study, and were included in the final analysis, representing a response rate of 86.1%.

Sampling technique: the study sites, Laquintinie Hospital and Presbyterian Health Complex Bepanda were purposefully selected based on their high volumes of T2DM patients. Respondents for the quantitative survey were chosen systematically, while focus group discussion participants were chosen purposively. On averagely 440 T2DM outpatients reported for care each month in both health facilities; according to the Hospital diabetes registry, 2025, Laquintinie Hospital had 300 outpatients approximately, and Presbyterian Health Complex Bepanda had 140 outpatients approximately per month. The sampling interval was three from, k =N/n, where N = study population =1000; n =planned sample size (261) k = 800/261. Systematic random sampling was used to select participants for the survey in each health facility. The first respondent was selected using a lottery method, and every 3rd T2DM outpatient was approached during follow-up visits in the diabetes clinics until the final sample size of 230 was achieved. Two focus group discussions took place, with a total of twelve participants (n=12), each with six participants who were purposively selected based on their ability to communicate in a discussion language commonly understood by the other members, and their willingness to discuss their experiences of self-care management practices and glycemic control in a group setting.

Data collection: for the quantitative data, the following were measured: self-care practices (diet, medications, physical activity, self-monitoring of blood glucose (SMBG), foot care, and smoking) using the revised summary of diabetes self-care activities questionnaire [15] adapted for the Douala context and glycemic control was assessed by testing glycated hemoglobin with the FinecareTM rapid quantitative kit, which employs fluorescence immunoassay technology. Calibration was carried out by inserting a unique ID chip that comes with each new box of test kit, and HbA1c results were expressed in percentage (%). Quality control was carried out internally, as each cartridge contains a built-in internal control that runs automatically with each patient sample. A semi-structured guide was used to collect the qualitative data. The length of each session was between 45-60 minutes, it was recorded and conducted in a discrete location. The participants shared their personal self-care habits, the motivating factors, and what made it difficult to engage in them. A note-taker who had received proper training monitored and described the non-verbal behavior and aided the session moderator.

Study variables: independent variables included self-care practices; diet, medication adherence, physical activity, self-monitoring of blood glucose, foot care, and smoking, measured using a modified version of the revised summary of diabetes self-care activities questionnaire. This questionnaire allowed the participants to indicate on a scale of 0-7 the number of days they had performed each of the activities during the past week, except items 4, 14, and 16. Self-care scores in each domain were categorized as Poor (under the 50^th percentile) or Good (at or above the 50^th percentile), and overall self-care as poor (under the 25^th percentile), fair (from the 25^th-75^th percentile), or good (at or above the 75^th percentile). The outcome measure, glycemic control, was categorized based on the glycated hemoglobin (HbA1c) level of the patient: good (<6.5%) and poor (?6.5%).

Data analysis: quantitative and qualitative findings were analyzed separately before their triangulation in the results' presentation. The methodological triangulation for the results' integration was employed through a convergent parallel mixed-methods design. Quantitative data were collected via KoBo Toolbox for Android by collecting data on sociodemographic and clinical characteristics, self-care practices, and glycated hemoglobin levels, through a questionnaire, translated into French and double-checked through reverse translation. The data were exported in comma-separated values format, cleaned, and analyzed using R software version 4.4.3. Participants' characteristics were summarized using descriptive statistics. Bivariate analysis was used to study the associations between self-care practices and glycemic control. Non-normally distributed variables were tested using Wilcoxon rank-sum test. The multivariable logistic regression model only included variables that attained p < 0.05 during the univariate logistic regressions. The statistically significant threshold was defined as p <0.05 at a 95% confidence interval (CI). The recorded FGD sessions underwent verbatim transcription to develop an authentic depiction of participant responses. The transcript data were compared to the audio recordings to verify accuracy and document complete information before starting the analysis. Thematic analysis was conducted on data obtained from the FGDs. Open coding was performed manually, generating twenty-three distinct codes when examining the initial transcripts. These codes led to the formation of clusters that resulted in five distinct themes agreed upon by all the authors. The themes combined various original codes into broader descriptions that highlighted vital dissimilarities and commonalities among focus group participants in managing their diabetes, along with the challenges they faced in self-care practices.

Ethical considerations: AMREF International University in Kenya approved the study, and the Littoral Regional Research Ethics Committee for Human Health provided ethical clearance (ref.: 2024/032/CE/CRERSH-LITTORAL). The protection of data was ensured through anonymized coding and secure storage. Participation was voluntary, and participants could withdraw and drop out of the study without issues. Professional intervention ensured that the potential risks of blood sampling were minimized.

Results     

Sociodemographic characteristics of participants: of the 230 respondents, the majority were females (67.0%). The median age was 62 years, the median Body Mass Index (BMI) was 29.0 kg/m². Over half of them were married, and most of them had primary (29.1%) or secondary (47.0%) levels of education. The majority were self-employed or were involved in informal jobs, and 63.5% earned below 100, 000 Central African Francs (XAF), approximately 167 United States Dollars (USD) per month. Most participants had been living with diabetes for more than three years, with 24.3% having diabetes for over 10 years (Table 1).

Levels of glycemic control: glycated hemoglobin test carried out was categorized based on respondents' glycated hemoglobin (HbA1c) levels, either good (HbA1c < 6.5%) or poor (HbA1c ≥ 6.5%) glycemic control. The majority of the respondents, 181 (78.7%), had poor glycemic control, while only 49 (21.3%) of participants achieved good glycemic control.

Distribution of self-care practices: different self-care practices were seen among the participants; 43.5% had good dietary practices, 35.7% monitored blood glucose, and 40.9% engaged in physical activities. All participants demonstrated poor medication adherence; 49.1% had good foot care, and 98.3% practiced good smoking habits. Overall, the majority had good or fair self-care practice (Table 2).

Association of self-care practices with glycemic control

Dietary practices: good dietary practices were observed in 57.0% of those with good glycemic control and 40.0% of those with poor glycemic control. The difference was statistically significant (p = 0.03), and thus, there is an association with glycemic control. In the focus group discussions, participants articulated various difficulties related to dietary practices. The first theme that emerged was “ Food choices, traditions, and restrictions.” most participants (n=10) reported difficulties in complying with recommended diets, especially due to the high prices of diabetic foods. Indeed, one focus group participant clearly stated that: “ Due to money issues, you eat what is cheap, and in two to three days, you try to fix the chaos.” The focus group discussions participants were in consensus that starchy food was a large part of their cultural traditions: “ All common foods are sweet, when told to avoid these, we ask what we are supposed to eat.”

Medication adherence: poor medication adherence was observed in all participants (49 with HbA1c < 6.5% and 181 with HbA1c ≥ 6.5%), with no differences between adherence categories. The focus group discussions explained the second theme, ´Medication dilemmas in daily Life,´ more than half (n=11) described the difficulties related to cost of medications, the use of medicines with complicated dosing schedules, and side effects. One participant noted, “On retirement, I cannot afford both living expenses and medication.”Another participant showed confusion, “I don´t know which time is actually morning,” and another stated, “I manipulate my dosage to economize.”Polypharmacy was seen as too many for someone to take at once: “ Taking ten pills in the morning and ten in the evening is unbearable.” Relying on herbs such as “folere” leaves was preferred by some participants, as one added, “Some people take natural herbs because traditional medicine does not impose as many restrictions on what we eat.”

Self-monitoring of blood glucose: good self-monitoring of blood glucose practices was observed in 39.0% of those with good glycemic control and 35.0% in those with poor glycemic control. The difference was not statistically significant (p = 0.60). A third theme emerged: “struggles and realities of blood sugar monitoring,” where some participants (n=5) of the FGD described their experiences of blood glucose monitoring. These lived experiences included the experience of frequent testing, knowledge on glucose targets, and testing system barriers, such as cost and shortages of supply. There was consensus regarding frequent testing: “I check my blood sugar levels every day at 6 am before I brush my teeth” Blood glucose results were also used for guidance on dietary changes, with one participant affirming that, “If it´s above 120, I change my diet.” Nevertheless, the availability of strips to carry out the test is not easy, and they are also costly for many people, one participant said, “You can have the machine, but the strip is hard to find.”Additionally, some participants did not know the correct normal glucose range, with one answering that, “I was told normal value is between 0.80 - 1.10.” It was also reported that when one has high readings frequently, it resulted in distress and avoidance of future checks, “I gave up checking my sugar because it was always high, and it discouraged me,” recounted one participant.

Physical activity: good physical activity practices were observed in 49.0% of those with good glycemic control and 39.0% of those with poor glycemic control. However, this difference was not statistically significant (p = 0.20). The fourth theme that emerged, 'the burden of exercise,' described participants´ experience with physical activity. Most of the participants (n=10) recognized that physical activity was important in the management of diabetes, yet many barriers were cited that hindered their capability of staying active. Some unpleasant experiences were mentioned, as one described, “When I do brutal sport, I feel my heart will stop, so I prefer walking.” Certain participants raised doubts on the effectiveness of physical activity, with one saying, “I was in the national athletics team but still developed diabetes.” Chronic pain and fatigue were mentioned as additional reasons for not being active, as one participant stated, “Even with the sports I do, I still have complications, I feel sharp pain behind my back.”

Foot care: good foot care was observed in 63.0% of those with good glycemic control and 45.0% of those with poor glycemic control. This difference was statistically significant (p = 0.02). The importance of foot care practices was explained during the FGD, with the fifth theme: ´ hidden dangers´ that exposed how some participants (n=8) frequently had peripheral neuropathy, which hid early symptoms of foot injury, putting them at risk of complications such as ulcers. One participant shared, “You wear your flip-flops, and you walk, and it goes out, and you don´t realize it, they say it´s when the feet are without sensation.” Others also reported having experienced regular physical problems, including “The soles of my feet feel like they are burning with pepper” and “I sometimes feel vibrations in my feet, or sometimes it is very hot.

Smoking: good smoking self-care was observed in 98.0% of those with good glycemic control and 48.0% in those with poor glycemic control. This difference, however, was not significant (p > 0.90).

Overall self-care practices: among the participants with good glycemic control, 47.0 % had fair self-care practices, 37.0% had good self-care practices, and 16.0% were classified in the poor self-care practices. Among individuals with poor glycemic control, 52.0 % were defined as having fair self-care, with 22.0 % having good practices and 25.0 % having poor practices. These differences were not statistically significant (p = 0.08).

Logistic regression model for self-care management practices and glycemic control: poor dietary self-care and poor foot care were statistically significant among the self-care assessed. The likelihood of poor glycemic control was more than twice as likely in participants with poor dietary habits and poor foot care than in participants with good glycemic control. Other self-care practices, self-monitoring of blood glucose, physical activity, smoking, and overall self-care displayed no statistically significant relationships. All the participants had poor medication adherence (Table 3). The independent effect of self-care practices and their association with glycemic control was determined via multivariate logistic regression using the enter method, where all variables were introduced at once. This approach enabled us to determine the adjusted impact of the various self-care practices on glycemic control, adjusting for the impact of the others. The multivariate logistic regression was developed using self-care practices that were significant (foot-care and dietary practices) at the univariate logistic regression level (p < 0.05). Table 4 shows the results of the univariate and multivariate logistic regression. Poor dietary self-care was significantly associated with poor glycemic control (adjusted odds ratio (AOR) = 2.01; 95% CI: 1.05-3.94; p = 0.03) in the multivariable logistic regression analysis. Similarly, poor glycemic control was also significantly associated with poor foot care (AOR = 1.95; 95% CI: 1.03-3.76; p = 0.04). These results show that persons who had poor dietary habits or inadequate foot care were about two times more likely to report experiencing poor glycemic control than those who engaged in good self-care habits in these domains (Table 4).

Discussion     

In this study, we found that T2DM patients in Douala generally had poor glycemic control. Based on the multivariable analysis, independent associations were found between poor dietary practices and poor foot care practices and increased risks of poor glycemic control. The focus group discussions revealed that the participants had emotional pressure and diabetes-related complications, such as neuropathy, that limited their compliance with medical advice on diet, medication adherence, and glucose monitoring. The findings indicate that individuals who consume a nutritious diet are more than twice as likely to have good glycemic control. This agrees with the findings of studies conducted in China and the guidelines of the Centers for Disease Control and Prevention [16,17], which emphasize that nutrition should be central to managing diabetes as outlined in the diabetes self-management education and support guidelines. The focus group discussions provided useful data, such as barriers to acquiring recommended diets, which are costly and difficult when substituting carbohydrate staples like cassava. This corresponds with views suggesting that there is a need for consideration of both local food systems and cultural food practices when it comes to engaging in proper dietary self-care [18,19]. Dietary restrictions made many participants feel stressed and led them to frequently miss their favorite foods, which indicates that dietary advice that is too restrictive may be challenging to follow for some individuals. However, most of them realized that they had to change their diet, but they could not find simple methods to adhere to it permanently.

An association between foot care and glycemic control was observed, as good foot care contributes to alleviating complications and ensuring a good state of mind among T2DM patients [20]. Other studies demonstrated that teaching diabetic foot care to families adds value to patient care [21]. Focus group discussions found that many people experience numbness, but do not have much knowledge on how to care for their feet. This is consistent with research carried out in Cameroon, which showed that patients lacked proper foot care knowledge because they are poorly educated by healthcare staff while also struggling with economic hardship and illiteracy [22]. Many participants did not notice neuropathy-related injuries, which is in line with studies indicating that proper education on the care of the feet can reduce hospitalization and amputation [23]. The findings emphasize that the absence of sensation can cause even more injuries, and it is necessary to wear correct shoes and examine one´s feet regularly. We used mixed methods to enhance the understanding of the factors that are associated with glycemic control. Reliability and validity of the findings were strengthened by the use of tools such as the revised summary of diabetes self-care activities questionnaire. One weakness of this study is self-reporting, which may have introduced recall bias.

Another is social desirability bias, which could have compelled participants in the focus group discussions to deliver standard responses, thereby creating incorrect assessments regarding their self-care practice commitment and the obstacles they experienced. In addition, due to the study design, the results may not be generalized to all regions of Cameroon, as it limits causal inference. Furthermore, model fit indicators were not reported, restricting full evaluation of the logistic regression model's performance. However, AORs and CIs were provided, showing the strength and significance of the associations. Regarding the important differences in results, self-monitoring of blood glucose, recognized by the Centers for Disease Control and Prevention as crucial for glycemic control [16,17], did not have a significant effect on glycemic control in this study. Cost, unavailability of monitoring supplies, and insufficient awareness were the main challenges raised in the focus group discussions.

The findings in this study suggest the need for education, which involves both self-care and cultural factors. In addition, the qualitative findings suggested that for many, lack of knowledge made it difficult to check their feet regularly. This indicates that there is a need for educating people on regular screening and proper footwear. In addition, insufficient funds and cultural differences make it difficult to stick to eating properly. Therefore, community-based organizations are invited to collaborate with healthcare providers and government bodies to create culturally appropriate programs and strategies promoting a healthy diet and foot-care, and to promote and implement financial support programs that can help subsidize diabetic drugs, inexpensive glucose monitoring materials, and diabetes-friendly foods. These results also aligns with the Transtheoretical Model, which explains how individuals progress through different stages when making behavioral changes and underlines the need for tailored support according to their needs.

Conclusion     

The majority of people living with T2DM in Douala had inadequate glycemic control, implying that specific strategies would be vital. Foot care and dietary self-care turned out as strong predictors of glycemic outcomes. Physical activity and self-monitoring of blood glucose are still relevant aspects in clinical terms, despite their non-statistical significance in the multivariate model. Medication adherence was found to be poor across all participants. In the focus group discussions, participants stated that the cost of food, lack of knowledge on foot care, and restricted monitoring options were the main obstacles. These findings underscore the importance of implementing context-specific diabetes education and culturally appropriate self-care guidelines in low-resource settings. Future research should consider cohort studies incorporating assessments of T2DM self-care practice consequences on patient health outcomes to understand their combined effects on long-term glycemic control.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Ngenche Comfort Tangang, Shiphrah Kuria-Ndiritu, Jean Claude Katte, Nyagero Josephat: conception and design, acquisition of data, analysis and interpretation of data, drafting the manuscript, and critical revision of the article for important intellectual content. All authors have read and agreed to the final manuscript.

Acknowledgments     

We thank AMREF International University Kenya, for institutional support, the Fungwa and Tangang families for financial assistance, and all participants, healthcare providers, community members, and colleagues from the Pan African Organization for their support throughout the study.

Tables     

Table 1: sociodemographic and clinical characteristics of study participants living with type 2 diabetes mellitus, recruited from the outpatient units of Laquintinie Hospital and Presbyterian Health Complex Bepanda, Douala, Cameroon, from January to March 2025 (n = 230)

Table 2: distribution of self-care practices among individuals living with type 2 diabetes mellitus, recruited from the outpatient units of Laquintinie Hospital and Presbyterian Health Complex Bepanda, Douala, Cameroon, from January to March 2025 (n = 230)

Table 3: logistic regression of self-care management practices and glycemic control among individuals living with type 2 diabetes mellitus, recruited from the outpatient units of Laquintinie Hospital and Presbyterian Health Complex Bepanda, Douala, Cameroon, from January to March 2025 (n = 230)

Table 4: multivariable logistic regression model of self-care practices independently associated with glycemic control among individuals living with type 2 diabetes mellitus, recruited from the outpatient units of Laquintinie Hospital and Presbyterian Health Complex Bepanda, Douala, Cameroon, from January to March 2025 (n = 230)

References