Predictors of diabetes self-management and glycemic control among patients with diabetes mellitus in conflict-affected Palestine

Abstract

Background

Poor glycemic control and inadequate self-management drive the global diabetes mellitus (DM) burden. In conflict-affected Palestine, structural barriers worsen outcomes, yet diabetes management self-efficacy (DMSE) remains understudied.

Aim

To assess diabetes management self-efficacy and glycemic control, their interrelationship, and the independent predictors of each among patients with diabetes mellitus in conflict-affected Palestine.

Methods

A cross-sectional study was conducted among 418 adults with type 1 or type 2 DM attending governmental diabetes clinics in the Hebron governorate, West Bank, Palestine. Data were collected using the Arabic-validated Diabetes Management Self-Efficacy Scale (DMSES) and a structured clinical questionnaire. Non-parametric tests and two multiple linear regression models were applied to identify independent predictors, with HbA1c and DMSE scores specified as the two dependent variables.

Results

Glycated hemoglobin (HbA1c) was suboptimal (median = 7.50, IQR = 2.30), while DMSE was moderate (median = 3.45/5, IQR = 1.10). Blood glucose management was the strongest subdomain, whereas physical exercise was the weakest. Former smoking (β = 0.202, 95% CI: 0.430, 1.494, p < .001), lower physical activity frequency (β = 0.156, 95% CI: 0.098, 0.455, p = .003), and diabetes-related complications (β =−0.154, 95% CI: −1.132, − 0.210, p = .004) independently predicted higher HbA1c. Lower physical activity (β =−0.177, 95% CI: −0.208, − 0.060, p < .001), longer disease duration (β =−0.135, 95% CI: −0.177, − 0.021, p = .013), lower income (β = 0.163, 95% CI: 0.081, 0.348, p = .002), and lower education (β = 0.127, 95% CI: 0.018, 0.227, p = .022) independently predicted lower DMSE.

Conclusion

Clinical, behavioral, and socioeconomic factors independently predicted glycemic control and self-efficacy. The DMSE–HbA1c disconnect indicates structural barriers impede self-management behavior. Integrating clinical education with economic and structural support is essential to improve diabetes outcomes in conflict-affected settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12875-026-03423-1.

Introduction

Diabetes mellitus (DM) is a chronic metabolic disorder marked by sustained hyperglycemia due to deficiencies in insulin secretion, insulin action, or both, resulting in severe complications such as nephropathy, retinopathy, neuropathy, and cardiovascular disease [1, 2]. In 2019, approximately 463 million people worldwide had DM, and projections indicate that this figure will exceed 700 million by 2045. Type 2 diabetes mellitus (T2DM) made up 85–90% of all cases [3]. Glycated hemoglobin (HbA1c) is the best way to keep track of glycemic control. A high HbA1c level is linked to more complications, a lower quality of life, and a higher cost [1, 4]. In Palestine, the prevalence of diabetes among adults is estimated at approximately 10–12%, with higher rates reported in the Gaza Strip compared to the West Bank. Studies among Palestinian patients with diabetes have documented mean HbA1c levels ranging from 8.0% to 9.5%, indicating persistently suboptimal glycemic control across the population [5, 6].

A central psychological determinant of diabetes outcomes is management self-efficacy, defined as an individual’s belief in their capacity to organize and execute self-management behaviors in the face of obstacles [7, 8]. The Diabetes Management Self-Efficacy Scale (DMSES), developed by van der Bijl et al. [9] for patients with T2DM, operationalizes self-efficacy across four subdomains: nutrition-specific and weight control, nutrition general and medical treatment, physical exercise, and blood glucose control. Because daily decisions about diet, medication, and physical activity directly determine clinical outcomes, patients’ self-management confidence is clinically inseparable from their glycemic control [4, 10].

Globally, researchers have consistently linked better self-efficacy in diabetes management to greater adherence to self-care activities and improved glycemic control. The DMSES has been validated across diverse cultural contexts, demonstrating that diabetes self-efficacy substantially predicts diabetes-related self-management behaviors [11–14]. Systematic reviews of diabetes self-management education (DSME) programs also found that self-efficacy-focused interventions achieve better HbA1c levels, greater adherence to medications, and increased levels of physical activity [15, 16]. Individuals who are physically inactive, smoke, and have diabetes complications have been consistently identified as having a higher likelihood of diminished glycemic control compared to those without these characteristics. For those individuals who are low income, have a lower educational attainment, and have had diabetes for a longer period of time than others, their risk of developing a poor level of glycemic control increases even further [2, 4, 17, 18].

In the Middle East, a systematic review of DSME programs across Iran, Turkey, the United Arab Emirates (UAE), Jordan, and Qatar found that structured education improved glycemic control and self-management behaviors, yet highlighted heterogeneity in program outcomes [19]. In Saudi Arabia and the UAE, complications and poor glycemic control were strongly associated with reduced quality of life, while higher income and education were protective [17, 20–22]. In Jordan, a family-centered empowerment intervention improved self-efficacy and glycemic control in adolescents with T1DM [23]. Shared regional barriers include low health literacy, fatalistic cultural beliefs, limited economic resources, and inadequate patient education [19].

Palestine presents a uniquely challenging context. The Palestinian healthcare system operates under constraints imposed by ongoing conflict, military occupation, restricted movement, and chronic medication shortages, all of which impede access to diabetes care [24, 25]. Palestinian patients experience a dual burden of inadequate glycemic control and diminished quality of life, with socioeconomic deprivation and complications identified as primary risk factors [5, 26, 27]. The sole Palestinian study analyzing diabetes management self-efficacy concentrated solely on T2DM and the correlation between self-efficacy and quality of life, neglecting independent predictors of glycemic control or self-efficacy as co-primary outcomes [10, 28]. Most current studies analyze HbA1c or self-efficacy independently, do not incorporate mixed T1DM and T2DM samples, and do not apply regression modeling for both outcomes concurrently.

Therefore, this study aimed to [1] assess diabetes management self-efficacy and glycemic control among patients with diabetes mellitus in conflict-affected Palestine [2], examine the relationship between diabetes management self-efficacy and glycemic control, and [3] identify their independent sociodemographic, clinical, and behavioral predictors.

Methods

Design and setting

A cross-sectional design was employed, a widely used approach for examining health-related associations within a defined population at a single point in time [29]. The study was conducted in the Hebron governorate, southern West Bank, Palestine, across urban centers, rural villages, and refugee camps. Participants were recruited from diabetes outpatient clinics in governmental hospitals and primary healthcare centers (PHCs), which provide routine follow-up for patients with T1DM and T2DM. MOH PHCs provide routine follow-up for stable diabetes patients; MOH hospitals manage complications and specialist consultations. Participants were recruited from both settings, but results were not disaggregated by facility type as many patients attend both over time.

Population and sampling

The study population comprised adults with T1DM or T2DM attending diabetes clinics and PHCs in the Hebron governorate between January and February 2026. As the exact number of registered patients was unavailable, the sample size was calculated using the Raosoft calculator (95% confidence level, 5% margin of error, 50% response distribution), yielding a minimum of 384 participants. To account for anticipated incomplete responses or missing data (estimated 15% non-response rate), an additional 66 questionnaires were added, resulting in a target of 450 distributed questionnaires.

Inclusion and exclusion criteria

Adults aged 18 years or older with a confirmed diagnosis of Type 1 Diabetes Mellitus (T1DM) or Type 2 Diabetes Mellitus (T2DM) who could independently read and write in Arabic were eligible. Participants were excluded if they had cognitive impairments, severe psychiatric disorders, or communication barriers preventing reliable questionnaire completion.

Instrumentation

A structured, paper-based, self-administered questionnaire comprised two sections. Section 1 collected demographic variables (age, sex, marital status, educational level, monthly income, and smoking status) and clinical variables (diabetes type, duration, HbA1c, presence of complications, and physical activity frequency). Physical activity frequency was assessed by asking participants: “How often do you engage in physical activity?” Response options were: (1) Regularly (≥ 3 times per week) (2), Sometimes (1–2 times per week) (3), Rarely (1–3 times per month) (4), Never. HbA1c values were extracted directly from participants’ clinical records at the attending diabetes clinic, representing the most recent value documented within the preceding three months.

Section 2 contained the Arabic version of the DMSES [9], which was based on Bandura’s Social Cognitive Theory (SCT). The DMSES includes 20 items rated on a five-point Likert scale from 1 (No, surely not) to 5 (Yes, surely), with total scores ranging from 20 to 100; higher scores reflect greater self-efficacy. Four subdomains are assessed: Nutrition Specific and Weight (5 items), Nutrition General and Medical Treatment (9 items), Physical Exercise (3 items), and Blood Sugar Control (3 items). Each item is prefaced with the stem “I,” operationalizing the strength dimension of self-efficacy as defined by Bandura [8].

Validity and reliability

The DMSES has demonstrated strong reliability across multiple cross-cultural adaptations: Australian/English (α = 0.91) [12], Turkish (α = 0.88) [11], Chinese (α = 0.93) [14], and Iranian (α = 0.92) [13]. The Arabic version used here was translated and validated by Eshtaya et al. [10] following the World Health Organization (WHO) translation protocol, with Cronbach’s alpha of 0.94 (pilot) and 0.92 (full sample). In the current study, internal consistency was α = 0.916, indicating excellent reliability. A pilot test on 30 patients confirmed item clarity and an estimated completion time of 15–20 min; no modifications were required.

Data collection

Following ethical approval, permission was obtained from the Palestinian Ministry of Health (MOH) and administrators of selected facilities. Data were collected between January 10 and February 20, 2026. The research team recruited eligible patients during scheduled follow-up visits. After providing written informed consent, participants completed the questionnaire in a private area within the clinic, with team members available for clarification only. Completed questionnaires were reviewed for completeness, securely stored, and entered into IBM SPSS Statistics version 28.

Data analysis

Data were analyzed using IBM SPSS Statistics for Windows, Version 28.0. Descriptive statistics were computed for all variables. Normality of HbA1c and DMSE were assessed using the Shapiro–Wilk test; significant departures were identified (all p < .05), so non-parametric tests were applied. Mann–Whitney U tests compared two-group variables; Kruskal–Wallis H tests were used for three or more groups. Spearman’s rank-order correlation examined associations between HbA1c and continuous variables. Two multiple linear regression analyses (enter method) were conducted with HbA1c and overall DMSE score as separate dependent variables. All variables listed in Table 1 (sociodemographic, clinical, and behavioral) were entered simultaneously into each regression model using the enter method, regardless of their significance in bivariate analyses. Smoking status and marital status were dummy-coded. All regression assumptions were verified prior to interpretation. No multicollinearity was detected (VIF < 5), and residual diagnostics confirmed normality and homoscedasticity. Statistical significance was set at p < .05.

Table 1.

Demographic, lifestyle, and clinical characteristics of study participants (N = 418)

Characteristic	Category	n	%
Gender	Male	199	47.6
	Female	219	52.4
Marital status	Single	56	13.4
	Married	343	82.1
	Other	19	4.5
Educational level	Elementary	207	49.5
	Secondary	134	32.1
	Bachelor’s degree	64	15.3
	Above bachelor’s	13	3.1
Monthly income (NIS)	< 2,000	242	57.9
	2,000–4,000	150	35.9
	> 4,000	26	6.2
Smoking status	Former smoker	83	19.9
	Current smoker	73	17.5
	Non-smoker	262	62.7
Physical activity	Regularly	146	34.9
	Sometimes	130	31.1
	Rarely	76	18.2
	Never	66	15.8
Type of diabetes	Type 1	127	30.4
	Type 2	291	69.6
Duration of diabetes	< 1 year	47	11.2
	1–5 years	100	23.9
	6–10 years	54	12.9
	> 10 years	217	51.9
Any complication	Yes	312	74.6
	No	106	25.4
Complication: Hypertension	Yes	226	54.1
Complication: Visual impairment	Yes	225	53.8
Complication: Renal failure	Yes	165	39.5
Complication: Diabetic foot	Yes	83	19.9
Age (years) Mean ± SD	56.0 ± 17.4
BMI (kg/m²) Mean ± SD	29.5 ± 6.98
HbA1c Mean ± SD	7.89 ± 1.89

Complication percentages do not sum to 100% as participants could report more than one

n  number, SD  Standard deviation, BMI  Body mass index, HbA1c  Glycated hemoglobin

Ethical consideration

Ethical approval for this study was obtained from the Education in Health and Scientific Research Unit, Palestinian Ministry of Health (MOH), West Bank, Palestine (Approval Ref: 2025/3493/162). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki. All participants were provided with comprehensive information regarding the study’s purpose, procedures, potential risks, and benefits. Written informed consent was obtained from every participant prior to data collection. Participants were informed that their involvement was entirely voluntary and that they could withdraw at any time without any penalty or impact on their healthcare access. No personally identifiable information was collected; participants were assigned numerical codes for anonymity. All data were stored in password-protected files accessible only to the research team.

Results

Sample characteristics

A total of 418 participants were included (mean age 56.0 years, SD = 17.4; 52.4% female). Nearly half (49.5%) had elementary education, and 57.9% reported monthly income below 2,000 NIS. T2DM was present in 69.6% of the sample, with 51.9% having a diabetes duration greater than 10 years. Mean body mass index (BMI) was 29.5 kg/m² (SD = 6.98), and median HbA1c was 7.50 (IQR = 2.30). Regular physical activity was reported by 34.9%; 62.7% were non-smokers. Diabetes-related complications were reported by 74.6% of participants, most commonly hypertension (54.1%), visual impairment (53.8%), renal failure (39.5%), and diabetic foot (19.9%). Full characteristics are presented in Table 1.

Diabetes management self-efficacy (DMSE) scores

The overall mean DMSE score was 3.47 (SD = 0.81) out of 5, with a sum score of 69.40 (SD = 16.15) out of 100. The median DMSE score was 3.45 (IQR = 1.10) out of 5. Blood glucose management yielded the highest subdomain score (mean = 3.93, SD = 1.11), followed by nutrition general and medical treatment (mean = 3.66, SD = 0.82). Nutrition specific and weight control (mean = 3.09, SD = 1.10) and physical exercise (mean = 3.07, SD = 1.14) were the lowest. See Table 2.

Table 2.

DMSE scores and subdomain means (N = 418)

Domain/Subdomain	Mean (SD)	Observed Range
Overall DMSE (mean of 20 items)	3.47 (0.81)	1.05–5.00
Overall DMSE (sum of 20 items)	69.40 (16.15)	21.00–100.00
Blood glucose (3 items)	3.93 (1.11)	1.00–5.00
Nutrition general and medical treatment (9 items)	3.66 (0.82)	1.00–5.00
Nutrition specific and weight (5 items)	3.09 (1.10)	1.00–5.00
Physical exercise (3 items)	3.07 (1.14)	1.00–5.00

Mean score is the average item score (1–5); sum score is the total of all 20 items (range 20–100)

DMSE Diabetes management self-efficacy

Association between variables and glycemic control (HbA1c)

Non-parametric tests were used to examine associations between participant characteristics and glycemic control (HbA1c) due to non-normal data distribution. As shown in Table 3, significant differences in HbA1c were observed for complication status, smoking status, physical activity, and diabetes duration. Participants with diabetes-related complications had higher HbA1c levels (Mdn = 8.00) than those without complications (Mdn = 7.00; U = 12,399.50, p = .001). Marital status was not significantly associated with HbA1c (H = 5.68, p = .059), though single and widowed/divorced participants showed numerically higher medians than married individuals. Educational level was not significantly associated with HbA1c (H = 3.45, p = .328).

Table 3.

Association between sociodemographic and clinical variables and HbA1c: Mann–Whitney U and Kruskal–Wallis H Tests (N = 418)

Variable	Group	N	Mdn (IQR)	Mean Rank	U / H	Z	p	Effect size
Mann–Whitney U Tests
Gender	Male	199	7.50 (2.20)	210.94	21503.5	−0.23	0.815	r = .01
	Female	219	7.50 (2.30)	208.19
Type of diabetes	Type 1	127	7.70 (2.40)	211.01	18286.5	−0.17	0.865	r = .01
	Type 2	291	7.50 (2.30)	208.84
Any complication	Yes	312	8.00 (2.00)	221.29	12399.5	−3.86	0.001**	r = .19
	No	106	7.00 (1.90)	174.78
Kruskal–Wallis H Tests
Marital status	Single	56	8.00 (2.45)	239.47	5.68	—	0.059	η² = 0.009
	Married	343	7.40 (2.40)	202.73
	Other	19	8.00 (2.00)	243.42
Educational level	Elementary	207	7.70 (2.00)	217.71	3.45	—	0.328	η² = 0.001
	Secondary	134	7.45 (2.40)	199.20
	Bachelor’s	64	7.20 (2.50)	208.95
	Above bachelor’s	13	7.20 (3.05)	187.69
Monthly income	< 2,000	242	7.50 (2.00)	218.98	2.16	—	0.340	η² = 0.003
	2,000–4,000	150	7.50 (2.80)	196.88
	> 4,000	26	7.05 (2.28)	194.10
Smoking status	Former smoker	83	8.00 (3.20)	243.55	7.58	—	0.023*	η² = 0.016
	Current smoker	73	7.40 (2.50)	204.32
	Non-smoker	262	7.40 (1.90)	200.15
Physical activity	Regularly	146	7.40 (2.00)	208.32	10.98	—	0.012*	η² = 0.019
	Sometimes	130	7.20 (2.22)	186.95
	Rarely	76	8.00 (2.83)	231.43
	Never	66	8.00 (2.00)	231.27
Duration of diabetes	< 1 year	47	8.00 (3.70)	241.19	8.45	—	0.038*	η² = 0.013
	1–5 years	100	7.20 (2.75)	189.97
	6–10 years	54	7.30 (2.53)	213.31
	> 10 years	217	7.60 (2.00)	210.69

IQR  Interquartile range, Mdn  Median

U = Mann–Whitney U statistic; H = Kruskal–Wallis statistic; Z = standardized test statistic. r = effect size for Mann–Whitney U (Z/√N); η² = eta-squared effect size for Kruskal–Wallis

^*p < .05, ^**p < .01

Behavioral factors were likewise significant. Former smokers had higher HbA1c (Mdn = 8.00) compared to current or non-smokers (H = 7.58, p = .023). Physical activity was associated with glycemic control (H = 10.98, p = .012), as participants who rarely or never exercised had higher HbA1c (Mdn = 8.00) than those exercising regularly or sometimes (Mdn = 7.20–7.40). Diabetes duration was significantly associated with HbA1c (H = 8.45, p = .038), with participants having diabetes for less than one year showing higher median HbA1c (8.00) compared to those with 1–5 years (7.20). In contrast, gender (p = .815), type of diabetes (p = .865), and monthly income (p = .167) were not significantly related to HbA1c. Spearman’s correlations revealed no significant associations with age (r_s = − 0.043, p = .382), BMI (r_s = − 0.043, p = .378), or overall DMSE scores (r_s = -0.063, p = .201).

Predictors of poor glycemic control (HbA1c)

The multiple linear regression model was statistically significant, F(14, 403) = 3.03, p < .001, explaining 9.5% of the variance in HbA1c (R² = 0.095, adjusted R² = 0.064). Three predictors reached statistical significance. Former smoking status was the strongest predictor (β = 0.202, p < .001), followed by lower physical activity frequency (β = 0.156, p = .003). Presence of diabetes-related complications was significantly associated with higher HbA1c levels (β = −0.154, p = .004). DMSE was not a significant predictor of HbA1c (p = .604). All other variables were non-significant. Full results are presented in Table 4.

Table 4.

Multiple linear regression predicting HbA1c (N = 418)

Predictor	B	SE	β	t	p	95% CI LL	95% CI UL
(Constant)	8.462	0.988	—	8.56	< 0.001	6.519	10.404
Age	−0.011	0.007	−0.103	−1.58	0.115	−0.025	0.003
BMI	−0.009	0.014	−0.033	−0.63	0.528	−0.037	0.019
Gender	0.285	0.227	0.075	1.26	0.210	−0.161	0.730
Educational level	−0.070	0.129	−0.031	−0.54	0.589	−0.324	0.184
Monthly income	−0.153	0.166	−0.049	−0.93	0.355	−0.479	0.172
Physical activity	0.276	0.091	0.156	3.04	0.003**	0.098	0.455
Type of diabetes	0.193	0.224	0.047	0.86	0.388	−0.247	0.633
Duration of diabetes	−0.035	0.096	−0.021	−0.37	0.712	−0.224	0.153
Diabetes complications	−0.671	0.235	−0.154	−2.86	0.004**	−1.132	−0.210
DMSE	0.062	0.120	0.027	0.52	0.604	−0.173	0.298
Current smoker (ref = non-smoker)	0.431	0.268	0.086	1.61	0.109	−0.096	0.959
Former smoker (ref = non-smoker)	0.962	0.271	0.202	3.56	< 0.001**	0.430	1.494
Married (vs. single)	−0.273	0.309	−0.055	−0.88	0.377	−0.880	0.334
Other marital status (vs. single)	0.152	0.524	0.017	0.29	0.773	−0.879	1.182

Physical activity coded from 1 = regular to 4 = never. Complications coded 1 = yes, 2 = no

B  Unstandardized coefficient, SE  Standard error, β  Standardized coefficient, CI  Confidence interval

^*p < .05, ^**p < .01

Bivariate analysis of factors associated with DMSE

Non-parametric tests were used to examine associations between participant characteristics and DMSE due to non-normal data distribution. Bivariate analyses (Table 5) demonstrated several significant associations between participant characteristics and diabetes management self-efficacy (DMSE). Male participants reported higher DMSE scores than females (U = 19,320.00, p = .045). Significant socioeconomic gradients were observed, with DMSE increasing across educational levels (H = 17.577, p = .001) and monthly income categories (H = 19.095, p < .001). Behavioral factors were also significant; higher DMSE was reported among participants engaging in regular physical activity compared to those with lower activity levels (H = 18.832, p < .001). Smoking status showed a significant association (H = 7.112, p = .029), with current smokers demonstrating higher DMSE scores. Additionally, DMSE differed significantly by duration of diabetes (H = 11.444, p = .010), with shorter disease duration associated with higher self-efficacy. In contrast, no significant differences in DMSE were observed by diabetes type (p = .160), complication status (p = .194), or marital status (p = .383).

Table 5.

Bivariate analysis of factors associated with DMSE (N = 418)

Variable	Group	N	Mdn (IQR)	Mean Rank	U / H	Z	p	Effect size
Mann–Whitney U Tests
Gender	Male	199	3.60 (1.20)	221.91	19320.00	−2.00	0.045*	r = .10
	Female	219	3.40 (1.05)	198.22
Diabetes Type	Type 1	127	3.50 (1.25)	222.07	16882.50	−1.41	0.160	r = .07
	Type 2	291	3.45 (1.05)	204.02
Complications	Yes	312	3.45 (1.20)	213.97	15140.50	−1.30	0.194	r = .06
	No	106	3.40 (0.79)	196.33
Kruskal–Wallis H Tests
Education	Elementary	207	3.35 (1.10)	188.52	17.58	—	0.001**	η² = 0.035
	Secondary	134	3.53 (1.01)	217.35
	Bachelor’s degree	64	3.68 (1.34)	245.67
	Above Bachelor’s	13	4.10 (0.92)	284.58
Income (NIS)	< 2,000	242	3.25 (1.15)	187.75	19.10	—	< 0.001**	η² = 0.041
	2,000–4,000	150	3.65 (0.86)	236.82
	> 4,000	26	3.78 (0.95)	254.31
Physical Activity	Regularly	146	3.75 (1.16)	235.98	18.83	—	< 0.001**	η² = 0.038
	Sometimes	130	3.53 (0.76)	216.93
	Rarely	76	3.15 (1.25)	177.53
	Never	66	3.25 (1.15)	173.09
Smoking	Current smoker	73	3.75 (1.05)	240.82	7.11	—	0.029*	η² = 0.012
	Former smoker	83	3.45 (1.25)	215.36
	Non-smoker	262	3.40 (1.05)	198.92
Duration of Diabetes	< 1 year	47	4.05 (1.45)	262.15	11.44	—	0.010*	η² = 0.020
	1–5 years	100	3.58 (1.05)	214.35
	6–10 years	54	3.35 (2.05)	193.13
	> 10 years	217	3.40 (1.05)	199.94
Marital Status	Single	56	3.63 (1.30)	225.74	1.92	—	0.383	η² = −0.003 ≈ 0.00
	Married	343	3.45 (1.10)	208.28
	Other	19	3.20 (0.90)	183.63

Mdn  Median, IQR  Interquartile range

U = Mann–Whitney U statistic; H = Kruskal–Wallis statistic; Z = standardized test statistic. r = effect size for Mann–Whitney U (Z/√N); η² = eta-squared effect size for Kruskal–Wallis

^*p < .05, ^**p < .01

Spearman’s correlations showed a modest positive association between age and DMSE (r_s = 0.124, p = .011), while BMI (r_s = − 0.051, p = .299) and HbA1c (r_s = − 0.063, p = .201) were not significantly correlated, indicating that higher self-efficacy does not necessarily correspond to better glycemic control in this context.

Predictors of DMSE

The regression model predicting DMSE was statistically significant, F(14, 404) = 4.83, p < .001, R² = 0.135, adjusted R² = 0.107. Four predictors were significant. Physical activity was the strongest predictor (β = −0.178, p < .001; B = − 0.134, 95% CI [− 0.207, − 0.062]). Longer diabetes duration was a significant negative predictor (β = −0.141, p = .010; B = − 0.104, 95% CI [− 0.182, − 0.025]). Higher monthly income (β = 0.158, p = .002) and educational level (β = 0.126, p = .023) were each independently associated with greater self-efficacy. All other predictors were non-significant (all p > .10). Full results are presented in Table 6.

Table 6.

Multiple linear regression predicting DMSE (N = 418)

Predictor	B	SE	β	t	p	95% CI LL	95% CI UL
(Constant)	4.037	0.398	—	10.13	< 0.001**	3.253	4.820
Age	0.001	0.003	0.031	0.485	0.628	-0.004	0.007
BMI	-0.006	0.006	-0.054	-1.070	0.285	-0.018	0.005
Gender	0.036	0.094	0.022	0.377	0.706	-0.150	0.221
Educational level	0.123	0.053	0.127	2.298	0.022 *	0.018	0.227
Monthly income	0.215	0.068	0.163	3.158	0.002 **	0.081	0.348
Physical activity frequency	-0.134	0.038	-0.177	-3.555	< 0.001 **	-0.208	-0.060
Type of diabetes	-0.065	0.093	-0.037	-0.699	0.485	-0.248	0.118
Duration of diabetes	-0.099	0.040	-0.135	-2.504	0.013 *	-0.177	-0.021
Last HbA1c test	0.011	0.021	0.025	0.520	0.604	-0.030	0.051
Diabetes complication(s)	-0.324	0.097	-0.175	-3.334	0.060	-0.515	-0.133
Current smoker (dummy)	0.184	0.111	0.086	1.648	0.100	-0.035	0.403
Former smoker (dummy)	-0.100	0.114	-0.049	-0.873	0.383	-0.324	0.125
Married (dummy)	-0.033	0.128	-0.016	-0.256	0.798	-0.285	0.220
Other marital status (dummy)	-0.213	0.218	-0.055	-0.980	0.328	-0.641	0.214

Physical activity coded from 1 = regular to 4 = never. Complications coded 1 = yes, 2 = no

B  Unstandardized coefficient, SE  Standard error, β  Standardized coefficient, CI  Confidence interval

^*p < .05, ^**p < .01

Discussion

This study investigated diabetes management self-efficacy, glycemic control, and their predictors among Palestinian patients with DM. The overall DMSE score showed that participants demonstrated a moderate level of confidence in their ability to manage their own diabetes, whereas median HbA1c levels indicated suboptimal glycemic control. Major independent predictors of HbA1c identified in regression analyses were complication status, smoking history, and physical activity. On the other hand, DMSE was shaped by physical activity, complications, diabetes duration, income, and educational level. These results are consistent with a convergence of clinical, behavioral, and socioeconomic factors that compromise outcomes in a healthcare setting impacted by conflict.

The high complication rate of 74.6% significantly exceeds the rate reported in Saudi Arabia (39.2%) [21] and cannot be attributed solely to disease biology. Continual conflict, military checkpoints, and persistent healthcare disruptions in Palestine [30–32], hinder early diagnosis, disrupt follow-up care, and cause sporadic medication shortages, conditions that perpetuate the uncontrolled hyperglycemia that leads to complications [5, 33, 34]. The prevalence of chronic disease (> 10 years in 51.9%) exacerbates this scenario [4].

The moderate DMSE score noted in this study is lower than the findings reported by Eshtaya et al. [10] among Palestinian T2DM patients, likely indicative of a broader and more complex sample. Blood glucose management was the subdomain with the highest score, which is expected, as it is a central component of diabetes management education for people who don’t stay in the hospital. Self-efficacy for physical exercise was the lowest. This is further exacerbated in Palestine by movement restrictions due to checkpoints and the separation barrier, chronic insecurity, and a scarcity of safe recreational spaces [30, 34]. The similarly low score for nutrition-specific situations shows how hard it is to stick to a healthy diet when you’re around other people, as shown in several DMSES cross-cultural adaptations [11, 14].

The lack of HbA1c disparities by gender or diabetes type probably indicates a homogenizing influence of common structural adversity. In contrast to income-stratified environments like Saudi Arabia, where socioeconomic disparities in glycemic control are significant [17], occupation-related restrictions in Palestine seem to uniformly hinder glycemic control across various patient subgroups [34]. Although marital status was not statistically significant in regression analyses, single and widowed/divorced participants showed numerically higher HbA1c levels than married individuals. This non-significant trend may suggest a possible role of family support in diabetes management, as spouses often act as medication reminders, dietary monitors, and appointment facilitators in Palestinian health culture [10]. However, this finding requires confirmation in larger studies.

The elevated HbA1c levels in former smokers compared to current smokers likely indicate the sick-quitter phenomenon, where individuals stop smoking due to declining health or the emergence of complications, suggesting that inadequate glycemic control occurred prior to cessation [35]. This finding warns against treating cessation as a complete glycemic intervention and stresses the need for ongoing metabolic monitoring after cessation. Physical inactivity consistently served as an independent predictor of elevated HbA1c in both group comparisons and regression analyses, corroborating substantial evidence that exercise enhances insulin sensitivity via direct metabolic pathways [4].

The insignificant correlation between DMSE and HbA1c is theoretically significant. In Palestine, patients may have self-management confidence but be unable to act on it due to poverty, insecurity, and lack of access to care [10, 33, 36]. This creates a disconnect between confidence and behavior, which Bandura’s [8] SCT recognizes when environmental affordances are lacking. The combination of physical inactivity, complications, long disease duration, low income, and low education as separate DMSE predictors identifies a vulnerable subgroup whose self-management confidence is steadily undermined by a confluence of these factors. With 57.9% of participants earning less than 2,000 NIS monthly, indicative of occupation-driven economic contraction, self-efficacy interventions devoid of economic support will rapidly encounter their limitations within this demographic [5, 10, 19].

Limitations

Several limitations should be considered. The cross-sectional design precludes causal inference, and convenience sampling from clinics in the Hebron governorate limits generalizability to other Palestinian regions and to patients with restricted healthcare access. The findings are therefore representative only of patients attending MOH diabetes clinics in the Hebron governorate.

Both regression models explained modest variance, suggesting that unmeasured factors, including diabetes-related distress, social support, medication adherence, and dietary quality, independently shape these outcomes. Social desirability may have inflated DMSE scores. Psychosocial variables and type-stratified analyses were not included. The unique Palestinian structural context limits transferability to stable healthcare settings.

The DMSES was originally validated for T2DM. However, 30.4% of our sample had T1DM. DMSES domains do not fully capture T1DM self-management tasks (insulin titration, carbohydrate counting, hypoglycemia management). Type-stratified analyses were not feasible due to sample size. This may affect validity of DMSE scores for T1DM participants.

Structured diabetes self-management education is not routinely available in the Hebron MOH setting. Patients receive fragmented, opportunistic counseling. This absence may have lowered DMSE scores independent of actual capacity.

Although HbA1c values were obtained from clinical records, variability in laboratory equipment and quality control across MOH facilities cannot be ruled out.

We did not disaggregate results by facility type (hospital vs. PHC), limiting setting-specific conclusions for primary care audiences.

Recommendations and implications

Organizational and hospital level

• Integrate routine DMSE screening alongside HbA1c monitoring to identify patients requiring targeted self-management support before glycemic control deteriorates.

• Establish nurse-led diabetes education programs prioritizing physical activity and dietary counseling, the two lowest-scoring DMSE subdomains.

• Strengthen complication screening pathways, given that 74.6% of participants reported at least one complication, a key independent predictor of both poor HbA1c and low DMSE.

• Establish or strengthen electronic documentation systems and patient registries to enable data linkage between primary healthcare centers and hospital diabetes clinics, reducing service duplication and supporting continuity of care.

Individual and practice level

• Tailor self-management goal-setting to each patient’s DMSE subdomain profile, targeting physical exercise and nutrition-specific confidence first.

• Provide enhanced post-cessation metabolic monitoring, as former smokers showed paradoxically higher HbA1c, consistent with the sick-quitter phenomenon.

• For single or unmarried patients, assess family support availability during routine visits; where support is limited, prioritize brief, structured self-management counseling during clinic encounters rather than assuming availability of external peer networks, given current fragmentation of community health services.

Educational and academic level

• Embed DMSES-based self-efficacy assessment in nursing and allied health curricula to equip future clinicians with validated tools for identifying self-management deficits.

• Prioritize longitudinal and mixed-methods research designs to establish causal pathways and capture psychosocial dimensions absent from this study.

Policy and societal level

• Address structural determinants of poor self-management—poverty, movement restrictions, and healthcare inaccessibility—through policy reform, recognizing that self-efficacy interventions have limited reach when environmental barriers remain unresolved.

• Guarantee uninterrupted supply of diabetes medications and monitoring supplies, and fund community-based education programs targeting low-income and low-education patients.

• Establish a governorate-level electronic diabetes registry within MOH Hebron facilities to enable data linkage between primary healthcare centers and hospital clinics, reducing duplication and supporting efficient resource allocation. Pilot this registry model in selected MOH or NGO settings before scaling.

• Address fragmentation of health service delivery by establishing formal coordination mechanisms between MOH facilities and other providers (including UNRWA clinics and NGO services), particularly for structured nutrition education.

• All recommended interventions should be piloted with formal evaluation to generate evidence for policymakers.

Conclusion

This study provides the first evidence from Palestine simultaneously examining diabetes management self-efficacy and glycemic control as co-primary outcomes across both T1DM and T2DM patients. Suboptimal HbA1c and moderate DMSE scores were identified, with former smoking, diabetes-related complications, and physical inactivity serving as independent predictors of HbA1c, while physical inactivity, longer disease duration, lower income, and lower education predicted reduced DMSE. The absence of a direct DMSE–HbA1c correlation underscores a distinctly Palestinian reality: patients may possess self-management confidence yet remain structurally prevented from acting on it when poverty, insecurity, and fragmented care intervene. Effective improvement in diabetes outcomes in Palestine therefore requires integrating behavioral and clinical interventions with structural reforms that address the socioeconomic and political determinants of health. Future research should adopt longitudinal designs, incorporate psychosocial measures, and extend across Palestinian governorates to build a comprehensive evidence base for diabetes care planning.

Abbreviations

ANOVA

Analysis of Variance

CI

Confidence Interval

DM

Diabetes Mellitus

DMSE

Diabetes Management Self—Efficacy

DMSES

Diabetes Management Self—Efficacy Scale

HbA1c

Glycated Hemoglobin

MOH

Ministry of Health

NIS

New Israeli Shekel

PHC

Primary Healthcare Center

QoL

Quality of Life

SD

Standard Deviation

SE

Standard Error

SPSS

Statistical Package for the Social Sciences

T1DM

Type 1 Diabetes Mellitus

T2DM

Type 2 Diabetes Mellitus

Authors’ contributions

**FF** : Conceptualization, methodology, supervision, formal analysis, investigation, data curation, writing – original draft preparation, writing – review and editing, project administration, and correspondence. **AH** : Conceptualization, methodology, investigation, and data collection. **AJA** : Methodology, investigation, writing – review and editing. **AS** : Conceptualization, methodology, investigation, and data collection. **MAA** : Conceptualization, methodology, investigation, and data collection. **MR** : Writing – review and editing. All authors read and approved the final manuscript and agreed to be accountable for all aspects of the work.

Funding

The authors have not received any funding for this study.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the Education in Health and Scientific Research Unit, Palestinian Ministry of Health (MOH), West Bank, Palestine (Approval Ref: 2025/3493/162). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki. All participants were provided with comprehensive information regarding the study’s purpose, procedures, potential risks, and benefits. Written informed consent was obtained from every participant prior to data collection. Participants were informed that their involvement was entirely voluntary and that they could withdraw at any time without any penalty or impact on their healthcare access. No personally identifiable information was collected; participants were assigned numerical codes for anonymity. All data were stored in password-protected files accessible only to the research team.

Consent for publication

Not applicable. This manuscript does not contain any individual person’s data in any form (including individual details, names, images, or videos).

Competing interests

The authors declare no competing interests.