Continuous Glucose Monitoring in Primary Care: Understanding and Supporting Clinicians’ Use to Enhance Diabetes Care

Tamara K. Oser; Tristen L. Hall; L. Miriam Dickinson; Elisabeth Callen; Jennifer K. Carroll; Donald E. Nease; LeAnn Michaels; Sean M. Oser

doi:10.1370/afm.2876

Article Figures & Data

Tables

Additional Files

View popup

Table 1.

Sample Characteristics

Respondent Characteristic	No. (%)
Professional role (n = 642)
Attending, faculty or community physician	460 (71.7)
Resident	122 (19.0)
PA or NP	50 (7.8)
Other	10 (1.6)
Medical specialty (n = 519)
Family medicine	495 (95.4)
Other^{^a}	24 (4.6)
Primary setting (n = 607)
FQHC or similar	165 (27.2)
Hospital-owned practice	161 (26.5)
Private practice	134 (22.1)
Academic medical center	83 (13.7)
Other	64 (10.5)
Years since training (n = 626)
0	73 (11.7)
1-5	130 (20.8)
6-10	153 (24.4)
11-15	109 (17.4)
16-20	28 (4.5)
>20	104 (16.6)
NA	29 (4.6)
Payer mix
>25% of patients have Medicare (n = 610)	403 (66.0)
>25% of patients have Medicaid (n = 609)	318 (52.2)
>25% of patients have private insurance (n = 609)	341 (55.9)
>25% of patients have no insurance (n = 590)	77 (13.0)
Practice region (n = 632)
West	317 (50.1)
South	141 (22.2)
Midwest	107 (17.0)
Northeast	64 (10.2)
US Territory	3 (0.5)
Miles from endocrinologist (n = 614)
≤10	405 (66.0)
11-40	102 (16.6)
>40	107 (17.4)
Diabetes care & education resource access
Diabetes care and education specialist within practice (at least part-time) (n = 614)	221 (36.0)
Can refer to diabetes care and education specialist within 10 miles (n = 612)	498 (81.4)
Can refer to other diabetes education resources within 10 miles (n = 613)	454 (74.1)
Experience with CGM (n = 629)
I have never heard of CGM	6 (1.0)
I have heard of CGM, but I have never had patients on it	87 (13.8)
I have had patients on CGM, but I have never prescribed it	293 (46.6)
I have prescribed a CGM	243 (38.6)
Likelihood to prescribe CGM (n = 603)
Not at all likely	40 (6.6)
Somewhat likely	187 (31.0)
Moderately likely	168 (27.9)
Very likely	208 (34.5)
Moderately/very likely to prescribe CGM with access to resources
Consultation on insurance issues (n = 603)	434 (72.0)
CGM education/training workshop (n = 603)	436 (72.3)
One-time consult with endocrinologist (n = 602)	380 (63.1)
Virtual specialty care center to refer patients for direct support (n = 602)	379 (63.0)
Educational website about CGM (n = 603)	375 (62.2)
Telementoring sessions with specialty team (n = 603)	372 (61.7)
Asynchronous e-consultations with an endocrinologist (n = 604)	324 (53.6)

CGM = continuous glucose monitoring; FQHC = federally qualified health center; NA = not applicable; NP = nurse practitioner; PA = physician assistant.
↵a General internal medicine, internal medicine-pediatrics, pediatrics.

View popup

Table 2.

Respondent and Practice Characteristics and CGM Prescribing (N = 570)

Characteristic	OR (95% CI)	P Value
Respondent role		<.001
Physician	Ref.
Resident	0.30 (0.16-0.58)
PA or NP	0.36 (0.16-0.77)
Primary setting		.13
Private practice	Ref.
Academic medical center	1.42 (0.73-2.76)
FQHC/similar	0.88 (0.50-1.53)
Hospital-owned	1.39 (0.84-2.29)
Other	1.82 (0.93-3.57)
Employment		.04
Full-time or resident	Ref.
Part-time	0.55 (0.30-0.98)
Time primary care, %		.03
≥75	Ref.
<75	0.60 (0.37-0.96)
Miles from endocrinologist		.03
≤10	Ref.
11-40	0.95 (0.58-1.53)
>40	1.94 (1.17-3.21)

CGM = continuous glucose monitoring; FQHC = federally qualified health center; NP = nurse practitioner; PA = physician assistant.
Note: C-statistic = 0.688. Years since training and payer mix did not meet threshold for inclusion in final model.

View popup

Table 3.

Respondent and Practice Characteristics and Future Likelihood to Prescribe CGM (N = 570)

Characteristic	OR (95% CI)	P Value
Previously prescribed CGM	7.44 (4.80-11.53)	<.001
Miles from endocrinologist		.17
≤10	Ref.
11-40	0.76 (0.46-1.25)
>40	1.43 (0.83-2.46)
Patients with medicare, %		.004
≤25	Ref.
26-50	1.37 (0.89-2.10)
>50	2.67 (1.50-4.78)
Patients with medicaid, %		.08
≤25	Ref.
26-50	0.63 (0.41-0.98)
>50	1.04 (0.62-1.75)

CGM = continuous glucose monitoring.
Note: C-statistic = 0.756. Respondent role, primary setting, part or full-time employment, years since training, percent of patients using private insurance, and percent of patients with no insurance did not meet threshold for inclusion in final model.