Documentation of Hypoglycemia Assessment among Adults with Diabetes during Clinical Encounters in Primary Care and Endocrinology Practices

Abstract

Purpose:

To examine the proportion of diabetes-focused clinical encounters in primary care and endocrinology practices where the evaluation for hypoglycemia is documented; and when it is, identify clinicians’ stated actions in response to patient-reported events.

Methods:

A total of 470 diabetes-focused encounters among 283 patients non-pregnant adults (≥18 years) with type 1 or type 2 diabetes mellitus in this retrospective cohort study. Participants were randomly identified in blocks of treatment strategy and care location (95 and 52 primary care encounters among hypoglycemia-prone medications [i.e. insulin, sulfonylurea] and others patients, respective; 94 and 42 endocrinology encounters among hypo-treated and others, respectively). Documentation of hypoglycemia and subsequent management plan in the electronic health record were evaluated.

Results:

Overall, 132 (46.6%) patients had documentation of hypoglycemia assessment, significantly more prevalent among hypo-treated patients seen in endocrinology than in primary care (72.3% vs 47.4%; P = .001). Hypoglycemia was identified by patient in 38.2% of encounters. Odds of hypoglycemia assessment documentation was highest among the hypo-treated (OR 13.6; 95% CI 5.5–33.74, vs. others) and patients seen in endocrine clinic (OR 4.48; 95% CI 2.3–8.6, vs. primary care). After documentation of hypoglycemia, treatment was modified in 30% primary care and 46% endocrine clinic encounters; P = .31. Few patients were referred to diabetes self-management education and support (DSMES).

Conclusions:

Continued efforts to improve hypoglycemia evaluation, documentation, and management are needed, particularly in primary care. This includes not only screening at-risk patients for hypoglycemia, but also modifying their treatment regimens and/or leveraging DSMES.

Introduction

The goals of diabetes care are to reduce immediate and long-term complications, alleviate symptom burden, and improve quality of life [1,2]. Severe (defined by neuroglycopenia requiring external assistance) and other clinically significant hypoglycemia are common, serious, and potentially preventable adverse events in the management of type 1 and type 2 diabetes [3]. Severe hypoglycemia is associated with increased risks of cardiovascular disease, cognitive decline, impaired quality of life, and death [4–8]. One of the strongest predictors of future hypoglycemia is prior hypoglycemia [9]. Other risk factors include insulin or sulfonylurea therapy, advanced age, presence of comorbidities, long duration of diabetes, and hypoglycemia unawareness [10,11]. Because hypoglycemia risk may be reduced by addressing key modifiable risk factors, particularly treatment modification and self-management education, clinical practice guidelines advise that all patients at risk for hypoglycemia be screened for it at each clinical encounter [3,12–16].

Most people with diabetes are cared for by primary care clinicians, with or without concurrent endocrine consultation [17]. Efforts to improve hypoglycemia detection and management therefore need to encompass both clinical settings [5]. Screening for hypoglycemia is particularly important because patients rarely inform their healthcare team of hypoglycemic events, even when severe [10]. Moreover, screening alone is not sufficient to prevent hypoglycemia recurrence if no actions are taken after it is identified. Prior studies suggest that treatment regimens are rarely modified in response to hypoglycemic events [10]. Thus, despite consistent recommendations to screen for hypoglycemia and address the risk factors for it, there is no evidence whether this occurs in U.S. primary care and endocrinology practices.

We therefore examined a random representative sample of diabetes-focused clinical encounters in primary care and endocrinology for documentation of hypoglycemia assessment and any subsequent management steps taken by clinicians. Identifying potential gaps in hypoglycemia assessment, documentation, and management are necessary first steps toward improving the quality, safety, and patient-centeredness of diabetes care.

Materials and methods

Study design

This is a retrospective cohort study of adult patients with diabetes seen in Mayo Clinic (Rochester, MN) primary care and endocrinology practices for diabetes management between January 1, 2015 and December 31, 2015. Mayo Clinic is a multi-specialty integrated health care delivery system providing comprehensive medical care through outpatient, inpatient, and eHealth platforms to a diverse population of local, regional, and national/international patients. In its Employee and Community Health (ECH) section, Mayo Clinic also delivers longitudinal primary care to employees, dependents, and people residing within the local catchment area. ECH is comprised of eight primary care clinics staffed by internal medicine and family medicine physicians and residents, physician assistants, and nurse practitioners. The endocrinology consultative practice serves both local patients (i.e. those who also receive primary care in ECH) and regional, national, and international referral patients. This study was approved by the Mayo Clinic Institutional Review Board.

Study Population

We identified adults (≥18 years old) with the diagnosis of diabetes who had a faceto-face encounter in either primary care or endocrinology clinic between January 1, 2015 and December 31, 2015. Only encounters with the primary diagnosis of diabetes were included. We excluded patients who did not grant authorization for their medical records to be reviewed for research; were deceased as of December 31, 2015; had a diagnosis of secondary diabetes, gestational diabetes, or pregnancy during the study period; or had no hemoglobin A1c (HbA1c) result available during the study period.

Eligible encounters were stratified on the basis of site of care (primary care vs. endocrinology) and diabetes treatment regimen at the time of the encounter (high-risk of hypoglycemia [i.e. insulin and/or sulfonylurea] vs. all other glucose-lowering medications or lifestyle management). We randomly selected 200 patients treated with hypoglycemia-prone medications (heretofore referred to as “hypo-treated”; 100 each from primary care and endocrinology) and 110 patients not treated with these medications (55 patients each from primary care and endocrinology). All diabetes-focused visits for these patients during the study period were analyzed. Seven patients had visits in both primary care and endocrinology during the study period and each specialty visit was analyzed separately for the visit-based analyses. However, for patient-level analyses, these patients were classified in accordance to the site of their first visit during the study period.

Outcomes

The primary outcome was documentation of a discussion pertaining to hypoglycemia during any visit in either primary care or endocrinology that took place during the study period. Secondary outcomes, ascertained only for the subset of patients with documented self-reported hypoglycemia were: (1) how the episode was managed by the patient (e.g. self-administered carbohydrates, glucagon, emergency medical services call, hospitalization, not assessed); (2) how the clinician responded to the reported event (e.g. provide education, change treatment regimen, provide or refer to diabetes self-management education/support [DSMES]); (3) if the treatment regimen was altered, what was the nature of that change (e.g. change to new medication at the same level of hypoglycemia risk, de-intensification vis-à-vis dose reduction or switch to a non-hypoprone regimen, intensification vis-à-vis dose escalation or addition of another hypo-prone medication, and/or prescription of glucagon and/or glucose tablets).

Independent Variables

All variables were abstracted manually, independently and in duplicate, from the electronic health record (EHR) into a web-based standardized electronic form. First, the abstraction form and methodology were iteratively tested and refined through pilot review of five EHR by all reviewers. Formal data extraction was initiated after perfect concordance between reviewers was achieved. Variables of interest were ascertained at the time of the first (unless specified otherwise) diabetes-focused visit during the study period (in the event of multiple visits per year) and included patient age, sex, BMI, HbA1c (most recent prior to all eligible encounters), diabetes type, treatment regimen (at any encounters), diabetes duration, and comorbidity burden quantified using the Charlson comorbidity index [18]. Encounters were categorized as new (clinician had not seen the patient before) vs. established. Assessment and documentation of hypoglycemia unawareness was defined as documentation of the patient experiencing glucose <70 mg/dL without associated symptoms.

Statistical Analysis

Data are presented as means and standard deviations (SD) for continuous variables and as frequencies and percentages for categorical variables. Univariate between-group comparisons were performed using χ2 tests for categorical and binary variables. Normality was determined using the Kolmogorov-Smirnov test. Unpaired Student’s t test or MannWhitney U test were used to compare continuous variable according to normality. Logistic regression analysis was performed to identify predictors of hypoglycemia documentation with independent covariates including clinician specialty, diabetes type, patient age, sex, BMI, new or established patient visit type, years of diabetes duration, and hypoglycemia risk profile of the patient’s treatment regimen. P values ≤ 0.05 were considered statistically significant. IBM SPSS 22.0 version (SPSS, Inc., Amonk, NY) was used to perform all analyses.

Results

Study Population

We reviewed 310 patient charts and excluded 27 who did not meet prespecified inclusion/exclusion criteria (Figure 1). The final study cohort was comprised of 283 individual patients: 95 hypo-treated and 52 non hypo-treated in primary care, and 94 hypo-treated and 42 non hypo-treated in endocrinology. Baseline characteristics of the 147 (51.9%) patients seen in primary care and 136 (48.1%) seen in endocrinology are detailed in Table 1. Per study design, the proportions of hypo-treated were similar between those seen in primary care and endocrinology clinics; P = 0.77. The proportion of patients with type 1 diabetes was also similar between the two groups (4.1% in the primary care group and 5.2% in the endocrinology group; P = .66).

Figure 1.

Study design and participants

Table 1.

Baseline characteristics of the study population.^a

Characteristics	Clinic Specialty
	Primary Care	Endocrinolog y
	147 (51.9)	136 (48.1)	P-value
Patient demographics
Age, years, mean (SD)	61.4 (±15.2)	65.1 (± 12.0)	0.02
Age, years, category
<45	20 (13.6)	5 (3.7)	0.02
45–64	63 (42.8)	58 (42.6)
65–74	33 (22.4)	42 (30.9)
≥75	31 (21.1)	31 (22.8)
Female, percent	50 (34)	45 (33.1)	0.86
BMI, kg/m2, mean (SD)	33.9 (±6.9)	32.0 (± 6.7)	0.006
Diabetes type
Type 1	6 (4.1)	7 (5.2)	0.66
Type 2	141 (95.9)	129 (94.9)
Diabetes duration, mean (SD), y	8.1 (±7.5)	11.2 (± 8.2)	0.001
HbA1c, %, mean (SD)	8.0 (±1.6)	7.7 (± 1.4)	0.31
HbA1c category
<7.0%	46 (31.2)	41 (29.6)	0.13
7.0 – 8.9%	67 (45.5)	75 (55.5)
≥9.0%	34 (23.1)	20 (14.8)
Diabetes treatment
Lifestyle therapy only	19 (12.9)	13 (8.9)	0.77
Non hypoglycemia-prone regimen	33 (22.4)	29 (21.5)
Sulfonylurea b	35 (23.8)	39 (28.9)
Insulin b	37 (25.2)	34 (25.2)
Sulfonylurea and Insulin b	23 (15.6)	21 (15.6)
Encounter characteristics
Type of encounter
Initial visit	21 (14.3)	42 (30.8)	0.001
Follow-up visit	126 (85.7)	94 (69.6)
Comorbidities
Charlson Comorbidity Index, mean (SD)	3.8 (±2.2)	3.9 (± 1.8)	0.17
Short-term diabetes complicationsc	6 (4.1)	11 (8.1)	0.15
Chronic kidney disease	10 (6.8)	10 (7.4)	0.85
End-stage kidney disease	0 (0)	3 (2.2)	0.07
Chronic diabetes complication(s)d	17 (11.6)	14 (9.8)	0.73
Advanced retinopathy	0 (0)	1 (0.7)	0.29
Diabetic Retinopathy	10 (6.8)	5 (3.7)	0.24
Amputation(s)	1 (0.7)	0 (0)	0.33
Coronary artery diseasee	19 (12.9)	23 (16.9)	0.34
Other atherosclerotic diseasef	22 (15)	28 (20.6)	0.21
Heart Failure	7 (4.8)	8 (5.6)	0.67
Dementia	3 (2.1)	0 (0)	0.09
Mild cognitive impairment	2 (1.4)	1 (0.7)	0.6
Major mental illness	13 (8.8)	2 (1.5)	0.006
Anxiety, depression	16 (10.9)	5 (3.7)	0.02

BMI body mass index, HbA_1c hemoglobin A_1c, PVD peripheral vascular disease

^aValues are numbers (percentages) unless otherwise stated.

^bWith or without concurrent non-hypoglycemia prone medications.

^cShort-term diabetes mellitus complications: Hypoglycemia, diabetes ketoacidosis, hyperosmolar hyperglycemic non-ketotic syndrome

^dChronic diabetes complications : Heart failure, chronic kidney disease, liver failure, diabetic retinopathy, cognitive impairment, dementia.

^eCoronary artery disease: Stable angina, instable angina, myocardial infarction.

^fOther atherosclerotic disease: ischemic stroke, peripheral artery disease.

Patients treated in primary care were younger (mean age 61.4 [SD ± 15.2] years) than those treated in endocrinology (mean age 65.1 [SD ± 12] years); P = .02. They also had shorter duration of diabetes (8.1 [SD ± 7.5] years vs. 11.2 [SD ± 8.2] years; P = .001). Mean HbA1c was 8.0% (SD ± 1.6) among patients seen in primary care and 7.7% (SD ± 1.4) in endocrinology; P = .31. Both groups had otherwise comparable comorbidity burden (CCI 3.8 [SD ± 2.2] in the primary care group and 3.9 [SD ± 1.8] in the endocrinology group; P = .17). The examined encounters were first-time encounters for 30.8% of patients in endocrinology and 14.3% of patients in primary care; P = .001.

Documentation of hypoglycemia assessment

We examined all 470 diabetes-focused encounters for the included patients that occurred in either primary care or endocrinology during the study period, with an average of 1.6 (SD ± 0.9) encounters per patient; Table 2. Overall, hypoglycemia was assessed in 40.6% (191/470) of encounters and was subsequently endorsed by patients 38.2% of the time (73/191). When examined by hypoglycemia risk profile, hypoglycemia was assessed in 53.1% (173/326) and subsequently identified in 42.1% (73/173) of the encounters among hypo-treated patients. Further stratified by site of care, hypoglycemia was assessed in 69.2% (108/156) of encounters among hypo-treated patients in endocrinology and 38.2% (65/170) in primary care; P < .001. There was also significantly more hypoglycemic events elicited from screened hypo-treated patients in endocrinology than primary care: 46.2% (50/108) of encounters in endocrinology and 35% (23/65) of encounters in primary care; P < .001 (Table 2).

Table 2.

Hypoglycemia Assessment and Documentation in the Electronic Health Record.^a

	Total	High Risk of Hypoglycemia			Low Risk of Hypoglycemia
		Primary Care	Endocrinology	P-value	Primary Care	Endocrinology	P-value
Number of patients	283	95	94		52	42
Number of encounters	470	170	156		90	54
Visits per patient, mean ( ± SD)	1.6 (0.9)	1.7 (1.6)	1.5 (0.8)	0.11	1.8 (1.2)	1.2 (0.4)	<0.001
Hypoglycemia assessment and documentation
N (%) of patients	132 (46.6)	45 (47.4)	68 (72.3)	<0.001	7 (13.2)	12 (28.6)	0.06
N (%) of encounters	191 (40.6)	65 (38.2)	108 (69.2)	<0.001	7 (7.8)	11 (20.4)	0.02
Confirmed hypoglycemia among those with hypoglycemia documentation
N (%) of patients	55 (19.4)	18 (18.9)	37 (39.4)	0.002	0	0
N (%) of encounters	73 (15.5)	23 (13.5)	50 (32)	<0.001	0	0
N (%) of encounters in which hypoglycemia was assessed	73 (38.2)	23 (35.3)	50 (46.2)	<0.001	0	0

^aValues are numbers (percentages) unless otherwise stated.

The most common modality of hypoglycemia ascertainment was based on patient self-report (56.5% of encounters in primary care and 70% in endocrinology), followed by review of glucometer data (4.3% of encounters in primary care and 12.0% in endocrinology), or both glucometer review and patient self-report (30.4% of encounters in primary care an 8.0% in endocrine). Continuous glucose monitor data was used by 2% of encounters in endocrinology and none in primary care. Modality was not specified in 8.7% of primary care and 8.0% of endocrinology encounters.

Hypoglycemia was also assessed in 20% of endocrinology encounters among nonhypo-treated patients, compared to 8% in primary care; P = .02. None of the screened patients endorsed experiencing hypoglycemia. Results were comparable when assessed on the patient level (Table 2).

In multivariate analysis, factors associated with hypoglycemia assessment/ documentation were the type of diabetes treatment regimen (OR 13.6; 95% CI 5.5–33.7 for hypoglycemia-prone regimen vs. not), clinician specialty (OR 4.4; 95% CI 2.3–8.6 for endocrinology vs. primary care), and encounter type (OR 2.3; 95% CI 1–5.1 for new consultation vs. established visit); Figure 2. There was no association with patient age, sex, duration of diabetes, BMI, or diabetes type.

Figure 2.

Influence of variables on Hypoglycemia Assessment and Documentation

Patient and Clinician Responses to Hypoglycemia

Clinicians did not routinely document their patients’ management of hypoglycemic events, particularly in endocrinology (Table 3). In the minority of encounters where hypoglycemia was documented and whose management of it by the patient was assessed, it was usually to self-administer carbohydrates (primary care n = 10/23 [43.5%] vs endocrinology n = 8/50 [16%]; P = .03). One patient in primary care required an emergency department visit, while three patients in endocrinology were hospitalized for hypoglycemia.

Table 3.

Patient and clinician response to hypoglycaemia events among encounters with documented hypoglycemia.^a

	Primary Care	Endocrinology	p-value
Patients with identified hypoglycemia episodes	23 (13.5)	50 (32.1)
Patient response to the episode
Self-administered carbohydrates	10 (43.5)	8 (16)	0.03
Glucagon	0 (0)	1 (2)
Emergency medical service called	1 (4.3)	0 (0.0)
Hospitalization	0 (0)	3 (6)
Not assessed	12 (52.2)	38 (76)
Clinician education and management response
Patients with previous diabetes education	15 (65.2)	18 (36)	0.02
In patients with previous diabetes education, time since last education
less to 1 year	11 (73.3)	5 (27.8)	0.009
1 to 5 years	4 (26.7)	13 (72.2)
Clinician provided education about hypoglycemia	8 (34.8)	21 (42)	0.55
During the encounter	8 (100)	12 (57.1)	0.08
Outside of the encounter	0 (0)	6 (28.6)
Both during and outside of the encounter	0 (0)	3 (14.3)
Not reported	15 (65.2)	29 (58)
Content of diabetes education
Signs and symptoms of hypoglycemia	0 (0)	1 (2)	0.51
Signs and symptoms of hypoglycemia, Diet and exercise counseling	1 (4.3)	1 (2)
Diet, exercise, and medication counseling	0 (0)	1 (2)
Exercise counseling	0 (0)	1 (2)
Medication counseling	2 (8.7)	8 (16)
Not specified	1 (4.3)	1 (2)
DSMES referrals
Dietitian	4 (17.4)	4 (8)
Diabetes educator	0 (0)	4 (8)
Changes in the treatment regimen	8 (34.8)	36 (72)	0.003
Type of treatment change
Same risk, different medication	1 (4.3)	5 (10)	0.31
De-Intensified	7 (30.4)	23 (46)
Intensified	0 (0)	8 (16)
Glucagon was prescribed	2 (8.7)	2 (4)	0.41
Glucose was prescribed	3 (13)	2 (4)	0.15

DSMES diabetes self-management education and support

^aValues are numbers (percentages) unless otherwise stated.

We also examined the clinicians’ responses to elicited hypoglycemia (Table 3). During the majority of encounters, clinicians did not document the delivery of hypoglycemia-focused education to their patients. In the 34.8% of primary care encounters and 42% of endocrinology encounters in which patients did receive education, it was most often delivered during the encounter (100% of primary care clinicians and 57.1% of endocrinologists); endocrinologists also documented providing education outside of the encounter (29% of those who provided education) or both (14%); P = .08. Evidence of prior diabetes self-management education was noted in 15/23 (65.2%) of primary care encounters and in just 18/50 (36%) of endocrinology encounters; P = 0.02.

Treatment regimens were modified in 72% of encounters in which patients reported hypoglycemia in endocrinology and 35% of encounters in primary care; P = .003. Few clinicians prescribed glucagon (primary care n=2 [8.7%] vs endocrinology n= 2 [4%]; P = .41) or glucose tablets (primary care n=3 [13%] vs endocrinology n=2 [4%]; P = .15).

Discussion

Efforts to prevent hypoglycemia are contingent on the ability to recognize at-risk individuals, identify potential precipitating and contributing factors, and take action toward hypoglycemia prevention by way of treatment modification, patient counseling, and/or referral for DSMES [9]. However, despite explicit recommendations for routine assessment of at-risk patients for hypoglycemia [3,12–16], we found major gaps in both endocrinology and primary care practices. Patients treated with hypoglycemia-prone medications were screened for hypoglycemia during just 69.2% of encounters in endocrinology and 38.2% in primary care. Furthermore, even when hypoglycemia was identified, clinicians did not consistently act to prevent its recurrence. Specifically, clinicians de-intensified treatment regimens in fewer than half of encounters were hypoglycemia was confirmed (46% in endocrinology and 30.4% in primary care) and did not routinely provide or refer patients for DSMES (34.8% in primary care and 42% in endocrinology). These findings reinforce the need for processes that facilitate screening of at-risk patients for hypoglycemia, deintensifying or otherwise modifying treatment regimens when hypoglycemia is identified, and referring patients for necessary self-management education and support.

Screening for and acting upon hypoglycemia was less prevalent during primary care than endocrinology encounters, though there were substantial gaps in all clinical settings. This is not surprising, as endocrinologists are specifically trained in the management of diabetes, see higher volumes of diabetes patients as a proportion of their overall practice, and can devote an entire clinical encounter to the discussion of diabetes management. In contrast, primary care clinicians often do not have diabetes-specific training, have no access to diabetes-management or education resources, and need to address multiple items per visit. The low yield of elicited hypoglycemia (i.e. few instances of hypoglycemia screening resulted in an identified event), particularly in primary care, makes it unlikely that the low rates of documented hypoglycemia assessment are due to gaps in documentation as opposed to screening itself. Ultimately, the fact that fewer than half of patients at risk for hypoglycemia seen in both primary care and endocrinology were screened for hypoglycemia is alarming and reinforces the need for continued efforts to improve hypoglycemia awareness, evaluation, documentation, and management.

The only factor other than clinician specialty and glucose-lowering treatment regimen that affected the likelihood of hypoglycemia screening was whether the encounter was an initial as opposed to established visit. Patients seen for first-time consultations were more than twice as likely to be screened, reflecting a more thorough assessment of disease management and control. In contrast, there was no independent association between hypoglycemia screening and either patient age or diabetes duration, despite data demonstrating that older patients and patients with longer duration of disease have an increased risk of hypoglycemia and should therefore be screened for its occurrence [19].

In addition to inadequate documentation of hypoglycemia assessment, clinicians rarely documented hypoglycemia severity, precipitating factors, or how patients treated their events. Yet, these are key to understanding what may have precipitated these events and gauging patient knowledge of self-management,[20] both necessary for subsequent treatment modification and/or provision of DSMES as a way of preventing recurrent events.

Finally, both primary care clinicians and endocrinology did not consistently act to prevent hypoglycemia recurrence. Specifically, they infrequently provided hypoglycemia-focused education during or outside the encounter or referred patients for DSMES (34.8% of the time in primary care and 42% of the time in endocrinology). Clinicians also de-intensified treatment in just 46% of endocrine and 30.4% of primary care encounters during which hypoglycemia was identified. Treatment de-intensification is critical to preventing hypoglycemia recurrence [21], and clinical inertia in the face of hypoglycemic events is likely to contribute to high rates of repeated events [22,23].

Our study has several important limitations, though they do not affect the main conclusions or their clinical relevance. We relied on the EHR to ascertain hypoglycemia screening and may have missed situations where hypoglycemia was discussed but not documented. However, documentation is an integral aspect of the care process, particularly in the era of team-based care where multiple healthcare providers are engaged in the care for each patient via the EHR [24,25]. Patient’s glucometer and/or CGM data was not routinely a way of identifying and confirming hypoglycemic events. This data was not routinely available in the EHR and therefore could not be analyzed/included. Furthermore, we do not have data about how patients monitored their blood glucose at home, and whether (for patients who reported hypoglycemia in the absence of confirmatory glucometer or CGM data) they self-diagnosed hypoglycemia on the basis of symptoms alone or substantiated glucose levels. Moreover, irrespective of documentation quality, the lack of actions toward hypoglycemia prevention in the context of documented events is striking. Furthermore, the study cohort may not be reflective of the general U.S. population of adults with diabetes, as it was an enriched sample of primary care and endocrinology patients in one academic integrated healthcare delivery center. However, it is likely that other healthcare organizations face the same challenges in caring for patients with diabetes, including appointment length constraints, gaps in awareness of the morbidity associated with hypoglycemia, financial pressures stemming from performance measurement, and clinical inertia. Our findings are also consistent with earlier literature suggesting gaps in hypoglycemia assessment and documentation, though most prior studies relied on questionnaire recall or prospective diaries which are subject to bias [6,10,11,26,27]. Moreover, previous studies that evaluated documentation of hypoglycemia events in the EHR only reported if those events occurred, without knowing if hypoglycemia assessment by the physician happened in the first place [20]. Our study is the first to examine whether physicians actively ask their patients about hypoglycemia, and to assess for potential differences between endocrinology and primary care.

In order to improve hypoglycemia screening and documentation, and ultimately its management, clinicians may want to consider it as a sixth vital sign of people with diabetes receiving hypoglycemia-prone medications [28]. Additionally, hypoglycemia assessment needs to be standardized, to include not only whether an event occurred, but also characterizing its severity, blood glucose level (if measured) and symptoms (if any) at onset, context of occurrence and potential triggers, and how it was treated. De-prescribing programs for patients at high risk for hypoglycemia, particularly those who have already experienced hypoglycemic events, can be implemented in primary care and specialty practices. Embedding pharmacists, diabetes educators, and social workers in primary care and endocrine practices can help clinicians manage patients experiencing hypoglycemia and reduce the burden of screening. This is particularly important given the lack of treatment modification or hypoglycemia-focused education after the majority of identified events. Finally, performance measurement can better support high quality and safe diabetes care by including hypoglycemia as a quality indicator [29]. Additional research is needed to determine whether efforts to enhance screening for hypoglycemia can spur appropriate treatment responses and reduce risk of recurrent hypoglycemia while simultaneously ensuring continued control of hyperglycemia.

Take home messages.

1. Hypoglycemia is not routinely assessed and documented in either primary care or endocrinology practices even among patients treated with medications at risk for hypoglycemia (i.e. insulin, sulfonylurea), despite guideline recommendations to do so.

2. Even amongst patients reporting hypoglycemia, treatment regimens are not routinely modified or de-intensified and only a minority of patients receive diabetes self-management education in response to the hypoglycemic events.

3. Our study reinforces the need for processes that facilitate screening of at-risk patients for hypoglycemia, de-intensifying or otherwise modifying treatment regimens when hypoglycemia is identified, and referring patients for necessary self-management education and support.