Prevalence, determinants and consequences of delayed treatment intensification among type 2 diabetes mellitus patients at the University of Gondar Comprehensive Specialised Hospital, Northwest Ethiopia, 2024: a mixed-methods study

Abstract

Objective

To ascertain the clinical impact, prevalence and associated determinants of delayed treatment intensification, defined as delaying the escalation of treatment plans for individuals with type 2 diabetes mellitus who fail to attain ideal glycaemic control, at the University of Gondar Comprehensive Specialised Hospital in Northwest Ethiopia.

Design

A mixed-methods study.

Setting

University of Gondar Comprehensive Specialised Hospital.

Participants

420 patients with type 2 diabetes mellitus with poor glycaemic control after the index date were included in this study. A simple random sampling technique was employed to select the required sample size. Data were collected retrospectively and entered into EpiData V.4.6 and exported to Stata V.14.2 for analysis.

Method

Multivariable logistic regression was used to identify factors associated with delayed treatment intensification. A p value of 0.05 in the multivariable analysis was considered statistically significant. Qualitative data were collected through in-depth interviews with eight selected healthcare providers, and thematic analysis was undertaken to identify the underlying barriers to timely treatment intensification.

Primary outcome

Delayed treatment intensification.

Results

The prevalence of delayed treatment intensification was 51.4% (95% CI 46.6% to 56.2%), with a median delay of 14 months (IQR: 7.5–42 months) from the index date. Among those experiencing delayed treatment intensification, 43.1% developed new chronic diabetic complications, including retinopathy (18.1%), neuropathy (14.4%) and nephropathy (6.0%). Other complications (hypertension, stroke, heart failure and diabetic foot ulcer) accounted for 4.64% of the cases. Significant predictors of delayed treatment intensification included longer duration of diabetes (adjusted ORs (AOR) 1.68; 95% CI 1.13 to 2.5), presence of comorbidities (AOR 1.83; 95% CI 1.04 to 3.2) and use of cardioprotective medications (AOR 1.59; 95% CI: 1.04 to 2.43). The qualitative findings revealed additional barriers contributing to delayed treatment intensification, including financial limitations, insufficient patient awareness and non-adherence among patients. Additionally, healthcare provider-related factors, including professional fatigue and knowledge gaps, as well as health institution-related factors such as inadequate healthcare infrastructure.

Conclusions

This study found a high prevalence of delayed treatment intensification (51.4%), associated with comorbidities, longer disease duration, low patient awareness, cardioprotective drug use and barriers related to the system and providers. To address these gaps, priorities should include strengthening patient education, scheduling regular reviews for high-risk patients and improving clinical decision support tools for timely treatment intensification. Enhancing healthcare infrastructure, such as medication supply and diagnostic services, and offering refresher training to reduce provider fatigue, are also crucial for improving the delivery of diabetes care.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• We had a large sample size.

• We used a mixed-methods design.

• We mainly used fasting blood sugar because glycated haemoglobin was not always available.

• The study was conducted only at one hospital, so the results might not be generalised.

• The quantitative part of the study design was retrospective in nature.

Introduction

Diabetes mellitus (DM) is a chronic metabolic disorder characterised by elevated blood glucose levels.¹ According to the International Diabetes Federation’s 2021 report, approximately 536.6 million people worldwide are living with diabetes, and this number is expected to rise to 783.2 million people by 2045.² Type 2 DM (T2DM) is a complex, progressive disease that significantly impacts healthcare utilisation, medical expenditure and global mortality rates, necessitating continuous medical management.³ Current international guidelines, including those from the American Diabetes Association (ADA) and European Association for the Study of Diabetes, recommend a stepwise approach to managing hyperglycaemia in patients with type 2 diabetes.⁴

Glycaemic control is considered poor if fasting blood sugar (FBS) is >152 mg/ dL, comparable to glycated haemoglobin (HbA1c) of 7%.⁵ Despite the availability of various antihyperglycaemic medications and evidence-based treatment guidelines, poor glycaemic control persists due to delays in treatment intensification (TI), whereby clinicians fail to adjust therapy in a timely manner when patients remain uncontrolled.⁶ TI is defined as either increasing the dose of an existing medication, adding additional oral antidiabetic agents or initiating insulin therapy.⁷ Failure of clinicians to intensify therapy in line with evidence-based clinical guidelines is referred to as therapeutic inertia.⁸ Timely and adequate TI plays a critical role in achieving target HbA1c or fasting blood glucose levels, thereby reducing the risk of diabetes-related complications.⁹ Factors affecting adequate TI include polypharmacy, lifestyle and health institution constraints.

Delays in TI are associated with prolonged hyperglycaemia, increasing the risk of diabetes-related complications such as cardiovascular disease, neuropathy, nephropathy and retinopathy.¹⁰ Studies have shown that patients with prolonged exposure to uncontrolled glycaemia are associated with increased hospitalisations, greater healthcare costs and reduced quality of life.¹¹ Additionally, delayed TI has also been linked to a higher risk of glucose toxicity, where chronic hyperglycaemia worsens insulin resistance and β cell dysfunction, making future treatment escalation less effective.¹² The rising global prevalence of delays in optimising treatment for T2DM presents significant health and economic challenges, particularly in low-income and middle-income countries (LMICs).¹⁰ In Ethiopia, suboptimal therapy remains common, with physician hesitation, poor patient follow-up and financial constraints being major contributors to treatment delays.¹³ These delays not only diminish patients’ quality of life for individuals but also lead to increased healthcare costs due to complications arising from poorly managed diabetes.¹⁴ Therapeutic ineffectiveness due to delayed TI is a major challenge in managing hyperglycaemia in T2DM, leading to poor glycaemic control and increased risk of complications. Evidence shows that many T2DM patients experience poor glycaemic control after diagnosis, with delays in TI ranging from 0.3 to 7.2 years.³ Globally, 26.2%–70% of patients with T2DM face such delays.¹⁵ In Ethiopia, one study reported that more than half of the patients had an unchanged medication regimen despite inadequate glycaemic control.¹⁶ The ADA recommends an HbA1c target of less than 7% for optimal diabetes care, but only 20% of patients achieve this goal¹⁷ to 17.6% in Ethiopia.¹⁸ Among patients on a single oral anti-hyperglycaemic drug (OAD), the median time to TI after exceeding glycaemic targets ranges from 0.3 to 2.7 years.¹⁹ A retrospective cohort study highlighted a consistent delay in TI across all stages of treatment, starting from the first OAD and continuing until the initiation of insulin, contributing to the progression of complications and reduced quality of life.²⁰ A study conducted in the USA found that nearly 50% of patients with T2DM with uncontrolled hyperglycaemia experienced delays of more than 1 year before TI, leading to worse long-term outcomes.²¹ Similarly, research from the UK and Germany showed that physicians often delay initiating insulin therapy, even when glycaemic targets are not met, due to patient reluctance and clinical uncertainty.²² In Africa, resource limitations, medication stockouts and inadequate patient education contribute significantly to delayed intensification.

A study from South Africa reported that over 60% of patients requiring insulin initiation faced significant delays due to stock shortages and poor patient education.²³ Despite extensive global research,^{24 25} limited evidence exists regarding the prevalence, determinants and consequences of delayed TI in Ethiopia. Given the rising burden of poor glycaemic control and the challenges faced by the Ethiopian healthcare system, context-specific evidence is urgently needed to inform clinical practice and policy. This study aims to fill this critical gap by examining the prevalence, causes and outcomes of delayed TI among patients with T2DM at the University of Gondar Comprehensive Specialised Hospital (UoGCSH). The findings will provide critical insights for optimising diabetes management and reducing the burden of complications in Ethiopia.

Methods and materials

Study design and period

This study employed an explanatory sequential mixed-methods design that combined both quantitative and qualitative approaches. The quantitative component used a hospital-based retrospective follow-up study to assess the prevalence, effect and predictors among patients with T2DM. The qualitative component adopted an interpretative phenomenological approach to explore additional underlying factors contributing to delayed TI from the perspectives of healthcare providers. The study was conducted over a four-month period from 1 July to 1 November 2024, with the quantitative part taking place from 1 July to 1 September 2024, and the qualitative part being carried out from 5 October to 1 November 2024 (see figure 1).

Figure 1. A visual expression of the study design and timeline.

Study area

The study was conducted at UoGCSH, located in Gondar town, Amhara region, northwest Ethiopia (see figure 2). Gondar lies approximately 738 km from Addis Ababa, the capital city of Ethiopia, and 173 km from Bahir Dar, the regional capital of Amhara. UoGCSH is a government-affiliated teaching hospital that provides comprehensive healthcare services, including preventive, diagnostic, curative and rehabilitative care, to millions of residents in the region. The hospital comprises 27 wards, 14 outpatient units and 960 inpatient beds. The hospital currently serves over 8000 patients with diabetes, of whom approximately 4300 are actively followed up for T2DM. It offers services such as medication supply and routine follow-up.

Figure 2. Map of the study area.

Population

Source population

For the quantitative component, the source population comprised all adult patients who were on antidiabetic medications and attending follow-up visits at UoGCSH.

For the qualitative component, the source population comprised healthcare providers working in the chronic outpatient department (OPD) at the UoGCSH.

Study population

For the quantitative component, the study population comprised adult patients with T2DM who had follow-up visits at UoGCSH from 2018 to 2024 and met the inclusion criteria.

For the qualitative component, key informants, including healthcare providers with direct experience in managing patients with T2DM at the chronic OPD of UoGCSH, were available during the data collection period.

Eligibility criteria 

Inclusion criteria

For the quantitative component of the study, patients were eligible if they were adults diagnosed with T2DM, had been receiving antidiabetic medications for at least 3 months, had a fasting blood glucose measurement above 152 mg/dL and had attended follow-up visits at UoGCSH between 2018 and 2024.

For the qualitative component of the study, healthcare providers working in the chronic OPD at UoGCSH who were selected as key informants were included.

Exclusion criteria

For the quantitative component of the study, patients were excluded if they had incomplete medical records (such as missing diagnosis date, type of diabetes or fasting blood glucose), had any chronic diabetic complications prior to the index date, were pregnant or breastfeeding or had multiple comorbidities with a Charlson Comorbidity Index (CCI) score above 3.

For the qualitative component of the study, healthcare providers working in the chronic OPD who were unwilling to participate as volunteers or who did not have sufficient detailed knowledge regarding diabetes management were excluded.

Sample size determination and sampling procedure

Sample size determination

The sample size required for the quantitative component was calculated using a formula for a single population proportion with the assumption of a 5% margin of error and a 95% confidence level. The sample size was calculated by considering the estimated proportion of patients who were found to have therapeutic inertia (53.8%), obtained from the previous study in Ethiopia¹⁶ and by adding a 10% non-response rate, corresponding to an expected response rate of 91%, the final sample size was 420 patients with type 2 diabetes who have follow-up from 2018 to 2024, at the UoGCSH.

$${\displaystyle {\displaystyle n=\frac{{\left(z_{\frac{a}{2}}\right)}^2\times p\times \left(1-p\right)}{d^2}=\frac{{\left(1.96\right)}^2\times 0.538\times \left(1-0.538\right)}{{\left(0.05\right)}^2}=382}}$$

The sample size for the qualitative part was not computed. The data were collected until the information for the study was saturated.

Sampling technique and procedure

Quantitative component

Eligible patients with T2DM who attended follow-ups at the UoGCSH from 2018 to 2024 were selected using diabetic follow-up logbooks as a sampling frame. The sampling frame comprised 603 eligible patients. A simple random sampling technique was employed to select the required sample using the lottery method. Each eligible patient’s medical registration number was placed in a container, and patients were selected randomly until the required sample size was reached. Medical charts of the selected patients were thoroughly reviewed to extract relevant data.

Qualitative component

Key informants were selected purposively to ensure the inclusion of healthcare providers with direct experience managing patients with diabetes. This targeted approach enabled a deeper exploration of insights relating to delayed TI.

Study variables

Dependent variable.

Delayed TI.

Independent variables.

Patient-related variables: Age, gender, income, duration of DM, comorbidities, financial constraints, lack of awareness and treatment non-adherence.

Healthcare provider-related variables: Fear of polypharmacy effect on adherence, lack of familiarity with guidelines, specialty of doctors, lack of team-based care, lack of information on new treatment options, fear of potential risks of hypoglycaemia, lack of training, professionals’ negligence and fatigue.

Healthcare-system-related factors: Time constraints, bureaucratic difficulties with new drugs, absence of clinical guidelines, lack of continuity of care, lack of structured educational programmes, shortage of resources and cost of newer medications.

Operational definitions

Index date: The first date on which a fasting blood glucose measurement exceeded 152 mg/dL following the initiation of antidiabetic treatment.⁴

Poor glycaemic control: FBS >152 mg/dL, which corresponds to an HbA1c of 7% according to the ADA Standards of Medical Care in Diabetes.⁵ FBS was used as a follow-up tool in our setting because HbA1c is not consistently available for patient diagnosis and monitoring.

Short duration: Diabetes duration less than 5 years.¹⁶

Long duration: Diabetes duration ≥5 years.¹⁶

Incomplete record: If important data such as fasting blood glucose and date of DM diagnostic care are not recorded.

Fasting: no caloric intake for at least 8 hours.

Comorbidities: Patients with T2DM and other chronic medical or psychiatric disorders.

Comorbidities score: Mild, CCI scores of 1–2; moderate, CCI scores of 3–4 and severe, CCI score >5.²⁶

TTTI: Calculated by subtracting the index date from the first date of TI.²⁷

Early intensifiers: Those intensifying within 6 months after the index date.²⁷

Delayed intensifiers: Those intensifying more than 6 months after the index date.²⁷

Chronic diabetic complications: A patient with diabetes who has at least one confirmed macrovascular complication by a physician (diabetic-related cardiovascular disease, peripheral vascular disease and stroke) or microvascular complications (including diabetic nephropathy, diabetic retinopathy, peripheral neuropathy and others) after being diagnosed with diabetes.²⁸

Data collection tools, procedures and quality management

Quantitative component

Data were collected using a structured data extraction tool. This tool captured information on patient demographics, clinical characteristics and treatment details from medical charts. Prior to the main data collection, the tool was pretested on 5% (21) of eligible patients at the UoGCSH, who were not included in the main study, and the contents were modified accordingly. The data collection team comprised one clinical pharmacist as supervisor, two clinical pharmacists as data collectors and two medical record clerks responsible for selecting the patient charts. All team members received training to ensure consistency and accuracy during data collection. The collected data were checked daily for completeness and correctness.

Qualitative component

Data were collected through in-depth interviews using a semistructured interview guide. An interview guide was developed by the investigator, focusing on healthcare professionals’ experiences of TI (online supplemental file 1). The interview guide was translated into Amharic to facilitate better communication with the participants. Interviews were conducted from 5 to 18 October 2024 and were audio recorded after obtaining verbal consent from participants. The interviews were held in private and convenient locations to encourage open dialogue. Notes were also taken to complement the recordings. Data collection was continued until thematic saturation was reached, meaning no new themes or insights emerged from additional interviews. The interviewer stopped when three interviews in a row did not give any new codes or topics, and when everything already seemed solid and complete. Through this procedure, the interviewer was able to decide that data collection could end because more interviews were unlikely to yield new insights.

Handling of the missing data

The study employed a rigorous data-handling technique, including structured questionnaires and thorough quality control procedures. This involved interviewer training, systematic completeness checks and instrument pretesting. Statistical analysis was conducted using tools such as STATA, and data entry and cleaning were handled using specialised software. Missing data were addressed using maximum likelihood estimation and multiple imputation.

Data processing and analysis

Quantitative component

Data were coded, cleaned and checked for completeness and consistency before being entered into Epi data V.4.6 and exported to Stata V.14.2 for analysis. Descriptive statistics were used to summarise the socio demographic, clinical and treatment-related characteristics. Continuous variables were presented as means, medians and IQRs, while categorical variables were reported as frequencies and percentages. Comparisons of categorical variables were conducted using χ² tests, and the Mann-Whitney U test was applied for continuous variables that were not normally distributed. To identify the factors associated with delayed TI, both bivariable and multivariable logistic regression analyses were conducted. Variables with a p value less than 0.25 in the bivariate analysis were included in the multivariate analysis.^{29 30} Adjusted ORs (AORs) with 95% CIs were calculated to estimate the strength of the associations. A p value of less than 0.05 in the multivariate analysis was considered statistically significant.

Multicollinearity was checked using the variance inflation factor (VIF=1.00–1.03, mean VIF=1.02), indicating no evidence of multicollinearity. The model goodness of fit test was also checked using the Hosmer-Lemeshow goodness of fit test, and the result was found to be p=0.2867.

Qualitative component

Audio recordings from the interviews were transcribed verbatim, translated and analysed thematically using Open Code software (V.4.03). Key themes and subthemes were generated, reflecting patient-level, provider-level and system-level barriers to timely TI.

Trustworthiness

To ensure the trustworthiness of the study, prolonged engagement with participants was undertaken. We also checked our first results with some of them to see if they agreed. Plus, we kept track of all our choices during the study. The team was also aware of its own biases and tried to keep them in check.

Patient and public involvement

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

Results

Sociodemographic and clinical characteristics

A total of 420 medical records of patients with T2DM who experienced poor glycaemic control following the index date, defined as the first date on which fasting blood glucose measurement exceeded 152 mg/dL following the initiation of antidiabetic treatment, were analysed. Of these, 204 (48.6%) patients received early TI, whereas 216 (51.4%) experienced delayed TI. The median age of the study participants was 55 years (IQR: 46–60). Among all participants, 224 (53.3%) were female, and the majority, 315 (75%), were urban residents. A total of 242 (57.6%) patients had a diabetes duration of 5 years or more. Approximately 66 (15.7%) participants had one or more comorbid conditions. Most patients, 315 (75%), were prescribed metformin monotherapy at the index date, while the remainder received insulin or other OADs, either as monotherapy or in combination (metformin+glibenclamide). Additionally, 137 (32.6%) participants were receiving cardioprotective medications (table 1).

Table 1. Sociodemographic and clinical characteristics of patients with T2DM followed from 2018 to 2024 at the University of Gondar Hospital, Northwest Ethiopia, 2024 (N=420).

Variables	Variable categories	All samples (N=420)	Early intensification (n=204)	Delayed treatment intensification (n=216)
Age (median, IQR)		55 (46, 60)	55 (44, 60)	55 (48, 59.5)
Sex, (%)	Male	196 (46.7%)	86 (42.2%)	110 (50.9%)
	Female	224 (53.3%)	118 (57.9%)	106 (49.1%)
Residence	Urban	315 (75%)	151 (74.0%)	164 (75.9%)
	Rural	105 (25%)	53 (26.0%)	52 (24.1%)
Duration of diabetes (years)	<5	178 (42.4%)	102 (50%)	76 (35.2%)
	>5	242 (57.6%)	102 (50%)	140 (64.8%)
Antidiabetic medication used on index date, %	Metformin	315 (75%)	162 (79.4%)	153 (70.8%)
	Glibenclamide	12 (2.9%)	5 (2.5%)	7 (3.2%)
	Insulin	49 (11.7%)	20 (9.8%)	29 (13.4%)
	Metformin+glibenclamide	44 (10.5%)	14 (6.9%)	30 (13.9%)
Comorbidities on the index date	Yes	66 (15.7%)	43 (19.9%)	23 (11.3%)
	No	354 (84.3%)	173 (80.1%)	181 (88.7%)
Use-cardioprotective drugs on the index date	Yes	137 (32.6%)	82 (38.0%)	55 (27.0%)
	No	283 (67.4%)	134 (62.0%)	149 (73.0%)

T2DM, type 2 diabetes mellitus.

Prevalence of delayed TI

Patients with uncontrolled blood glucose levels of those who did not receive any treatment changes within 6 months after documented poor glycaemic control (index date) were assessed for delayed TI. Among the 420 patients with poor glycaemic control, 216 (51.4%; 95% CI 46.6% to 56.2%) did not receive treatment changes within 6 months of the index date, meeting the criteria for delayed TI (figure 3). Of these: 104 (48.2%) had no initiation of insulin, 68 (31.5 %) did not receive a second OAD, 42 (19.4%) had no dose adjustment and 2 (0.9%) did not have frequency modified (table 2).

Figure 3. Prevalence of delayed treatment intensification among patients with type 2 diabetes followed from 2018 to 2024 at the University of Gondar Hospital, Northwest Ethiopia, 2024.

Table 2. Prevalence of types of delayed treatment intensification among patients with type 2 diabetes followed from 2018 to 2024 at the University of Gondar Hospital, Northwest Ethiopia, 2024.

Types of delayed treatment adjustment	Number of patients	Percentage
Failure to initiate insulin	104	48.2
Failure to add a second OAD	68	31.5
Failure to increase the dosage of a drug	42	19.4
Failure to increase the frequency of a drug	2	0.9

OAD, oral antihyperglycaemic drug.

Time to TI

The time to TI was calculated by subtracting the index date from the first date of TI. For participants who experienced delayed TI, the median time was 14 months (IQR: 7.5–42), whereas for patients who received early TI, it was 1 month (IQR: 1–2.04) after the index date. The Mann-Whitney U test revealed a Z value of 17.821 and a p value of 0.0000 (p<0.001), as shown in online supplemental table 1. This indicates a strong association between delayed TI and prolonged time to intervention, underscoring the need for timely adjustments in diabetes management.

Prevalence of chronic diabetic complications in patients with delayed

This study showed that, among patients with delayed TI, 43.1% developed at least one chronic diabetic complication, compared with 23.5% in the early TI group (p<0.001). This suggests a strong association between delayed TI and complication burden. Furthermore, the OR for patients in the delayed group was 2.28 (95% CI 1.48 to 3.53), indicating that they were approximately 2.28 times more likely to develop complications compared with those in the early group (online supplemental table 2). The most prevalent chronic complications among patients with delayed TI were retinopathy (18.1%), neuropathy (14.4%) and nephropathy (6.0%). Other complications (hypertension, stroke, heart failure and diabetic foot ulcer) accounted for 4.6% of the cases (see figure 4). Factors associated with delayed TI.

Figure 4. Prevalence of chronic diabetic complications in patients with delayed treatment intensification among patients with type 2 diabetes followed from 2018 to 2024 at the University of Gondar Hospital, Northwest Ethiopia, 2024.

Factors associated with delayed TI

Quantitative component

To quantitatively assess the factors associated with delayed TI, both bivariable and multivariate logistic regression analyses were conducted. Variables including age, sex, duration of T2DM diagnosis, presence of comorbidities and taking cardioprotective drugs had a p value of less than 0.25 in the bivariable analysis. However, only the duration of T2DM diagnosis, presence of comorbidities and taking cardioprotective drugs were significant in the final multivariable logistic regression analysis.

Patients with a diabetes duration of ≥5 years had 1.68 times higher odds of delayed intensification (AOR 1.68; 95% CI 1.13 to 2.5). Those with comorbidities and those on cardioprotective drugs were also significantly more likely to experience delays (AORs=1.83 and 1.59, respectively) (table 3). The full model of the association is shown in online supplemental table 3.

Table 3. Multivariable logistic regression showing factors associated with delayed treatment intensification among patients with type 2 diabetes followed from 2018 to 2024 at the University of Gondar Hospital, Northwest Ethiopia, 2024.

Variables	Variable categories	Delayed-treatment intensification		OR (95% CI)		P value
		Yes (%)	No (%)	COR (95% CI)	AOR (95% CI)
Age (median, IQR)		55 (48, 59.7)	55 (44, 60)	1.02 (0.99 to 1.04)	1.00 (0.99 to 1.04)	0.247
Gender	Male	110 (50.9)	86 (42.2)	1.00	1.00
	Female	106 (49.1)	118 (57.8)	0.7 (0.48 to 1.03)	0.75 (0.5 to 1.11)	0.150
Duration of T2DM	<5	76 (35.2)	102 (50)	1.00	1.00
	≥5	140 (64.8)	102 (50)	1.84 (1.25 to 2.73)	1.68 (1.13 to 2.5)	0.011*
Comorbidities	No	173 (80.1)	181 (88.7)	1.00	1.00
	Yes	43 (19.9)	23 (11.3)	1.96 (1.13 to 3.38)	1.83 (1.043 to 3.2)	0.035*
Cardioprotective drug use	No	134 (62.0)	149 (73.0)	1.00	1.00
	Yes	82 (38.0 %)	55 (30.0)	1.66 (1.096 to 2.5)	1.59 (1.04 to 2.43)	0.033*

^*Variables with statistically significant p value.

AOR, adjusted OR; COR, crude OR; T2DM, type 2 diabetes mellitus.

Qualitative component

The qualitative component of this study aimed to explore the factors contributing to delayed TI among patients with type 2 diabetes who are not easily captured through quantitative data by interviewing eight healthcare professionals (table 4). Patients-related, providers-related and healthcare system-related factors were identified as the primary influences linked to the delayed intensification of treatment.

Table 4. Sociodemographic characteristics of the participants interviewed working at the chronic outpatient department of UoGCSH, Gondar, Ethiopia, 2024, (N=8).

Sociodemographic characteristics	Number (8)
Sex
Male	5
Female	3
Year of practice at the diabetic clinic
<5 years	8
5–10 years	_
>10 years	_
Type of profession
Physician	6
Nurse	2
Academic qualification
BSc nurse	2
Doctor of medicine/bachelor’s degree	5
Internal medicine specialist	1

UoGCSH, University of Gondar Comprehensive Specialised Hospital.

Thematic areas

Identifying the predictors of delayed TI among patients with type 2 diabetes was categorised with three themes and nine subthemes (table 5).

Table 5. Themes and subthemes developed from the data on predictors for delayed treatment intensification among patients with type 2 diabetes at the University of Gondar Hospital, Northwest Ethiopia, 2024.

Themes	Subthemes
Patient-related factors	Financial constraints Lack of awareness and understanding about the disease and its treatment Non-adherence to the treatment
Healthcare provider-related factors	Knowledge gaps Lack of collaborative care Professional fatigue and negligence
Health institution-related factors	Limited access to health education Inadequate healthcare infrastructure Shortage of resources

Theme 1: patient-related factors

This theme highlights barriers resulting from patients’ social and personal situations that contributed to delayed TI. Under this theme, three subthemes were identified: financial constraints, lack of awareness and understanding of the disease and its treatment, and non-adherence to treatment.

Subtheme 1: financial constraints

Interviewed participants reported that many patients face out-of-pocket expenses for medications and diagnostic tests. The cost of insulin and other medications can lead to delayed treatment adjustments, worsening the patient’s condition and leading to more severe complications.

Inability to afford

Many patients struggle with the cost of medications and diagnostic tests (such as HbA1c). Even when I recommend adjusting their treatment, they often delay or avoid making changes because they simply cannot afford it. Especially insulin is expensive. It depends on the brand you are buying. Unfortunately, this can intensify exacerbate the patient’s condition and lead to more severe complications. M, General Physician.

Subtheme 2: lack of awareness about the disease and treatment

One of the critical factors contributing to delayed TI among patients with diabetes was their negative beliefs about the efficacy of anti-hyperglycaemic drugs, fear of injections from starting insulin therapy and poor knowledge on the progressive nature of T2DM.

Negative beliefs

Patients believe that all anti-hyperglycaemic drugs are the same in efficacy.

Many patients think that it does not matter which diabetes medication they take because they believe they all work the same way. This can lead to frustration when they do not see the expected results, and they might not be open to changing their treatment when necessary. M, General Physician.

Fear of injections

Fear of injections is a significant barrier to TI for many patients with T2DM, particularly those who require insulin therapy or injectable medications.

I have observed many patients who refuse to start insulin therapy simply because they are terrified of needles. Even when they need it, that fear keeps them from getting the care they need. F, General Physician.

Limited diabetes knowledge

Many patients did not fully understand the importance of timely adjustments in their medication regimen.

Majority of patients think that once they are prescribed a medication, they do not need to change it, regardless of how their condition progresses. They often do not realize that diabetes is a progressive disease that requires ongoing treatment adjustments to prevent complications. F, General Physician.

Subtheme 3: treatment non-adherence

Many patients who were already on anti-hyperglycaemic drugs were non-adherent to their prescribed treatment plans due to missing follow-up appointments.

Missing follow-up appointments

Most healthcare providers emphasised the challenges posed by patients missing their scheduled follow-up appointments. This non-adherence causes patients to remain on the initial treatments without modifications, hindering effective diabetic management.

In my earlier years of practice, most patients had a problem with not being present at their scheduled appointments. They would miss the follow-ups, and without these visits, we could not adjust their medications as needed. F, Nurse.

‘The follow-up issue caused patients to remain on their initial anti-hyperglycaemic drugs without necessary adjustments, even when the intensification of treatment was warranted.’ M, General Physician.

Theme 2: healthcare provider-related factors

This theme reflects how healthcare provider-related factors knowledge leads to delays in TI. Under this theme, three subthemes were identified: knowledge gaps, lack of collaborative care and professionals, lack of proactive management and fatigue. Gaps in training and reliance on outdated information limited providers’ ability to make medication adjustments early. Weak collaboration and communication between professionals further worsened the failure to intensify medications early. In addition, negligence, lack of proactive management and fatigue from heavy patient loads contributed to a ‘wait and see’ approach, delaying timely TI.

Subtheme 1: knowledge gaps

Participants believed that inadequate training and outdated information about diabetes management led to underestimating the importance of regular monitoring and the need for timely adjustments in treatment plans.

Outdated information

‘It is critical for us as healthcare providers to stay informed about the latest guidelines and treatment options.’ I have observed, some clinicians who used references which were outdated “M, General Physician.

Lack of training

The participants’ major concern now is the absence of continuous and adequate training on the current updated treatment practice and protocol about diabetes.

Lack of continuous and adequate training on updated treatment practice can hinder our ability to recognize when a patient’s condition necessitates intensification, resulting in missed opportunities for intervention. M, Nurse.

Subtheme 2: lack of collaborative care

All participants reported that poor professional integration and communication among healthcare providers hindered effective diabetes management.

Poor professional integration

Managing diabetes requires a team effort, but in our setup, there is often a disconnect between healthcare providers. I have seen patients miss out on proper care because their treatment is not well coordinated. If the integration between general practitioners, specialists, and diabetes educators were stronger, we would be able to adjust treatment regimens more effectively and prevent complications. M, nurse.

Poor communication

Physicians, mainly the residents, lack communication skills that must interact with patients regarding the disease and the required medications.

During the initiation of insulin therapy, I provide patients with a concise overview of the symptoms they might experience and basic strategies for managing hypoglycaemia. But nurses gave the details. M, General Physician.

Subtheme 3: professionals’ lack of proactive management and fatigue

Study participants shared their deeply troubling experiences on predictors for delayed TI. Professionals’ negligence and fatigue due to high patient flows were the most prevalent factors explained by most participants.

Professionals’ lack of proactive management

The most mentioned factor for delayed TI, as reported by the participants, is the healthcare providers being negligent.

In some cases, giving a soft excuse for the high blood glucose level may overshadow the need for long-term management of T2DM. This can lead to a “wait and see” approach, where treatment intensification is postponed until more serious complications arise, potentially compromising patient outcomes. M, General Physician.

Overload from high patient flow

Healthcare providers may become fatigued or overwhelmed by high patient flows, resulting in missed opportunities for timely intervention.

We see so many patients in a day, and sometimes it’s easy to overlook the need for treatment adjustments, especially if the patient seems stable at first glance. Fatigue plays a role, and unfortunately, that can mean we do not always intensify treatment when it is needed. M, General Physician.

Theme 3: healthcare system-related factors

Healthcare providers have shared their deeper concerns about the systemic barriers that often delay essential treatments for patients. Many of them noted that limited access to structured health education, inadequate infrastructure and a shortage of crucial resources, like diagnostic tests and medications, leave patients feeling confused and unsupported in managing their diabetes. As a result, many patients find themselves stuck in a cycle of poor glycaemic control, which can lead to serious complications.

Subtheme 1: limited access to health education

The study participants shared and expressed the lived experience of the deep impact of the lack of public diabetic education programmes and limited resources for patient education sessions on TI.

Lack of public diabetic education programmes

Public education programs about diabetes are far too limited. Many patients are unaware of the severity of the condition or how it develops over time. Patients would be more aware of the dangers of postponing care and the necessity of frequent treatment adjustments if public health campaigns and community initiatives were more widely implemented. M, General Physician.

Limited resources for patient education sessions

One issue is the lack of sufficient educational resources on the general concepts of diabetes and its treatment that are prepared for the patient in the local language.

One doctor noted, ‘Majority of patients lack access to appropriate diabetes education resources. They found it difficult to understand the significance of modifying their treatment in the absence of a clear direction.’

Subtheme 2: inadequate healthcare infrastructure

High patient loads, insufficient time and lack of specialised rooms for diabetic care impede individualised attention and effective treatment.

Inadequate staff for diabetes care and education

Many healthcare professionals expressed concern about the challenges posed by high patient loads and inadequate staffing for diabetes care and education. They noted that these constraints make it difficult to provide individualized attention, leaving patients without a clear understanding of their treatment options and the importance of timely adjustments.

Time constraints

In particular, the initiation of insulin was found to be delayed by the lack of sufficient time for follow-up, some residents who rotate frequently as part of their residency programme, as well as their avoidance of responsibility.

In actuality, residents are coming to the diabetes clinic for a brief period as part of their rotation, and they are unwilling to take on the duty of starting insulin therapy during this time. All they want to do is see the patient and send them home with their current prescription. M, Specialist in Internal Medicine (CSIM).

In my experience, the number of patients seen per day is high since most diabetic patients come here as UoGCSH is known to have a better supply of medicines. Nurse, F.

Lack of specialised rooms for diabetic care

The absence of a specialized room for diabetic care in our set-up results in delays in treatment intensification because there is no suitable environment for focused care, particularly for complex interventions such as insulin initiation. M, General Physician.

Subtheme 3: shortage of resources

Absence of HbA1c tests and lack of access to essential diabetic medicines delays necessary adjustments to the treatment plan.

Absence of the HbA1c test

According to the healthcare providers interviewed, ‘Patients cannot afford or obtain the HbA1c test, and it is not available at the hospital laboratory. As a result, the doctors are forced to base their decisions about adjusting the therapy solely on the FBG results.’

Shortages of diabetic medicines

Lack of access to diabetic medicines is a significant barrier to timely and effective TI.

Healthcare professionals highlighted the lack of access to essential diabetic medicines as a major barrier to timely and effective treatment intensification. ‘They noted that many patients face challenges due to medication shortages or prohibitive costs, which delays necessary adjustments to their treatment plans’.

Discussion

This mixed-methods study assessed the prevalence of delayed TI among individuals with T2DM, identified contributing factors and evaluated its association with the development of chronic diabetic complications. The results revealed that over half (51.4%) of patients with uncontrolled blood glucose did not receive timely treatment adjustments. This delay was associated with significantly higher rates of chronic complications, particularly retinopathy, neuropathy and nephropathy. Both quantitative and qualitative data identified patient-level, professional-level and system-level factors contributing to these delays.

In this study, the prevalence of delayed TI among patients with type 2 diabetes was 51.4% with 95% CI of (46.6% to 56.2%), which is similar to findings from Malaysia (53.6%)³¹ and Jimma, Ethiopia (53.8).¹⁶ However, it is higher than reports from Taiwan (38.5%),³² France (36.6%)³³ and Thailand (26.2%).³⁴ This discrepancy may be potentially attributed to differences in healthcare infrastructure, treatment protocols and follow-up periods, as our study had a longer follow-up of 7 years compared with 5 years in some of the other studies, potentially capturing more cases of delayed intensification. Similarly, this study also revealed that the highest proportion of people with delayed TI was due to failure to insulin initiation (48.2%). The median time to TI among patients who experienced delays was 14 months, consistent with a study from Sydney (14.4 months) but longer than the 8.1 months reported in Japan.¹⁴ The extended duration observed in our study may reflect infrastructural limitations and staffing shortages, which were also highlighted in the qualitative component. In contrast, the shorter delay compared with the UK study (2 years) could be attributed to differences in patient population characteristics and comorbidity burden.²⁶

Patients who experienced delayed TI were significantly more likely to develop chronic diabetic complications (43.1%) compared with those who received an early intensification. This prevalence was notably higher than the rates reported in San Antonio (35%)³⁵ and Spain (28.1%),³⁶ suggesting a strong association between prolonged poor glycaemic control and complication development. Although our design was a retrospective cohort, it provides temporal evidence that prolonged uncontrolled hyperglycaemia is associated with a greater risk of complications. This is similar to evidence, including the UK Prospective Diabetes Study (UKPDS) and Diabetes Control and Complications Trial (DCCT), which stated the role of hyperglycaemia in the development of microvascular complications.^{37 38}

The observed prevalence of microvascular complications (retinopathy, 18.1%; neuropathy, 14.4% and nephropathy, 6.0%) was lower than that reported in Switzerland, possibly due to differences in the sample size or disease severity.²⁶

The prevalence rates of diabetic foot ulcer, heart failure, stroke and hypertension among patients with T2DM who experienced delayed TI were 0.9%, 0.9%, 1.4% and 1.4%, respectively. These findings were lower than a study conducted in Spain, which reported prevalence rates of 14.4% for heart failure and 6.8% for stroke, respectively.³⁹ This discrepancy may be attributable to the high prevalence of delayed TI, which was approximately 60% in the study conducted in Spain, which could lead to prolonged hyperglycaemia and an increased risk of these complications.¹⁴ Furthermore, delays in treatment may prevent the timely use of more effective therapies that could reduce these risks.⁴⁰

The study identified that longer diabetes duration (AOR=1.68), presence of comorbidities (AOR=1.83) and use of cardioprotective drugs (AOR=1.59) were independent predictors of delayed intensification. A long duration of diagnosis may create a perceived sense of stability, which leads to delayed TI through misinterpretation of stable glycaemic levels. Communication gaps between patients and providers may reinforce the false perception of stability and the need for intensification. The qualitative component of this study also supports this possible reason as poor communication between patients and healthcare providers hinders effective and timely diabetes management.³

Comorbidities increase management complexity, as healthcare providers are reluctant to intensify treatment in a timely manner due to potential risks and patients may experience a higher psychological burden or fear of side effects.^{4 41} Similarly, taking cardioprotective drugs can also contribute to delayed TI, as balancing cardiovascular protection with glucose control complicates decision-making, and patients who feel well or notice improvements in their health status may perceive their current regimen as sufficient, leading them to resist further treatment adjustments.³⁵ The qualitative component of this study also supports this reason as patients with poor diabetic knowledge often do not realise that diabetes is a progressive disease that requires ongoing treatment adjustments.

The qualitative findings complemented the quantitative results by revealing the underlying causes for delayed TI. Patient-related factors, including financial constraints, lack of awareness and understanding of the disease and its treatment, and non-adherence to treatment, help explain why patients with longer disease duration or comorbidities may resist treatment adjustments, supporting quantitative associations.^{3 42}

Similarly, healthcare provider-related factors, including knowledge gaps, professional fatigue and negligence, and lack of collaborative care, aligned with a higher likelihood of delayed TI seen among patients with complex management. Inadequate communication and coordination among providers, as well as healthcare-system constraints such as high patient flow, insufficient staffing, inefficient appointment systems and the absence of HbA1c testing, further worsen delays in treatment adjustments, particularly for patients with multiple comorbidities or those on cardioprotective drugs.^{43 44} The absence of HbA1c testing, which may be specific to this study, was another significant factor. Because the patient must provide multiple FBS results, the physicians may not be sure about a single result and send the patient home with a follow-up appointment. Consequently, it causes the physician to delay making decisions, which delays the intensification of treatment. These findings mirror the barriers reported in other LMIC settings and underscore the need for a multilevel intervention approach. Addressing these barriers by setting up regular education, promoting healthcare professionals’ training and improving healthcare infrastructure could significantly improve TI and patient outcomes.

The use of FBS as a follow-up tool was one of the study’s limitations because HbA1c is not always accessible for diagnosis and follow-up. Other limitations of this study were the absence of patient perspectives, which may have provided additional insights into the barriers and facilitators of timely TI. The results may also face issues with generalisability, as the study was conducted in a single, specialised comprehensive hospital.

Therefore, coordinated action is paramount. The Amhara Health Bureau, in collaboration with the Federal Ministry of Health, should fulfil appropriate laboratories and medications, alongside continuous and comprehensive training about therapeutic inertia for healthcare professionals working in the chronic OPD. At the hospital level, strengthening public diabetes health education and optimising the use of available resources will help bridge current gaps in clinical practice. Similarly, healthcare professionals should foster a culture of continuous learning through regular reading and updating knowledge on diabetes care to improve diabetes management. Finally, future research should include a multicentre, joint study to assess the long-term impact of delayed TI on the incidence of chronic diabetic complications in Ethiopia and other resource-limited settings.

Data availability statement

Data are available on reasonable request.