Knowledge, Attitudes, and Practices Toward Continuous Glucose Monitoring Among Diabetes Educators, Nurses, and Pharmacists in Saudi Arabia: A Cross-Sectional Survey

Abstract

Purpose

Continuous glucose monitoring (CGM) is increasingly used in diabetes care. However, data on CGM-related knowledge, attitudes, and practices among non-physician healthcare professionals in Saudi Arabia are limited. This study evaluated these domains among diabetes educators, nurses, and pharmacists.

Patients and Methods

We conducted a cross-sectional web-based survey of licensed diabetes educators, nurses, and pharmacists involved in diabetes care in Saudi Arabia. The survey assessed knowledge of guideline-aligned CGM metrics, attitudes toward CGM use, practice patterns, perceived barriers, and training needs. Bivariate analyses and Firth penalized logistic regression were used to identify predictors of routine ambulatory glucose profile (AGP) review.

Results

Fifty-three healthcare professionals participated, including 24 nurses, 20 diabetes educators, and 9 pharmacists. Knowledge of CGM metrics was modest overall. Correct responses ranged from 24.5% for glucose variability metrics to 53.1% for limitations of glycated hemoglobin interpretation. Only 26.5% achieved adequate knowledge scores. Overall, 61.4% agreed that CGM improves safety and quality of life in older adults using insulin, and 63.6% reported confidence using time in range and trend arrows during patient counseling. Routine AGP review was reported by 76.2% of respondents and was most common among diabetes educators. In multivariable analysis, diabetes educator role remained independently associated with routine AGP review (adjusted odds ratio 23.57; 95% confidence interval 5.32–82.49). Cost and coverage were the most commonly reported barriers to CGM use. Case-based workshops and workplace teaching were the preferred training formats.

Conclusion

Healthcare professionals demonstrated positive attitudes toward CGM, but important knowledge gaps remain. Diabetes educators reported greater confidence and more frequent use of CGM reports. Targeted multidisciplinary training and improved access to CGM may support wider implementation in routine diabetes care. Given the small sample size, these findings should be interpreted cautiously and may not be generalizable to the broader healthcare workforce in Saudi Arabia.

Introduction

For many years, measuring blood glucose at home using a glucometer in addition to A1C testing was the standard for monitoring diabetes.¹ In recent years, continuous glucose monitoring (CGM) devices have become widespread, and technology has improved significantly.^1,2 These devices track glucose in interstitial fluid continuously producing near real-time glucose values and trends. The most used devices are the real-time CGM (rtCGM) with continuous display and alerts and the intermittently scanned CGM (iCGM), where data are shown when the user scans the sensor.^2,3

The adoption of CGM in managing type 1 diabetes mellites (T1DM) is common and might be considered a standard of care in many healthcare systems.^1,3,4 The use among type 2 diabetes mellites (T2DM) patients has increased but remains below that of T1DM. Recent clinical guidelines and consensus reports recommend that all people with T1DM should be offered CGM. It is also strongly considered for patients with T2DM on intensive insulin regimens such as multiple daily injections or insulin pump when hypoglycemia is a concern.^1–3 This also extends to patients with hypoglycemia unawareness, recurrent severe hypoglycemia, or marked glycemic variability. Important to note that this should be in addition to A1C testing and not as a replacement.^1,2

CGM provides set of metrics that go beyond a single HbA1c value. Time in range (TIR), time above and below range (TAR and TBR), mean glucose, and glucose variability often expressed as standard deviation or coefficient of variation are now recognized as core measures for routine practice.^1,2 International consensus reports recommend using these metrics, particularly TIR and TBR, as day-to-day therapeutic targets. They relate to both overall control and the risk of hypoglycemia and complications.¹ To make these data easier to interpret, the ambulatory glucose profile (AGP) has been proposed as a standardized single-page report that summarizes several days to weeks of CGM data into a typical 24-hour profile with key statistics.^2,5

This is relevant in Saudi Arabia and the wider Gulf region, where the burden of T2DM is very high and glycemic control remains suboptimal.^3,5,6 International data have reported that CGM uptake remains limited among healthcare professionals outside endocrinology settings.^7–9 Recent survey-based studies in the United States, United Kingdom, Singapore, South Korea, and multi-country cohorts have highlighted critical knowledge and implementation gaps among healthcare professionals.^7–11 Limited healthcare professionals training and confidence remain common barriers to CGM use. Also, cost, workflow constraints, and lack of institutional support have limited usage across practice settings. However, much of the existing literature has focused on perceived familiarity, confidence, and implementation barriers, with relatively limited objective assessment of healthcare professionals’ knowledge of guideline-aligned CGM metrics.

Managing diabetes requires a multidisciplinary team approach which includes physicians, nurses, pharmacists, health care educators or diabetes educators and others. Given their direct involvement in diabetes care and patient education, diabetes educators, nurses, and pharmacists represent important non-physician professional groups for understanding multidisciplinary CGM implementation. Nurses report strong acceptance of inpatient CGM, and pharmacists are emerging as key implementers in some healthcare systems.^10,12 Despite this growing international evidence, to our knowledge no published study has evaluated and compared objective knowledge of guideline-aligned CGM metrics, attitudes, and practice patterns among diabetes educators, nurses, and pharmacists in Saudi Arabia. Accordingly, we conducted a cross-sectional survey of diabetes educators, nurses, and pharmacists practicing in Saudi Arabia to characterize knowledge, attitudes, and practices toward CGM. The primary objectives were to estimate the prevalence of guideline-aligned CGM knowledge, to characterize professional attitudes toward CGM in routine clinical care, and to quantify the routine use of CGM-derived reports, specifically the AGP and TIR, in patient counseling. Secondary objectives were to identify respondent-level predictors of routine CGM report use, including professional role, years in practice, geographic region, care setting, recent CGM exposure, self-rated confidence, and adequate composite knowledge; to characterize perceived barriers to CGM implementation; and to determine the educational and training formats preferred by respondents for future implementation.

Methods

Study Design and Setting

This was a cross-sectional, web-based anonymous survey study directed to licensed diabetes educators, nurses, and pharmacists currently practicing in Saudi Arabia who provided diabetes care.

Survey Development and Structure

The survey was adapted from previously published healthcare provider CGM surveys^7,8,13 and aligned with the 2025 standards of diabetes care and international CGM consensus standards.^1,2,14

The screening section confirmed eligibility by assessing professional status and involvement in diabetes care. The background section collected demographic and professional characteristics, including profession, years in practice, geographic region of practice, workplace setting, and recent exposure to CGM technology.

The knowledge section included four single-best-answer questions assessing guideline-concordant understanding of CGM metrics, specifically TIR targets, level 2 hypoglycemia thresholds, limitations of glycated hemoglobin (HbA1c), and the appropriate metric for daily glucose variability. These four items were regarded as a criterion-referenced, brief, nonvalidated knowledge index that was based on prespecified guideline-aligned responses. The index was designed to evaluate the knowledge of these specific CGM metrics and does not provide a comprehensive assessment of the broader CGM knowledge construct.

The attitude section used a 5-point Likert scale to evaluate perceptions regarding CGM benefits, confidence in interpreting CGM data, and perceived patient-related barriers. The practice and barriers section assessed the frequency of AGP review, major barriers to CGM implementation, preferred educational or training formats, and perceived professional responsibilities in CGM data interpretation. At the end of the survey, respondents were also invited to provide a single open-ended suggestion regarding what one change would increase CGM use at their institution.

Content validity, clarity, and relevance were assessed. The survey was reviewed by two health educators working in an endocrinology clinic, one consultant diabetologist, and one clinical pharmacist. The survey underwent consecutive rounds of review until the final version was reached.

Participant Recruitment and Data Collection

The survey was distributed electronically to eligible professionals through the Saudi Commission for Health Specialties, reaching the maximum number of professionals their distribution system permitted. Eligible healthcare professionals received an Email invitation containing the survey link. At least three reminder emails were distributed at one-week intervals to improve the response rate. Participation was voluntary and anonymous. Before accessing the questionnaire, participants were presented with an electronic study information page describing the study objectives, eligibility criteria, estimated completion time, confidentiality, and ethical approval. Participants were informed that by proceeding to the survey they confirmed their eligibility and provided informed consent to participate voluntarily in the study. No personally identifiable information was collected. Data were stored securely in a password-protected Microsoft Excel file accessible only to the research team. Because survey distribution was conducted through the Saudi Commission for Health Specialties, the exact number of healthcare professionals who received the survey invitation was not available to the authors. Therefore, a formal response rate could not be calculated.

Statistical Analysis

Continuous variables were summarized as median (interquartile range) or mean (standard deviation), as appropriate, while categorical variables were summarized as frequencies and percentages. Bivariate associations across the three professions were assessed using the Fisher exact test for categorical variables, in preference to the asymptotic chi-square test, given the small expected cell counts at this sample size, and using the Kruskal–Wallis test for ordinal and non-normally distributed continuous variables. Wilson 95% confidence intervals were reported for proportions, in preference to the Wald (normal-approximation) interval, as Wilson intervals retain accurate coverage at small sample sizes.

A composite knowledge score was constructed as the count of correct responses across the four knowledge items (range, 0–4), and adequate knowledge was defined a priori as a score of ≥3 of 4 correct. Prior CGM exposure (Q7) was operationalized as a three-level hierarchical categorical variable (real-time CGM, intermittently scanned CGM, or professional CGM, with respondents reporting multiple modalities assigned to the highest tier), with intermittently scanned CGM retained as the reference category. Geographic region of practice was collapsed a priori from five categories to two (Central versus non-Central) to maintain stable estimates given the small expected counts in the four non-Central regions.

Independent predictors of routine AGP review (the primary practice outcome, defined as ≥2 reviews per month) were examined using Firth penalized maximum-likelihood logistic regression, which provides finite, asymptotically unbiased coefficient estimates under complete separation and small-sample conditions.^15–17 Five of nine candidate predictors met the pre-specified bivariate entry threshold of P < 0.20. Two of those, formal CGM training and patient CGM volume, were excluded a priori because of substantial collinearity with professional role at this sample size and complete separation with the outcome, retaining profession as the principal predictor.

The final multivariable model retained profession, self-rated confidence, and adequate knowledge. Confidence intervals for the adjusted odds ratios were estimated by bias-corrected and accelerated (BCa) nonparametric bootstrap with 1000 resamples (random seed 12345 for reproducibility),¹⁸ in preference to the asymptotic profile-likelihood interval, given the small analytic sample and the complete separation observed in the role-by-outcome cross-tabulation.

Free-text responses to the open-ended question on facility-level CGM uptake were summarized descriptively by frequency of theme; no formal qualitative analysis was performed. Two-sided P values < 0.05 were considered statistically significant. All analyses were performed using Stata version 19.0 (StataCorp, College Station, Texas), with the Firth model implemented via the user-written firthlogit package.¹⁹ No a priori sample-size calculation was performed for this study.

Ethical Considerations

The study was conducted in accordance with the ethical principles of the Declaration of Helsinki and local institutional research regulations. The study protocol was reviewed and approved by the Regional Research Ethics Committee, Qassim Province, Ministry of Health, Saudi Arabia, registered with the National Committee of Bio and Med. Ethics (NCBE) (Registration No. H-04-Q-001), under Ethics Approval No. 607–47-5879, on October 29, 2025. Before accessing the questionnaire, participants were presented with an electronic study information page and informed that proceeding to the survey indicated their informed consent to participate voluntarily and anonymously. No personally identifiable information was collected.

Results

Respondent Characteristics and Survey Completion Rate

A total of 53 eligible healthcare professionals participated in the survey, including 24 nurses (45.3%), 20 diabetes educators (37.7%), and 9 pharmacists (17.0%) (Table 1). Respondents were distributed across several geographic regions. The Central region accounted for 35.8% of the sample, followed by the Western (24.5%) and Eastern (20.8%) regions. Regional distribution differed across professions (P = 0.046).

Table 1.

Characteristics of the Respondents^a

Characteristic	Overall (N = 53)	Nurse (N = 24)	Pharmacist (N = 9)	Diabetes Educator (N = 20)	P Valueb
Region — no. (%)					0.046
Central	19 (35.8)	9 (37.5)	3 (33.3)	7 (35.0)
Eastern	11 (20.8)	6 (25.0)	2 (22.2)	3 (15.0)
Western	13 (24.5)	3 (12.5)	1 (11.1)	9 (45.0)
Northern	9 (17.0)	6 (25.0)	3 (33.3)	0
Southern	1 (1.9)	0	0	1 (5.0)
Years in practice — no. (%)					0.15
<2 yr	5 (9.4)	0	3 (33.3)	2 (10.0)
2–5 yr	33 (62.3)	15 (62.5)	5 (55.6)	13 (65.0)
6–10 yr	11 (20.8)	7 (29.2)	1 (11.1)	3 (15.0)
>10 yr	4 (7.5)	2 (8.3)	0	2 (10.0)
Workplace — no. (%)					0.003
Governmental hospital	17 (32.1)	11 (45.8)	3 (33.3)	3 (15.0)
Private hospital	18 (34.0)	10 (41.7)	1 (11.1)	7 (35.0)
Diabetes center	6 (11.3)	0	0	6 (30.0)
Other	12 (22.6)	3 (12.5)	5 (55.6)	4 (20.0)
CGM type encountered in past 3 mo — no./total no. (%)c
Real-time CGM	22/39 (56.4)	1/13 (7.7)	4/7 (57.1)	17/19 (89.5)	<0.001
Intermittently scanned CGM	27/39 (69.2)	9/13 (69.2)	5/7 (71.4)	13/19 (68.4)	1.00
Professional CGM	7/39 (17.9)	3/13 (23.1)	1/7 (14.3)	3/19 (15.8)	0.86
None	0/39	0/13	0/7	0/19	—
Formal CGM training — no. (%)					<0.001
Yes	35 (66.0)	14 (58.3)	2 (22.2)	19 (95.0)
No	18 (34.0)	10 (41.7)	7 (77.8)	1 (5.0)
Estimated patients currently using CGM — no. (%)					<0.001
<10%	10 (18.9)	6 (25.0)	3 (33.3)	1 (5.0)
10–30%	11 (20.8)	8 (33.3)	2 (22.2)	1 (5.0)
>30%	22 (41.5)	4 (16.7)	1 (11.1)	17 (85.0)
Not sure	10 (18.9)	6 (25.0)	3 (33.3)	1 (5.0)

Notes: ^aCGM: continuous glucose monitoring. Percentages may not total 100 because of rounding. ^bP values were calculated with Fisher’s exact test for differences in the distribution across the three professional roles. ^cThe denominators reflect the number of respondents who reached this question. This was a multi-select item; each option is reported independently and percentages do not sum to 100.

Most respondents had 2 to 5 years of professional experience (62.3%). Eleven respondents (20.8%) had 6 to 10 years of experience, and 4 (7.5%) had more than 10 years in practice. Years in practice did not differ significantly across professions (P = 0.15). Workplace setting differed across professional groups (P = 0.003). Governmental hospitals accounted for 32.1% of practice settings, followed by private hospitals (34.0%), diabetes centers (11.3%), and other settings (22.6%).

Exposure to CGM technologies during the previous 3 months varied across professions. Overall, 56.4% of the 39 respondents who answered this item reported real-time CGM exposure, with significant differences across groups (P < 0.001; Table 1). Exposure to intermittently scanned CGM was reported by 69.2% of the 39 respondents and did not differ significantly across professions (P = 1.00). Professional CGM exposure was reported by 17.9% of the 39 respondents who answered this item (P = 0.86). Overall, 66.0% of respondents reported prior formal CGM training. Formal training differed significantly across professions (P < 0.001). The estimated proportion of patients currently using CGM also differed across professional groups (P < 0.001).

Survey completion was incomplete for a substantial proportion of respondents and varied across professional roles. Of the 53 eligible respondents, 38 (71.7%) completed the full survey, with the remaining 15 (28.3%) discontinuing at varying points across the four sections. Completion rates were considerably higher among diabetes educators (19 of 20; 95.0%) than among nurses (14 of 24; 58.3%) or pharmacists (5 of 9; 55.6%). Per-section and per-item analytic denominators, which vary because of progressive non-response, are reported in Tables 1–6 and reflect this differential pattern of attrition.

Table 2.

Knowledge of CGM Metrics and Glycemic Targets, Overall and According to Professional Role^a

Knowledge Item	Overall (N = 49)	Nurse (N = 22)	Pharmacist (N = 7)	Diabetes Educator (N = 20)	P Valueb
Correct response to individual knowledge items — no. (%)
Q10. TIR target (≥70% of readings)c	24 (49.0)	7 (31.8)	2 (28.6)	15 (75.0)	0.012
Q11. Level 2 hypoglycemia (<54 mg/dL)	21 (42.9)	11 (50.0)	2 (28.6)	8 (40.0)	0.54
Q12. HbA1c inaccurate in anemia or CKD	26 (53.1)	7 (31.8)	6 (85.7)	13 (65.0)	0.022
Q13. Glucose variability metric (SD)d	12 (24.5)	4 (18.2)	2 (28.6)	6 (30.0)	0.66
Composite knowledge score (0–4)d
Mean ± SD	1.69 ± 1.18	1.32 ± 1.21	1.71 ± 1.38	2.10 ± 0.97	0.081e
Median (IQR)	2 (1–3)	1 (0–2)	1 (1–3)	2 (1.5–3)
Score distribution — no. (%)
0 correct	9 (18.4)	6 (27.3)	2 (28.6)	1 (5.0)
1 correct	13 (26.5)	7 (31.8)	1 (14.3)	5 (25.0)
2 correct	14 (28.6)	5 (22.7)	2 (28.6)	7 (35.0)
3 correct	10 (20.4)	3 (13.6)	1 (14.3)	6 (30.0)
4 correct	3 (6.1)	1 (4.5)	1 (14.3)	1 (5.0)
Adequate knowledge (≥3 of 4) — no. (%)	13 (26.5)	4 (18.2)	2 (28.6)	7 (35.0)	0.46

Notes: ^aThe knowledge analytic cohort comprised the 49 of 53 respondents who answered Section B; 2 nurses and 2 pharmacists did not reach this section. CGM denotes continuous glucose monitoring; CKD chronic kidney disease; HbA1c glycated hemoglobin; IQR interquartile range; SD standard deviation; TIR time in range. Percentages may not total 100 because of rounding. ^bFor binary item correctness and the binary adequate-knowledge variable, P values were calculated with Fisher’s exact test. For the composite score, the P value was calculated with the Kruskal–Wallis test (see ^e). ^cTIR refers to the proportion of CGM readings in the target glycemic range of 70 to 180 mg/dL. ^dThe original survey instrument keyed Q13 (best CGM metric for daily glucose variability) to TIR. However, the International Consensus on Time in Range (Battelino et al¹) and the ADA Standards of Care identify the coefficient of variation (CV) as the metric of choice for glycemic variability, with SD as the next-best alternative; TIR is a time-at-target metric, not a variability metric. Because CV was not among the response options offered, SD was used as the keyed-correct answer in this analysis. This decision was made before any inferential analysis was performed (see Methods). ^eP value for the composite knowledge score is from the Kruskal–Wallis equality-of-populations rank test, with correction for ties.

Table 3.

Attitudes Toward CGM Use, Overall and According to Professional Role^a

Item/Response	Overall (N = 44)	Nurse (N = 19)	Pharmacist (N = 5)	Diabetes Educator (N = 20)	P Valueb
Q14. CGM improves safety and quality of life in adults ≥65 yr who use insulin
Full distribution — no. (%)					0.26
Strongly disagree	7 (15.9)	2 (10.5)	2 (40.0)	3 (15.0)
Somewhat disagree	1 (2.3)	0	0	1 (5.0)
Neutral	9 (20.5)	7 (36.8)	0	2 (10.0)
Somewhat agree	5 (11.4)	3 (15.8)	0	2 (10.0)
Strongly agree	22 (50.0)	7 (36.8)	3 (60.0)	12 (60.0)
Median (IQR)	4.5 (3–5)	4 (3–5)	5 (1–5)	5 (3–5)	0.62c
Agree (somewhat/strongly) — no. (%)	27 (61.4)	10 (52.6)	3 (60.0)	14 (70.0)	0.57
Q15. I feel confident using trend arrows/TIR to guide patient counselingd
Full distribution — no. (%)					0.007
Strongly disagree	4 (9.1)	1 (5.3)	2 (40.0)	1 (5.0)
Somewhat disagree	2 (4.5)	1 (5.3)	1 (20.0)	0
Neutral	10 (22.7)	7 (36.8)	0	3 (15.0)
Somewhat agree	14 (31.8)	8 (42.1)	1 (20.0)	5 (25.0)
Strongly agree	14 (31.8)	2 (10.5)	1 (20.0)	11 (55.0)
Median (IQR)	4 (3–5)	4 (3–4)	2 (1–4)	5 (4–5)	0.013c
Agree (somewhat/strongly) — no. (%)	28 (63.6)	10 (52.6)	2 (40.0)	16 (80.0)	0.096
Q16. Alarm fatigue and skin issues are significant barriers for my patients
Full distribution — no. (%)					0.68
Strongly disagree	4 (9.1)	3 (15.8)	0	1 (5.0)
Somewhat disagree	2 (4.5)	0	1 (20.0)	1 (5.0)
Neutral	12 (27.3)	5 (26.3)	1 (20.0)	6 (30.0)
Somewhat agree	19 (43.2)	7 (36.8)	3 (60.0)	9 (45.0)
Strongly agree	7 (15.9)	4 (21.1)	0	3 (15.0)
Median (IQR)	4 (3–4)	4 (3–4)	4 (3–4)	4 (3–4)	0.92c
Agree (somewhat/strongly) — no. (%)	26 (59.1)	11 (57.9)	3 (60.0)	12 (60.0)	1.00
Q17. I find CGM data too complex to integrate into my patient counselinge
Full distribution — no. (%)					0.003
Strongly disagree	16 (36.4)	2 (10.5)	2 (40.0)	12 (60.0)
Somewhat disagree	11 (25.0)	6 (31.6)	0	5 (25.0)
Neutral	6 (13.6)	5 (26.3)	0	1 (5.0)
Somewhat agree	8 (18.2)	5 (26.3)	1 (20.0)	2 (10.0)
Strongly agree	3 (6.8)	1 (5.3)	2 (40.0)	0
Median (IQR)	2 (1–3.5)	3 (2–4)	4 (1–5)	1 (1–2)	0.005c
Disagree (somewhat/strongly) — no. (%)e	27 (61.4)	8 (42.1)	2 (40.0)	17 (85.0)	0.010

Notes: ^aThe attitudes analytic cohort comprised the 44 of 53 respondents who answered Section C; 5 nurses and 4 pharmacists did not reach this section. Items were rated on a 5-point Likert scale (strongly disagree to strongly agree). Higher Likert scores indicate stronger agreement with the statement as written. CGM denotes continuous glucose monitoring; IQR interquartile range; TIR time in range. Percentages may not total 100 because of rounding. ^bFor full Likert distributions and dichotomized agree-rates, P values were calculated with Fisher’s exact test. ^cP values for medians are from the Kruskal–Wallis equality-of-populations rank test, with correction for ties. ^dThe dichotomized version of Q15 (Agree vs other) was used as a candidate predictor in the multivariable model of routine AGP review (Table 6). ^eQ17 is reverse-coded relative to the other items: disagreement with the statement “CGM data are too complex” represents the favorable attitude. The dichotomized row therefore reports the proportion who disagreed (somewhat or strongly) — ie, who do NOT find CGM data too complex.

Table 4.

Practice Patterns and Perceived Barriers Related to CGM, Overall and According to Professional Role^a

Item/Response	Overall	Nurse	Pharmacist	Diabetes Educator	P Valueb
Q18. AGP review frequency in past month — no. (%)	(N = 42)	(N = 17)	(N = 5)	(N = 20)	<0.001
Never	8 (19.0)	6 (35.3)	2 (40.0)	0
Once	2 (4.8)	1 (5.9)	1 (20.0)	0
2–3 times	10 (23.8)	7 (41.2)	1 (20.0)	2 (10.0)
>3 times	22 (52.4)	3 (17.6)	1 (20.0)	18 (90.0)
Median (IQR)	3 (2–3)	2 (0–2)	1 (0–2)	3 (3–3)	<0.001c
Routine review (≥2 times) no. (%)d	32 (76.2)	10 (58.8)	2 (40.0)	20 (100)	0.001
Q19. Biggest barrier to CGM use — no. (%)	(N = 42)	(N = 17)	(N = 5)	(N = 20)	0.51
Coverage or cost	13 (31.0)	4 (23.5)	1 (20.0)	8 (40.0)
Device supply or availability	8 (19.0)	4 (23.5)	1 (20.0)	3 (15.0)
Other	7 (16.7)	4 (23.5)	2 (40.0)	1 (5.0)
Time or workflow limitations	5 (11.9)	1 (5.9)	1 (20.0)	3 (15.0)
Patient acceptance or visibility	5 (11.9)	1 (5.9)	0	4 (20.0)
Skin or adhesive issues	3 (7.1)	2 (11.8)	0	1 (5.0)
Training to interpret CGM reports	1 (2.4)	1 (5.9)	0	0
Q20. Preferred training format — no. (%)	(N = 42)	(N = 17)	(N = 5)	(N = 20)	0.18
Case-based workshop	15 (35.7)	4 (23.5)	2 (40.0)	9 (45.0)
Short in-person teaching at workplace	12 (28.6)	5 (29.4)	0	7 (35.0)
Self-paced e-learning	6 (14.3)	2 (11.8)	2 (40.0)	2 (10.0)
1-hour webinar	4 (9.5)	4 (23.5)	0	0
Not needed	3 (7.1)	1 (5.9)	1 (20.0)	1 (5.0)
Arabic pocket guide	2 (4.8)	1 (5.9)	0	1 (5.0)
Q21. Person mainly reviewing CGM reports — no. (%)	(N = 41)	(N = 16)	(N = 5)	(N = 20)	0.13
Shared responsibility	16 (39.0)	6 (37.5)	2 (40.0)	8 (40.0)
Physician	15 (36.6)	8 (50.0)	3 (60.0)	4 (20.0)
Diabetes educator	10 (24.4)	2 (12.5)	0	8 (40.0)
Q22. Comfort counseling without physician — no. (%)	(N = 38)	(N = 14)	(N = 5)	(N = 19)	0.16
Yes, following a protocol	24 (63.2)	11 (78.6)	2 (40.0)	11 (57.9)
Yes, independently	10 (26.3)	1 (7.1)	2 (40.0)	7 (36.8)
No, prefer physician input	4 (10.5)	2 (14.3)	1 (20.0)	1 (5.3)
Q23. Training topics needed — no. (%)e	(N = 42)	(N = 17)	(N = 5)	(N = 20)
Interpreting AGP	21 (50.0)	7 (41.2)	3 (60.0)	11 (55.0)	0.68
Patient device-use training	21 (50.0)	11 (64.7)	1 (20.0)	9 (45.0)	0.21
Managing skin or adhesive issues	19 (45.2)	7 (41.2)	3 (60.0)	9 (45.0)	0.83
Trend arrows and TIR in counseling	17 (40.5)	9 (52.9)	2 (40.0)	6 (30.0)	0.37
Setting and managing alarms	14 (33.3)	7 (41.2)	2 (40.0)	5 (25.0)	0.58
Mean number of topics selected (range 0–5)	2.2 ± 1.6	2.4 ± 2.0	2.2 ± 1.6	2.0 ± 1.2	0.92c

Notes: ^aThe practice analytic cohort comprised respondents who reached Section D; cohort sizes vary by question because of progressive non-response. AGP denotes ambulatory glucose profile; CGM continuous glucose monitoring; IQR interquartile range; TIR time in range. Percentages may not total 100 because of rounding. ^bP values were calculated with Fisher’s exact test, except where noted. ^cP values for medians and the mean number of training topics are from the Kruskal–Wallis equality-of-populations rank test, with correction for ties. ^dRoutine AGP review (defined as ≥2 reviews of an AGP/CGM summary in the past month) is the primary outcome of this study and is the dependent variable in the multivariable model in Table 6. ^eQ23 was a select-all-that-apply question; respondents could select more than one training topic. Each option is reported independently and percentages do not sum to 100. The mean number of topics selected per respondent is reported as a summary measure of training-need scope.

Table 5.

Bivariate Associations Between Respondent Characteristics and Routine AGP Review (N = 42)^a

Predictor/Subgroup	No. of Respondents	Routine AGP Review, no. (%)b	P Valuec	Candidate for Multivariable Modeld
Professional role			0.001	Yes
Nurse	17	10 (58.8)
Pharmacist	5	2 (40.0)
Diabetes educator	20	20 (100)
Years in practice			1.00	No
<6 yr	28	21 (75.0)
≥6 yr	14	11 (78.6)
Workplace			1.00	No
Diabetes center or other	14	11 (78.6)
Hospital (governmental or private)	28	21 (75.0)
Region			0.47	No
Non-Central	26	21 (80.8)
Central	16	11 (68.8)
Formal CGM training			0.002	Yese
No	12	5 (41.7)
Yes	30	27 (90.0)
CGM exposure type			0.31	No
is-CGM only	10	7 (70.0)
rt-CGM (± is-CGM)	18	16 (88.9)
Professional CGM (± others)	6	4 (66.7)
Patients currently using CGM			<0.001	Yese
<10%	8	4 (50.0)
10–30%	9	4 (44.4)
>30%	22	22 (100)
Not sure	3	2 (66.7)
Confident using trend arrows/TIR (Q15)			0.020	Yes
Disagree or neutral	15	8 (53.3)
Agree (somewhat or strongly)	27	24 (88.9)
Adequate knowledge (≥3 of 4 correct)			0.13	Yes
Inadequate (<3 of 4)	29	20 (69.0)
Adequate (≥3 of 4)	13	12 (92.3)

Notes: ^aThe bivariate analytic cohort comprised the 42 of 53 respondents who answered the AGP-review-frequency question (Q18). AGP denotes ambulatory glucose profile; CGM continuous glucose monitoring; TIR time in range. ^bRoutine AGP review is defined as ≥2 reviews of an AGP/CGM summary in the past month (vs never or once). The denominator for each percentage is the total number of respondents in that subgroup. ^cP values were calculated with Fisher’s exact test. ^dVariables with bivariate P < 0.20 were pre-specified as candidates for the multivariable Firth penalized logistic regression in Table 6. ^eAlthough these variables met the bivariate P < 0.20 entry threshold, they were not retained in the final multivariable model in Table 6 because of substantial empirical collinearity with professional role at this sample size. See Methods and Table 6 footnotes for details.

Table 6.

Crude and Adjusted Odds Ratios for Routine AGP Review, from Firth Penalized Logistic Regression with Bootstrap-Derived Confidence Intervals (N = 42)^a

Predictor	Crude OR (95% CI)b	Crude P value	Adjusted OR (95% CI)c	Adjusted P value
Professional role (reference: Nurse)
Pharmacist	0.51 (0.08–3.32)	0.48	0.41 (0.02–6.23)	0.42
Diabetes educatord	29.29 (1.52–563.89)	0.025	23.57 (5.32–82.49)	0.039
Self-rated confidence (Q15) (reference: Disagree/Neutral)
Agree (Somewhat or Strongly)	6.18 (1.39–27.39)	0.017	3.00 (0.24–24.86)	0.21
Adequate knowledge (≥3 of 4 correct) (reference: <3 of 4)
Adequate (≥3 of 4)	3.86 (0.60–24.81)	0.16	4.41 (0.04–20.99)	0.18

Notes: ^aOutcome: routine AGP review, defined as ≥2 reviews of an AGP/CGM summary in the past month. The analytic cohort comprised the 42 of 53 respondents with non-missing data on the outcome and all candidate predictors. Of nine predictors of the bivariate analysis, five met the pre-specified entry threshold (P < 0.20); formal CGM training and patient CGM volume were excluded from the multivariable model because of substantial collinearity with professional role and complete separation with the outcome, which would have destabilized the estimates. AGP denotes ambulatory glucose profile; CGM continuous glucose monitoring; CI confidence interval; OR odds ratio; TIR time in range. ^bCrude (unadjusted) odds ratios were estimated by separate Firth penalized logistic regression models, one per predictor, with 95% confidence intervals derived from the penalized profile likelihood. ^cAdjusted odds ratios are from a single Firth penalized logistic regression model containing all three predictors (professional role, self-rated confidence, and adequate knowledge) simultaneously (Wald χ²[4] = 7.87, overall P = 0.097). The 95% confidence intervals on the adjusted odds ratios were estimated by bias-corrected and accelerated (BCa) nonparametric bootstrap with 1000 resamples (random seed 12345 for reproducibility), in preference to the asymptotic profile-likelihood interval, given the small analytic sample and the complete separation observed in the role-by-outcome cross-tabulation. P values are from the Wald test on the original Firth penalized estimates and are therefore unaffected by the choice of confidence-interval method. ^dAll 20 diabetes educators reported routine AGP review, producing complete separation in the unpenalized data. Standard maximum-likelihood logistic regression would yield an infinite point estimate for this coefficient. Firth’s penalty (a Jeffreys-prior bias correction) returns a finite point estimate, and the BCa bootstrap distribution further provides empirical 95% confidence bounds that are not contingent on the asymptotic profile-likelihood approximation.

Knowledge of Guideline-Aligned CGM Metrics

The knowledge analysis included 49 respondents who completed Section B of the survey (Table 2). Overall, correct response rates varied substantially across the four knowledge items. The most frequently answered item was the limitation of HbA1c interpretation in anemia or chronic kidney disease, which was answered correctly by 26 respondents (53.1%). Correct identification of the recommended TIR target was reported in 24 respondents (49.0%), whereas 21 respondents (42.9%) correctly identified the threshold for level 2 hypoglycemia. The glucose variability metric (SD) question had the lowest proportion of correct responses, with 12 respondents (24.5%) answering correctly (Table 2).

Correct responses to the TIR target question differed significantly across professions (P = 0.012). Diabetes educators had the highest proportion of correct responses (75.0%), compared with nurses (31.8%) and pharmacists (28.6%). Correct identification of HbA1c limitations in anemia or chronic kidney disease also differed across professions (P = 0.022), with correct responses reported by 85.7% of pharmacists, 65.0% of diabetes educators, and 31.8% of nurses. In contrast, correct responses for level 2 hypoglycemia and glucose variability metrics did not differ significantly between groups (Table 2 and Figure 1).

Figure 1.

Item-level knowledge accuracy among diabetes educators, nurses, and pharmacists in Saudi Arabia, on four guideline-concordant CGM knowledge items. Bars are colored by professional role (blue, nurses; green, pharmacists; burgundy, diabetes educators). Horizontal black lines indicate 95% Wilson confidence intervals around each percentage.

Note: The analytic cohort comprised the 49 of 53 respondents who completed Section B of the survey.

Abbreviations: CGM, continuous glucose monitoring; CKD, chronic kidney disease; HbA1c, glycated hemoglobin; SD, standard deviation; TIR, time in range.

The overall mean composite knowledge score was 1.69 ± 1.18, with a median score of 2 (IQR, 1–3). Diabetes educators had the highest mean composite score (2.10 ± 0.97), followed by pharmacists (1.71 ± 1.38) and nurses (1.32 ± 1.21), although the difference did not reach statistical significance (P = 0.081). Adequate knowledge, defined as a score of at least 3 of 4 correct responses, was observed in 13 respondents (26.5%) and did not differ significantly across professions (P = 0.46) (Table 2).

Attitudes Toward CGM in Clinical Practice

The attitudes analysis included 44 respondents who completed Section C of the survey (Table 3). Overall, 61.4% of respondents agreed that CGM improves safety and quality of life in adults aged 65 years or older using insulin. Agreement rates did not differ significantly across professions (P = 0.57) (Table 3).

Overall, 63.6% of respondents somewhat or strongly agreed that they felt confident using trend arrows and TIR to guide patient counseling. Median confidence scores differed across professional groups (P = 0.013). Diabetes educators reported the highest median score (5 [IQR, 4–5]), compared with nurses (4 [IQR, 3–4]) and pharmacists (2 [IQR, 1–4]). The overall distribution of responses also differed significantly across professions (P = 0.007) (Table 3 and Figure 2).

Figure 2.

Distribution of Likert responses on four attitude items, visualized as a diverging stacked bar chart. Bars show percentage of respondents at each Likert level. Question 17 (Q17) retained in original direction; disagreement is the favorable response for this item.

Notes: The analytic cohort comprised the 44 of 53 respondents who completed Section C of the survey. The bars in this figure retain the original Likert direction (so that agreement extends rightward for all items) for visual consistency; this means that for Q17, leftward (red) bars represent the favorable response, in contrast to the other three items for which rightward (blue) bars are favorable.

Abbreviations: CGM, continuous glucose monitoring; TIR, time in range.

Alarm fatigue and skin-related issues were identified as important patient barriers by 59.1% of respondents overall. Agreement rates and median scores did not differ significantly between professional groups (Table 3). Overall, 61.4% of respondents disagreed with the statement that CGM data were too complex to integrate into patient counseling (indicating that most did not perceive CGM data as overly complex). Disagreement rates differed significantly across professions (P = 0.010). Diabetes educators had the highest proportion of disagreement responses (85.0%), compared with nurses (42.1%) and pharmacists (40.0%). Median scores also differed significantly across groups (P = 0.005) (Table 3 and Figure 2).

Practice Patterns and Perceived Barriers

The practice analysis included respondents who completed Section D of the survey (Table 4). Overall, 22 of 42 respondents (52.4%) reported reviewing an AGP or CGM summary more than 3 times during the previous month. Eight respondents (19.0%) reported never reviewing an AGP report. AGP review frequency differed significantly across professional groups (P < 0.001). Median AGP review frequency was highest among diabetes educators (median, 3 [IQR, 3–3]), compared with nurses (median, 2 [IQR, 0–2]) and pharmacists (median, 1 [IQR, 0–2]).

Routine AGP review, defined as reviewing an AGP or CGM summary at least twice during the previous month, was reported by 32 respondents (76.2%). This was driven largely by universal routine review among diabetes educators (20 of 20; 100%), compared with 58.8% of nurses and 40.0% of pharmacists. Routine review rates differed significantly across professions (P = 0.001) (Table 4).

Coverage or cost was the most reported barrier to CGM use (13 of 42 respondents; 31.0%), followed by device supply or availability (8 of 42; 19.0%). Training to interpret CGM reports was selected by 1 respondent (2.4%). Reported barriers did not differ significantly across professions (P = 0.51) (Table 4 and Figure 3A).

Figure 3.

Perceived barriers to CGM use (A) and self-reported training needs (B), as reported by Saudi diabetes educators, nurses, and pharmacists. (A) shows the single-select biggest barrier (Q19, n=42). (B) shows training topics needed (Q23, multi-select; n=42).

Notes: Both panels use the same analytic cohort of 42 respondents who completed Section D of the survey. The mean number of training topics selected per respondent in Q23 was 2.2 ± 1.6, with no significant difference across professional roles (P=0.92 by Kruskal–Wallis test; see Table 4).

Abbreviations: AGP, ambulatory glucose profile; CGM, continuous glucose monitoring; TIR, time in range.

The most preferred training format was a case-based workshop (35.7%), followed by short in-person workplace teaching (28.6%). Preferred training format did not differ significantly across professional groups (P = 0.18). Shared responsibility was the most reported model for CGM report review (39.0%), followed by physician-led review (36.6%). Ten respondents (24.4%) identified diabetes educators as the primary reviewers of CGM reports. Responses did not differ significantly across professions (P = 0.13) (Table 4).

Overall, 24 of 38 respondents (63.2%) reported being comfortable counseling patients regarding CGM following a protocol without physician presence, whereas 10 respondents (26.3%) reported being comfortable independently. Four respondents (10.5%) preferred physician input. Responses did not differ significantly across professions (P = 0.16).

The most frequently selected training topics were AGP interpretation and patient device-use training, each selected by 50.0% of respondents. Managing skin or adhesive issues was selected by 45.2% of respondents, followed by trend arrows and TIR in counseling (40.5%) and setting or managing alarms (33.3%). Training topic selection did not differ significantly across professions. The mean number of selected training topics per respondent was 2.2 ± 1.6 (Table 4 and Figure 3B).

At the end of the survey, respondents were invited to provide a single free-text response identifying the change they considered most likely to increase CGM use at their facility. A response was provided by 16 of the 53 eligible respondents (30.2%). Comments related to cost were the most frequent and were offered by 8 of the 16 respondents (50.0%). Including references to direct cost, insurance coverage, pricing, and access through the governmental health system. The remaining responses addressed workforce capacity (n = 2), educational materials (n = 1), device accuracy (n = 1), multidisciplinary physician engagement in CGM prescribing (n = 1), patient acceptance (n = 1), device availability for adults with type 2 diabetes (n = 1), and integration of CGM data with hospital information systems (n = 1).

Bivariate Predictors of Routine AGP Review

The bivariate analysis included 42 respondents who completed the AGP review frequency question (Table 5). A total of 32 respondents (76.2%) reported routine AGP review which was defined as reviewing an AGP or CGM summary at least twice during the previous month.

Professional role was significantly associated with routine AGP review (P = 0.001). Routine review was reported by all diabetes educators (100%), compared with 58.8% of nurses and 40.0% of pharmacists. Formal CGM training was also associated with routine AGP review (P = 0.002). Routine review was reported by 90% of respondents with formal CGM training, compared with 41.7% of those without training. The estimated proportion of patients currently using CGM differed significantly according to routine AGP review status (P < 0.001). Routine review was reported by all respondents who estimated that more than 30% of their patients were using CGM. Confidence using trend arrows and TIR for patient counseling was also associated with routine AGP review (P = 0.020). Routine review was reported by 88.9% of respondents who agreed that they felt confident using trend arrows and TIR, compared with 53.3% of those who disagreed or were neutral. These bivariate results indicate that professional role, formal CGM training, higher estimated patient CGM uptake, and greater confidence with trend arrows/TIR are associated with routine AGP review.

Adequate knowledge was not significantly associated with routine AGP review (P = 0.13). Years in practice, workplace setting, geographic region, and CGM exposure type were also not significantly associated with routine AGP review (Table 5).

Multivariable Predictors of Routine AGP Review

The multivariable analysis included 42 respondents with complete data on the outcome and candidate predictors. The Firth penalized logistic regression model simultaneously included professional role, adequate knowledge (≥3 of 4 correct responses), and confidence using trend arrows and TIR for patient counseling. In the adjusted model, diabetes educator role remained independently associated with routine AGP review. Compared with nurses, diabetes educators had higher odds of routine AGP review (adjusted OR, 23.57; 95% CI, 5.32–82.49; P = 0.039). The wide confidence interval reflects the small sample size (Table 6 and Figure 4).

Figure 4.

Forest plot of adjusted odds ratios for routine AGP review, from a firth penalized logistic regression model with bootstrap-derived 95% confidence intervals (N = 42). Adjusted odds ratios from a single Firth penalized logistic regression model (complete cases, n=42). Bootstrap BCa 95% confidence intervals from 1000 resamples.

Notes: Reference categories for the predictors are: Nurse (for the role comparisons), Disagree or Neutral (for the confidence comparison), and Inadequate knowledge (less than 3 of 4 correct, for the knowledge comparison). The model also included an intercept; the reported odds ratios are conditional on all other predictors in the model. P values reported in Table 6 are from the Wald test on the original Firth penalized estimates and are unaffected by the choice of confidence-interval method.

Abbreviations: AGP, ambulatory glucose profile; BCa, bias-corrected and accelerated (bootstrap method); CGM, continuous glucose monitoring; CI, confidence interval; OR, odds ratio; TIR, time in range.

Pharmacist role (adjusted OR, 0.41; 95% CI, 0.02–6.23; P = 0.42), self-rated confidence using trend arrows and TIR (adjusted OR, 3.00; 95% CI, 0.24–24.86; P = 0.21), and adequate knowledge (adjusted OR, 4.41; 95% CI, 0.04–20.99; P = 0.18) were not independently associated with routine AGP review in the adjusted model (Table 6).

Discussion

This cross-sectional survey evaluated CGM-related knowledge, attitudes, and practice patterns among diabetes educators, nurses, and pharmacists in Saudi Arabia. Knowledge of guideline-aligned CGM metrics was modest overall and varied among healthcare professionals, particularly for glucose variability metrics and TIR targets. Overall, respondents showed positive attitudes toward CGM use in clinical practice. Routine AGP review was relatively common. Diabetes educators consistently demonstrated higher levels of CGM exposure, formal training, confidence, and AGP review frequency compared with nurses and pharmacists. Cost and coverage were the most reported barriers to CGM use, while case-based workshops and workplace teaching were the preferred educational formats.

International consensus recommendations have emphasized the importance of standardized CGM metrics for routine CGM interpretation and clinical decision-making. This includes TIR targets, level 2 hypoglycemia thresholds, and daily glucose variability assessment.^1,2 This study found a variation among respondents regarding the knowledge of guideline-aligned CGM metrics. Knowledge was modest overall and about one-quarter of respondents met the predefined threshold for adequate knowledge. Previous studies have explored healthcare professionals’ perceived familiarity, confidence, and implementation readiness related to CGM use but have rarely included objective testing of guideline-aligned CGM metrics.^{7–9,20–22}

In this study, 49% correctly identified the recommended TIR targets among respondents, of which 75% were diabetes educators. De Block et al conducted a survey distributed to seven countries to assess healthcare professionals’ knowledge of and attitudes toward the use of TIR. The study included physicians and allied health care professionals, which included nurses, physician assistants, and diabetes educators. Interestingly, 53% of the allied healthcare professionals reported that “they know a lot or fair amount about TIR”.⁷ It’s important to note that in their study no pharmacists were included. These findings with ours suggest that familiarity with TIR does not necessarily translate into guideline-aligned knowledge of TIR targets, particularly outside specialist groups. They have also assessed benefits of TIR that healthcare professionals have identified. These benefits included optimizing diabetes therapy, supporting informed clinical decision-making, and helping them in empowering people with diabetes.⁷

In our study, daily glucose variability measures such as SD were the lowest-performing knowledge items. Several factors may explain this result. TIR has become the dominant CGM metric in clinical communication, and glucose variability metrics are more technically complex and less emphasized during routine patient counseling.^1,2,4 Moreover, current CGM education focuses on TIR and AGP understanding, and many healthcare professionals interpret CGM visually rather than mathematically.^8,9 Nonetheless, familiarity with glucose variability measures by healthcare professionals in clinical practice is important, as these metrics provide additional important information regarding glycemic stability and hypoglycemia risk.²³

In this study, approximately two thirds have agreed that CGM improves safety and quality of life in older adults and that CGM data was not too complex to integrate in patient counseling. Favorable perception of CGM was also reported by other studies.^{7,8,11,13,21,24} For instance, De Block et al reported that approximately 90% of healthcare professionals including nurses and diabetes educators believed TIR would become a standard diabetes metric.⁷ Another study of ICU nurses found that CGM provided safer patient care.¹⁰ Yap et al found that diabetes care providers in Singapore generally recognized CGM as useful for optimizing glycemic control.¹³ Moreover, several studies from different countries many of which primarily involve physicians, have found a favorable perception of CGM.^8,9,11,22 These findings suggest that CGM is increasingly recognized as a valuable component of routine diabetes care across diverse healthcare settings and professional groups.

In our study, diabetes educators reported greater exposure to CGM, higher confidence, and more frequent AGP reviews. This might be partially explained by the fact that diabetes educators in Saudi Arabia are usually working in specialized diabetes clinics such as endocrinology clinics. In such settings, they may have greater exposure to CGM patients, counsel patients directly and receive diabetes technology training. In our study, diabetes educators have scored highest in formal CGM training and exposure to real-time CGM and have the highest confidence in using trend arrows and TIR. Moreover, in the adjusted analysis, the diabetes educator role remained independently associated with routine AGP review. Our findings extend the existing literature by demonstrating substantial differences in CGM readiness across non-physician professional groups within routine diabetes care settings.^7,8,21

Adequate knowledge of guideline-aligned CGM metrics was not independently associated with routine AGP review. The professional role remained a strong predictor of AGP review, and this suggests that structural and workflow factors may play a larger role in driving routine CGM data review. These findings suggest that improving CGM implementation may require more than increasing individual knowledge; integrating CGM interpretation into professional roles, routine workflows, and multidisciplinary care pathways may be equally important.

The specialty and scope of practice may also explain the differences seen among healthcare professionals. For example, ambulatory care clinical pharmacists working within diabetes-focused clinics may have greater familiarity with CGM interpretation.^12,20,25 Similar variability may also exist among nurses depending on their clinics.^26,27 In some healthcare systems, including the United States, pharmacist-led diabetes clinics may contribute to multidisciplinary diabetes technology management.^12,20,28 Also, some pharmacists and nurses involved in diabetes care may also be certified diabetes educators. In Saudi Arabia, these specialized multidisciplinary roles may currently be less widespread, and healthcare professionals heavily involved in CGM-focused diabetes care may have been relatively underrepresented in our survey sample.

In our study, the participants identified coverage and cost as the main barriers to CGM use, followed by device supply or availability. Previous studies have also reported similar barriers.^9,13,24 In our study, the need for training to interpret CGM reports was only identified by one participant, which is a notable difference from many other studies that reported the need for additional education and training.^7,9 Furthermore, respondents in our study when asked about the training format; they identified case-based workshops followed by workflow-based educational approaches as the preferred methods. These findings suggest that efforts to expand CGM use should address system-level barriers to access, including cost, coverage, and device availability, alongside educational initiatives for healthcare professionals.

This study has some limitations that should be acknowledged. The study used a cross-sectional design and relied on self-reported responses, which may introduce recall and response bias. The sample size was relatively small, particularly within the pharmacist subgroup. The relatively small sample size and voluntary participation may limit the generalizability of the findings to all healthcare professionals involved in diabetes care. Moreover, the knowledge index is not a comprehensive measure for the broader CGM knowledge construct. Differential survey completion across professions may also have introduced selective attrition bias. Approximately 95% of diabetes educators completed the full survey, compared with 58.3% of nurses and 55.6% of pharmacists. The completion rate for the primary outcome (routine AGP review, Q18) was 79.2%. Section C (attitudes) was a notable dropout point for pharmacists. In particular, only 5 of 9 pharmacists completed the full survey, which further limits the reliability and interpretability of pharmacist-specific estimates and warrants particular caution when drawing conclusions for this subgroup. In addition, the survey was distributed electronically through professional channels, which may have preferentially included respondents with greater interest or familiarity with CGM technology. Furthermore, several adjusted odds ratios were estimated with wide confidence intervals, reflecting limited statistical precision attributable to the small sample. These estimates should therefore be interpreted with caution. Future multicenter studies with larger and more representative samples are needed to confirm these findings, while longitudinal studies could evaluate the effectiveness of structured CGM educational interventions.

Conclusion

Healthcare professionals in Saudi Arabia demonstrated positive attitudes toward CGM use in clinical practice, although important differences were observed across professional groups. Diabetes educators showed higher levels of CGM exposure, confidence, and routine AGP review compared with nurses and pharmacists. Despite frequently reported use of CGM in clinical practice, knowledge gaps remained in several guideline-aligned CGM metrics, particularly TIR targets and glucose variability measures. Cost, coverage, and workflow-related barriers were also commonly reported. Providing multidisciplinary teams with better access to CGM, clearer roles, and targeted training could make CGM use consistent in providing routine diabetes care. Future training should combine clear CGM workflows with hands-on CGM practice. These findings should be interpreted in light of the small sample size and voluntary response sampling through a registry-based distribution, which limit their generalizability to the broader healthcare workforce in Saudi Arabia.

Funding Statement

This study received no external funding for study design, data collection, analysis, or manuscript preparation.

Abbreviations

A1C, glycated hemoglobin; AGP, ambulatory glucose profile; BCa, bias-corrected and accelerated; CGM, continuous glucose monitoring; CI, confidence interval; CKD, chronic kidney disease; HbA1c, glycated hemoglobin; iCGM, intermittently scanned continuous glucose monitoring; IQR, interquartile range; NCBE, National Committee of Bio & Med. Ethics; OR, odds ratio; rtCGM, real-time continuous glucose monitoring; SD, standard deviation; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus; TAR, time above range; TBR, time below range; TIR, time in range.

Data Sharing Statement

All data generated or analyzed during this paper is included in this published article.

Ethics Approval and Informed Consent

The study was conducted in accordance with the ethical principles of the Declaration of Helsinki and local institutional research regulations. The study protocol was reviewed and approved by the Regional Research Ethics Committee, Qassim Province, Ministry of Health, Saudi Arabia, registered with the National Committee of Bio & Med. Ethics (NCBE) (Registration No. H-04-Q-001), under Ethics Approval No. 607-47-5879, on October 29, 2025.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

No potential conflict of interest relevant to this article was reported.