Introduction

Hypoglycemia is a common complication of insulin-treated diabetes that impacts on quality of life and care and is a direct cause of mortality [1,2,3,4,5,6]. Its management determines prognosis. It is a common cause of visits to emergency departments and of hospital admissions, but also a cause of ambulance call outs. Among 90,435 emergency calls recorded by the East Midlands Ambulance Trust in England, 523 (0.6%) were for severe hypoglycemia, equating to an incidence of 2.76 per 100 patient-years [7]. Most of these hypoglycemia episodes were successfully treated on the spot. Hypoglycemia requiring an ambulance has been associated with an increased mortality in Minnesota (USA) [8].

Hypoglycemia does not only have major medical repercussions but also have an important economic impact [2]. Both direct and indirect costs of care and hospital admissions are high, ranging from €533 to 691 per severe hypoglycemic event in three European countries for direct costs [9, 10]. Hospital costs were ten times higher than ambulatory treatment costs in a study conducted in Sweden [11]. The estimated annual cost of emergency calls for severe hypoglycemia is 15 million € in England [12]. Indirect financial and social costs are also major [13]. Lost productivity and lost working hours, resulting in lost income to the individual, are important among indirect consequences.

Hypoglycemia is related to morbidity, and mortality in diabetic patients has a major economic and social cost and, moreover, also decreases both quality of life and diabetes control. The onset of repeated and/or severe hypoglycemia is clearly responsible for a decreased quality of life [14]. Furthermore, past history of hypoglycemia decreases diabetes control. Induced fear of hypoglycemia appears to a major issue in diabetic patient’s management [15].

Diabetes education of patients and their families is a key factor in the reduction in diabetes-related morbidity and mortality [16]. Hypoglycemia can be self-managed by the patient or managed by his/her entourage, providing appropriate previous education. However, the factors determining optimal diabetes education have not been clearly identified [16]. Our experience as emergency services (EMS) has suggested that the education of patients and their entourage falls short of the level required to prevent hypoglycemic events and emergency call outs [17]. The reasons related to the patients themselves, their environment and disease history and management have been poorly evaluated The aim of this study conducted in out-of-hospital setting was to determine the contributing factors in the successful diabetes education of patients and their families.

Research design and methods

Setting

This was a prospective, multicenter, observational study conducted by 17 EMS centers (SAMU) across urban and rural France. French EMS operate as follows: an emergency physician receives the emergency call at the dispatching call center. Then the physician chooses the most appropriate action, in particular sending an ambulance manned by an emergency physician, critical care nurse and driver with basic life support training. The ambulances are mobile intensive care units fully equipped with appropriate drugs (anesthetics and catecholamines), biological testing facilities and ultrasound devices [18].

The study was approved by the local ethics committee (Committee for the Protection of Persons—CPP Ile de France, Hôpital Robert Ballanger, Aulnay-sous-Bois, France). Because patient care was not altered in any way, French law required no signed informed consent. Patients and their entourage were informed de facto of their inclusion in the study.

Inclusion criteria

All insulin-treated patients over 18 years of age were included in the study provided that at least one person (family member or close entourage) living under the same roof was present and could be questioned. If several persons were present, all were questioned. There were no exclusion criteria.

Data collection

The investigators (physicians and/or critical care nurses) prospectively collected the following data using a semi-structured questionnaire: (1) gender, age, whether French mother tongue, mastery of the French language as recorded on a visual analogic scale from 0 (perfect) to 10 (none) and highest educational qualification of the patient and of family members; (2) patient’s medical history, treatments, total number of drugs taken and whether a current smoker; (3) time on insulin therapy, insulin type, awareness of prior hypoglycemic episodes and of their frequency, availability of log sheets, whether log sheets completed, time lapse since last recorded blood sugar level, awareness of glycated hemoglobin (HbA1c) test and knowledge of last recorded value, possession of prescriptions for a glucose meter and glucagon and ability to administer glucagon; (4) whether the patient was cared for by a general practitioner or a diabetologist (physician’s office or hospital) and was visited by a practice nurse who performed blood glucose tests and insulin injections; (5) whether they knew how to measure capillary blood sugar and were aware of the symptoms of hypoglycemia (the first three mentioned were recorded) and of what to do should the patient experience a malaise (the first three mentioned were recorded).

Blood sugar level was measured on the spot. The glycemic threshold for hypoglycemia was <2.5 mmol/L.

End points

The main end point in patient education was ability to measure capillary blood sugar. The main end points in the education of the entourage were recognition of the first signs of hypoglycemia and ability to administer glucagon. Accepted symptoms of hypoglycemia were confusion, impaired consciousness, behavioral changes, visual disturbance, speech disturbance and malaise.

Statistics

At least ten patients and ten family members were to be recruited per variable (the questionnaire included 45 items). Total planned recruitment in each case, including a 15% safety margin, was 520.

Quantitative data were compared using the Mann–Whitney test, and qualitative data using the Chi-square test. We performed a multivariate analysis using a multivariate logistic regression model (StatView 5.0, SAS Institute, Cary, NC, USA). The variable to explain was “education” according to the three definitions reported above in the patient and his entourage. Potential explicative variables with a p value of <0.1 after univariate analysis were entered into the model. Odds ratios (OR) were calculated. p values of 0.05 or less were considered significant. Results were expressed as frequencies (with percentages) or medians (with 25–75‰).

Results

Patient characteristics

From September 15, 2009, to January 31, 2011, we included 561 patients and 736 family members in the study [1 [1–1] family member/patient; 258 (38%) spouses, 221 (32%) children, 69 (10%) parents, 35 (5%) siblings and 105 (15%) others] (Table 1). Among the 561 patients, 155 (27%) were current smokers. The main ongoing treatments were cardiovascular drugs [379 patients (68%)], psychotropic drugs [102 patients (18%)] and neurology drugs [53 patients (9%)], with prescriptions available in 423 cases (75%). Median number of drugs per prescription was 5 [2–6]. Oral antidiabetics were being taken by 62 patients (11%), beta-blockers by 12 patients (2%) and antiplatelet drugs by 11 patients (2%).

Table 1 Characteristics of diabetic patients and of entourage
Full size table

Diabetes care

Patients had been using insulin for a median of 12 [5–20] years. The main products were insulin glargine (Lantus®) [n = 234 (42%)], insulin aspart (Novorapid®) [n = 147 (26%)], insulin aspart (NovoMix®) [n = 142 (25%)], insulin lispro (Humalog®) [n = 92 (16%)], insulin detemir (Levemir®) [n = 38 (5%)], human insulin isophane (Insulatard®) [n = 37 (7%)], insulin glulisine (Apidra®) [n = 20 (4%)], human insulin (Umuline®) [n = 27 (5%)] and 3 other insulin types [n = 43 (<5% per type)]. Six patients (1%) were injecting exenatide (Byetta®), a glucagon-like peptide-1 agonist.

Overall, 480 patients (86%) possessed a glucose meter, and 402 (71%) had log sheets. Log sheets were well kept by 345 patients (61%). The median time elapsed since the last recorded blood sugar level exceeded 24 h [0–24] h (n = 387). The result of the HbA1c test was known by 156 patients (28%) (median HbA1c 7.4% [6.8–8.1]). Thirty-five patients (6%) recalled no previous hypoglycemic episode. The last hypoglycemic episode had occurred less than 1 year ago in 153 patients (27%). Frequency of episodes was <1/month in 167 patients (30%), >1/month in 96 patients (17%) and >1/week in 107 patients (19%). Overall, 186 (33%) patients had glucagon on the premises, and 259 (22%) family members were aware that glucagon was available.

The physician caring for the patient was a general practitioner [n = 377 patients (67%)] and/or a diabetologist [n = 348 patients (62%)]. The patient attended either the physician’s office [n = 353 (63%)] or the hospital outpatients department [n = 277 (49%)]. A practice nurse visited 194 patients (34%) and measured capillary blood sugar and injected insulin in 184 patients (33%).

Median capillary blood sugar measured by EMS was 2.1 [1.5–4.2] mmol/L. Overall, 343 patients (61%) were experiencing a hypoglycemic episode (<2.5 mmol/L), with 90 further patients (16%) presenting a blood sugar level <5.5 mmol/L.

Knowledge of diabetes

Overall, 141 (75%) patients and 343 (50%) members of the patients’ family knew how to measure capillary blood sugar. They could name a median of 2 [0–3] symptoms of hypoglycemia with no prompting although 217 (39%) patients and 262 (39%) family members could name no symptom. The three most mentioned symptoms by patients were sweating [n = 188 (34%)], malaise [n = 154 (27%)] and fatigue [n = 90 (16%)] and by their entourage were sweating [n = 236 (36%)], coma [n = 206 (28%)] and malaise [n = 173 (24%)].

Patients and their entourage were, respectively, aware of a median of 1 [0–2] and 2 [1–3] actions to be taken in the event of hypoglycemia. The three most mentioned actions by patients were taking glucose [n = 239 (43%)], eating [n = 178 (32%)] and drinking [n = 123 (22%)] and by their entourage were glucose administration [n = 358 (49%)], calling EMS (SAMU) [n = 178 (24%)] and giving something to drink [n = 158 (21%)]. Few had ever administered glucagon [80 (19%) patients and 84 (21%) family members].

Factors associated with diabetes education

A comparison of educated and non-educated diabetic patients (criterion: ability to measure capillary blood sugar) and of educated and non-educated entourage (criterion: awareness of hypoglycemia symptoms) is given in Table 2. No center effect was observed for either population. Significant factors in these univariate analysis (p < 0.1) were entered into multivariate analysis with the following exceptions: we excluded prior hypoglycemic episodes and HbA1c test results as directly related to education; we included French mother tongue but excluded VAS score for mastery of the French language; we included visits by a practice nurse but excluded their interventions (i.e., blood tests and injections). The factors associated with diabetes education in multivariate analysis are given in Table 3 according to end point. The main factor associated with better patient education was care by a diabetologist. Lack of patient’s educational qualification and visits by a practice nurse were associated with poor patient education. French mother tongue and, depending upon end point, care by a diabetologist were factors associated with better education of entourage.

Table 2 Factors related to diabetes education in patients (in relation to ability to measure blood sugar) and the patients’ entourage (in relation to awareness of hypoglycemia symptoms)
Full size table
Table 3 Multivariate analysis of factors associated with diabetes education of patients and of their entourage
Full size table

Discussion

Management of hypoglycemia is a major medical, psychological and economic issue in patients with diabetes. However, our study found that the diabetes education of patients was inadequate and that of their entourage more than inadequate. Patients and their entourage had little awareness and knowledge of the warning signs of hypoglycemia and of the action to be taken in the event of a hypoglycemic episode. A quarter of patients and half of the family members did not know how to measure blood sugar.

Prevention of the frequency and impact of hypoglycemic episodes relies on the ability of patients and of their entourage to anticipate onset and, consequently, on their ability to control blood sugar levels and to recognize warning signs [19]. It has been clearly established that capillary blood sugar monitoring enables early detection [20]. Brain disorders are early warning signs needing awareness but were not among the first to be mentioned by patients and their entourage in our study [21]. Sweating, which occurs later, was mentioned first. Early detection enables self-management. A call to the EMS signals a severe hypoglycemic episode in a patient who is no longer capable of measuring blood sugar and of self-management [22, 23]. Furthermore, hypoglycemia has been showed to decrease diabetes control. Because patients with history of hypoglycemia are afraid of hypoglycemia recurrence, they accept higher glucose blood level [15].

In France, management of diabetic patients largely devolves to the general practitioner [24]. However, in our study, care by a diabetologist rather than by a general practitioner was associated with better diabetes education of patients and of their entourage. This may be partly because diabetes education campaigns in France have been driven by diabetologists [25]. Care by a visiting practice nurse was strongly associated with poorly educated patients. There are several possible reasons for this: (1) although diabetic patients’ care is part of a nurse’s remit in France, patients’ education is not necessarily so, unlike in countries where diabetes specialist nurses are at the forefront of education and have a key role in patients’ care [26,27,28]; (2) nurses may lack the necessary time for patients’ education because of the low fee (3.15 €) for an intervention during a visit to a diabetic patient in France; (3) general practitioners may order home visiting by a practice nurse for the less well-educated patients. This French practice dramatically contrasts with the way diabetic patients are managed in many countries such as in the USA. Nurses are efficiently involved in diabetic patients’ management. Globally, multidisciplinary management should be promoted [29]. How to reach high-quality education in diabetic patient has been largely study, included in randomized studies. Recent guidelines focus on the crucial impact of an organized dedicated structure, a coordinated, comprehensive, personalized and evaluated program [16]. These guidelines are coherent with our results that place the specialist in the center of the education process. Similar guidelines exist on the diagnostic and the management of hypoglycemia [30]. Specific education and training programs have been suggested.

The main personal factors associated with diabetes patients’ education in our study were French mother tongue and educational qualification. Patient’ self-management and HbA1c level have been shown to be significantly correlated with educational attainment [31]. Poorer control of diabetes has been reported in populations with language difficulties in a study of 21 international pediatrics centers and in immigrants [32, 33]. We previously reported such result regarding education of patients with acute coronary syndrome [34]. Information provided by health professionals and diabetes education offer need to be tailored to each patient and their environment [35].

Glucagon is the only available home treatment when an altered state of consciousness prevents a patient from eating or drinking. It is thus a relevant and stringent criterion of diabetes education. Its underuse in the treatment of hypoglycemia is well established [21]. In our study, it was available in only 33% homes. Use of glucagon by the entourage is quite feasible if appropriate education is offered and also accepted. This is not always the case [9, 36]. Physicians may be unwilling to educate family members as family presence increases consultation time [36]. Moreover, the family member or entourage present daily may not necessarily be the person accompanying the patient to the physician’s office or educated by the physician. Globally, education of the entourage that can contribute to improve diabetes control and complications’ management has to be developed [29].

Being able to question patients and their families in their home was a distinct advantage in our study on self-management, but our study also has limitations. It was based on data reported by patients and their entourage, some of whom may have overestimated their ability to measure capillary blood sugar or administer glucagon. We did not check their ability to perform these tasks. Unfortunately, we did not distinguish type I and type II diabetes, whereas it could be related to the education level. Conclusions drawn in France may not apply to other countries because of differences in diabetes management according to health system. However, our results strongly suggest that the key to successful education results in coordinated network multidisciplinary management and specific training dedicated to patients and family members. Although networks taking charge of diabetic patients vary across France, no center effect was detected.

In conclusion, in France, diabetic patients and their entourage are inadequately educated. There was a major benefit in receiving care from a diabetologist whose actions may serve as a model for improvement initiatives. The nurse’s role needs reappraisal. Diabetes education should be tailored to the age, educational qualification and mother tongue of patients and their entourage.