Adherence to retinal screening recommendations and its associated factors among patients with diabetes in Ethiopian public hospitals: a cross-sectional study

Abstract

Objective

To assess adherence to retinal screening recommendations and its associated factors among patients with diabetes receiving follow-up care at public hospitals in Ethiopia.

Design

A cross-sectional, institution-based study was conducted at public hospitals in Addis Ababa from 27 February to 27 March 2023.

Setting

Five of the fourteen public hospitals in Addis Ababa were selected for this study.

Participants

A systematic random sample of 405 adult patients with diabetes mellitus (receiving follow-up care at public hospitals in Addis Ababa was selected.

Outcome measures

Bivariate and multivariate logistic regression analyses were conducted to assess statistical associations. Statistical significance between the outcome variable and independent variables was determined at a p value of less than 0.05 and a 95% CI.

Result

223 participants (55.1%) adhered to the recommended retinal screening guidelines. Factors significantly associated with adherence to retinal screening recommendations included being an urban resident (adjusted OR (AOR)=2.783; 95% CI 1.022 to 7.583), lack of formal education (AOR=0.242; 95% CI 0.100 to 0.589) and poor knowledge regarding diabetic retinopathy (AOR=0.281; 95% CI 0.171 to 0.462).

Conclusion and recommendation

Slightly more than half (55%) of the participants in this study adhered to the recommended retinal screening guidelines. Urban residence, higher educational attainment and good knowledge of diabetic retinopathy were significant predictors of adherence to retinal screening recommendations. Healthcare professionals should take a more proactive role in educating patients with diabetes about diabetic retinopathy and the importance of retinal screening, with particular emphasis on rural residents and individuals without formal education.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study instrument was pretested and modified before data collection.

• The study participants were recruited from a variety of healthcare facilities.

• The study employed face-to-face interview techniques, resulting in a high response rate.

• Diabetes patients who are on follow-up care at private healthcare institutions are not addressed.

• The cross-sectional design limits the ability to establish a causal relationship.

Introduction

Diabetes mellitus (DM) is a chronic metabolic disorder and a rapidly growing global health concern with significant social and economic impacts.¹ DM can lead to several ocular complications, including diabetic retinopathy (DR), diabetic papilloedema, glaucoma, cataracts and ocular surface diseases.² The most common ocular complication of diabetes is DR.^{3 4} DR is the leading cause of vision loss among adults with diabetes aged 20–74 years. Nearly all individuals with type 1 diabetes develop signs of retinopathy after 20 years of disease duration, while more than half of those with type 2 diabetes show some signs of retinopathy after the same period.^{4 5} In its early stages, DR often does not cause noticeable vision problems; however, without regular retinal screening, it can progressively worsen and ultimately lead to vision loss.^{4 6}

Since DR is asymptomatic in its early stages, regular retinal screening examinations are essential for its early detection and prevention.^{7 8} The American Diabetes Association (ADA) recommends that patients with type 1 diabetes receive their first dilated eye examination 5 years after diagnosis, followed by annual examinations. For individuals with type 2 diabetes, the initial examination should occur at the time of diagnosis, with annual examinations thereafter.9,11 Along with regular eye examinations, maintaining good control of risk factors such as blood glucose, blood pressure and cholesterol levels is crucial for preventing the progression of DR.¹¹

Poor adherence to this ADA recommendation has led to an increase in the prevalence of diabetes-related vision-threatening diseases. This has ultimately become a worldwide epidemic, resulting in greater disability and higher healthcare costs.¹² However, early detection of this disorder through screening, followed by appropriate treatment, may provide a practical means of preventing blindness.13,17 Regular retinal screening not only facilitates the early detection and treatment of diseases but also reduces inappropriate referrals to ophthalmologists for non-sight-threatening conditions that are better managed through ongoing observation in a primary healthcare setting.¹⁸

According to studies conducted in Switzerland, Turkey and Northwest Ethiopia, 70.5%, 77.3% and 40.6% of participants, respectively, reported having undergone a retinal examination by an ophthalmologist within the past year.^{8 19 20} Studies have also shown that older age,^{21 22} longer duration of diabetes,^{20 23} good knowledge of diabetes^{24 25} and a history of eye disease^{20 23} are positively associated with adherence to retinal screening recommendations.

Determining adherence to retinal screening recommendations and identifying associated factors are crucial for preventing modifiable risk factors of DR and visual impairment in individuals with DM. However, data on adherence to these recommendations among patients with diabetes in Ethiopia remain limited. Therefore, this study aimed to assess adherence to retinal screening recommendations and its associated factors among patients with diabetes receiving follow-up care at public hospitals in Ethiopia.

Methods

Study design and participants

This institution-based cross-sectional study was conducted in Addis Ababa, the capital city of Ethiopia and the headquarters of the African Union. The estimated population of the city was 6.6 million inhabitants in 2016.²⁶ Five of the fourteen tertiary-level public hospitals in Addis Ababa were included in this study. Data were collected between 27 February 2023 and 27 March 2023. All adults aged 18 years and older with DM who attended diabetic follow-up clinics at these public hospitals and were willing to participate were included in the study. Patients with diabetes who were critically ill, those diagnosed with mental health problems unable to respond to the questionnaire, and patients with type 1 diabetes who had been diagnosed for less than 5 years (as recommended by the ADA to begin dilated eye examinations 5 years after diagnosis) were excluded.

Sample size determination and sampling procedure

The sample size was calculated using a single population proportion formula, considering a standard normal distribution (Z) at a 95% confidence level and a 5% margin of error. A 39.6% proportion of good eye check-up practice, as reported in a study conducted at Debark Hospital in Northwest Ethiopia, was used.²⁰

$${\displaystyle {\displaystyle n=\frac{Z_{\alpha /2}^2\times p(1-p)}{d^2}}}$$

$${\displaystyle {\displaystyle n=\frac{(1.96{)}^2\times 0.39(1-0.39)}{(0.05{)}^2}}}$$

$${\displaystyle {\displaystyle n=368}}$$

Considering a 10% non-response rate, the final sample size was 405.

Five public hospitals were selected from a total of 14 hospitals located in Addis Ababa based on the presence of specialised diabetes care units and a high volume of patient flow. The sample size was proportionally allocated to each hospital according to client flow, determined by the average number of clients who received follow-up care at each facility during the month preceding the data collection period. Systematic random sampling was employed to select study participants from the chosen hospitals. The selection interval was calculated using the formula Kth value=N/nf, where Kth represents the interval at which respondents were interviewed, N denotes the expected total number of diabetic patients per month across all selected hospitals, and nf indicates the final calculated sample size. The total number of patients with diabetes at the five selected hospitals during the month prior to data collection was 3468. As shown in figure 1, Tikur Ambesa Hospital had 1208 patients; St. Paul Hospital had 756; Yekatit 12 Hospital had 569; Ras Desta Hospital had 402; and Zewditu Hospital had 533 patients with diabetes during that period. The Kth value was calculated by dividing the total number of patients (N) by the final sample size (nf), resulting in 3468/405=8. Using the sequence of patients’ medical records from follow-up appointments as the sampling frame, the first participant was randomly selected from numbers 1–8. Subsequently, every eighth patient was chosen until the required sample size was reached. Figure 1 provides a detailed illustration of the sampling procedure for each hospital.

Figure 1. Schematic representation of sampling procedure in a selected public hospital in Addis Ababa, Ethiopia, 2023.

Operational definitions

The ADA recommends annual dilated eye examinations for all adults with diabetes. For individuals with type 1 diabetes, the initial examination is recommended 5 years after diagnosis, whereas those with type 2 diabetes should have their first examination at the time of diagnosis. In the present study, adherence to retinal screening is defined as having had a dilated eye examination within the previous 12 months (the last year). Otherwise, individuals are considered non-adherent.¹⁰ Knowledge of DR: respondents who scored greater than or equal to the mean (≥7) on the knowledge questions were considered to have good knowledge of DR, while those who scored below the mean were considered to have poor knowledge of DR.

Data collection instruments and procedures

The data collection tool consists of 34 items divided into four sections: 9 items on the sociodemographic questionnaire, 7 items on the participant’s clinical profile, 12 items assessing knowledge of DR and 5 items on adherence to retinal screening. The retinal screening items were developed based on existing literature.^{20 27 28} This tool includes questions about previous eye examination practices, eye examinations following a diagnosis of DM, the frequency of examinations in the past year and the types of eye examinations received. Additionally, individuals who reported not having undergone a diabetes eye examination were asked to provide reasons for non-adherence. The tool demonstrated good reliability, as indicated by a Cronbach’s alpha coefficient of 0.8. Additionally, a panel of subject matter experts evaluated the content validity of the tool. All items were deemed relevant, with item-level content validity index scores between 0.8 and 1.

The knowledge items on DR were developed based on previous literature.^{20 21 29} The 12 multiple-choice questions designed to assess knowledge of DR cover several dimensions, including its definition, associated risk factors, available treatment options and the importance of eye screening for patients with DM. Each correctly answered item was assigned a value of 1, while incorrect answers received a value of 0. Participants with knowledge scores above the mean were considered to have good knowledge of DR, while those scoring below the mean were classified as having poor knowledge. The internal validity of the knowledge assessment tool was evaluated, yielding a Cronbach’s alpha of 0.84, indicating strong internal consistency. Additionally, subject matter experts assessed the content validity of the tool. The questionnaire was initially prepared in English, then translated into Amharic and subsequently back-translated into English to verify the accuracy of the translation. Data were collected by three nurses holding BSc degrees and supervised by one nurse with an MSc degree. Both the data collectors and the supervisor received 2 days of training on data collection procedures and the use of the Kobo Toolbox application.

Data quality assurance

Pretesting of the questionnaire was conducted on 5% of the sample size at Alert Hospital to ensure the tool’s reliability and clarity before the actual data collection. Based on the results obtained, necessary modifications were made to the questionnaire before commencing the main data collection. Data management and analysis procedure. The data collected using Kobo Collect V.2022.4.4 were exported to SPSS V.27 for analysis. Completeness was assessed using frequency tables to identify any missing values. The sample was described using descriptive statistics, with numerical data reported as means, SD, proportions or percentages. Bivariate and multivariate logistic regression analyses were conducted to examine the associations between independent and dependent variables. Independent variables with a p value less than 0.25 were included in the multivariate analysis. The Hosmer-Lemeshow statistic was used to assess the model’s goodness of fit. The variance inflation factor (VIF) was employed to evaluate multicollinearity. No multicollinearity was detected, as all VIF values were below 5. Adjusted ORs (AORs) with 95% CIs and p values ≤0.05 were used to assess the strength and statistical significance of associations. Results were presented using tables, figures and text.

Patient and public involvement

The patient and the public were not involved in the planning and design of the study.

Result

Sociodemographic characteristics

405 patients with diabetes attending follow-up appointments at five public hospitals in Addis Ababa participated in the study. The mean age of the participants was 53.58±14.92 years (see table 1). The majority were female (216, 53.3%) and urban residents (367, 90.6%). 120 participants (29.6%) had a college or university-level education. The median monthly family income was 4000 ETB (IQR: 2000–5000 ETB).

Table 1. Sociodemographic characteristics of participants in Addis Ababa, Ethiopia, 2023.

Variable	Frequency	Per cent
Age in years
18–35	57	13.3
36–50	119	29.7
51–62	105	26.1
≥63	124	30.9
Residence
Urban	367	90.6
Rural	38	9.4
Marital status
Single	45	11.1
Married	318	78.5
Divorced	18	4.5
Widowed	24	5.9
Sex
Female	216	53.3
Male	189	46.7
Religion
Orthodox	258	63.7
Muslim	53	13.1
Protestant	84	20.7
Other	10	2.5
Educational level
No formal education	58	14.1
Primary level	118	29.1
Secondary level	109	26.9
Tertiary	120	29.6
Occupation
Farmer	11	2.4
Daily labourer	9	1.2
Government employed	71	17.8
Housewife	133	33.0
Retired	66	18.1
Merchant	43	16.6
Other	46	10.9
Average monthly income (ETB)
≤2000	112	26.6
2001–3577	86	20.4
3578–6500	146	34.7
≥6501	77	18.3

Clinical profile of study participants

The majority of participants in the study (73.8%) were patients with type 2 diabetes (see table 2). The median duration of DM since diagnosis was 8 years (IQR: 4–15). 223 participants (54.3%) had a history of eye disease, 53.8% reported symptoms of vision problems and 44.7% demonstrated poor knowledge of DR.

Table 2. Clinically related factors of participants in Addis Ababa, Ethiopia, 2023.

Variable	Frequency	Per cent
Duration of diabetes
<10 years	215	52.5
≥10 years	190	47.5
Type of DM
Type 1	105	25.7
Type 2	300	74.3
Hypertension
Yes	220	53.9
No	185	46.1
Family history of DM
Yes	165	40.6
No	240	59.4
Systemic complications of DM
Kidney complication	25	7.1
Cardiovascular complication	30	8.3
Other complications of DM	3	0.7
I do not know	4	1.0
I do not have any complication	343	82.9
Previous eye disease
Yes	223	54.4
No	182	45.5
Presence of visual symptoms
Yes	219	53.2
No	186	46.8

DM, Diabetes Mellitus.

Adherence to retinal screening recommendation

The adherence rate to retinal screening recommendations among the study participants was 55.1%. This indicates that 252 participants (55%) underwent retinal screening at least once in the past year (see table 3). Conversely, 44.9% were non-adherent to retinal screening guidelines. 152 participants (37.5%) had never had their eyes examined, even after being diagnosed with DM. The primary reason cited for non-adherence was the absence of visual symptoms, which led 74 participants (48.68%) to believe that regular retinal screening was unnecessary.

Table 3. Adherence to retinal screening recommendations among the study participants in Addis Ababa, Ethiopia, 2023.

Variable	Frequency	Per cent
Have you ever been referred to check your eyes?
Yes	327	80.7
No	78	19.3
Have your eyes been examined after the diagnosis of DM
Yes	253	62.5
No	152	37.5
How many times in the last 1 year?
None	30	7.4
Once	134	33.1
Twice	67	16.5
Three times or more	22	5.4
What type of eye examination did you get?
Vision	16	6.3
Checking for eyeglasses	24	9.5
Dilated examination of the eye	205	81
Slit lamp/torch	4	1.6
Other	4	1.6
Why did you not get an eye examination?
Lack of visual symptom	74	48.7
Not referred by a physician	56	36.8
Financial restriction	4	2.6
Lack of convenient facility	4	2.6
Other	14	9.2

DM, Diabetes Mellitus.

Factors associated with retinal screening recommendation

Binary logistic regression was employed to identify factors significantly associated with adherence to retinal screening recommendations. The analysis indicated that age, residence, education level, monthly income and clinical variables—including duration of diabetes, family history of diabetes, systemic complications of diabetes, presence of visual symptoms and knowledge of DR were candidates for multivariable analysis at a p value less than 0.25.

In the multivariable logistic regression analysis, higher educational level, urban residence and knowledge of DR were significantly associated with adherence to retinal screening recommendations (p<0.05) (see table 4). Patients with diabetes residing in urban areas were 2.78 times more likely to adhere to retinal screening recommendations compared with those from rural areas (AOR=2.783; 95% CI 1.022 to 7.583). Likewise, patients with diabetes who had no formal education were 76% less likely to adhere to retinal screening recommendations compared with those with a college education or higher (AOR=0.242; 95% CI 0.100 to 0.589). Additionally, patients with diabetes who had poor knowledge of DR were 72% less likely to adhere to retinal screening recommendations than those with a good knowledge of DR (AOR=0.281; 95% CI 0.171 to 0.462).

Table 4. Adherence to retinal screening recommendation among the study participants in Addis Ababa, Ethiopia, 2023.

Factors	Adherence to retinal screening recommendation		COR	AOR (95% CI)	P value
	Yes	No
Residence
Urban	188	179	5.601	2.783 (1.022 to 7.583)	0.045
Rural	6	32	1	1
Educational level
No formal education	12	44	0.154	0.242 (0.100 to 0.589)	0.002
Primary (1–8)	50	67	0.421	0.599 (0.326 to 1.101)	0.099
Secondary	54	56	0.544	0.987 (0.512 to 1.905)	0.970
Tertiary (college/university)	78	44	1	1
Knowledge of diabetic retinopathy
Poor	49	132	0.177	0.281 (0.171 to 0.462)	0.000
Good	145	79	1	1

AOR, Adjusted Odds Ratio; COR, Crude Odds Ratio.

Discussion

This study aimed to assess adherence to retinal screening recommendations and the associated factors among adult patients with diabetes in public hospitals in Addis Ababa. The findings revealed that the adherence rate to the recommended retinal screening schedule among patients with diabetes at the studied hospitals was 55.1%. This indicates that, out of every 10 patients with diabetes in the study, approximately five to six adhered to the retinal screening recommendations.

The results of this study are lower than those reported in studies conducted in the USA, which found a rate of 72.2%,³⁰ and Australia, which reported 77.5%.²⁷ Differences in the sociodemographic and socioeconomic characteristics of the study populations may account for this discrepancy. Additionally, the USA study population was drawn from Kaiser Permanente Southern California (KPSC), a non-profit, integrated healthcare delivery organisation that provides comprehensive, coordinated medical services. Patients with diabetes at KPSC are managed according to ADA guidelines and benefit from a systematic outreach programme that employs proactive strategies to engage patients and ensure adherence to their screening schedules, rather than relying on patients to initiate appointments. This essential component of KPSC’s integrated healthcare model, which improves adherence to recommended retinal screening among patients with diabetes, is absent in our context and likely contributes to the lower adherence observed in our study.

The results of our study are higher than those reported in Ethiopia (26.6%)²⁰ and Ghana (21.7%).³¹ The Ethiopian study highlighted a lack of sufficient eye care centres in the study area, which may be the reason for the lower adherence rate to the recommended DR screening. In contrast, our study was conducted in the capital city, Addis Ababa, where participants can get relatively more eye care centres, which might be the reason for the discrepancy.

According to our study, 37.5% of participants had never undergone a retinal examination, even after being diagnosed with DM, which is higher than the result reported in a study conducted in rural China, which found rates of 68.7%.³² This discrepancy may be attributed to the study in rural China being conducted in a hospital located in northern Guangdong, an economically underdeveloped region. Residents in such areas likely have less disposable income to spend on healthcare, including routine eye examinations. Additionally, healthcare facilities and specialised services such as ophthalmology are often scarce in rural communities. Lower levels of health literacy and public awareness in these areas may also contribute to participants not seeking regular retinal screening as recommended. In contrast, the majority of participants in our study are from the urban population of Ethiopia’s capital city.

According to this study, the primary reason for not adhering to eye examinations was the absence of visual symptoms (48.65%). This finding is supported by previous studies conducted in India,³³ Switzerland¹⁹ and Jordan.³⁴ This may be attributed to a lack of awareness among patients with diabetes about the silent progression of DR. The results clearly indicate a critical misconception among the study participants. Many believed that clear vision signified healthy eyes, assuming that eye or retinal examinations were only necessary when visual problems occurred. The results clearly reveal a critical misconception among the study participants. Many believed that clear vision indicated healthy eyes, assuming that eye or retinal examinations were only necessary when visual problems occurred. This reflects insufficient knowledge regarding the importance of regular retinal screening, which may explain the lower adherence observed in this study. However, another study conducted in Ethiopia²⁸ found that long distances from healthcare facilities were the primary reason for poor adherence to the recommended retinal screening among participants.

According to this study, 55.3% of the participants demonstrated good knowledge of DR, which aligns with findings from a study conducted in Bangladesh.³⁵ However, the results of this study are lower than those reported in studies conducted in Japan³⁶ and Australia,³⁷ which found rates of 98% and 96%, respectively. This discrepancy may be attributed to the Australian study being conducted among patients who were members of a diabetes association, providing them with more recent and updated information about DR. In contrast, the current study randomly selected all patients with diabetes receiving follow-up care in public hospitals, regardless of their membership status.

Patients with diabetes living in urban areas are three times more likely to adhere to retinal screening recommendations than those in rural areas. This finding is supported by studies conducted in Bangladesh³⁵ and Ethiopia.^{20 38} The higher adherence to retinal screening in urban areas may be explained by the better availability of eye care services and professionals in urban Ethiopia compared with rural regions. A recent study in Ethiopia reported a significant urban-rural disparity in the distribution of eye care professionals and specialised eye care centres, with several remote and rural areas lacking these services.³⁹ The centralisation of ophthalmic services in urban areas improves access to retinal screening for patients with diabetes residing there, while limiting opportunities for rural residents, ultimately reducing their adherence. Furthermore, the higher adherence among urban residents is attributed to their greater health literacy and healthcare-seeking behaviour compared with rural residents.⁴⁰ Urban areas also have better access to diverse information sources. Urban residents benefit from multiple channels, such as healthcare centres and mass media, which equip them with a better understanding of DR and contribute to higher screening rates compared with rural residents.

Additionally, patients with diabetes with no formal education were 76% less likely to adhere to retinal screening recommendations compared with those with a college education or higher. Similar findings have been reported in previous studies conducted in Malaysia,⁴¹ South Africa⁴² and Ethiopia.³⁸ This may be because higher educational attainment leads to greater health literacy, which ultimately promotes healthy behaviours, including regular retinal screening. Furthermore, educated individuals may have better access to various forms of mass communication, making them more empowered and aware of the importance of undergoing regular retinal examinations.

Finally, patients with diabetes who had poor knowledge of DR were 72% less likely to adhere to retinal screening recommendations compared with those with a good understanding of the condition. This finding is consistent with previous studies conducted in Ethiopia,³⁸ South Africa,⁴³ India,³³ Switzerland¹⁹ and Taiwan.⁴⁴ These results suggest that a solid understanding of DR likely promotes the belief in the importance of regular retinal screening and improves adherence to screening recommendations.

This study did not find an association between adherence to retinal screening recommendations and factors such as age, duration of diabetes or the presence of visual symptoms. However, other studies have demonstrated such associations. Research conducted in Bangladesh,⁴⁵ India³³ and Turkey⁸ found a relationship between the duration of diabetes and adherence to retinal screening. Additionally, previous studies in Ethiopia^{31 46} and the USA⁴⁷ reported that age was significantly associated with adherence to retinal screening recommendations. The presence of visual symptoms was also associated with an increased likelihood of individuals seeking an eye examination in a study conducted in Pennsylvania, USA.⁴⁸

Implication of the study

The clinical relevance of this study lies in providing valuable information about the level of adherence among patients with diabetes to retinal screening recommendations. It enables healthcare professionals to move beyond passive patient education toward proactive clinical intervention. This study empowers healthcare providers to educate individuals with diabetes about the silent progression of DR and the necessity of regular retinal screening, even in the absence of symptoms. From a public health perspective, the study promotes early detection of DR through increased screening, ultimately preventing the progression of DR to advanced stages that require costly and complex treatments. This, in turn, reduces the financial burden on both patients and the public healthcare system, leading to significant economic benefits. Additionally, the results of this study support the development of public health programmes aimed at improving adherence to screening guidelines.

Conclusion

Slightly more than half (55%) of the participants in the current study adhered to the recommended retinal screening. Nearly half demonstrated poor knowledge of DR. Urban residence, higher educational status and good knowledge of DR were significant predictors of adherence to retinal screening recommendations. The primary reason cited for non-adherence was the absence of visual symptoms.

Recommendations

Based on the findings of this study, we recommend that health authorities develop and implement a national strategy for DR screening, shifting from a passive approach to a proactive, population-based model. Additionally, they should invest in establishing and equipping more dedicated eye care centres, particularly in rural areas and integrate retinal screening into primary healthcare services. Healthcare professionals must take a more proactive role in educating and managing patients with diabetes about DR and the importance of retinal screening, with special attention to rural and less-educated populations. At every patient visit, they should emphasise the significance of regular retinal screenings and dispel the misconception that clear vision necessarily indicates healthy eyes. Future research should focus on developing and evaluating the effectiveness of culturally tailored educational interventions for DR and retinal screening, particularly targeting rural and less-educated communities.

Data availability statement

Data are available upon reasonable request.